PS_ON M81 incorporated into the emergency stop functions?
-
As the heaters are the real hazard, they are easy to power separately, and relays are readily available for the sort of mains consumption the PSUs use I agree that seperate heaters PSU is best.
-
You can attempt to achieve passive (or intrinsic) safety, or use active safety. I choose passive safety as far as possible.
Passive safety means that if something goes wrong, nothing disastrous happens even if no action is taken. So in the case of a heater, it means that if the heater gets stuck on, the ultimate temperature rise doesn't cause anything to melt or catch fire, although there may be minor damage e.g. burned out thermistor. You can achieve passive safety by limiting the heater power. For example, use a 25W hot end heater instead of 40W. If you do need more power, use a copper heater block instead of aluminium (because it has a higher melting point) and a high temperature heater cartridge (so that it can take higher temperatures without shorting out). Likewise, choose the bed heater power so that if left on at full power, nothing bad happens. My 350W 300mm diameter silicone bed heater tops out at about 190C, which is below its temperature rating. I accidentally had a hot end heater get stuck on several years ago and the temperature topped out at about 300C.
Active safety means having something detect there is a problem and take some corrective action. As a general rule, the simpler the active safety system is, the more likely it is to work. So a thermal cutout attached to the bed and connected in series with the bed heater is a good example.
Engineers building critical systems usually try hard to avoid having any software in active safety systems.
Duet WiFi hardware designer and firmware engineer
Please do not ask me for Duet support via PM or email, use the forum
http://www.escher3d.com, https://miscsolutions.wordpress.com -
@dc42 I complete agree with you but still - it is always better to have protection in software too then not to have it.
-
@dragonn true
And in some cases there is no possible "passive safety", so active safety is the only option.
-
A good example of that would be a machine designed to process anything from PLAs up to high temp materials.
-
As a mechanical engineer (retired) I'm struggling to think of examples where passive safety isn't an option. For example, if the heater could get hot because it falls out of the heater block, design the block so that the heater fits into a vertical blind hole, then it can't fall out. Simple.
Or if for some other reason, a high power heater heater runs out of control, why is it not possible to design the machine such that in that eventuality, the heater of part being heated, isn't in contact with any combustible material? It could get hot and even melt but it doesn't have to cause a fire and burn the house down.
On cars, you can have "active" means of monitoring tyre pressure but that won't save your neck if you hit a lump of concrete on the M25 and get a blow out (been there, done that). But "passive" run flat tyres will (save your neck that is).
Maybe there are some examples where passive safety can't be implemented and "active" or software controlled safety is the only option, but I'm struggling to think of any.
I know I'm just an old guy, but I prefer the passive, mechanical approach wherever possible. Every time I read one of the multiple threads about bed levelling or mesh compensation, I feel exonerated in my approach that it's better just to build the printer with a flat level build surface and not have to bother with any of this bed probing malarkey. But I digress............
-
@deckingman said in PS_ON M81 incorporated into the emergency stop functions?:
Every time I read one of the multiple threads about bed levelling or mesh compensation, I feel exonerated in my approach that it's better just to build the printer with a flat level build surface and not have to bother with any of this bed probing malarkey. But I digress............
True, but I think sometimes people prefer to use 5-30$ on a sensor + some software to compensate, instead of rebuilding their whole printer for +100-200$.
-
This isn't about discouraging passive safety. Bog standard risk management process is as follows once risks have been established:
- Eliminate the hazard.
- Reduce the chance of hazard.
- Protect against the hazard causing issues. (eg guarding etc)
- Train users of a hazard.
- Accept the risk as is.
You're stopping at the first point. Fair enough. I'd rather the system reliably stopped itself as soon as it detects a serious problem. My main issue was the potential for heater faults while the system was idle when the user assumes is cool. A single short along the ground wire from either heater could do that, or perhaps less probably a mains power spike has damaged the heater control circuit. In this situation there is a conceivable chance user could burn themselves, or a fire is started if the hotend is resting on a print (eg unbraked belt z).
Auto ignition temperature of PLA 388C (https://shop.3dfilaprint.com/ekmps/shops/billyboyclyde/resources/Other/datasheet-pla-extrafill.compressed.pdf)
Melt point of PEEK ~350-400C (https://www.medicaldesignandoutsourcing.com/peek-hard-extrude/)
Therefore single point of failure could cause the ignition of the extruded material.
Looking at the standard V6 system the stock heater blocks begin to loose mechanical strength at 350C and the silicone socks have issues at 300C. Trying to size a heater to avoid the potential of hitting 300C while still achieving a responsive system and not going loopy if the part geometry defects cooling air onto the hot end is not easy.
