Burned H-bridge with Arduino Due

[titous]

Hey Adafruit,

This is another burned h-bridge topic (no I didn't use the supplied plug in pieces ); this time I was running the shield on the new Due with a library that I whipped together to do some testing (am I the first to blow the thing on a Due? do I get a prize? )

Anyway, the stepper I was driving only pulls 1A/phase AND I had two H-bridges stacked AND I had a clever little heatsink with thermal paste; so, needless to say I took all the necessary precautions. So, there I am, running my tests, and I finally get the stepper to run; but not after 20sec of run time, I get the dreaded puff of smoke. I've managed to pull the H-bridges off with a dremel tool and I think I can salvage the shield.

So my question is, should I try stacking three H-bridges? (I've got some extra ones laying around and I could try the other motor port) Or should I consider the big easy driver from Sparkfun which has a current ceiling of about 2A?

As always, thanks for your thoughts!

-----------------EDIT-------------------------

I stacked a third H-bridge onto the shield and it seems to run OK for now.... I'd still like to hear your thoughts though about whether this is a good idea?

[adafruit_support_mike]

(am I the first to blow the thing on a Due? do I get a prize? )

Yep.. one pair of de-smoked H-bridges. Wear them with pride.

I stacked a third H-bridge onto the shield and it seems to run OK for now.... I'd still like to hear your thoughts though about whether this is a good idea?

Three is okay, but about as far as you want to go. The chip in the middle is the one most at risk. It's getting heat from the ones above and below it. Ideally you'd run a heatsink between the chips to keep them from cooking each other.

.. well, ideally you'd break them out onto a separate board so each one has plenty of breathing room, but you get the point.

The apparent current draw concerns me, since you obviously took reasonable precautions. A pair of piggybacked L293Ds should be able to handle 1.2A/channel of continuous current with 2.4A peaks.

I'm making assumptions though, since on rereading I see you didn't list a part number. Were you using L293Ds (the ones with the built-in flyback protection diodes) or something else?

[titous]

I'm making assumptions though, since on rereading I see you didn't list a part number. Were you using L293Ds (the ones with the built-in flyback protection diodes) or something else?

I'm using the L293Ds that are supplied with the motorshield. I guess I"ll stick with three of them stacked for the time being; I'll be adding some fans as well for active cooling

[adafruit_support_mike]

I'm using the L293Ds that are supplied with the motorshield. I guess I"ll stick with three of them stacked for the time being; I'll be adding some fans as well for active cooling

Hmm..

Do you have an oscilloscope? If so, you might want to put a current sense resistor in line with the output and see exactly how much current is going through the lines. A bundle of ten 1/4 Watt 1 ohm resistors in parallel will be equivalent to a 2-1/2 Watt 100 milliohm current sense resistor, at about a fifth the price. Actual current-sense resistors cost about 50c each, where 1/4 Watt carbon film resistors cost $2 per 200.

If you put that between the shield and your motor, you'll see 100 mV for every amp that goes to the motor.. enough to measure but not enough to change the way anything works. You might just have a motor with lower impedance than the spec sheet says, but you might also see something that suggests a problem with the motor.. more current through one coil than the other, unusually large spikes of flyback current, or something like that.

[titous]

Hmm..

Do you have an oscilloscope? If so, you might want to put a current sense resistor in line with the output and see exactly how much current is going through the lines. A bundle of ten 1/4 Watt 1 ohm resistors in parallel will be equivalent to a 2-1/2 Watt 100 milliohm current sense resistor, at about a fifth the price. Actual current-sense resistors cost about 50c each, where 1/4 Watt carbon film resistors cost $2 per 200.

If you put that between the shield and your motor, you'll see 100 mV for every amp that goes to the motor.. enough to measure but not enough to change the way anything works. You might just have a motor with lower impedance than the spec sheet says, but you might also see something that suggests a problem with the motor.. more current through one coil than the other, unusually large spikes of flyback current, or something like that.

I do not have an oscilloscope

I'm going to run this setup with the three chips and see if I run into any problems. I've bought some extra h-bridges so that I can setup to drive two of these steppers; we'll see what happens. Cross your fingers!

Thanks for the input though, I'll keep my eye out on that stepper...

[adafruit_support_mike]

I do not have an oscilloscope

They are a hefty investment for a hobby.

You can get the same general effect with a cheap analog multimeter.. the needle-and-dial kind you can get for $10 at Walmart. A galvanometer reacts to changes more or less instantly while a digital meter needs the signal to remain stable long enough for it to count a given number of pulses.. usually about 2000. For signals that change quickly, you can at least get a general idea of what's going on by watching the needle pop back and forth.

** EDIT **

Let me qualify that statement about 'scopes: a new digital scope is a 'swallow and write the check' investment.. the Rigol DS1052E is a good performer for its price, but still costs about $400.

You can get old analog scopes for a song though.. stuff in the 25MHz range that's 25-40 years old and getting cleaned out of the closet in a lab somewhere. Those go for $25-50 on eBay, and some people will send them to you for the cost of shipping, and if they work at all they'll probably be a fantastic investment.

If/when you get to the point where you want to acquire a scope, don't make the mistake of assuming digital is better than analog.. they're actually different machines that do similar things, but neither is a complete replacement for the other. A good digital scope is the electronic equivalent of a camera.. it can store a sequence of samples thousands of readings long, so you can see everything that happened in a fairly long window of time. A good analog scope is the electronic equivalent of a microscope.. it lets you see what's happening immediately, and can clue you in to tiny details that a digital scope can't resolve. It's the same as being able to see a moving signal with an 18th-century galvanometer better than with a late 20th century pulse counting meter.