Hi all... Interesting thread.

I am wondering since this thread dates back over a year now... perhaps they were using the Sanguinololu 1.2 PCB The current ver. = 1.3a
I am hoping this is correct and that 1.3a does not suffer from mosfet melting.

I'm just about to start constructing a Prusa i3 with a Sanguinololu 1.3A PCB (1284p) attached to a 12v 20amp power supply. (not ATX)

The kit I ordered comes with an aluminium heat sink that's not attached to the heat-bed mosfet.... I must attach it myself.

After reading the above I'm now under the impression that the PCB is simply a mosfet fryer. (especially the Heat-bed mosfet. )

If this is so... why are there no warning signs regarding the setup of the PCB ? ( I sort of wish I had simply ordered RAMPS 1.4 or does it also suffer )

So no turning back now... If I simply attach the heat-sink to the mosfet & power up the PCB is there a high probability the heat-bed mosfet will melt ?

???

Edited 4 time(s). Last edit at 06/05/2013 02:36PM by Mickman.

Hi Mickman,

My issue wasn't due to the version of the Sanguinololu; rather the quality of the PCB Heatbed.

My PCB has a considerably lower resistance than would normally be required (0.85 ohm rather than 1.5 ohm); so to raise it to a high enough temperature, considerably more current has to be pumped through. While the current is within the specs for the MOSFET, it had no way to deal with the heat that this current generates; which is why my first MOSFET melted.

To be fair, this PCB Heatbed needs just over 14amps; which is very very high; well beyond the specs for most printed circuit boards. It wouldn't really have made a difference if I'd been using a RAMPs board.

I'll be looking for a replacement heatbed at some point, I've just been too busy to look.

Cheers,

T.

Yeps,

ordered a board from reprapworld, build according to spec I suppose.....Mosfet went up in smoke on the first print. Now these babies have operating temps of 175 degrees, so well no suppries that if they fail they do that catastrophically. So replaced it will be (no problems there). My heated bed circuit runs at 1.64 Ohm.....not especially low, nice in spec , I hope .

Questions du jour:

- What is the best way to a create a (sort of reliable) sollution without to drastic modifications to the Sanguinololu?
- If this happens all the time, why isn't this taken into account in the basic specs of the Sanguinololu?
- What 'bout this Aussi mod?

[create3d.com.au]

I have another board (that does not communicate over USB ) from a different supplier and the mosfets on that board are an 1C53AF for the heated bed and a 1C35AD for the hot end. Now I can't find anything on the specs of these little rascals and since I've seen what happend to it's freind at the other Sanguinololu I'm a teensy bit apprehensive.

- Does anyone have the specs on these?

Greets, Ed.

1C53AF and 1C35AD are probably batch numbers the Fairchild Semiconductor denomination seems to be BUZ11A (okay, I don't have a rotten trout at hand, please)....

Edited 2 time(s). Last edit at 10/11/2013 02:34AM by Ed.

Hi, hi,

Same for me......trick I discovered.....clamp the board in a vice (plastic Dremel vice in my case).....then apply heat from above....and pul the part out from below....if you only have residual solder use heat from above and apply the desoledering pump from below, until you have a concave "hole" where the hole in the board should be.....then use a 1 or 1,2 mm drill in a Dremel en thus remove access solder.

BTW, I discovered that after I rampaged the same lead as you did and after I fried a perfectly innocent (original spec) Mosfet to pieces.....whooosh.....I which someone told the Sanguino guys about the existence of PCB hotbeds in combination with ABS ......greets, Ed.

Next step.....IRLB8743PBF and some diodes....gonna Gen 7 a bit....maybe I even add the odd optocoupler (4N25) into teh mix and completely seperate the hotbed circiut and the rest of the arduino electronic.....how did "the offspring" say it.....righ "You gotta keep 'm seperated" but I definitly don't think they where refering to a heated bed.

Greets, Ed.

The easiest modification would be to attach the Sanguinololu to a big relay coil, and the heatbed will be switched by said relay. That way, the Sanguinololu sees less than 1A, and you can use whatever you want (even low voltage AC) on the heatbed .

I didn't go all the way to an optocoupler, but I made a separate circuit using a low-end PNP transistor and two power NMOS. The switching circuit from the Sanguinololu pulls the PNP's base to ground, and the PNP sends +12V to the the gates of the NMOS. It didn't heat up beyond ambient temperatures according to my cheap IR thermometer + prodding with fingers. This was with a 1.5 Ohm heatbed and 11.8V measured at the wire going from my circuit to the bed.

Don't forget to shield your MOSFET from the nasty stuff coming back from the heatbed when you turn it off:
[reprap.org]
I use an antiparallel diode and a 0.33uF, 50V capacitor soldered to the heatbed's contacts. The first trick is from the Wiki article up there, and the second is from hobby R/C (pair up any switched inductive thingie with small ceramic capacitors).

Quote

The easiest modification would be to attach the Sanguinololu to a big relay coil

Sanguinololus miss a flyback diode, so you'd destroy the MOSFET when attaching a coil of noticeable value.

---

Generation 7 Electronics	Teacup Firmware	RepRap DIY

Hmm that's true that the relay coil is an inductor too that you have to protect.

I soldered diodes on the underside of the connectors for heatbed and extruder despite my heatbed actually switching a PNP transistor, just in case I ever switch to relay without remembering that bit.

Hi all,

I have the same board and the same problems with the hotbed. I also have a weak PS powering the printer, so decided to add another one just for the bed. Hooked a modified car relay to the heatbed connector of the Sanguinololu to power the bed using that second power supply - worked quite well, modded the relay a bit (capacitor + diode) so that it survives the constant switching, but the rattle was very annoying.
Then I made myself a Sevenswitch board (thanks Traumflug!) and tried to use it in place of the relay - no luck, the two MOSFETS seem to cancel each other . Now, I can imagine it's possible to reassign the hotbed to another pin, one present on the extension connector for example - but frankly I got lost in the Repetier firmware trying to find the right pin assignment, and plus I would like to connect an LCD one day..
So, need to use the original pin but - without the board's MOSFET, just as a straight GND and +5V pair that I would connect to the Sevenswitch. But, since I have a limited knowledge of electronics - need some answers to dumb questions: How should I go about it? Where is the processor pin connected, to the gate of the MOSFET?

Thanks! Igor

Edited 1 time(s). Last edit at 04/06/2014 02:18PM by bujholm.

Hi Igor,

The overheating problem isn't just the MOSFET, but also the PCB isn't designed very well to manage the excess heat. Simply adding a heat-sink to the board was a bit risky because there isn't a lot of space for it. We decided moving the MOSFET off the board was a good solution. Similar to what you're doing. Our solution is:

- Created a cable harness with housing for the the MOSFET and added a heat sync.
- Use the same signal line for the MOSFET off-board. A single wire is run to the housing from the board.
- Added a resistor (220 ohms) to provide minimal back-current protection for the processor.
- Bypass the board for heat bed power.
- The harness also bundles the thermistor wires as well.

Some photos are attached. One of the cable harness (the cover snaps in snugly)
The only modification to the board was the addition of the resistor.

Send me a private message if you're interested, and are in Australia (I am from Emblem Robotics and we only post within Australia), I have a number of extra harnesses here. Otherwise I would be happy to post the housing STL file to print your own.

EDIT: I just realised the scad file for the MOSFET mount is already available on our bitbucket repository

Cheers,

slip

Edited 2 time(s). Last edit at 04/07/2014 11:11PM by slip.