Author: siencilabs

Next Big SLB Update

Posted on by

It’s SuperLongBoard update time!

Hi everyone, Chris here. It’s great to be speaking with y’all again, it’s not often I get around to writing blog posts since I tend to be more behind the curtain on the tech development we do here at Sienci. As Andy mentioned in his November Production Update post, there’s a lot to update on so I’ll spread as many juicy details as I can. We’re now at over 400 people interested in buying the new SLB system we’ve been working very hard at so thank you for your support! It’s very clear that our community is very excited about this massive upgrade to the LongMill and for the future of the hobby CNC industry.

To read past updates about the SLB, see our first major blog post (First look at the SuperLongBoard) as well as all the Production Updates that Andy has been putting out to keep you guys updated on our progress: May Update 1, Aug Update 2, Sept Update 3, Oct Update 4.

There’s also a video I filmed which you can feel free to watch if you prefer to see more of my face 🙂

Recap

Let’s start with a quick refresher about how the SLB came to be Sienci’s biggest project to-date.

As many of you know, Sienci Labs has historically been built up by our mechanical solutions – an expected direction seeing as myself and Andy co-founded the company while studying Mechanical Engineering. Our goal with Sienci was always to take advanced, industry technology and try to find a way to simplify it in ways that made it less expensive and easier to use so that everyone could have access to making complex things at home. We feel like that goal of price-point and ease-of-access is something we’ve been proud and successful at to-date with our big projects like the LongMill turned LongMill MK2 CNC, LaserBeam laser system, Vortex rotary axis, AutoZero touch plate, and our unique MK2 Magnetic Dust Shoe.

What’s been clear to us since the start though is that the CNC experience won’t be able to be improved with mechanics alone. Most hobby CNC communities have quirks with their machines, but once those quirks are understood the majority of confusion comes from everything else you see day-to-day when you use it: designing projects, tool selection, feeds and speeds, machine control, and reliable cutting. This is why we started working hard on non-mechanical solutions that we’re also very proud of like gSender for open-source CNC control, our in-house feeds and speeds charts, our expansive resources for learning and understanding CNC, ever-growing roster of tutorial videos, and now the SLB.

Why the SLB?

The SuperLongBoard is a next-generation CNC control board we’ve been dreaming of making for years now and is getting closer to completion. We believe that at this current stage, integrating smarter, more reliable, and more capable CNC control electronics will make the biggest improvement to the CNC experience. If you look at options on the market there’s been a large divide between the capabilities and reliability of the typical hobby CNC boards in the $40-$200 price range and the far more expensive, semi-industrial options that can reach the several-hundred-dollar range to over $1000. With chips getting more powerful and lower cost, we thought that gap in ability could be reduced while still staying in the lower price range by making a board with those new technologies. We knew this was going to be a big challenge – there are reasons why it’s typically easier to up-sell older technologies than it is to develop new ones – but we felt strongly in our decision. We felt that creating a new, improved, and unique solution could bring new innovation to the market to benefit everyone by reducing barriers to entry, keeping the industry progressing, and making more reliable and capable CNCs at lower price-points.

Working alongside Andrew and his team at Expatria Technologies, and building off of Terje Io’s amazing grblHAL project, all these tools need to work hand-in-hand so we can set our sights on resolving two major areas that we think hobby CNCs could benefit from:

  1. Reliability: this is such a broad topic to solve, but it can be more-or less divided into a couple categories.
    • Board wiring and protections: designing a board with proper grounding and plenty of voltage protection and isolation takes a lot of time and experience but is needed to shelter the electronics from CNCing – an electrically ‘noisy’ environment from all the moving motors and static electricity buildup from cutting and vacuuming materials
    • USB protocol: this is implemented in firmware and grblHAL comes with far more checks with data communication mechanisms in place to guarantee delivery over USB and know when to resend lost information
    • Backup Communications: adding Ethernet on top of grblHAL’s already robust USB implementation isn’t necessary but having an alternate communication method to turn to is always a great option if you’re ever experiencing problems with the primary style – an onboard SD card also gives you that flexibility
    • Processor speed: a slower processor can’t buffer as much information when dealing with short and complex movements, meaning better chips can reduce the chance of errors popping up
  2. Room for growth & “Wow factor”
    • Motor drivers: the same stepper motor controlled by different motor drivers can see notably better performance with all else being equal – so why not upgrade to newer-age drivers to get better speed with less motor noise out of the same CNC
    • Smarter CNC: having the CNC better tell you when something has gone wrong, automatically fix a problem for you, or make one of your existing processes easier creates far less room for error
    • More control options: there are typically specific limitations on what a CNC can do, so why not expand it’s options to support all commonly used CNC accessories plus leave space for even more customizable outputs. This can include things like independent Macros buttons, independent Spindle and Laser control, Modbus over RS485 for more closed-loop VFD control, tool length sensor support, ring and rail lights, door sensor, CANBUS, CNC pendant options to control the machine more easily, adapting to plasma cutting, and more
    • Multi-axis control: benefits those who have started to dive into cutting rotary projects to create ornate, fully 3D projects on their CNC and want either a simpler transition between cutting styles or are interested in full 4-axis cutting
    • Onboard computer: if you’ve ever run into problems with a Windows update, USB port falling asleep, or having hardware too underpowered to run your CNC, these problems can all be solved with a dedicated, purpose-built, and built-in computing solution running something like gSender onboard

We think we can address most of these aspects with the SLB at the time it ships so it can play a part of the next big step in hobby CNC technology. It’s advanced electronics and software will bring not just new features and functionality to the LongMill, but at a price point that we believe will be affordable for hobbyists. With it being a Sienci project, it can also guarantee that the SLB will be backward compatible with ALL LONGMILL CNC MACHINES OF ALL GENERATIONS, which means that users can upgrade their machine’s capabilities by simply replacing the controller. It will also support all our current plug-in add-ons. We also plan, just like with gSender, to make our board well documented and able to be adapted to many other CNC form-factors when we start to work through rolling it out in production.

How it’s coming along

Very well! Since Andy’s last update, our SLB team has been working hard to prioritize checking the functionality of the physical circuitry on the V2 prototype board are we’re now feeling confident that the following circuity is good to go:

Newest version of the SLB (V2)
  • 24V 12.5A brick power supply input with power switch (no more power over computer USB)
  • External, detachable E-stop with light, controller feedback, plus 3 customizable buttons
  • More reliable implementation of serial communication over USB-C (no drivers needed) as well as Ethernet
  • 4 higher-end, independent motor drivers rated for 2.8A RMS (TMC2660C, higher efficiency, less noise and resonance)
  • RGB Machine status LED
  • ‘Contact’ style touch plate input
  • 4 independent limit switches with two connector styles and 5V or 24V output
  • Coolant digital output set via M8 and M9 commands (5V 40mA)
  • Independent rotary axis support with dedicated limit switch
  • Tool length sensor input for easier tool changes
  • Modbus over RS485 to control spindles and other accessories with two connector styles, terminal breakout and CAT3 cable with RJ11-6C/RJ25 and supports daisy-chaining
  • Very dust resilient enclosure with quick assembly
  • Many status LEDs for troubleshooting
  • Tons of opto-coupling for optimal protection against incorrect wiring and electrical noise
  • Less heavy boards with less fragile components should lead to lower failure rate
  • Extendable RGB LED ring light and rail strip output
  • 4 customizable outputs to external circuits, controllable via M62/64/63/65, M7/M9, M8/M9, and M3/M5 (2 SSR pin shorting and 2 relay/solenoid/motor driving)
  • SD card to store onboard information and possible g-code file overflow (min 512MB)
  • Door sensor for input to pause cutting
  • Special CANBUS for pendant communication
  • Extra breakout IO to access spare MCU pins and other alternate outputs
  • ADC input for future sensor input
  • Communication connections (e.g. Rx and Tx) available for future control via external controller / pendant

These could still change a little on our way to production but I’m very confident that this will give an accurate image of what the SLB will be capable of. The gray items on the list might not be ready at the time of launch but the great news is that we have verified all their circuitry and put them on the board so all that’d be missing is a firmware update we plan to put out after shipping boards to enable those features as well.

