PINDA not fail-safe by design, but why?
Hi all,
first of all I love my Original Prusa MK3. Recently my PINDA got faulty because of a broken wire and the nozzle crashed into the bed. What a shame. 😥
Anyway, it can't be undone, but maybe prevented in future.
Now I wonder why the PINDA Probe isn't designed to be fail-safe. If the signal of the PINDA would be inverted this could not happen.
In my case the PINDA is still powered up and the LED is working as expected, this means the supply wire and the ground wire are still intact, but the signal wire is broken. If the PINDA would pull the signal LOW (instead of HIGH) when detecting the metal surface of the bed, the printer would also stop moving towards the bed when the signal wire breaks. Even before getting close to the bed. Pulling the signal wire LOW to be fail-safe is a usual design practice in electronics and automated stuff.
Maybe there is any reason with higher priority, selecting the PINDA model to pull the signal HIGH when detecting a metal surface. But i can't imagine any reason to do so. Please tell me if I am completely wrong with my assumption or not.
Happy discussion and thanks for participation!
Greetings,
Andi
Re: PINDA not fail-safe by design, but why?
I could be wrong...but maybe it has something todo with the way an inductive probe can fault.
I have a Prusa,therefore I research.
Re: PINDA not fail-safe by design, but why?
Thanks for your reply!
From my experience I can say, that you can buy these inductive sensors in two variants.
Assuming cable break as the common fault, any other way it can get faulty has lower priority in my opinion. Also these sensors are designed to be used in rough industrial environment (heat,EMV,...), wether the are inverted or not, so i can't imaging any other reason the type of fault inside the sensors electronics would affect the decision...
Re: PINDA not fail-safe by design, but why?
now that im thinking way too much about something I don't need to. (it is fun theorizing, I love fixing stuff)
I would bet the fast flipping of the poles is the issue.
I have a Prusa,therefore I research.
Re: PINDA not fail-safe by design, but why?
sorry, but I do not understand what you mean. It does not matter wether the signal is invertet on the output of the analysis circuit inside the sensor or not.
This only invertes a possible fault state as well 😉 but this has nothing to to with cable fail-safe (in german "Drahtbruchsicherheit").
I will not offend you, but this is not funny theory, this is typical practice.
As always, maybe somewone knows better than me
Re: PINDA not fail-safe by design, but why?
I think we are hitting a language barrier..
best of luck.
I have a Prusa,therefore I research.
Re: PINDA not fail-safe by design, but why?
Whether the PINDA is NO or NC, NPN or PNP, a failure will still result in a crash. It is lack of a change that creates the condition for a crash, not the polarity of the change.
About all one could do to prevent a bed crash is to have a test condition that is safe: a metal object at the top of the rails (the only physically safe location) the extruder can move to to test the PINDA is working as expected. But it'd be pretty daunting to have the printer move the full range of top right to bottom left every print. But it could be safer. Even then, you may need a mechanical arm that removes the "trigger" when the test is done so it doesn't interfere with the Z-Axis. After all, the PINDA has to be within 2 mm to sense metal.
Re: PINDA not fail-safe by design, but why?
I think we are hitting a language barrier..
best of luck.
Sorry toaf, maybe you are right, as stated i didn't want to offend you.
Whether the PINDA is NO or NC, NPN or PNP, a failure will still result in a crash. It is lack of a change that creates the condition for a crash, not the polarity of the change.
About all one could do to prevent a bed crash is to have a test condition that is safe: a metal object at the top of the rails (the only physically safe location) the extruder can move to to test the PINDA is working as expected. But it'd be pretty daunting to have the printer move the full range of top right to bottom left every print. But it could be safer. Even then, you may need a mechanical arm that removes the "trigger" when the test is done so it doesn't interfere with the Z-Axis. After all, the PINDA has to be within 2 mm to sense metal.
Thanks for your reply tim.m30
You are right, if the PINDA internal electronics itself fails, a crash could not be avoided.
BUT when the cable breaks (which is the most common fault according to the forum), the printer gets the same signal from the PINDA it would get when its detecting the metal surface. And then the Z-Axis stops and the printer is in a defined safe state.
