Heatbed, Powersupply and Z Axis issues.

tesseract - Number on the polyfuse F2, is X(or logo) 16 GF1100 which I am not finding an exact manufacturer reference for. Faces are not exactly parallel and would require some heat sink compound to fill the voids. Maybe use two thin plates of aluminum 1/2" x 2" one on each side. Make two small clothes pin like printed clips to sandwich plates in place? Can also L bend the plates for more surface area.

I punted and installed a 15A mini fuse holder from OReilly Autoparts FHM200BP right on the RAMPs board, using some wire wrap socket pins for rigidity. Not the best solution, Radio Shack had nada. Longer term will look for a low profile fuse socket. I could be talked into going back to the Polyfuses, possibly two in parallel or maybe the heat sink. The Polyfuses called out on the RAMPs board are just too close to the tripping point for my taste. I haven't tried it, but expect neither the 15A mini or the Polyfuses are fast enough to keep the power fet from popping with a dead short to the power supply. They are there to prevent a fire if somehow the the FET faulted short.

Bed temp is still only getting to 74C where the extruder fan is blowing on it, and not 100C. It might be warmer directly under the part. Annoying part is the heating bed message staying on.

 

I'm not sure how you'd really attach a heatsink to the fuses... it's not a flat enough surface so really the best bet would be airflow over them to provide cooling. I've pushed my fuses apart from one another and away from the other parts which might generate heat, to try and allow as much airflow around then as possible, but don't have a fan currently pointed at my RAMPS. It's a 12.5 hour print, and is the same print that killed my hotend previously. Robo sent me a new one (
FedEx Saturday Delivery even, natch!) to replace the one that died so I'm printing again finally, yay!!! It's been plugging away for an hour and a half so far... so 11 hours to go! Currently printing layer 5 of 191 of the 1.2 MILLION lines of GCode. Hopefully this works... othewise I'll have to start printing all the parts individually rather than all nine at once... yeah, all my eggs in one basket... {shrug} Maybe I'm just stupid. LOL

 

Danggit!!!! X axis slip!!! Guess I'll need to tighten and start trying to print the parts one at a time now... {sigh}

Oh yeah, I've also tried taping a sheet of paper on both sides of my housing in an effort to try and trap some of the heat in there and contain it so maybe the heated bed wouldn't be worked as hard. Plus I'm keeping the door to the deck closed. I have to look out every minute or so and check to see if the dog is trying to get back in or not... as the wifey would NEVER let me hear the end of it if I left the dingus boy trapped out on the deck with me not in the room watching him!

 

I'm printing with PLA. I know I shouldn't NEED a heated bed, but I was having a lot of adhesion issues without it... so now I'm just using it just because it's working for me. After tightening the Y axis it's not slipping. Of course I'm also just printing the four LM8UU bushings/bearings. I've got some real ones on order from China but that won't arrive for two weeks so am printing some to use in the meanwhile so i can get my hotwire foamcutter built... The generic NEMA23 axis will form the two X and Y axis. I tried building out out of drawer slides, but was having serious issues with resolution (I'm building 3 CNC routers) for the foam HBLB34's... So, I scrapped the idea, ordered some 8mm ground and polished rods from MSC, the bearings from china, and will print the rest of the axis and build the foamcutter CNC from that...

 

Hi Guys,

My second day with the printer and I ran straight into the Overheating Polyfuse switching off the heat bed problem - while printing Mr. Jaws in ABS

After looking at the problem - I've come up with a quick "Heath Robinson" to fix the issue, it appears to be working the heatbed and extruder have now been running for 2 hours at 90deg on heat bed and 210deg on extruder without the issue occurring prior to the mod it would last 30mins at most.

Images are below (couldn't figure out adding them here):

https://www.dropbox.com/sh/opvu51jpzvhrtsx/5n_jM2PeV_/Heath-Robo1.jpg

https://www.dropbox.com/sh/opvu51jpzvhrtsx/XWbMJETsuR/Heath-Robo2.jpg

It's a pretty simple mod and took 20mins with a dremel, a 40mm PC 12VDC fan and a soldering iron(optional) and some heat shrink.

Not sure if my quick fix will work for anyone else but it's solved my issue !

All the best

Matt

 

Matt - How do you REEEEAAAALLLLLLLY know you fixed the problem? it might only run for 2:00:01 before it fails? Seriously nice hack, but I am still sticking with the 15A auto fuse. Also less power loss with the 15A. These Poly's use about 1-4W before they trip. Blow too much air over the Poly, and its trip current goes way up. Plus if I have a bed heater fault, I want to know about it, because at that point the output control FET is probably toast as well (and most likely protected the fuse). I love the fact that these Poly's have a max rated current of around 40A depending on vendor (maybe consider adding a Slo blow fuse to protect the Poly). This link illustrates temp sensitivity of Poly's vs fast /slo blow. Its a known problem that this Poly is not sized correctly for the heater bed current. From across the pond, I had to wiki Heath Robinson. Cheers!