The main items absent from the list are twofold:

  1. Firstly, though we aimed for the V2 board to be our second and last prototype, we did find some areas that we wanted to tweak. These were: changing the Spindle 0-10V output back to 5V PWM to support the new spindle we’re working on, changing the Laser PWM and Flood output circuitry to deliver a signal that’s more in-line with the previous LongBoard, tweaking the RGB LED driver buffer, fixing some onboard status LEDs that were floating, and adding some more EMI improvements. Most of these changes are quite small, but we decided that getting a V3 prototype made will help us feel confident in finalizing the design for production. This is a hard decision as it’s likely going to add another 1 month to our board release timeline, but we really wanted to be sure to support one of the key features we wanted the board to have “Independant 5V PWM and EN Spindle and Laser control set via M3, M4, M5, and S g-code commands, plus SpinDir“. The V3 design should be sent out to get made in the next week and should now lock-down the full SLB design. While we wait for it to arrive we’ll be focusing fully on testing and firmware improvements.
  2. Secondly, for those who might not have been following Andy’s production updates, we’d originally envisioned the SLB as being a system of two different parts working together. The first being the board itself, containing all of the core CNC functionality controlling motors and handling g-code, and second being an optional onboard compute module that would act to replace a computer or laptop and instead be integrated. Users could connect a keyboard, mouse, and monitor to control all functions of the machine directly through the SLB. This was very exciting to us given the considerably low price of the compute module over a computer, around $40-80 dollars plus the cost of the monitor, keyboard, and mouse, as well as the extra speed, user experience, and reliability of an onboard system.

    In our extended tests with this idea in mind, we weren’t finding the success we’d hoped in creating a seamless user experience with this solution. Despite trying many Linux kernels, drivers, GPU acceleration, and bringing many more efficiencies to gSender, the Broadcom and Rockwell-based processors used on smaller compute modules were not powerful enough to accommodate the visualization of g-code directly onboard. This also meant they didn’t have extra headroom if in the future we wanted to implement other features such as having a camera monitoring system or other sensor inputs. With many months delay trying to chip away at a resolution we decided to split the development of SLB back into its two parts; prioritize improving the baseline machine performance first so CNCers don’t have to wait any longer for the SLB to make better machines, and strip out all the on-board connectors and switch to a higher-power off-board solution that we’ll implement at a later stage. This will mean anyone could still upgrade at a later date. Higher-power Fanless PCs will cost more, from our initial budget of around $80CAD/60USD for the compute module, to somewhere around $100-$200USD depending on the specs and configuration, but would ensure a smooth and seamless experience as well as provide headroom for future applications.

    To summarize, we decided to take the concept of the onboard computer and divide and conquer on it at a later date. It would’ve been really cool to have a fully integrated system but it kept pushing our board delivery back further and we felt it would benefit everyone if we just pushed ahead on the other features that will all still bring great benefit to the CNC experience. This means that users will still need to connect their computers to the board to control their machines when the first batch of SLBs release, but should still see drastic improvement.

To-date, SLB development costs have tallied to more than $300k, making it easily our most expensive project to-date if you don’t count gSender which is I’d guess is around $500k at this point. These costs are pretty standard and have mostly come from paying salaries to do in-depth research on board infrastructure options, decisions on features that CNCers would like, designing robust PCBs, lots of custom firmware development, and testing. The new board uses a large number of components, adding to the challenge and complexity in manufacturing, but this was expected and we’re working through it. We feel confident that we’ve done something useful here and we’re excited to see the interest in the SLB on launch and continue moving development forward.

grblHAL support in gSender

Another big effort we’re having to undertake to support the SLB has been with gSender. Though grblHAL has “grbl” in the name, it might as well be a completely new firmware with all the new work that our development team has had to put in so far to make a seamless CNC control experience and UI. Since grblHAL is such a new firmware on the hobby CNC scene, it’s still growing and not widely adapted, and we want to help change that. There’s so much documentation that has already been made but there’s also a lot more we have to figure out. We’ve been working on an all-new firmware flashing utility for the new board, and had to build in a lot of modularity so gSender can be compatible with old and new boards alike

Look at an example of all these new EEPROM settings!

You can see this progress being made live, with each new gSender Edge version we’ve been releasing since the start of this year.

Enclosure

Iteration on the SLB enclosure I think is coming along well, the design is now pretty much locked in 🙂

Back in April I put out a post on our Forum with the leading design we had at the time to get peoples feedback on what they thought (https://forum.sienci.com/t/official-call-for-feedback-on-new-board-in-development/8519) and the results were positive but still mixed. I wanted to take a new approach to the enclosure design to accommodate all the new plugs without them all being exposed, as well as improve cable management on the LongMill in particular. Though many people were on board, many still expressed concern with my idea behind mounting the board to the Y-axis rail due to their fear of dust, vibration, or accommodating enclosures.

After going through way more iterations internally, this is now our revised design which I feel much more confident will suit everyone’s needs:

#image_title

The main perks of this enclosure is it’s mounting versatility and wire management. With this new setup, anything that you’d expect to have more common access to faces toward the front and all other wires can be routed out the rear. The RGB status LED will be visible through the faceplate put the panel will also be easily removable for troubleshooting or servicing. All the components on the board face upwards so everything is accessible and the board has been tested to run very cool so even in hot environments and with the bit of venting we’ve added in we expect everything to operate as normal.

Rail mounting bracket

The rail mounting is also a separate piece that can be used or not. This gives versatility to mount the SLB on any surface (rail, vertical, or horizontal) with 2 or 4 screws. We’ve already begun production of the parts and should be getting them in Nov-Dec.

E-stop Design

This is also mostly completed, just waiting for some last few checks. You’ll notice on the earlier feature list that we wanted to move away from the 3 buttons attached directly to our board. This is because there was feedback that people didn’t tend to use the existing buttons because of their fixed function, and when people mounted their board out of the way they couldn’t access the buttons anyway. We wanted to fix this by making the buttons separate and customizable, and this lead to a new E-stop design with 3 other customizable buttons integrated into it! Now you’ll have the freedom to have 3 physical buttons perform a myriad of custom functions to suit your workflow right next to your E-stop.

New E-stop design

Beta Testing progress

As we’re turning the corner on finalizing the board design and implementing our initial set of firmware features, our main goal over the next month will be: test, test, test! We’re now in Phase 2 of Beta testing where our main goal is to make sure that all our new features play nicely with each other, since we already know they work well in isolation.

We currently have 3 boards with testers and 2 dedicated machines in our own shop running CNC jobs every chance we can get and have already found unique situations that cause problems which we can now work toward resolving. This also involves comparisons between the old and new board and real-life durability and performance testing. One example is Ian’s Onefinity where his former setup used our original Longboard and with the SLB he’s seen drastically reduced noise and nearly doubled movement speeds. This hasn’t been as drastic for our other testers but we’re also already noticing some of the stability and perks that the SLB has to offer. Once the batch of V3 boards arrive, the plan will be to redistribute them as well as contact a new wave of testers to complete one final round of feedback. Some people we reach out to will be those with persistent and existing EMI and disconnection problems to see if the SLB is able to resolve those issues. At that point we expect to feel confident in the performance and features we can promise with the board that we can prepare for release.

