L298 Problem

Thread Starter

hensle

Joined Dec 31, 2011
46
I am wondering if anyone has run into a similar problem.

I was testing an L298 by having the output being a simple resistive load of 10 k-ohms. This was to make sure it was switching correctly before I put a large amount of current through it.

I tested each of the 2 channels separately and tried it with three different devices to make sure it was not a defect in the ic.

When using input 1 or 4, the device worked perfectly. The output voltage matched the source voltage and matched the triggering even at high frequency.

However, when using input 2 or 3, it did not work. Instead, as soon as it was triggered there was just a bunch of noise and the device began drawing excessive current (not completely shorted but more than could be explained).

It is simple to hook up and one of the tested components had worked in the past, while the other two were brand new. I have the 100n caps at their required locations.

Any suggestions would be appreciated.
 

SgtWookie

Joined Jul 17, 2007
22,230
Hello Hensle,
Welcome to AAC!

Here is the ST Microelectronics' datasheet for the L298:
http://www.st.com/internet/com/TECHNICAL_RESOURCES/TECHNICAL_LITERATURE/DATASHEET/CD00000240.pdf

I was testing an L298 by having the output being a simple resistive load of 10 k-ohms. This was to make sure it was switching correctly before I put a large amount of current through it.
That is not a bad idea.

I tested each of the 2 channels separately and tried it with three different devices to make sure it was not a defect in the ic.
OK.

When using input 1 or 4, the device worked perfectly. The output voltage matched the source voltage and matched the triggering even at high frequency.
Now this I have a hard time believing, as the high side is basically a Darlington follower. You'd lose at least 0.6v even with just a light load on it.

However, when using input 2 or 3, it did not work. Instead, as soon as it was triggered there was just a bunch of noise and the device began drawing excessive current (not completely shorted but more than could be explained).
Where did you get these L298 drivers? Can you take a very clear photo of the front and back of them? You may have forgeries, particularly if you bought them from other than an authorized distributor.

It is simple to hook up and one of the tested components had worked in the past, while the other two were brand new. I have the 100n caps at their required locations.
You should show a schematic of exactly how you have it connected. Schematics are pretty much a basic requirement here, as without them, all we can do is make guesses. Guessing frequently does nothing to solve a problem.
 

thatoneguy

Joined Feb 19, 2009
6,359
Inputs 2 and 3 shouldn't be used together.

When testing, put enable high, then Input 1 High, Input 2 Low, out should show voltage across resistor across Out1 and Out 2 pins. If you then switch Input 1 Low and input 2 High, the voltage across the resistor on Out1/Out2 should reverse polarity.

Same thing for Input3/Input4 and Output3/Output4 & enable.

The IC contains 2 Full bridges, so you should test them as a bridge, rather than individual outputs. i.e. You shouldn't expect anything to happen when enabling only one input of each bridge (pin 2 is for 1st bridge, Pin 3 is for 2nd bridge)

The inputs should have a value of high or low, not left floating/disconnected, which is what your test sounds like was done.

Test as shown in the diagram below, except put the resistor where the motor is, and don't install the motor protection diodes if you are using a non-wirewound resistor.

 

Attachments

Thread Starter

hensle

Joined Dec 31, 2011
46
Hello Hensle,
Welcome to AAC!
Thanks for the response.

I thought about providing a scematic, but it is merely a matter of connecting the pins. I followed the example in one of the datasheets.

the pins are: VSource=0-50V, VLogic=7V, sensing is grounded, Enable=7V
that is all there is to it, but to turn on an input which I did with 6V

The point is, that if I wired it incorrectly, input 1 and 4 would not have worked either as it is a symmetrical/complimentary arrangement.

I don't know, maybe all 6 channels of the three L298's had defects.
I bought them from Sparkfun.
 

Thread Starter

hensle

Joined Dec 31, 2011
46
Thanks for your response.

