UCC3895: Power management forum

sachin mc

Part Number: UCC3895
Other Parts Discussed in Thread: UCC2895, UCC28950

Dead time between gate pulses is set to 250ns but when measured using DSO I am getting around 500ns. after replacing the ucc3895 controller problem resolved, the same is happening 2nd time during the development process pleas help regarding the issue

equation (1) and (2) from the data sheet are used to set the above deadtime, looks like everything is correct from this part because replacing the controller solved the problem.

over 3 years ago

+1 Mike O' over 3 years ago

TI__Guru* 78306 points

Hello,

I reviewed the graph in figure 3 of the data sheet. In order for delay to be off by 200% the impendence would have to double on the DELAYAB and DELAYCD pins.

Equation 3 shows how the delay moves with the CS voltage if you are using adaptive delay. This delay will move with loading.

You may want to study the CS, ADS, and your gate drives to see if there are any differences in the test conditions that are causing the difference in delay times.

This will help to figure out what me causing the timing descrepancies.

Please note that turn on delay has an accuracy of + 10.7%/-19.6%.

Regards,

0 sachin mc over 3 years ago

Prodigy 20 points

Thank you for the response, I will take a look and get back soon

0 sachin mc over 3 years ago in reply to Mike O'

Prodigy 20 points

Im not using the adaptive delay feature. By connecting the ads pin directly to cs pin, so that's not an issue. And the resistors im using are 4.7kohm for DELAB and 3.8kohm for DELCD.

AND TO BE CLEAR.. after trying everything to solve this at last I replaced the ucc controller and that solved the problem. This tells that the issue was with that particular unit of the ucc3895 IC. So my doubt is if is it a manufacturing faulty pice or something happened to it after I used it. So I s there anything that can result in this kind of IC failure....

0 Mike O' over 3 years ago in reply to sachin mc

TI__Guru* 78306 points

Hello,

Please note that DELAB and DELCB are generally set to the same value. You might want to check your delay setting with 10% load.

I am just checking, but you do you have ADS biased with a voltage divider off of Vref? If not this might be part of the issue.

Regards,

0 sachin mc over 3 years ago in reply to Mike O'

Prodigy 20 points

Ads is connected directly to cs and to nothing else

0 Mike O' over 3 years ago in reply to sachin mc

TI__Guru* 78306 points

Hello.

This will still produce and adaptive delay. Try setting it to 2 V with a resistor divider off of Vref to see if you get more predictable turn on delays.

Regards,

0 sachin mc over 3 years ago in reply to Mike O'

Prodigy 20 points

And regarding the resistor values of the DELAB and DELCD it was set using the guide provided on the 8.2.2.10.5 section of the data sheet(figure 21 and 22). How can that cause a issue.. can please elaborate on this

0 sachin mc over 3 years ago in reply to Mike O'

Prodigy 20 points

I will consider trying this..

But still my doubt is how can you explain the fact that the issue was with a particular controller. But not with a new replacement Ic.

0 Mike O' over 3 years ago in reply to sachin mc

TI__Guru* 78306 points

Generally the LC tank circuits for each side of the H Bridge are similar. So generally the delay time constant is the same.

0 Mike O' over 3 years ago in reply to sachin mc

TI__Guru* 78306 points

If you are not using adaptive delay the ADS pin should be at a fixed voltage.

In regards to troubleshooting this you should study the CS pin, ADS pin, and outputs. There may be noise in the system on the CS pin and ADS that is contributing to the variation in delay times you are seeing.

The devices are tested before they leave the factory to match data sheet specifications. I am thinking that connecting the ADS pin to the CS pin may be causing some of the misbehavior.

One last thought if you have any pins of the UCC2895 ringing more than 300 mV below ground can cause the IC to misbehave. If this does happen it will turn on a parasitic path that will cause the device to misbehave. Also the behavior can vary between devices. I would study CS, ADS, Out Xs to make sure this is not occurring,

If you do have ringing below ground going greater than 300 mV below ground. This can be resolved with clamping Shootky diodes, filtering and some cases layout.

Regards,

0 sachin mc over 3 years ago in reply to Mike O'

Prodigy 20 points

Thank you for support. I will do further testing on this and come back soon.

0 sachin mc over 3 years ago in reply to Mike O'

Prodigy 20 points

" Setting RADSH = 0 ohm defeats the adaptive delay and a fixed value for t_DELAB and t_DELCD is used." This is what is written in section 8.2.2.10.5.

And also from equation (6) it implies that ADS should be connected to CS to keep the V_DEL constant(ie 0.5V) in turn to keep the t_DELAY constant.

0 Mike O' over 3 years ago in reply to sachin mc

TI__Guru* 78306 points

I reviewed the data sheet and it appears that you are correct. I thought this device did adaptive delay similar to the UCC28950/1 devices.

So that should be O.K.

I would still suggest studying the CS, ADS and outputs looking for noise and negative voltages below ground. It will be interesting to see what these waveforms look like.

Regards.

0 sachin mc over 3 years ago in reply to Mike O'

Prodigy 20 points

Ok ill come back soon with my findings