Here’s a video from Andy’s last post of us testing the new programmable macro buttons. These allow you to program a specific code or function to 3 unique buttons, rather than just start, pause, and stop, which is hardwired into the current control board. In this case, it’s being used to move the machine to a specific position to assist with getting the machine out of the way for changing tools and materials.

Pricing and Timeline

If you’re interested in our previous evaluation on price, you can look back to Andy’s original big post under the “Pricing” header (https://sienci.com/2023/04/10/first-look-at-the-superlongboard/).

The pricing is likely to be refined as we’ve made new decisions and near production, but I think no matter how you slice it you’ll be able to expect that the SuperLongBoard will be giving you every bang for your buck.

As far as timeline, many followers of this project would know that we’ve had to push our originally anticipated late-summer, early-fall launch. We’ll be putting out another update soon on the roll-out strategy we have in mind moving forward because we want to give people a guarantee to be in line for the SLB before Christmas, so stay tuned to hear more about that. You can get a glimpse of what some of the steps will be with the roll-out in Andy’s original post as well under “What’s Next?”.

As far as ongoing project schedule, much of it has been laid out in the sections above which I’ll try to summarize here:

  • Last tweaks will be made in the next week to start a batch of V3 SLB prototypes
  • Firmware, Beta testing, and gSender support will continue over the coming month while we wait for the V3s to arrive
  • SLB Enclosure and E-stop designs should also become finalized and be ready for production in the next month or so
  • Once V3s arrive, final checks can be made to feel confident about the circuit design and begin board production
  • Phase 3 of Beta testing will continue to refine Firmware and gSender support while all other production is underway

I’d say that this means the boards won’t be ready to ship in December / Christmas and are instead more likely to be able to ship around Feb-March. The first major steps we took until June 2023 involved much of the initial product churn that we expected to see, though with delays in board prototypes and unexpected difficulty interfacing with the drivers we lost about 2 additional months. Add to that the outcome of the compute module testing meant that we had to pull out a lot of components for a full redesign between V1 and V2 and now we have to address some of the missed items between V2 and V3. This is the reality of developing such a complex product and is why we now feel much more confident with the recent testing and steps that we’ve taken that the completion of the SLB for launch is just around the corner 😀

Thanks

Thanks for sticking along for the ride on this big update, I hope it answers most of your questions on where the SLB is currently at and gets you excited about what’s to come. I expect the SLB to be very unique in it’s abilities to exceed the capabilities of many hobby CNC boards, perfect even for the DIY hobby CNCer for all it’s additional IO, and still have many more plans for it in the future. Of course it’ll be an Open-source design so hopefully all this time and effort that our team has put in will be able to return to the community and benefit other efforts toward easier CNCs.

Leave any comments you have that I missed and I’ll try to answer them. As I mentioned I don’t tend to put out updates because I’m a very slow writer but I’ll ensure to keep Andy in the loop as progress continues so he can update y’all as things keep moving forward. Just remember to read the Production Updates!

Catch ya later

January 2022 Production Updates

Posted on by

Hi everyone. We are now back from the holidays and full force into getting orders shipped out. Here are some news and updates for this month.

Air Assists being assembled for the LaserBeam

LongMill MK2 is now open for pre-order

LongMill MK2s are now open for preorder. For complete info about the updated version of the LongMill, please see the update.

An example project sent in by Andy Mctaggart, one of our beta testers

Pending LongMill orders

If you placed an order for the original version of the LongMill and you’re in the queue here, we are working on packing and shipping your order. We have been delayed in shipping the remaining LongMill orders as the couplers which were estimated to arrive on December 30, 2021 have been delayed. Due to the delay, we have ordered a rush batch of couplers which are expected to arrive around Jan 12th, and shipping for LongMills will be temporarily halted until the couplers arrive.

We’ve sent customers who are expected to be affected by this delay emails over the holidays to let them know ahead of time. We are currently assessing estimated delivery times for the remaining orders and will provide customers with updates soon. We’ll be updating order statuses on our Forum and our new Order Status page.

Response times for customer service (emails, tickets, and phone calls)

We have been overwhelmed with emails, tickets, and phone calls from customers, especially as we get caught up with messages that have come in over the past week while our offices were closed. We are working hard to respond to all of our customers, however, it may take longer than usual to get back to everyone. Thank you for everyone’s patience.

COVID situation

COVID 19 cases in the Waterloo region have seen a large increase in recent weeks. We recently had one of our employees coming back from vacation test positive for COVID 19. Fortunately, they were not exposed to everyone else at the shop as they tested positive before they came back to work, but it feels that a potential outbreak in our workplace becomes more and more likely as the pandemic becomes more severe.

To help keep our employees safe, we are pushing more of our staff to work from home. We are continuing to work to help set more of our staff to work from home and reduce the number of people at our office.

We continue to ask folks that are picking up orders from our office to stay inside their vehicles for pick-ups instead of attempting to enter our building without permission.

Lead times for LongMills and LaserBeams

We continue to wait for parts to arrive for the LongMill, LongMill MK2, and LaserBeam kits and ship products out as materials come in and get processed. We’ve created a new system for people to see the status of their order and see where they are in line for shipping.

Doing customer service when lead times are long for your products is particularly challenging for a number of reasons, including:

  • Folks are antsy about when they are going to get their machine, especially since they spent hundreds if not thousands of dollars on something from a random company on the internet (us)
  • It’s hard to keep track of when the order was placed and when it’s supposed to show up, as well as remember the estimated delivery time that was promised
  • There’s a lot of anticipation and excitement built up, which quickly turns into frustration when there are delays
  • Folks want to get their things to line up with their personal lives, such as with holidays and vacation time

We made some mistakes in terms of communication and setting expectations with our customers on when their machines would get delivered. For example, we said that we could ship a certain number of machines per day, so customers were expecting us to be shipping that number every day. However, because the number we can ship is limited to other factors, such as parts availability and the number of machines that the couriers can take that day, we were not clear on the fact that the number that we said we could ship was not always the number that we actually shipped. Some of our customers were calculating the number of days they would be waiting to get their machine based on the best-case scenario.

More realistically, customers should be planning their delivery based on the number of weeks that we advertised on our order page at the time the order was placed. Typically, the lead time will be stated as a range of weeks, such as 4-6 weeks.

The second mistake was how I sent the email out about the delay on the couplers. As some folks may know, the shipment of couplers that were slated to arrive on Dec 30, 2021, but was delayed until the middle of January. I emailed customers that this may affect the delivery time of their order. This email was only sent out to a small handful of customers who would potentially see that their orders would ship outside of the initial lead time that was promised, however, customers who the delay would not be affected also were upset that delays were happening, even though it would not affect the delivery time of their order as it is scheduled to ship later.

Our team is working to find new ways to improve transparency and communication with customers, but in the meantime, we’ve created a simple auto-updating list for new orders.

Our goal with this system is:

  • Customers can see what the status of their order is at any time, reducing the need to contact us directly about the order
  • They can see what the lead time was and how much time had elapsed, providing clarity on what delivery time was promised and if we are meeting them

Creating this system was actually a bit eye-opening as well. Given how much pressure we were putting on ourselves to ship quickly and manage customers with high expectations, we were feeling that we were always missing shipping deadlines. Now that we can visually see how much time had passed, it appears that we’ve been shipping almost all of our LongMills within our estimated times.

gSender 1.0 is now here. Simple and powerful control of GRBL CNCs.

