I have the PWM driver v1.1 board, and when I plug in the high(er) voltage, every light on it lights up. Even for the PWM channel that isn't hooked up. It provides power too - I attached it to my nichrome and it did indeed get hot.

I am running 24 volts instead of 12, so I switched out all the 560ohm LED resistors for 1Kohm resistors. That was on the advice of an EE, so I figure it's good. But I can't figure out why all channels are "default on" instead of "default off".

That's not normal, is it?

I don't think it is, as when I tell the heater to start it doesn't change. It's possible that I did a poor enough job soldering that I caused a problem... But to get all three bad in the same way would be extraordinary even for me. Maybe I installed the transistors backwards? Something systematicly wrong seems likely.

I've attached a couple of pictures for your perusal. Just a cell phone camera, but maybe that'll be good enough? My good camera is in storage...

---

--
I'm building it with Baling Wire

The pictures are too blurred to be sure.

I can't see any insulation on the transistors. The metal tab is connected to the centre pin which is the collector. So you may have connected all the outputs together, but I can't see why that would make them default on unless the heastsink is also grounded.

Reversing the diodes or the tranistors would create an always on fault, but as far as I can tell they look right.

---

[www.hydraraptor.blogspot.com]

I did the exact same thing and had the same problem. The heatsinks are *apparently* connected to something internal, so when one goes live the current travels through the heatsink and switches the others on.

I unrolled an aluminium PB heatsink and bolted it through all three holes. Same problem - one live, all live.

After I cut the heatsink into 3 pieces so they weren't touching, it all worked much better.

Try removing the HS and see if that improves things. It looks like yours is more substantial than mine (mine snipped with some tin snips).

[renoirsrants.blogspot.com]

DaveR

I did the exact same thing (connecting the three transistors with a single heat sink) and noticed the same behavior. Cutting the heat sink so they weren't connected fixed it.

I've also ran it for short periods of time without heat sinks. So, you should be able to test things out by just removing the heat sink.

I AM sorry for the ultra-blurry pictures - my cell phone camera is even worse than I'd thought.

Heat sink it is! Give me a moment to try that...

OK, now only PWM 2 has a lit LED. That makes more sense, the Diode and the Transistor are the same orientation as the others, so it's probably a soldering error. I'm guessing I fried the transistor. Just a guess, though.

I see nothing obviously wrong, so I'm going to just ignore that channel and use PWM channels 1 and 3 instead. Since those work now, that should be OK. I thought I had some mica around here somewhere, but I can't find it right now. But the screws have plastic bushings on them, and I've still got some heat conductive epoxy that claims to be rated for 600V, I'll give that a try...

---

--
I'm building it with Baling Wire

My stepper drivers seem to get a *lot* hotter than the PWM drivers. I have put big heatsinks on the steppers, and smaller, seperate HS on the PWM.

The PWM driver v1.1 circuit is incorrect. The back EMF protection diodes should be wired across the load, not across the transistor.

It will be prone to destroying the transistor or the diode when driving inductive loads.

---

[www.hydraraptor.blogspot.com]

Thanks, nop! That's good info to have. Now, that should be "fixable" by sticking a diode into the screw terminals along with the wires to the load. The stripe goes towards the "minus" side, I think.

---

--
I'm building it with Baling Wire

NO the stripe is the cathode. The diode will conduct when the cathode is more negative than the anode. You don't want it to conduct when the load is on, so you wire the cathode to +12V and the anode to the collector of the transistor.

When an inductive load turns off, the back emf tries to take the collector above the supply rail. That will make the diode conduct because its anode will then be more positive that its cathode. The effect is to clamp the voltage to about 0.8V above the 12V rail. Without a diode it could reach 100's of volts and breakdown the transistor.

The way the diodes are on the board they stop the load pin going negative. a) the back emf takes it positive, not negative and b) there is already a diode inside the transistor to stop it going negative.

---

[www.hydraraptor.blogspot.com]

Doesn't matter. I just checked and I didn't order any extra diodes. Unless LED's count. So it'll have to wait until I place another mouser order. Unless LED's count, but I'm to sleepy to check their data sheet right now. Time for bed.

Actually, I knew about the back EMF problems with inductive loads from (your?) previous posts on the subject, so I knew exactly why you where saying that the diodes where in the wrong place. On that same subject, don't we also need a capacitor pair? Electrolytic and ceramic disk? Or is that only for motor loads, which are VERY inductive? And does that mean the PWM board could be used intead of the DC motor driver board if: 1. You moved the Diode to the proper place; 2. Added a pair of capacitors; 3. You didn't care about running the motor backwards?

That may be useful, as I'm not sure the motor driver's chip will run the gearmotor I have at all. Huh. DARN! I though I was down to only three major tasks before my first print. (1. Attach my improvised pump to my improvised thermal barrier. 2. re-write the new g-code firmware for "coil1/coil2" instead of "step/direction". 3. Move the z-axis limit switch to the BOTTOM, and adjust for head height.) I have a couple extra dual-h-bridge chips in case (for when) I blow them up, nice to know that I can try to re-purpose the PWM board too.

---

--
I'm building it with Baling Wire

No you can't use LEDs as they have a very low reverse breakdown voltage (~5V) and very low max forward current (~30ma).

You could move the original diodes if you can de-solder them, they aren't doing anything where they are now.

A diode is sufficient to protect the transistor from back EMF. Capacitors are added to reduce RFI and that is usually only an issue with DC motors, and best added at the motor end. See [blog.reprap.org].

Yes the PWM board will drive a DC motor in one direction only. It will handle a higher current and drop less voltage than the h-bridge solution.

Replacing the Darlingtons with protected MOSFETs would eliminate the need for the diodes and heatsink, would waste less power, handle higher currents and be short circuit proof, bit I am sure you have heard me say that before.

---

[www.hydraraptor.blogspot.com]