-
The art is providing a reliable minimally intrusive risk management system where a single point of failure does not drop the machine into a dangerous state.
It's one thing designing a machine for yourself, but when supplying one to others you tend to go a few steps further. I wouldn't let other people use my machines at the moment (isolating relays due to arrive today!), and never leave them running overnight or in the building alone unattended.
Interesting to see Ultimaker have certificated their machines for unattended operation, and I feel that this is a key requirement in educational centres. I wonder how far they have gone? I noted the Up Printer (that I gutted the OEM control board from in favour of a Duet WIFI) had a mechanical relay to isolate power to the heaters.
Having directed the control system development on much more hazardous commercial Additive Manufacturing processes I can say that users often find ways to bypass safety features. After all if they are Duet based they can wipe the SD/Firmware and reinstall stock firmware with safeties disabled. But if these are in the manual with comments about not interfering with safety measures then due diligence can be shown, and the chance of direct prosecution significantly reduced.
-
...I also have a nagging doubt about how my home insurance companies will behave if I have a fire and can't prove it wasn't started by my home built printers.
If my safety measures are over the top I'm giving them less wiggle room.
Insurance companies are wriggly sods!
Running 3 P3Steel with Duet 2. Duet 3 on the shelf looking for a suitable machine. One first generation Duet in a Logo/Turtle style robot!
-
@martin1454 said in PS_ON M81 incorporated into the emergency stop functions?:
@deckingman said in PS_ON M81 incorporated into the emergency stop functions?:
Every time I read one of the multiple threads about bed levelling or mesh compensation, I feel exonerated in my approach that it's better just to build the printer with a flat level build surface and not have to bother with any of this bed probing malarkey. But I digress............
True, but I think sometimes people prefer to use 5-30$ on a sensor + some software to compensate, instead of rebuilding their whole printer for +100-200$.
Yes fair point. I forget that many people start with a cheap kit.
But my comment was more about people who build their own printers from the ground up. In those circumstances, it really isn't difficult or expensive to build a printer such that the bed will be inherently flat, level and will stay that way, without software intervention. Yet they still choose to use software to compensate for (to my mind) simple inaccuracies for which there is a cheap and easy mechanical solution. That always puzzles me. I guess it's a "generation" thing.........
-
@doctrucker said in PS_ON M81 incorporated into the emergency stop functions?:
...I also have a nagging doubt about how my home insurance companies will behave if I have a fire and can't prove it wasn't started by my home built printers.
If my safety measures are over the top I'm giving them less wiggle room.
Insurance companies are wriggly sods!
You are right to have those nagging doubts Wes. "Wriggly sods" isn't how I describe them - I use a much stronger term.
Some time ago I had a ridge tile blow off my roof in a storm. On it's way down the roof, it caught a few other tiles and took them out too. The whole lot ended up on my car which was parked below. The house insurance company said they would pay the cost of replacing the tiles but as that was likely to be less than my excess, then they'd pay bugger all. As for the car - "that's consequently damage which isn't covered - try you car insurance company". This I did. "Act of God" was their answer so tough s**t. Cost me hundreds - didn't get a bean back.
More recently. Flat roof developed a crack. Rain came in, took out the bedroom ceiling. Water ran down the back of our fitted wardrobes, ruined them and the stud wall behind. Took out the bedroom carpet and chipboard floor. Water continued on down, took out the dining room ceiling and found it's way into the consumer unit taking out all the electrics. Insurance company said the roof problem was "fair wear and tear" so they wouldn't pay. All the rest was "consequential damage" which wasn't covered, so they wouldn't pay for any of that either. This time it cost me thousands and the insurance company wouldn't pay a bean.
Moral of the story - never, ever leave the house with the printer running and always fit a smoke alarm above it, regardless of what safety features the printer might have (active or passive). Oh, and buy the cheapest house insurance you can (or have none at all) because the bar...ds won't pay out whatever the reason for claiming.
-
Like i mentioned in my case, I do have passive and active protection going on... There is a thermal fuse in line with the heater bed so if that fails its obviously cutting power... But having software controlled stuff that could save me a fuse and is free that would work 90% of the time. Why not?
So for people like me who have a mechanical relay as a first gate for the mains powered heat bed, having the option to shut off the power via any IO, (ps_on just happens to be the one used as thats what seems to be recommended at the time) would be hugely beneficial.
Simply put, you cant have enough safety.
Oh and I have that 700 w keenovo matt. So I think i read it tops out about 250c...
The only safety feature I haven't done yet is earth the frame of my printer.... Must do that...