Posted on by

Hi everyone, Chris here.

It’s not often I post here, most often it’s Andy or Kelly keeping y’all up to date on the status of things happening here at Sienci Labs Headquarters; I’m much more behind-the-scenes. Today I’ve come out of the woodwork for a very exciting and momentous announcement for us, the release of gSender 1.0.

Since much of gSenders development hasn’t made much of an appearance on the company blog, let me bring you up to speed on where this project started and what the release to 1.0 now means:

Origin

Back in 2016 when Andy and I first started Sienci Labs around our Mill One CNC, the range of free and paid g-code senders specific to operating hobby CNCs was surprising wide. Options like UGS and Candle were quite widely used and more were in active development at the time such as UGSPlatform, CNCjs, bCNC, and more. We were able to make a wide range of recommendations to our CNC users as well as made documentation around specifically UGS at the time since we felt it was the most intuitive to use.

Jumping to 2020 we’ve now got more machines in the wild then ever before. If you’re familiar with working at scale, you know that a 5% failure rate, for example, is very manageable when you’re working in the hundreds but starts to be more of a headache in the thousands. This is what we were starting to encounter with the g-code senders available at the time. And I’ll clarify, when I say “failure rate” this isn’t me saying that the software didn’t work. There were outlying bugs, quirks, and stable versions that were more stable than others; but what also stuck out was that many of the available options were made for more technically savvy people.

What were our options?

With more and more calls coming in, we could tell our users were getting confused with existing sender options. We heard questions about basic functions, confusion about behaviour, and one of the hardest things for us to hear was that our users were blaming the functionality and abilities of our CNCs in cases where we knew it was coming down to the sender. I remember personally dealing with situations of broken bits because of probing operations, or even a time I remember vividly where I helped someone out for a whole hour over the phone because he thought his file was visualizing correctly and we found that he’d just accidentally closed the visualization window within the sender.

I want to make it clear that I have the utmost respect and thanks for all those who took the time to make their own g-code senders for hobby use. These contributors to the hobby CNC community as a whole have taken the time to take what would otherwise be a personal project and have released it for the world to use. Not only that, but they continue to support updates, bug fixes, new features, and much more – all on their own time and out of the passion that they themselves have for CNC. But ultimately, the origin of these projects is what made me realize around that time that the goals of these open-source projects were never going to meet the goals that Andy and I have always had for the company: to make CNC routing fully accessible to the public. As long as open-source developers were at the helm, customization and a diverse range of features would take priority over UI design, streamlined use, and catering towards beginners by using more understandable language.

This left us with two options. Either we help fund an existing project or put more man-power toward it and get it moving in the direction we felt would be best for the next generation of g-code sender, or we begin making our own. I sent some emails around, and spoke with some other developers I knew and the consensus was that we’d go with the latter. Seeing as these projects were the ‘babies’ of each respective group of open-source developers, we felt it wouldn’t be our place to ask for more from people who wanted to do development on their own schedule, or try to push our vision onto others. Thus gSender was begun.

Goals and Principals of gSender

Starting ideation in October 2020, we began testing existing senders and outlined our core goals and design principals that we felt would guide development:

Goals

  • Act as new default machine interface / sender for LongMill customers (this means including all the functions and features that users are already used to from there existing sending options)
  • Supercharge our customers’ experience with their LongMill by streamlining / simplifying machine interaction and making even more machine operations easily available (a.k.a. nice-to-haves)
  • Keep design open for use by other hobby CNCers to provide benefit outside of just our own community (this means leaving some wiggle room for broader customization and adaptability to other CNC machines on the market as well as even further advanced features)

Principals

  • If it’s set once it doesn’t need to be visible 100% of the time
  • Options most users don’t need should not be obvious or take up much screen real estate
  • Most actions should be 1 click
  • “Someone off the street should be able to figure out how to get started”
  • Order of widgets should be close to order of operations/steps to proceed
  • Machine state should be obvious at a glance
  • Use colour patterns to subconsciously communicate button functions
  • Substitute CNC jargon with simple language when possible

We knew the other primary holes we wanted to fill roughly included:

  • Easy interface for beginners yet advanced features available for more experienced hobbyists
  • Smart handling of error cases, probing, jogging, firmware editing, machine connection, etc.
  • Built-in tools for convenient CNC management: stock/wasteboard flattening, motor calibration, machine squaring, etc.
  • Extensive keymapping and built-in joystick compatibility
  • Support of tool changing, start/stop g-code, macros, coding variables

The idea was to branch off of the well-known and provably reliable open-source codebase so we could be confident that we were starting with a good foundation to build off of. We also wanted the software to be made available easily for all systems so that future implementations could support standalone machines. After a wide range of tests performed over many weeks, CNCjs’ codebase was chosen for the previous reasons and because testing showed reliable operation on a variety of computer systems.

Present Day Release

I’m very excited to announce that gSender 1.0 is now available for free to anyone who wants it!

After nearly a year in development, the first couple months being in Alpha and the remaining ones being in BETA, we’ve now become confident enough to declare our first reliable version. We’re really excited to see more people begin to use it as we’ve now got a whole swath of great features and about 2000 users using it to run jobs every day.

Some things that you can expect from gSender in its current state:

  • Reliability of operation
  • Accommodates a wide range of computing systems via its ‘Lightweight mode’
  • Easy to understand and use due to its interface layout
  • Takes care of a lot of complexities behind the scenes, making mistakes and errors less likely and easier to deal with when they do happen
  • Built-in tools for surfacing, machine calibration, firmware editing
  • Ability to ‘start from line’, tool change, run custom macros, quick-travel with limit switches, and much more

If you’d like to learn more, we’ve got a primary download page, a documentation area, our full Github page, a discussion forum, and even though we’ve reached 1.0 we’ll still be releasing new dev builds for the foreseeable future as we continue development on gSender:

Download here: https://sienci.com/gsender/
Documentation: https://resources.sienci.com/view/gsender/
Github: https://github.com/Sienci-Labs/gsender
Forum: https://forum.sienci.com/c/gsender/
Feedback: https://sienci.com/gsender-feedback/

The team and I are very excited that we’ve reached this point. We set out to make our very own open-source and FREE g-code sender after gaining inspiration from some of the other commonly available options and we feel we’ve made a strong start towards those goals. To get here we owe a huge amount of thanks to our original 10 Alpha testers (you know who you are 😉 ), all those who participated in BETA testing and sending us suggestions and feedback, and importantly our greatest thanks to Will Winder and Cheton Wu for driving the UGS and CNCjs projects that not only were there for our initial users but also acted as a indispensable step towards us embarking on this project ourselves.

We want gSender to be the next-level open-source sender that’s available for all green CNCers to turn to whether they own a LongMill or otherwise: guided and easy to approach as a beginner yet customizable and feature-rich for higher-tier hobby CNCers. This means that it not only works for our Mill One and LongMill CNCs, but has already been proven to work on other machines like X-Carve, Shapeoko, Bob’s CNC, 3018, PROVer, and many more!

