Prusa temperature starts to drop mid print

I'm trying to debug a problem with my Prusa. It worked fine for months, with no evidence of problems. suddenly last week, it started having problems. Sometime during the print, usually 30 to 50% of the way through the temperature gradually drops to a level where the plastic wont melt anymore. Ive tried to print a part around 10 times now and the same problem occurs every time.
At first I thought it could be a power problem. I'm running it on a PC power supply. at first it had 3 yellow and 3 black wires connecting to a 12v male DC connector, and the printer has the female version of the connector, wired to 2 pairs of yellow and black wires to the Sanguino board. I heard you should use 4 pairs of wires to get adequate current.

So, I changed it that so that on both sides of the male/female connectors are 4 pairs of yellow and black wires to make sure there's no bottleneck. This didn't make much difference except it may have lasted a bit longer before the problem occurred. maybe the dc connector is the problem?

I'm not sure what could be causing the problem. One thing that Im not sure about is how the temperature of the hot end varies during print. I'm thinking maybe not enough current is getting to it. Ive uploaded a picture here. Is that graph normal? Is the temperature meant to vary like that, sometimes it varies way more than that.

I changed the wires hoping that would help but it hasnt. I read somewhere the wires used in power supplies cant carry enough current, that's why you have to wire up multiple wires together but maybe thats not enough?

Also, my motors are working fine, barely warm to the touch. The Polulos however are getting pretty hot (too hot to hold your finger on them for more than 2 seconds), I was thinking, could they be drawing too much current and "stealing" some from the hot end, or is that not possible?

I'm not sure if this is important, but the temperature I print with seems to be much higher than other people. I get best results at 240°C with PLA. with the machine off, but plugged in to usb, the temperature reading is 20 degrees, which seems relatively normal, but I don't have a thermometer to check.

Sorry if this is a rambly post, Im just giving all the information I can think of. ill summarize here:
Could the dc connector (got cheaply off ebay) between the pc power supply (500W) cause a problem?

Is the temperature of the hot end meant to vary like this?
[dl.dropbox.com]

Are the Polulo drivers taking too much current from the system causing the variations in temperature of the hot end and eventually stopping it from working completely?

Could a thermistor issue cause the temperature variation and in turn cause the hot end to give up and stop heating up?

I hope I gave enough information. Im not sure where to go from here. Thanks for your time

No, the temperature should not be falling over time. It may fluctuate but aside from pre print warm up and post print cool down, it should worst case oscillate about the target temperature if the PID isn't optimized, or IFF (I'm not even sure if this is an option for the hot-end, but) you've chosen to use bang bang for the hot-end.

Presuming you don't have losses in other parts of your **system, then it sorta looks like you're trying to print faster than the temperature control is set to compensate for losses from the heated material.

**(wire, for instance, has a positive coefficent of resistance, ie, as it heats up it has more resistance per unit length, which leads to lost voltage downstream if it's too high a gauge (too skinny) for the current it's providing, and too low a duty cycle means more power being dissipated thru the FET, since it's basically a resistor when it's switching on or off, which leads to heating and increased resistance there)

Have you tried slowing things down a bit to see if that helps? If so, you may need to change your PID settings to react faster (here I'm presuming you're using the PID rather than bang bang)

Regarding your "higher temperature", idk. that *could be the table you've chosen for the thermistor you're using. If the table isn't specifically for the thermistor beta value, it may not be calculating the actual temperature correctly. Not saying you've done anything wrong, just pointing out a possible reason for the discrepancy.

The others I can't answer with any authority, since I'm still building my first printer so have no experience with exactly how hot the drivers get, and my power supplies are linear, not computer supplies.

What thermistor and firmware are you using?

---

- akhlut

Just remember - Iterate, Iterate, Iterate!

[myhomelessmind.blogspot.com]

Sorry for late response, ive been playing with it all day yesterday. I had Sprinter installed and a 100k thermistor (Im not sure of the exact specs for this as I got it in a ParCan kit on ebay, it also came with the connecting wire for the resistor, which is quite dense, so I assume it is the correct gauge?)

Looking through the config file, I noticed PID wasnt enabled and was apparently an experimental feature.
I enabled it to see if it worked and it didnt so I ended up installing Marlin firmware instead since it appears to give better print results and has PID enable by default.
I was still having issues along with some new ones. Pronterface losing connection with the printer (caused by the higher baud rate used by marlin maybe?), sometimes crashing (never used to happen..).
I thought maybe the temperature spiking may be caused by a connection issue at the thermistor so and while touching the capton tape around the wires, I inadvertently ripped the wires off the thermistor...

I read something about thermistor resolution and Im thinking maybe it wasn't a very good at high temperatures which would partly explain the temperature spikes and also why the temperature stops mid print.
The spikes probably reach the safety value in the firmware, causing the machine to cut the heat so it doesnt burn out the resistor and thermistor.
Anyway I'll get back to this as soon as I get a new thermistor.

Edited 1 time(s). Last edit at 09/19/2012 10:56AM by gordon13.

Seems like it wasnt a thermistor problem. Still got a wildly varying temperature ( which I am now sure is what caused the printer to stop). Even with the PID setup I still couldn't get a print to finish. I noticed the stepper motor controllers are getting hot (they dont have heat sinks), so i put some cpu thermal compound and placed a small screw or bolt on each of the controllers. this made a massive difference and I was able to print a part without any issues.

Was this a fluke? I dont know, but the temperature seems to be way more stable than it was before so it shouldn't abort the print any more.


off topic: Now I am getting some different problem. The printer sometime crashes Pronterface when the hot end is heating up and I try to extrude plastic, or move it manually. The only way to fix it, is to unplug the printer from the usb.
I read something about power traveling up the usb from the board and causing communication errors.

I heard the ways to fix this would be to reduce the length of the power supply output cable or cut the 5v power rail on the sanguino. Is this worth doing? Has anyone had problems like this and fixed it like that or is there another way? Id rather not make a permanent modification to my board

I prevented both of these problems before I even started..

I have an 80mm fan blowing across my electronics.

I have a Ferrite Ring with my USB cable wrapped 3 times through it, and bite the end of the usb connector so it holds snuggly into my Laptop.

I also have a 1200 uF Capaciator across my 13.5vdc termials on my power supply. I use 13.5 for a few reasons, and just incase.. did you measure your power supply while your printer was running, or while it was off? You need to measure your power supply while the printer is running, perferably when the Z axis is going up, make it go up 50mm and read your voltage, set accordingly.. I say 13.5, but what do I know.