Filiment out shut down switch idea

[Jack Crow]

Hi all,
Wrote an extensive entry last night and when I went to check for answers today, could not find it.

Had a fun pm exchange with Kirk about last nights post. He is of the opinion that went 'poof'.

The idea was fairly simple.

In a nut shell was watching the roll of plastic trying to guess if it would run out before my print job would finish.

Then it dawned on me there is no protect device on the Simple Metal for this.

Was looking over other posts, and extracting cooked plastic out of a hot end is no joy at all.

Im the type who sets a print job and goes to bed, expecting it to be ready in the morning.
Life is too short for that kind of failure.

So why let it happen.

Was thinking of using a micro switch, the kind with a little roller on the end of a tab.
That roller will be in contact with the filament as it is routed between two small blocks to hold it straight.
So that when the filament is in place the switch is pressed.

When the roll ends and the tension comes off the switch, it changes state.
Forcing a shut down.

My Simple Metal came with the ATX PSU.
Seemed a good place to start.
The OEM placed some jumpers on the control connector.
Bet if I open one of those paths, the PSU shuts down, and the printer goes dark.

That will prevent cooking plastic in the hot end.

Made some sample parts for a bracket and switch holder, and a rookie error to go with it.

5-30-2017 001.JPG

5-30-2017 002.JPG

5-30-2017 004.JPG

Operational lesson learned.
Don't make complex posts to this list and design complex parts when exhausted.

Comments welcome.
Jack Crow

[RetireeJay]

Im the type who sets a print job and goes to bed, expecting it to be ready in the morning.
Life is too short for that kind of failure.

There are other kinds of failures, too. Like the 'bot vibrating itself off the table (been reported here) or a jam where the filament stops feeding and just sits in the hot end and cooks. Etcetera. There's always the possibility of fire (search in the forum for that), such as from connectors that are not properly seated and are pushed to the limit of their rating (it doesn't have to start from the Hot End). So bottom line, unless you have a smoke alarm in the immediate vicinity, and possibly other safety measures in place, Kirk and I are of the opinion that no one should let their printer run totally unattended for hours on end.

[Mooselake]

Search the forum for (iirc) "printrbot on fire", and the web for 3D printer fires. An automatic fire extinguisher is a good idea. There's been a number of fires caused by 3D printers, and some possible fatalities. The usual cause is the hotend thermistor fails/falls out, or the driver MOSFET fails on, and the hot end overheats and melts (as in molten aluminum) down. I've left it running and watched TV or done other things, but never left the house - hundred and twenty year old hand hewn tamarack logs burn like a torch. We've used some removed for remodeling for kindling, really easy to light. I'd be especially hesitant to leave it running while sleeping even with a smoke detector, don't trust technology to wake me up in an emergency.

And as I mentioned in the PM suddenly dropping power can cause firmware and other problems. Not every time, but there have been stories of bricked boards from power failures. Using a signal lead and the firmware to do an orderly shutdown (including moving the head to prevent melt marks) has been done elsewhere in the past.

Running out of filament isn't necessarily a disaster. I've measured the layer height (you could count the layers, but too much work...), edited the gcode, and picked up where the print left off. Pretty good success, although it can leave a cosmetic line in the print.

Kirk

[Jack Crow]

Guys
Valid points.
May need to re think some bad habits.

Thought that the 'fire' post was a one of a kind, fluke.
Was not aware that these things are so poorly protected.
Did notice SMT thermal switches on the board I replaced. Guess that was not enough.

From your post....

"Using a signal lead and the firmware to do an orderly shutdown (including moving the head to prevent melt marks) has been done elsewhere in the past."

Let's a assume that Im a knuckle dragging high voltage analog guy who works with computers simply because there is not choice.

Is this feature already in the board or will it require a custom firmware hack?

Jack Crow

[Mooselake]

There's unused inputs on the board that can be used on the expansion connector. You'll need to check to see if any of the code made into the official Marlin.

Kirk

[RetireeJay]

If you wired your switch to short out the thermistor lead from the hot end (or make it go open-circuit, if you prefer; either would work), that would make the board think there was a serious temperature control problem and it would shut down. It would just "freeze" in place, without politely withdrawing the printhead to a point removed from the model.

I see Kirk has already answered that it's possible. If you're willing to compile your firmware, the option for a filament runout sensor is already present in Marlin 1.1 Configuration.h; all you have to do is enable it. And figure out where to connect the sensor, which may be the more complicated issue. It's doable.

[RetireeJay]

Thought that the 'fire' post was a one of a kind, fluke.
Was not aware that these things are so poorly protected.

The Printrboard that caught fire (or came very close) was a Rev D, and there were several others where connectors were charred or melted. I must say that the Rev F board is considerably safer because the connectors for both incoming 12V power and the outgoing power to the Heated Bed have been significantly upgraded. Thanks to PxT for paying attention to the weak points in the original design. I don't think the Rev F board represents an unusual fire hazard in itself, but still, you are working with a machine that intentionally heats certain parts to rather high temperatures.

[Jack Crow]

Team
Looked up printer fires on the web, kinda grim what happens.
All valid points on how these things are made and used.

I lack the ability to do a custom hack of the firmware.
Code is simply black magic to me.
Will have to do this some other way.

I like RJ's idea about dinking with a thermistor sensor.
If it shuts down the system properly, so be it.
An aborted print is cheap at a price.
Should not clog the hot end with cooked plastic.
That is the main goal, not to spend even more money I don't have to feed this project.
If the firmware stays put, so much the better.

Based on your recommendations, I will drastically scale back my operations here to when Im home or can keep an eye on the printer while doing something else in the shop.
No gain in testing how the insurance on this apartment is going to work.
I do keep extinguishers in the apartment.
One dry chemical, and a CO2 bottle in the kitchen.

On a side note about coding...
Years ago I ran into a similar problem with a determined code writer and his inability to make sense of a transmitter.
Blew his mind that there were negative voltages. Analog electronics impressed the heck out of him.

To sum up what your saying about system reliability.

For the most part home printers are still in the project class and less in the product class.
Treat them accordingly and stupid stuff has less of a chance of breaking out.

Thanks for the continuing education.

Jack Crow

[Mooselake]

Don't negative voltages make the meter run backwards as the power company sucks power out of your appliances?

Gosh, I must be old. I took analog computer classes in college. Impressed how you first figured out the solution and worked backwards to the problem, or so it looked like

Kirk

[Mooselake]

Yep, that's hobby 3D printers

[PxT]

Thanks to PxT for paying attention to the weak points in the original design.

Would love to take credit but hardware isn't really my area of expertise. Laine is the brain behind the board design.

[RetireeJay]

Thanks to PxT for paying attention to the weak points in the original design.

Would love to take credit but hardware isn't really my area of expertise. Laine is the brain behind the board design.

Oops! I knew that a member of the forum did the design, but my memory slipped a cog. Thanks for giving us the right name!

I should have known you were the software guy; you're the one who taught me how to flash the firmware and then taught me how to compile it from source using PlatformIO. Which, by the way, still works even with Marlin 1.1. (Only trouble is, the PlatformIO package keeps updating itself, and sometimes "leaves behind" source code that compiled successfully with an earlier generation of the compiler.)
-J

[Mooselake]

the PlatformIO package keeps updating itself, and sometimes "leaves behind" source code that compiled successfully with an earlier generation of the compiler.)

If it ain't broke yet try again?

Kirk