If you’re interested, the full list of features up to this point is as follows 😀 :

  • GRBL controllers supported
  • Smart machine connection
  • 3-axis digital readout (DRO) with manual value entry
  • All-directional jogging with XY diagonals, jog presets, and incremental/continuous single-button handling
  • Zero-setting and gotos (independent and combined)
  • Probing in any direction plus safe continuity detection ensures no broken cutting tools
  • Full imperial/metric compatibility
  • Responsive screen design and workspace customizations including visualizer light and dark theme
  • 3D toolpath visualization (no machine connection required)
  • File insight on load (feed range, spindle range, tools used, estimated cutting time, and overall, max, and min dimensions)
  • Feed override and active job status indicators
  • Fully exposed keyboard shortcuts for external keyboard/keypad control
  • Joystick support built-in for a variety of controllers
  • Safe height movements – accommodates machines with or without endstops
  • Homing cycle and quick-movement locations available for machines with homing hardware
  • Full spindle/laser support via manual control widgets, active alerting, and live overrides
  • Full mist/flood coolant support via manual control widgets and active alerting
  • Macros buttons (rearrangeable) with enhanced macro variables and individually assignable keyboard shortcuts
  • Lightweight mode reduces processing intensity on less powerful hardware or when running larger files
  • Easy workspace swapping for more advanced jigging or alignment work
  • Optional automatic handling for common error throwing g-code
  • Firmware tool for easier GRBL EEPROM changes, loading defaults, and GRBL flashing
  • Surfacing tool auto-generates surfacing g-code based on machine cutting area and other preferences, ready to execute
  • Calibration tool for axis alignment – a step by step process to make sure your CNC is square
  • Movement tuning tool for calibrating motor axis movements
  • Tool change functionality – pause, ignore, or run code blocks on M6 commands
  • Start-from-line functionality to resume jobs part-way through in case of failure of abort
  • Outline functionality indicates the rough bounds of the job before cutting
  • Customizable g-code injection at job start & end
  • Tooltips for data entry points
  • Alarm warning explanations to better contextualize CNC errors
  • Sleep management to keep PC awake during g-code sending
  • Pre-built machine profiles, including: LongMill
    • Shapeoko
    • X-carve
    • OpenBuilds CNCs
    • 3018 CNC & PROVer
    • BobsCNC CNCs
    • CNC4Newbie CNCs
    • Mill Right CNCs
    • Ooznest WorkBee
    • Nomad
    • Carvey
    • Mill One, and more…

What’s the plan moving forward?

Even though we’ll be taking a quick breather after release of 1.0, we’ll still be working toward even more features into the coming future. Some of these will include:

  • Headless Pi operation
  • Surface probing and other ‘fancier’ probing types
  • Built-in g-code editing
  • 3D cutting visualization

We left these off to the side ‘till now since we know that many of them would be more appealing to the more advanced users, but we’re confident that with everything currently built-in it’ll satisfy the vast majority of y’all.

Whatever comes next, we’ll be developing in two distinct branches. One that is for stable releases only, and another that you can optionally follow if you’re okay with a little less reliability but are wanting to be a part of shaping the new features that come to gSender.

Closing

It’s been great to spill this story out to y’all, there’s been a lot going on behind the scenes here at Sienci Labs this past year and I’m excited for you all to not only get more out of your machines via gSender but also via some more upcoming projects that we’ll be announcing in the coming months. Cheers everyone!

September/October Production Updates

Posted on by

Hi everyone. Here’s a production update for September/October 2021.

We’ve seen a small uptick in LongMill sales over the last couple of months. Overall, shipping and production have been moving along smoothly, and most of the LongMill orders have been shipping out within a few days. We’ve reached the middle of Batch 5 a little sooner than we expected, and we may start to see us run out of a few parts as we wait on parts to complete the second half of Batch 5.

Just as a side note, we build machines in batches, and Batch 5 consists of 1000 LongMills. With smaller or inexpensive parts, parts with long lead times, and parts that we need to make large quantities of to take advantage of the economies of scale, we order 1000 sets of parts. On the other hand, larger or more expensive parts, parts that have short lead times, and parts that are made locally, we produce in smaller batches. In this case, we produced 500 sets of gantries, control box parts, and rails.

It’s important for us as a business to try to time the production and purchasing for each of our components. Over-producing means having to pay for inventory that needs to sit on shelves, while under-producing means having longer wait times for completed products. We of course try to balance these things and make production as seamless as possible.

Due to the uptick in sales, we’re running out of parts sooner than we expected, which may lead to shortages of parts. Currently, we are expecting to run out of gantries in the next few days. The laser cutting for the gantries is expected to be completed on Monday, with painting and finishing of the gantries to take about 2 weeks to complete. We will be working with our manufacturers to coat the parts that we are lowest first so that we can keep production going, but regardless, you should expect the lead time to increase until the rest of the gantries are completed.

Other parts that are still in production, but are expected to arrive before we run out include:

  • Rails
  • Control boards
  • Control box parts
  • Drivers
  • Flange bearings

Expected lead times

With the gantries expected to be our bottleneck, we expect machines, once they are out of stock, to start shipping LongMills again around the third week of October.

If you are planning to order a LongMill, please check the estimated shipping date on the LongMill order page for the most accurate information.

Surging ocean freight prices and continued instability in the shipping industry

Increased consumer spending and pandemic woes have continued to cause instability in the shipping industry. Recent surges in ocean freight prices now mean that shipping products between China and North America have roughly quadrupled in price compared to the start of this year. The chart below represents a pricing index to compare shipping rates.

This of course impacts us. A reasonable percentage of parts that we use for the LongMill comes from China, so we are also at the mercy of ever-changing shipping prices.

Luckily, the LongMill is a non-commodity product with a reasonably large profit margin, allowing us to adjust to changes in fluctuation prices more easily than commodity items that have low margins, such as general household items. We have kept the price of the LongMill the same for the past two years, but inflation and current events may eventually warrant increasing our prices as well.

In the longer term, I hope that shipping prices do stabilize. Realistically though, I expect that shipping prices will stay high in the near future, and other unexpected factors may come up, so we will have to keep our eyes peeled for that.

Batch 6 is now currently in production as well, with this batch being 1500 LongMills. By increasing our batch sizes, I hope to smooth out the supply chain and give ourselves more time and flexibility between batches.

Vectric VCarve Pro now available on our store

Posted on by

Hey everyone, Andy here. I’m here to announce that Vectric VCarve Pro is now available for sale on our store.

We first offered Vectric VCarve Desktop for purchase through our store after many LongMill users praised its ease of use, a wide array of resources, and powerful features that made it worth the price. Over the past year it has become one of my go-to programs for creating gcode, as it has been proven to be easy to use and come with many presets and features that I have found useful.

One very significant downside to using VCarve Desktop is its size limitation. The software limits you to projects to up to 25″x25″. This of course leaves several inches of wasted potential for LongMill users that have a 30″x30″ work area or requires users to split or tile projects. Here’s an example of one of my personal projects making a trim piece for my old Volvo out of plywood that required splitting the project into several pieces.

Vectric Pro does not have a size limitation. This, of course, comes at a price. While Desktop is priced at around $460CAD, Pro is priced at $896, a +$400 difference for the convenience of having a little extra working area and a couple of other extra features.

I still firmly believe that all new users should try out free software, as discussed in my previous article. But for advanced users who are making their bread and butter with their machines, it is easier to justify paying for software if it can save time and improve productivity. Initially, when the LongMill first was released, all of our users were new to CNC, which meant that we were focused mostly on helping beginners by focusing on supporting and recommending free and low cost software options. Now, with many users having had used the LongMill for a long time, we now have a growing community of advanced users who demand more functionality from their machines, some of which already have VCarve Pro and other paid software. So to cater to this growing community, as well as prepare for future machines such as the AltMill, which will focus on production and larger scale CNCing, VCarve Pro is now available for purchase on our website.