 

Thanks tesseract - it does indeed seem to solve the symptom - 6hours so far with no issue but I may run out of abs soon

Nickster I agree it's far from perfect hence the "Heath Robinson" moniker - however my concerns are more around slaving a DC fan to a binary 12V 10Amp draw which does seem to switch a lot... I'm thinking of taking a signal pin link from the arduino and driving the fan from the supply via a TIP120.. But I'd like a few days before I take a soldering iron to my new shines ramps /taurino. On the cooling effect on poly fuses there is no way the airflow on a 40mm fan will drop the temp on the poly below ambient so it will be operating in its design sweet spot of around 24-26degC.

Interesting idea on the car fuse - you could extend this to a good old fashioned breaker with a reset button externally accessible. Have you thought about using a bigger more ruggedised polyswitch instead of the cartridge -something like a bd240 - that way you maintain the plolyies fire protection ? Be interested in your thoughts as I'm sure the soldering iron will be out again soon !

Thanks

Matt

 

Let me see if I understand this correctly... you hooked the fan leads up to the same for the heated bed? So when the heated bed comes on, the fan comes on, and when it turns off, the fan turns off? This would result in the fan going on/off/on/off/on/off as the heated bed clicks on and off to maintain temperature... am I getting this correctly?

Why not have just tapped onto the power supply directly so that when it's on, the fan is on, full time... regardless of heated bed. You could always mount a switch to turn it on/off if you desired if you didn't want to listen to it when the heated bed isn't being used. Otherwise, you're going to have to monkey with the firmware to control the fan. Not saying that's a big deal, just more rube goldburg than I'd wanna deal with personally. I wired my LEDs for the hotend direct to the power supply.

 

I opt'ed for pulling waste heat from the system as I wanted a general airflow not a specific cooled spot. If the problem returns ill think about swinging the fan around to specifically cool (push) air at the polyfuses and board.

At the minute the fan is connected to Heated bed by a nice juicy iN5355BG zener diode and a tiny little 1K resistor, I think adding a very large decoupling Electolytic capacitor may help decouple the two elements a bit more.

The reason the fan is connected on the heated bed outputs is I over think problems and have therefore probably over engineered the solution to only be on when the problem occurs

I shall look at moving it to the PSU in my next adventure under the cover!

On the heatbed the cycle at 90degC in my environment seems to be a pretty square wave with a wave length of about 10seconds peak to peak. so the Fan's duty cycle is actually nicer than it's previous life on a motherboard northbridge !

 

Under worst case conditions, the thermal fuse can only degrade the power going to the bed by about 4% or 2.5C. When its resistance increases so that you are seeing 0.25V drop across the Polyfuse, it is probably a done deal, dissipating like 2.5W and is going to tip soon. It just can't significantly affect the bed power without tripping.

Polyfuses have a limited life time. Manufacturer's data sheet talk about things like guaranteeing they won't burst into flames after 100 cycles. Every time they are tripped, there is wear and the resistance doesn't go quite back to the original value. (at least what I can see from reading the literature). Also the UL and CSA approvals are contingent on the fuse not being used in an application where it repetitively cycles (defined as many thousands). Fall out is that if the Polyfuse trips because of cooling issues, it will continue to trip and probably at an increasing rate.

We nominally have 100W going to the bed and if you turn off the extruder fan, my guess is most Robo's will hit 100C bed temp. If you position the extruder over the the part of the bed near the thermistor pick up, even that small amount of air is able to suck a lot of power from the bed and you see temps like 80C. Another variable is to look at the bottom of your bed and see if the thermistor is directly over a PCB heater trace or in between. I suspect the temp is a lot warmer and easier to maintain directly over the trace. I am thinking about trying to add a thermal blanket to the bottom of my bed as an experiment.

Blowing air, even very warm air, directly over the Polyfuse will have an enormous effect at dropping its temperature and raising its trip point. Data sheets and white papers don't come right out say it, but the trip current appears to be proportional to the initial resistance. In addition to the 10% hold current derating at 35C ambient in the Robo, manufacturers call out +/-5% variation for passive aging, +/-5% for Humidity aging, +/- 5% thermal shock (80C/-40C cycle). Recovery resistance 1 hour after tripping is spec'ed at not more than 2.4x initial resistance. A data sheet here.

So I still think the best controlled solution is a Slo-blow fuse. Other options are the fan with or without external cooling. Another possibility is a higher current Polyfuse like 16R1200 or 16R1400 from Digikey ($1.09 +$500 S&H ). One app note talked about running Polyfuses in parallel, but current sharing is not guaranteed (which seemed strange) but cautioned that heatsinking thru the leads into the PCB needs to be balanced. Seems like you'd want to widely separate two Poly's in parallel so they have their own independent heat zone which would balance the current.

Now if my eBay replacement Hobbed Bolt would just get here....