Problems with TB6560 Driver Chip

Thread Starter

jwells02

Joined Jul 27, 2017
22
I designed a driver board using the TB6560 chip which just drives some coils, not a motor. I've made about 200 of these boards over the last two years and there have been no failures. Over the last few weeks I've experienced about 50% failure rate. I thought maybe it was a bad batch of chips so ordered the chips from another supplier but still experiencing a high failure rate. The chip is susceptible to static and back emf which I suspect could be an issue but don't understand why I'm just now having problems. The chip can fail immediately or within a few hours. I've gone back and read the datasheet and searched for sample circuits to see if I left out some termination or maybe flyback diodes on the outputs but nothing is telling what may be wrong. The chip is rated at 3.5 amps but I have it limited to 2a for 10 seconds then drops to 800 ma for about 4 minutes. It doesn't run hot.

I've attached the circuit if someone could look at it see what I may be doing wrong I really appreciate it. Or suggest a reason why I'm just now experiencing issues.

Thanks.
 

Attachments

Thread Starter

jwells02

Joined Jul 27, 2017
22
I followed an example circuit to create this. It didn't have diodes and I can't find anything that would be significantly different. I've read on other threads that you have to pay attention to protecting it because it's sensitive but they didn't elaborate. I went back and reread the datasheet a few times to see if I missed something.

Other than reordered all the parts to make more nothing has changed. Is there something you can see about the circuit that makes it susceptible to damage? Worked fine for a couple years.
 

Alec_t

Joined Sep 17, 2013
10,458
Why is U3 pin 21 floating? MOS circuits can behave strangely unless all inputs are properly terminated.
 
Last edited:

Thread Starter

jwells02

Joined Jul 27, 2017
22
The chip has internal pull down resistors so I left it. It's 100k but I think I should have grounded it.

I just found another circuit that is using flyback diodes. I had read that if its driving a motor and the power is shut down the back emf could damage the chip. I've only been using electromagnetic coils. Could be it? Doesn't explain why I had zero failures for two years though.
 

Alec_t

Joined Sep 17, 2013
10,458
Could be a back-emf problem with the coils, but (a) as you say, it's odd that the problem is only recent and (b) if the chip has MOS output transistors in a H-bridge configuration their inherent body diodes should provide some protection.
Could it be a software glitch affecting the current-setting?
 

Thread Starter

jwells02

Joined Jul 27, 2017
22
It’s bizarre. After two years. The data sheet seems to indicate it can handle a bit of back emf. I redesigned the board to ground the floating pins and added 8 fly back diodes. Should get the boards mid week. In the mean time going to reprogram to start at 75% power instead of 100%. It runs 10 seconds at 2a which is the rated power supply so could be the latest power supplies I received could be out of spec.

Thanks for the help.
 

Thread Starter

jwells02

Joined Jul 27, 2017
22
Was waiting for the new boards to come but decided to wire in fly back diodes to test. Chip burned out in 5 hours. Struggling with this.
 

TeeKay6

Joined Apr 20, 2019
457
Was waiting for the new boards to come but decided to wire in fly back diodes to test. Chip burned out in 5 hours. Struggling with this.
Rather than rely solely on guessing, why not offer some details? For example, how do you know the chip burned out; what were the symptoms? Have all the chips failed in the same manner? Did anything else fail along with "the chip"? Were the power supply voltages okay before and after chip failure?
 

Thread Starter

jwells02

Joined Jul 27, 2017
22
Was actually hoping someone could look at the circuit and point out what I have wrong. Even though it’s worked great for two years. The chip runs perfect for a period of time. Runs cool then burns out. I replace the chip and it may run perfect or may burn out. The chip shorts and the power supply trips and resets on overload. Didn’t have a recorder on the power supply to see what it was doing right before burn out. I’ve changed the circuit to include flyback diodes and grounded the floating pins. If the circuit looks pretty solid then I’ll start looking at where I’m sourcing the components.
 

TeeKay6

Joined Apr 20, 2019
457
Was actually hoping someone could look at the circuit and point out what I have wrong. Even though it’s worked great for two years. The chip runs perfect for a period of time. Runs cool then burns out. I replace the chip and it may run perfect or may burn out. The chip shorts and the power supply trips and resets on overload. Didn’t have a recorder on the power supply to see what it was doing right before burn out. I’ve changed the circuit to include flyback diodes and grounded the floating pins. If the circuit looks pretty solid then I’ll start looking at where I’m sourcing the components.
You have boards failing, you say due to a "burned out" IC (bad terminology if chip didn't get hot; "failed" would be better). You offer a schematic that you say is unchanged from 200 units that did not fail. You claim that nothing in the hardware has changed. Well, surprise! Something did change either in the schematic or hardware because the hardware is now failing. So, get over the fixation of "nothing changed" and devote all your efforts to finding out what in fact did change. Is any programming of U1 or U3 involved? If yes, has that changed? Have the MOSFETs always survived or failed (both?)?

A powerful technique in troubleshooting is to simplify the situation. In this case, what happens if you connect only one coil? Or no coils? How long does the circuit then operate without failure? If you replace coils with comparable resistors, does failure still occur as before? If you power the circuit from a 5V bench supply with current limiting (or a fuse inline), rather than using the usual 5V regulator, does failure occur as before? What is the part number of U3? Can we view a datasheet?

If I give you a book that has been selling for many years without complaint and tell you that within the past two weeks dozens of complaints have been received about an error in the book...and I tell you nothing more. How long do you think it would take you to find the error? What if I told you the complaints cited page 23? Would you find the error more quickly?
 

Thread Starter

jwells02

Joined Jul 27, 2017
22
Let me ‘simplify this the best I can. First the part number is in the subject line. 2nd if I use a power supply with enough current it actually ‘burns’ up. It shorts out. 3rd if I replace the chip half the time the board will function fine instead of failing within a few hours.

The first thing comes to mind is I could have a bad set of chips so I ordered from a new supplier. No good. I assume the chips are within spec so the there must be something wrong my circuit design probably lacking something needed to protect the chip. They are sensitive from what I’ve read. The purpose of posting this is to see if someone familiar with this type of chip/ circuit could point out what I may have wrong.

I will upload the latest schematic shortly and hopefully someone would be so kind to do that.
 

atferrari

Joined Jan 6, 2004
3,489
Could you retrieve one of the "good" old ones and put it to work side by side with one of the recent batch?
A scope checking both in parallel, point by point?
 
Top