Just as a side note, we’ve updated the license distribution process so that all licenses get sent out automatically. This means that if you place an order for any Vectric software on our store, you will get an email with the license details right away.

Going Bigger – Announcing the Development of the AltMill and Extended Versions of the LongMill

Posted on by

One of the most frequently asked questions at Sienci Labs is “Can you build a bigger version of the LongMill?”. Well, I just want to assure everyone we have been actively working on the development of 1) an extended version of the LongMill and 2) the Altmill! Although we’ve been working on these projects for a couple of months now, these projects are still quite early in their development and we don’t have a ton of details to share. The purpose of this announcement is to start getting our community involved by learning what sort of machines and features folks are looking for. If you’re interested in being involved in this process, please make sure to fill out the survey.

What is the extended version of the LongMill?

Well, I guess it’s in the name. We’re working on a version of the LongMill that uses the core components of the original machine, but extends the rails and leadscrews to add more working area to the machine. The goal is to have a LongMill 30×48 or LongMill 48×48 machine. We expect to be working in a price point of around $2000-$2500 for a full extended version of the LongMill, with kits available for adapting pre-existing models of the LongMill to the larger size as well around $800 to $1000. Please note that pricing is an estimate at this point and may change.

What is the AltMill?

The AltMill is a new machine that we’ve been working on that focuses more on the more industrial/production end of the spectrum of hobby CNCing. This means linear rails and ball screws, more powerful motors, and other features that are designed for more intense CNCing. We expect to be working in a price point of around $3000-4000 for a 48×48 inch working area.

Timeline

Extended version of the LongMill

We are currently in the early stages of manufacturing prototypes of the rails for the extended version of the LongMill. We will be conducting testing between October and November, to push for a December or early 2022 launch.

AltMill

We are currently building to scale prototypes with wood, with plans to start producing prototypes from aluminum in the coming months. Due to the scale and complexity of the project, we expect to have working versions of the AltMill at the start of 2022 with a launch for the AltMill in mid-2022.

Beta testing

If you wish to be part of our beta testing program, please fill out the survey. You will be able to provide your information at the end of the survey.

General challenges of the project

Building larger machines also prevents new challenges. Here are some things that we’re working on addressing. We also discuss this topic specifically about the LongMill here: https://sienci.com/2020/06/05/things-to-consider-when-making-a-longer-longmill/

Rigidity

Longer rails have more flex, which means that we need stronger rails to compensate. For the LongMill, we are currently working on a new rail design that improves rail rigidity while keeping overall weight down. This should help keep similar levels of overall rigidity in the machine and allow users to run their machines with the same speeds and feeds as any smaller LongMill.

The AltMill on the other hand will use stronger linear rails and bearings, as well as a solid aluminum structure to ensure a high degree of rigidity.

Squaring

Our machines rely on both Y rails to be parallel with each other and perpendicular to the X rail to ensure that the machine doesn’t rack or cut out of square. With the LongMill, we can generally rely on our table mounting procedure to ensure that the machine is square, but the larger and heavier the machines become, the harder it becomes to square the machine properly.

Squaring and calibration tool for the LongMill

To account for this, there are a couple of options:

  • A table which uses precision cut parts to help square the machine
  • A tool or measuring device included in with the machine
  • Making the machine smaller

A table, in my opinion, offers the most precise way of keeping the machine square, which is why we are developing additional structures to support the machine that can use similar or same designs between both the larger LongMill or AltMill.

Beyond this, gSender also offers a calibration tool that will play a more integral role in working with larger machines.

Mounting

Although a LongMill 30×30 fits perfectly on a 4×4 ft sheet of MDF, it is generally difficult to find sheets larger than this for mounting larger machines. One option is to cut and join multiple smaller sheets into a 5ft or 6ft square base to mount a machine on or have a pre-built bench or table that the machine mounts to, with space in the middle to put a larger wasteboard.

In terms of a 30×48 LongMill size, customers could purchase a 4×8 ft sheet and cut it down to 4×6 ft size to mount the machine to. However, 48×48 machines and the AltMill would need to use an alternative method.

In this case, having a table would also offer a good solution to this issue.

Power

The larger a machine gets, the more power it needs. This means larger motors and drivers. This is because:

  • The parts that make up the machine that need to move are larger and heavier
  • We want to cut faster so that larger projects don’t take forever

I generally use the rule of thumb that no matter how large or small the machine is, you want the machine to be able to travel between the lowest left corner to the highest right corner in the same amount of time. So this means that the machine needs to travel faster the larger it is.

We are currently working on either using larger motors as well as optimizing the power from the stock LongMill NEMA 23 motors.

Spindle and router choices

The bigger and more powerful the machine becomes, the router or spindle power becomes a limiting factor. Although I believe that the Makita router we recommend for our LongMills should be able to handle anything for the extended versions of the LongMill, a spindle may be necessary on an AltMill. Here are some hurdles to get over with spindles:

  • They are larger and weigh more, thus needing more hardware to support on a machine
  • Have higher power requirements, which means that users will also need to make sure their workplace can support it
  • Require additional wiring, which adds additional complexity
  • Generally not available in retail, which means that we have to source a spindle manufacturer and ensure we do proper QA and testing

Seeing as spindles could be used interchangeably between the AltMill and LongMill, this opens up the opportunity to offer spindles for both machines as well.

Shipping

Larger machines are larger and heavier, making it harder and more challenging to ship. The current shipping weight of the LongMill 30×30 is around 60lbs. Although fairly manageable, any heavier and larger than this, I feel would be unwieldy for the average user. Not only that, larger, heavier packages are more prone to being damaged during shipping, which is something we definitely want to avoid.

I expect our larger machines to be way bigger and heavier than this, and I estimate that weights will start to exceed 100lbs. This means either shipping the machine in several separate boxes, as well as figuring out the best way to handle the tracking and logistics on this.

Our goal is to continue to make it possible for us to ship by courier (UPS/Canada Post) to ensure that customers don’t need to worry about freighting. This should make our machines more accessible for the general public.

Pricing

Because of all of the factors discussed above, larger machines cost more. As with the current LongMill, our primary goal is to provide the best possible value by lowering manufacturing costs with the most optimal designs. We do make certain decisions, some good and some tradeoffs to achieve competitive prices. Here are some examples:

  • Using standard extrusions for building tables – easy to source and build with
  • Reducing the number of variations of the machine to take advantage of economies of scale (no custom sizes) – reduces the amount of different types of support and resources we need to create as well as reduces machine complexity by not needing to design high customizability, but means customers have less choice in the size of their machine.
  • Assembly required by the customer – better understanding on how the machine works and saves costs on in house assembly labour, but would take longer for customers to get up and running

On the other hand, there are some changes that will add costs that we feel are worthwhile to spend money on:

  • Partial assembly of the AltMill to ensure proper assembly of linear motion components – we are able to create jigs and tooling to make in house assembly and calibration easier than most customers
  • Tables and other mounting options for larger wasteboards and machines – ensure proper squaring and make it easier for the user to set up their machines
  • Larger lead screws and ball screws – Although more expensive, larger screw drives are needed to prevent whip which are more apparent in larger machines

Inductive Limit Switches – Production Update

Posted on by

Hey everyone here’s an update on the development of the inductive limit switches for the LongMill! If you haven’t read the last post, you can read it here: https://sienci.com/2021/07/30/longmill-limit-switches-coming-soon/

I know a lot of people are excited about this kit, and I assure everyone we’re working really hard on this. Over the last couple of weeks, we’ve been working on a couple of different things, including video and written info and content, continual testing, assembly instructions, packaging, and the supply chain for the kit. We are now waiting on our first batch of sensors and a couple of other parts to arrive from our manufacturers, and we will be starting packing and assembling the kits as soon as parts start to trickle in. All of the parts for the kit have been ordered and are expected to arrive in mid-September. We expect to start shipping kits a couple days after we’ve received all of the parts. Kits will be $60CAD or around $48USD each.