Inputs 2 and 3 shouldn't be used together.
I didn't. Though you can use them together if you wanted to have more current available by using both h bridges in parallel.
When testing, put enable high, then Input 1 High, Input 2 Low, out should show voltage across resistor across Out1 and Out 2 pins. If you then switch Input 1 Low and input 2 High, the voltage across the resistor on Out1/Out2 should reverse polarity.
I did this. Input 1 High worked fine. Input 2 yielded noise and excessive current.
Same thing for Input3/Input4 and Output3/Output4 & enable.

The IC contains 2 Full bridges, so you should test them as a bridge, rather than individual outputs. i.e. You shouldn't expect anything to happen when enabling only one input of each bridge (pin 2 is for 1st bridge, Pin 3 is for 2nd bridge)
I don't know how you could test individual outputs. I tested each bridge separately, by measuring voltage accross out1 and out2 and then out3 and out4.
The inputs should have a value of high or low, not left floating/disconnected, which is what your test sounds like was done.
First time I tested I did leave the other input open, but then thought that might be the problem, so when I grounded the other input which I was not testing at the time, the problem was not resolved.
Test as shown in the diagram below, except put the resistor where the motor is, and don't install the motor protection diodes if you are using a non-wirewound resistor.
This is exactly what I did.
 

thatoneguy

Joined Feb 19, 2009
6,359
This happens on all 3 of your L298's or just one of the three?

If you tested as stated above, and 2/3 worked fine, then the 1/3 would be bad.

Be sure to touch grounded surfaces when working with IC's. Many people say "I've never had anything die from static damage", but they've introduced noise in amplifiers, increased switching times on MOSFETs, and other things they are unaware of, since it "still worked". ESD Damage is the difference between "working" and "Working optimally" and sometimes "doesn't work".

I don't think many of us put a ground strap on for every little thing at home, but it is a good idea to have your bench grounded, and always touch something grounded before opening any bags. Keep your workspace on cement or tile rather than carpet, and humidity over 50% (hard to do in winter time)

Don't use a pure rubber mat that is not conductive. Most self-sealing mats with the cutting grid are OK to use, as it is hard to build up a charge on one, especially if you are in contact with ground a lot..
 

Thread Starter

hensle

Joined Dec 31, 2011
46
This happens on all 3 of your L298's or just one of the three?

If you tested as stated above, and 2/3 worked fine, then the 1/3 would be bad.
That's what has me perplexed. All three (actually 6) bridges failed in exactly the same way. So I thought for sure that it was something weird about my setup, but these were basic tests just as you described in your post.
Be sure to touch grounded surfaces when working with IC's. Many people say "I've never had anything die from static damage", but they've introduced noise in amplifiers, increased switching times on MOSFETs, and other things they are unaware of, since it "still worked". ESD Damage is the difference between "working" and "Working optimally" and sometimes "doesn't work".

I don't think many of us put a ground strap on for every little thing at home, but it is a good idea to have your bench grounded, and always touch something grounded before opening any bags. Keep your workspace on cement or tile rather than carpet, and humidity over 50% (hard to do in winter time)

Don't use a pure rubber mat that is not conductive. Most self-sealing mats with the cutting grid are OK to use, as it is hard to build up a charge on one, especially if you are in contact with ground a lot..
This may be an issue, but why the defect would be so targeted and consistent.

I think I will just build an H-bridge from transistors.

Thanks again for your help.
 

SgtWookie

Joined Jul 17, 2007
22,230
Did you have a load on them while you were testing them, and (most importantly) did you have a large heat sink installed?

Those L298 ICs dissipate a LOT of power as heat, because they can drop over 4v across the entire bridge when they have a load on their output.

Although the L298 is convenient, it is very inefficient.
 

Thread Starter

hensle

Joined Dec 31, 2011
46
Did you have a load on them while you were testing them, and (most importantly) did you have a large heat sink installed?

Those L298 ICs dissipate a LOT of power as heat, because they can drop over 4v across the entire bridge when they have a load on their output.

Although the L298 is convenient, it is very inefficient.
Thats why I was testing with 10kohms. Before, I did manage to blow one up. It was rather spectacular. So I obtained very large heat sinks and was going to put a couple of them in parallel so I could manage more current.

However I think I will go with the homemade bridge. It will give me more options for power requirements and heat management.
 
Top