Safety Question - grbl running even with power off? Got a nasty surprise

I had an exprience the other day that alarmed me that I wanted to ask about because I think it is a potential safety issue and it certainly caught me off guard.

I’ve suspect that the Arduino micro controller inside the long board is “on” even if the power switch power is “off” on the top of the Longboard, is this correct? So the power switch is only controlling whether the motor drivers are active? The fact that I can “connect” to it in UGS even when it is “off” suggests yes.

The safety issue I ran in to, I believe, is related to this if it is true. If not, there is another avenue for this that we also need to understand. I had opened UGS and was getting ready to start a project so I jogged the router from the back right corner to the front right corner by doing a 700mm Y movement. But the machine didn’t respond. I looked over and the Longboard wasn’t powered on. So of course it isn’t going to work, right? Well not exactly.

I exited UGS and relaunched it and then reached over and powered on the Longboard. Immediately the system started “screaming” like when you hit the end of an axis. But UGS wasn’t even running at this point. I thought WTH and cut the power. Then I realized that UGS had send the jog command to grbl on the Arduino, which did NOT know that the stepper drivers were not powered on and was actively sending signals to the drivers. Since it was a long transit time, and I turned the power to the Longboard on fairly quickly after sending the command to the Longboard, it appears that the Arduino was still firing the motor movement commands at the stepper drivers so as soon as they energized the mahine started responding to them. Except it didn’t seem to have good coorindates because I would have expected it to move correctly down the table. Even that is dangerous and unexpected as I had reasonable expectation that the controller was “OFF” and the jog I sent shouldn’t have buffered in a machine that is “off”.

Assuming my understanding is correct, I have a few thoughts:

  1. Is there a status bus on the stepper driver boards that you can monitor to know if they are energized and believe they are operating normally? That feedback should really be collected and monitored in the Arduino so dangerous situations can be detected and avoided. If there is no such support (i2c bus?) on the driver board, then at the least the main Longboard should have a power monitoring chip and detect whether voltage and/or load are present on the power lines going to the stepper drivers to provide at least some visibility. Once you have this, whether you modify grbl or bolt a little piece of proxy code on the front of it (so you don’t have to merge with grbl’s repo) you should reject commands from the gcode sender if the power states are unexpected or other alarms are detected.

  2. Are there any fuses in the Longboard? Given the high amperage involved and the potential for a jam or failure causing an overheat condition it seems like a good idea to have some user serviable fuses in the design.

  3. Given that the potential is present to cause a major problem if the polarity is inadvertently reversed on the power input, it is not only good practice but I think it may be regulation to have the polaity of the power connection labeled in decent point size, high contrast text. E.G. Input 24V/10A/60 HZ, and a diagram showing the two powers with + and - labeled, similar to what you find on any wall wart charger. If the cable comes out of the connector or has to be removed, the operator should have all the information they need right at the device to put it back properly, and NOT need to hunt around online.

In reality, I would really like to see a more standard IEC (the 3 prong one on a PC power supply) connection to the unit or some other standard connection. I understand from Andy that there were issues locating a barrel connection that was rated for the amperage needed, but the current solution is sub-optimal. I notice OpenBuilds appears to use the same connectors, so I realize this isn’t unusual, but I moving to a more robust and better insulated power connection to the board is advisable.

  1. I haven’t seen the updated Longboard (but am expecting a replacement this week) however I find the Play/Pause/Stop buttons (at least on the V1) way too small and hard to see. Admittedly, I have a b it of colour blindness that makes tone on tone harder for me, but all the more reason that the STOP button should be larger and RED as it is the only included physical emergency stop. I have only needed it twice, but the first time induced Panic as I fumbled to figure out which was which. I then marked it teporarily with some blue painters tape. For safety reasons I strongly suggest the STOP button be larger and a different color, possibly even illuminated

  2. The green industrial connectors used for the power and various other connections could shake looks. It’s likely not a problem with the motors, but that beefy 24V power supply has quite a heavy cable and if the brick is on the floor it has a continuous strain on the cable. Given that it is screwed down it may weaken the electrical connection over time as well. Some form of strain relief, even as simple as some zip ties/slots for zip ties and a basic loop in the cord at the unit would go a long way to reducing risky strain on the cable.

  3. Again, a small feature for safety but worthwhile is to use an illuminated POWER switch. They are only $4 from Princess Auto, like this one. https://www.princessauto.com/en/detail/20a-on-off-amber-illuminated-rocker-switch/A-p8803397e but they allow you to see from a distance from any angle that the machine is definitively (or at least the stepper drivers?) are on or off.

6a) I notice that the power switch appears to be installed “through hole” right in to the PCB. This is unusual and something I woudn’t recommend. The normal practice is to have the power switch connected with standard spade or other connections and have short wire leads. This way a $4 switch that fails (as it did in my small shop vac) can be sent out and replaced without any soldering tools or skills.

  1. While brainstorming on the topic of safety, it would be a good idea to include four yellow stickers with red Caution - Moving Parts - Do not touch or similar icons with the unit that can be placed on the Y axis aluminum extrustions at both ends. I have had fascinatied visitors watching a cut and without thinking they naturally rest their hands on the rails while watching or leaning in to get a better look - not realizing the lead screws are turning on the other side or the machine could bump in to them quickly without warning. In volume the stickers would be trivial in cost but go a long way to help people understand where the risk zone is.

I love the Longmill and I think it’s got tremendous potential but I do think a focused review of the safety items is a good idea. These machines have the potential to cause a lot of damage or harm quickly if something goes wrong and there are a number of small, affordable and easy to do changes that would improve things on this front considerably.

Stay safe,

-Jeff

3 Likes

@jwoody18 I’m happy to revisit this original feedback item you provided us several years ago since I feel that many of your original points of feedback have now been implemented with the SLB. This includes:

  1. The controller being on and the motors being on are now both in sync, and if the E-stop is pressed the motors are de-powered and the whole system is paused so there won’t be any more surprises
  2. Chip-based fuses are now installed on all important board IO so that there’s protection against over-voltage, over-current, improper polarity, etc. This is even more convenient than using physical fuses since these chip-based fuses can auto-reset and don’t need user-servicing
  3. The #2 point mostly already addresses this, but the text and rating for power input are now larger on the board to make hookup of the power supply more obvious
  4. We now ship an E-stop as standard, also we now have 3 other customizable buttons all as a part of one HMI unit that can be placed anywhere closer to where you operate your machine which makes seeing and using them a whole lot better
  5. Much of the cabling for the SLB now runs inside the unit so it has some better cable routing for cable relief
  6. Many more lights to indicate board is powered now, with the main light also showing machine status and an additional light on the separately placed HMI unit. Regarding 6a, I agree that it’d be nice to be able to ship a switch replacement rather than requiring soldering or other skills, though with our new configuration the switch no longer has to route hardly any power through it and it’s seemed to work very well in long-term testing - still something I’ll continue to keep in mind
  7. This is the only piece of feedback I feel we should still be looking at more closely and I’m glad you bring up - since safety is important and shipping sticker to indicate that I think would be good to do.

I hope to hear back from you on your thoughts of the improvements we made!

Chris

1 Like