Our initial timeline for this project was to have a product released at the end of August. However, we had a minor setback due to some changes in part price and availability from one of the sensor suppliers that we initially ordered and tested samples from, so we have acquired samples from two additional suppliers, one of which we’ve fully tested and have decided to move forward with to use for production.

Some a bit more specifics that we’re working on to provide users include information about using different workspace coordinates, returning to a certain part of your job after a power outage or shutoff, and using jigs, which should add some extra tricks and functionality users can add to their machines.

Making your own

As promised in the last post, here are some instructions on making and assembling your own mounts and sensors if you prefer making your own over buying the kit from us.

Please note that these instructions are still in development, and additional resources and videos will be available for users soon. These instructions should help the general user population if they wish to make their own mounts and sensors.

Choosing sensors

The sensors we recommend using are:

Model: LJ12A3-4-Z/BX

NPN Detection

Detection distance: 2mm-4mm

Normally open

Supply Voltage: 5V* 

Choosing the correct voltage option is very important, as this particular type of sensor is more commonly available in a working voltage of 6-36V, which requires additional wiring to make work with the LongBoard. For using higher voltage sensors, you may need to use either the 12V auxiliary power from the board, 24V from the power supply, or from an external power source. That being said, I highly recommend sourcing the 5V variant of the sensor as this will make installation much more simple.

There are many variations of the LJ12A3-4-Z/BX, as well as other M12 sized barrel sensors that come in different lengths. In my experience, most seem to be more than accurate enough for this application, with a repeatability of 1 thou or better.

Most sensors also come with a set of nuts and washers, which can be used for mounting.

Making the mounts

All mounts can be 3D printed. The models can be found on our public Onshape document for the LongMill. The models can be found under Electronics -> Limit/Homing Switches. Right click on the model to export as an STL or your preferred 3D model file format.

These parts can be printed with most FDM printers. If you’re interested in reading about our 3D printing process, please check out this post. I would recommend using a higher infill for these parts since a more rigid part generally helps mounting.

All of the mounts use a pair of M3 heatset inserts. CNC Kitchen has a couple of videos on using threaded inserts on 3D printing that are awesome which talk about them in general as well as how to install them.

For our application, we found a fairly inexpensive and commonly available insert that works great. A drawing of the insert can be found below.

Here is an exploded view of the inserts.

Assembling the mounts

Here is a view of everything assembled before mounting to the machine.

And here is the exploded view:

Attaching to the machine

The mounts slide onto different areas of the machine as shown in the images below. Use the M3 screws to secure them. You will need to position the sensor to a position that lines up the tip (usually blue or orange) with the gantry you are sensing for. Loosen and adjust the mounts as necessary.

Wiring

Although the sensors for our kit will come with pre-wired JST connectors with a 2.5m wire for running through the drag chain, it’s likely that off-the-shelf options will not. You will likely need to extend the wires to be able to run the wires through the drag chains.

The LongBoard comes with ports to connect limit switches via JST4 connectors or with the detachable screw terminal block. Here is a diagram of wiring the inductive sensor using the screw terminal. Note that the 5V and ground lines are shared between all of the sensors, and each black signal wire is connected to their separate axis.

More info on wiring can be found on our resources for limit switches.

Firmware settings

Once your sensors are installed you may need to update your firmware settings to enable the limit and homing functionality. A full outline of all of the related firmware settings can also be found in our resources.

Conclusion

I hope that this information helps some of our more ambition users who don’t want to wait to get a kit from us set up limit switches on their machine. I also hope that this will give you guys a head start in exploring all of the functionality in adding limit switches to your design. Over the next couple weeks, our team will continue developing the resources for the installation of the switches, so I highly recommend staying in tune on our social media and our blog, and check back on our resources page to check for updated resources!

LongMill limit switches coming soon

Posted on by

Hey everyone. One highly requested add-on for the LongMill has been limit switches. For the uninitiated, limit switches are often used on CNC machines for 1) homing the machine 2) preventing the machine from reaching the limits of its travel. If you’re interested in reading more about what limit switches are and what they can do, I recommend reading the article in the Resources.

Please note that in this post, we are using the term “limit switches” and “homing switches” interchangeably. I do understand that there is a small distinction for both, but for this application, they are basically the same.

At the beginning of LongMill development, limit switches were not a priority as a feature when focusing on beginner hobby CNCers. This primarily came down to a few factors. First was the added complexity of having limit switches, which means additional setup and assembly for the user, as well as adding to the learning curve of learning how to use limit switches. Secondly, with the LongMill set up so that crashing the machine will not damage itself, limit switches are not necessary to protect itself. For customers still adamant about having limit switches, we still provided full hardware support to plugin or wire in switches directly into the controller, which would take care of a small population of more advanced users.

For those who want to read more about community made limit switch solutions, this is a great thread to read: https://forum.sienci.com/t/homing-limit-switches/99/41

We still hold our opinion that beginner users do not need limit switches with their machine to get the full functionality of the machine, and we recommend starting out without them until a better understanding of the machine and its use is achieved. However, as our community has grown and along with that their experience, more and more users are now exploring new ways to bring advanced features to their machines. Not only that, the development of our very own gSender now allows us to integrate software and hardware more closely than ever before. With these things in mind, we’ve spent some time creating our own plug-and-play solution for the LongMill.

Creating a limit switch solution specific to the LongMill came with several challenges.

First was the lack of foresight on providing mounting points for limit switches. This simply came down to the fact that we did not integrate mounting points on the LongMill for adding limit switches. Later versions of the LongMill did come with holes and other features that could mount sensors, however, with so many different versions of the LongMill, it would be difficult to document and provide resources for installing limit switches for every single version of our machine.

Second was the voltage support of the sensors we need to use for the limit switches. We are using a variant of the LJ12A3-4-Z sensor as our limit switches, a very common and widely used sensor. However, almost all variants of this sensor are designed for a 6-36V input voltage. Although it is possible to pull 12V power from the LongBoard, the JST 4 pin connectors already integrated into the board which was designed to be used for a plug and play solution were designed for 5V only. In hindsight, it may have been a better idea to route the 12V power to the JST connectors, but this meant that we would need to purchase 5V compatible sensors, which do exist but are more difficult to source, to be compatible with the LongBoard. Our first supplier for the sensors created the proper wiring and plug set up for the LongBoard, but unfortunately, they were only able to provide 6-36V sensors which meant that we had to start looking for a new supplier.

The new design overcomes these two challenges. First of all, the mounting hardware for the limit switches will allow users to install their sensors to any version of the LongMill, as well as allowing the flexibility to choose which side of their axis they want to mount to. For example, some users may want to home from the bottom left corner of their machine and some may want to home from the upper left corner of their machine. Users only need to move their sensor from the front of the machine and remount it to the back and specify the change in the software to make the change. Second, we have re-sourced and tested a 5V variant of the LJ12A3-4-Z sensor, which will provide proper voltage compatibility with the LongBoard. This supplier will also be providing us with the proper wiring for a plug-and-play installation of the limit switches.