Greetings,
Andi
Re: PINDA not fail-safe by design, but why?
So what happens when the printer starts up and was last sitting right next to the bed? Using your conditions, that would look like a fault.
Re: PINDA not fail-safe by design, but why?
The printer would move up, as this is always a safe movement. then the singnal would change (assuming everything is allright) and the printer would recognize this...
Re: PINDA not fail-safe by design, but why?
I'm not seeing data that suggests cable breaks are the most common cause of failure. Most cases of bed pounding on this site are incorrect PINDA height after kit build. And there are three wires; you can't plan for only the active wire to break, it just as likely the power or ground wires. You can get just about any output state depending on failure mode.
You are still left with the fact if the sensor doesn't change state on the way down you will get a crunch. No simple way to avoid it.
Re: PINDA not fail-safe by design, but why?
First of all, a incorrect PINDA height is not a electrical failure. If builders do not follow the instructions correctly, we cannot avoid bed pounding, sure.
A in/correct PINDA height can easily be identified as written down in the great manuals Prusa provides and also in several posts on the internet.
Assumption: Metal detectet Signal LOW, no metal detectet signal HIGH
OK then we discuss step by step for every single wire:
0. Valid for all cases: Printer is sitting right net to bed on startup:
a) Signal HIGH: Movement in both directions is safe
b) Signal LOW: moving upwards is safe, downwards is unsafe/forbidden
1. Signal wire is broken:
Signal always LOW: see 0.b)
2. PWR wire is broken:
Signal always LOW: see 0.b)
3. GND wire is broken:
Signal always LOW: see 0.b)
Last case is an electrical failure inside the electronics of the PINDA with all wires working correctly: Then we do not know anything and the printer might crash. But this never can be avoided. The other cases can easily be avoided as described above.
PS the board does not trigger on a signal change, (HIGH->LOW or LOW to HIGH) it only triggers on a falling edge (in my suggested mode).
(In the current configuration the board triggeres on a rising edge and this will never happen if any of the wires is broken)
Re: PINDA not fail-safe by design, but why?
3. GND wire is broken:
Signal always LOW: see 0.b)
Not true - if ground breaks signal floats high.
Re: PINDA not fail-safe by design, but why?
The only safe way to detect a broken cable is to put a micro controller in the PINDA probe and communicate with it. Otherwise you can always find a scenario where it will fail.
Often linked posts:
Going small with MMU2
Real Multi Material
My prints on Instagram
Re: PINDA not fail-safe by design, but why?
Thanks for your comment, thats why we're discussing.
Not true - if ground breaks signal floats high.
it depends, we can discuss this further... but here we need to know the datasheet of the sensor and the schematic of the Rambo board.
BUT in case of loosing ground connection also the temp sensor looses ground and can be used for detection, right?
Re: PINDA not fail-safe by design, but why?
sorry, but I do not understand what you mean. It does not matter wether the signal is invertet on the output of the analysis circuit inside the sensor or not.
Inverting the signal costs time. In case of distance over time mesaurment You're loosing precision while doing that.
Often linked posts:
Going small with MMU2
Real Multi Material
My prints on Instagram
Re: PINDA not fail-safe by design, but why?
Suggest you create a design and submit it to github for inclusion as a future upgrade. Or, if you aren't a designer, submit a feature request to github.
Re: PINDA not fail-safe by design, but why?
Inverting the signal costs time. In case of distance over time mesaurment You're loosing precision while doing that.
I dont get what you mean. here we talk about max 10ns. also it is just a dead time which is constant and can be considered.
a sensor with an already invertet output does not show this behavior.
Suggest you create a design and submit it to github for inclusion as a future upgrade. Or, if you aren't a designer, submit a feature request to github.
I think i get my new PINDA on monday, then i will do some tests and will see if the temp is based on PWR or GND. good suggestion
edit:
The only safe way to detect a broken cable is to put a micro controller in the PINDA probe and communicate with it. Otherwise you can always find a scenario where it will fail.
sorry, i missed that post somehow.
are you still insist in this statement after the further progress in this thread?
This uC still cann not detect a internal failure of the sensor?!