UCC256404: Startup cycling - Damaged Component

Hello,

Your questions has been received.  However do to US holiday I will not be able to look into until Tuesday.

Regards,

Hello,

You claim the unmodified EVM has this behavior.  Could you tell me the test conditions that you put the EVM under to cause this behavior? 

Do you have waveforms to share? 

Regards,

Hello Mike,

Just for clarification, the EVM was tested in the configuration received as purchased from Digikey, with a minor resistor value change to lower the BLK detection threshold to achieve 1V when the Vin bulk voltage was 250Vdc. The EVM was operated from 300Vdc instead of 390Vdc to match my design.  Once functional testing was completed, the UCC256404 (U4) was replaced with a device that seemed to have a start up cycling from my design. The EVM then exhibited the same cycling characteristic as seen on my design. The original U4 has not been reinstalled on the EVM yet to confirm the design is still functional with the original component. I will do that when time allows.

The EVM was tested with Vin = 300Vdc and 5A load in both cases.

I currently don't have scope plots that correlate the waveforms as seen on each pin with the same time base. Below are some of the captures made at individual pins while the EVM was cycling.

The RVcc pin cycles every second:

The LO pin switches for a shirt period of time: 

The HO pin switches for a short period of time:

One could assume there is an overcurrent fault but how could that be if the only change to the functional EVM was U4?

Thanks,

Gord

Hello,

There are several issues that could be causing your design to fail.  In your design if components are failing then they may not be rated for the correct voltage and currents that they are exposed too.  If the devices are getting hot they could fail thermally.  Your over current protection circuit on the ISENSE pin may not be setup correctly.  So you might want to revisit that with the excel tool.

I can't find the UCC25640EVM-202.  However, I did find the UCC25640EVM-020 and believe this is the EVM you are working with.

The LLC stage of this design was designed to work at 365 V to 410 V.  More than likely if you test this design with this input range you will not have any issues. 

The transformer turns ratio of the transformers and ISNS was setup for an input of  365 V to 410 V.  If you change the resistor divider on the BLK so the design would work down to 300 V.  More than likely the design is going into over current protection when the design is at 300 V.  

For the EVM to work at 300 V, you would need to change the transformer turns ratio to support the 300 V input.  Also the peak and RMS currents in all the electronic components are higher so you would have to change them as well.  However, the over current protection if not modified will protect your design from an over current. 

If your design is having a similar issue to the EVM you might want to study ISNS, SS/LL pin, the switch of the EVM to see if the design is indeed going into over current protection.  Most likely it is.

You had mentioned the design was based on the EVM.  Did you use a different transformer turns ratio compared to the EVM?

The excel tool should help you choose the correct transformer turns ratio to use in your design.

Regards,

Hello Mike,

You are correct regarding the EVM ... it is a UCC25640EVM-020 (-202 was a typo). My apologies.

All issues I have seen thus far with my design have occurred at ambient temperature (25°C) and an on-board temperature rise of about 10°C so the operating temperature is approximately 35°C. No other components in the circuit have had issues, neither thermal or voltage stress... only the UCC256404. The ISNS pin is setup with the values derived in the excel tool.

I understand that making an adjustment to the EVM operating voltage to 300Vdc may have an impact of the OCP, but with the original UCC256404 installed,  the EVM worked as expected, and didn't with the replacement UCC256404 that exhibited the startup cycling characteristic on my design. Note that the replacement part on the EVM was taken from an initially functional design that was working for sometime at the test bench level. The EVM discussion was intended to illustrate the different functionality of two UCC256404s applied to the exact same operational environment.

Are you suggesting that the since one UCC256404 worked on the EVM and a second didn't under the exact same test conditions, there is a variation in the functionality with respect to OCP from one device to another?

A new transformer was required for my design and the ratio was taken from the excel tool. For 300V input, the recommended ratio was 12.5 and this value was rounded down to 12. The Bias Winding ratio is 8. It is a custom planar design with very low leakage inductance so a Lr inductor (47uH) was used. 

Could you provide details of any setup\filtering\protection that may be required on the ISNS, BW, LL/SS, & VCR that could be helpful?

May I request a review of my schematic as well as the excel tool was used to create schematic to determine if my interpretation of the excel tool's inputs\outputs are correct and implemented properly? 

Thanks.

Mike O' said:

UCC25640EVM-020

Hello Gord, 

Mike is out of office today, but I believe he will be back in Monday Sept 9th.  Please wait for him to return for a chance to assess your latest reply information and respond appropriately. 

Meanwhile, I see no harm in providing your schematic diagram and Excel tool ahead of time, if you are willing to do so.  They will be available to him when he gets back. 

Regards,
Ulrich

Hello,

Your inquiry has been received and is under review.

Regards,

Hello,

Thankyou for the information. I had suggested that the designs transformer turns ratio was not setup for 300 V.  There could be slight differences in the OCP trip points of the design due to parameter variation.  You would need to study ISNS, SS, VCR of the failing device and compare it to the passing device.  The differences may give you an indication of what the issue might be.

I wonder if the UCC256404 was damaged in your design.  Do you have other UCC256404s that you could try in your design in the EVM to see if that is the case?

Regards,

Hello,

I believe that the UCC256404 is being degraded or damaged on my design since full functionality returns when the UCC256404 is replaced with a new component. The question is what is causing the damage or degraded performance of the UCC256404.

I have attached the excel design tool and a copy of the current schematic for review. Could you please have a look at both and let me know if I made any errors that could account for this issue.

 780-00206_REV2 8.pdf UCC256404 - 12V @ 10A - 6Sep2024.xlsx

Note that all low level signals are routed on internal layers with ground layers on adjacent layers. These signals were also routed away from noisy switching nodes and any filter capacitors were placed at the pins of the IC. 

All Burst mode options were tried.
Several ISNS resistor values were tried.
A variety of LL setting were tried.

The only remedy to resolve the 1 second cycling issue was to change the part to a new component.

Thanks.

Hello,

I have received your inquiry and it is under review and I will get back to you shortly.

Regards,

Hello,

I reviewed the excel design file and schematic and they look O.K.

You might want to double check the transformer turns ratio in your design to see if it is correct?

Regards,