We expect the kit to be ready for sale and shipping around the end of August. Each kit will come with three sensors with a plug and play wiring harness which should have an installation time of around 15-20 minutes. The price for each kit will be around $60CAD or $48USD. Additional resources and software setup support will also be provided with the kit. We’ll also be publicly releasing the designs and specs for the kit for users that want to make their own setups. Please check our blog, email, and social media for further announcements.

Today’s testing of the sensors have shown repeatably of over 1 thou which should offer a very precise way to home the LongMill.

I’m excited to see the limit switch kit in the hands of LongMill users soon and look forward to seeing the rest of the development team and the community come up with ways to utilize homing on the LongMill!

July and August 2021 Production Update

Posted on by

Hi everyone, this is a production update for July and August. I’m happy to announce that we are expecting the lead times for the next 400 machines to be 1 to 2 weeks (update, as of August 10, lead times for LongMills is under 1 week). For the most part, we expect most machines to ship within a few days. Any orders that are still pending shipment at the time of writing should be shipped by Tuesday of next week.

All of the parts that we were waiting on for this part of the batch have now arrived and we should not be seeing any major delays in shipping for roughly the next 8 weeks.

During part of July, we ran into some part shortages, such as with our control boards, Delrin anti-backlash nuts, and drivers. We also had a short period of time where our print farm was also shut down due to a shortage of filament. These shortages were caused primarily due to shipping delays. We’ve finally have gotten all of our parts without much issue, and can continue production.

At this point, we will be able to produce another 400 units without running into supply shortages. There are still a number of components we expect to start running out of once we ship the 400 units, including:

  • Rails
  • Gantries and other steel components
  • Control boards

We have placed orders for these items at the time or writing or are in the process of ordering these parts.

In any case, that’s our update for this month (and August). Happy making!

June Production Updates

Posted on by

Hey everyone, this is our June production update. For previous production updates and other company news, please check our blog.

It’s continued to be a busy month for April, but as we talked a little bit in our previous update for April/May, we have continued to shorten our lead times. We’ve taken the time to train some of our staff on new responsibilities and reorganize and plan production for the coming months. We are near the end of our run of Batch 4 machines and are starting to prepare for shipping Batch 5 in June.

In terms of COVID, Ontario has slowly seen a decrease in cases and more of our staff are becoming vaccinated. We are fortunate to have had no cases so far, and hopefully none until the end of the pandemic.

A new batch of motors, lead screws, and drag chains

Lead Times

Lead times are expected to average around 1-2 weeks for this month, however we are starting to face shortages in parts that will rely on the timely arrival to keep up with production. Some of these parts include:

  • E-stops
  • Touch plates
  • Arduinos
  • Delrin V-wheels

We expect these parts to arrive in 1-2 weeks. However, this may change if we face delays in transit. We will keep lead times updated on the product page to account for these changes.

Supply Chain

There have been some minor bumps along the way in terms of supply chain especially due to current worldwide events, but luckily with early planning and dedication from the manufacturers we work with, the supply chain process for Batch 5 has been relatively smooth.

One area that we’ve seen a large spike in prices have been with drivers and Arduinos. Due to the chip shortage, many of the components that go into the LongBoard controller have gone up in price. Most ICs that go into this production have doubled in price, and new products that we are working on that involve chips may be delayed due to the unavailability of chips. We have acquired parts for the next 500 controllers with approximately another 100 controllers in stock, but we may need to be cautious of continuing shortages for the rest of 2021.

On the topic of spiking prices, steel prices have gone up more than double since the start of the pandemic, affecting the price of gantries and other steel components that go into making the LongMill. On a lesser level, copper, tungsten carbide, and other raw materials have increased overall prices for many components as well such as E-stops and end mills. Cardboard shortages with our packaging manufacturer have also affected costs and lead times a few times over the last few months too.

Luckily due to improving processes and increasing batch sizes, we have been able to find other ways to save costs and so we don’t expect to have major changes in pricing for our products, however, it is a reality that we may need to face at some point that our company will have to account for changing material prices by increasing the prices of our products.

We have also been affected by the shipping fiascos that have been happening around the world as well. Although we weren’t directly involved in the Suez Canal crisis, we have experienced slowdown in some shipments due to this situation. At the time of writing, most of the parts that we need for Batch 5 production are in transit within Canada (by rail) or are in production with local manufacturers. A couple of parts that we are still waiting on that are in transit by sea include:

  • Router mounts
  • Couplers
  • Delrin nuts
  • 3D printer filament

The remaining components for Batch 5 are expected to arrive this month but won’t be complete for shipping until these parts arrive.

Manufacturing

There have been a few changes in manufacturing at Sienci Labs. Here are some of the things that have been going on.

One small change is the material that we have made the ACME nuts from, switching from stainless steel to brass. Brass has shown to be easier to work with in terms of manufacturing and forming threads. In previous manufacturing batches, a portion of nuts were rejected due to rough threading that made it difficult to thread onto the lead screws. The new brass nuts are of much better quality.

As part of the transition from steel shoulder brackets and drag chain mounts, Batch 5 kits will use M8-16mm bolts instead of M8-25mm bolts to mount these parts. There is no functional change, as the longer bolts are a carryover from when longer bolts were needed on the 3D printed parts.

Next, we are switching to e-coating our gantries from powder coating. We believe that e-coating is an excellent alternative to powder coating as it provides a cleaner, more consistent surface which is important for our XZ gantry assemblies. In some of our recent batches of powder-coated steel, we were running into issues where paint contamination and dripping would either produce cosmetic defects or affect the assembly of the parts because of the unevenness of the surface. E-coating does have a thinner surface, which theoretically means that it offers less scratch resistance on parts. However, based on samples that we have been provided of our parts after being e-coated, we have seen significantly better resistance to chipping and surface quality, without much difference in scratch resistance. This change should decrease manufacturing costs while improving overall quality.

In the last batch, we switched to using M3 screws with captive washers to help keep the screws from coming loose. For these screws, we have switched from stainless steel screws to Class 12.9 alloy, which is a much stronger screw that will prevent head stripping. Head stripping has been a minor inconvenience as removing stripped screws takes a while.

We’ve added three new CR30s (3D Print Mills), a belt based 3D printer. These machines will add additional 3D printing capacity with the benefit of being able to continually print repeated parts. We are currently in the stage of testing and tuning these machines, but we expect each printer to do the work of 4 standard 3D printers, increasing our print capacity by approximately 25%.

Design

Batch 5 comes with some very minor design changes to the LongMill.

First to mention is that motor shafts on the X and Y will be fully round. This is due to a very small number of customers reporting their motor shafts breaking off. The engineers at LDO Motors and us have confirmed that the full shafts will prevent this.

We are manufacturing a slightly modified 65mm router mount to eliminate the need to use M5-12mm screws. Because of the drill tap depth of the four mounting screws in the back of the router, shorter M5 screws were needed compared to the rest of the machine assembly. With the new router mounts, M5-25mm screws can be used on all parts of the router mount. We have also relocated the additional tapped holes that are used for mounting to the front of the mount for easier installation of accessories such as the LaserBeam.

New ACME Delrin nuts have been manufactured without the counterbore, which were an unnecessary feature for our application.

Shipping

Shipping to US and Canada have been reliable overall and seems to have returned to pre-COVID speeds.

We had some delays with Canada Post shipments going to the US via US Air Parcel, so as an alternative we would recommend using UPS.

We have had several issues with customs for shipments going to Mexico this past month that are new. If you have an order that needs to go to Mexico, please let us know in case we need to make other arrangements.