UCC28070 Soft- and cold start problem

Ulrich<
From looking at the scope plots you attached, there is no clean, monotonic type of startup for the power supply.
The VAO signal is not a clean, slowly rising demand and the CAO current output is the same. This results in the PWM turning on/off and a general unpredictable startup.
You did not attach the full schematic but I feel the problem is due to the strange connection between the SS pin and the PKLMT pin.
The soft start for the UCC28070 is an adaptive style of start start that sources a variable current of 10uA ,1mA to the SS pin.
Loading this pin as in the customers application will result in unpredictable behavior.
I have never seen this done before. I recommend that they remove diode D11 and rerun the tests.
Do you know what is the situation with RSYNTH? Why is there a capacitor on this pin ?

Regards,
John Griffin

Hi John,
please find here our feedback:
For me the problem is not a non clean slowly rising of VAO but the non-starting over more than 5 seconds and in turn a static and no dynamic problem.

The Diode D11 is only a limitation for the softstart voltage to disable slew rate correction (SS limited to <4V) and to get a softer regulation at load steps and input voltage dips..
I wanted to test it without this diode but first I have to find a new sample with this failure.

The reason for the capacitor at Rsynth is the susceptibility against ESD, Burst and Imission.
There also is no specification of the maximum ratings for this pin. I would clamp this pin against Ref and Gnd.
Is this a solution you would recommend and and consider beneficial also with respect to avoid other problems?

Thanks & best regards,
Uli

Hi Ulrich,
In re-reading your initial question, I notice that the input voltage is specified as 12 to 138Vdc.
I'm sorry I missed that because the multiplier of the UCC28070 requires that the VINAC signal falls below 0.7V for at least 50uS.
This is because the quantized nature of the multiplier needs to be reset on the zero crossings of the VINAC signal. If this does not happen then the multiplier gets stuck in an indeterminate state of any the 1 to 8 levels. This is detailed in page 21 of the UCC28070 data sheet.
This could possibly result in a failure to startup.
The UCC28070 requires that the input voltage be essentially sinusoidal.

Also is there some reason for suspecting the RYSNTH pin as being any more susceptible to ESD than any other pin ?
A surge can manifest itself on any of the pins. An ESD surge could, for example, be capacitively coupled from the enclosure and a line surge could be coupled to the input or output pins. The effects are best mitigated by reviewing the pcb layout.
Is there a scope plot of the voltage on RSYNTH ? If the part is damaged it should be returned to TI Quality for analysis.

Regards,
John

Softstart and cold start problem of UCC28070_Rsynth.pdfHi John,

you said, UCC28070 requires sinusoidal input voltage for the current multiplier to function properly. However, for this boost converter application with DC input voltage, this problem is overcome by giving a fixed voltage from VREF to VINAC of the IC instead of sensing the input voltage. It has been observed that irrespective of DC voltage on VINAC pin, the current multiplier of the IC always switches on with Level 8 (kvff =3.857) and if the DC voltage is fixed (as done in this case), the current multiplier remains in Level 8. Thus this is not the likely cause of the delay in start during operation at -35 °C.

Please find the scope plot of Rsynth voltage during cold start attached.  Further, the sensitivity of Rsynth pin is suspected because, a few times the IC was damaged while trying to measure the voltage on Rsynth pin.

Best regards,

Uli

Hi Uli,

I spoke to one of my colleagues who has a great deal of experience with this part and he reminded me that the Rsynth input is a high impedance input.

The pin is very sensitive to noise coupling and it is not recommended to connect a scope probe here during operation.

Also adding a capacitor here provides a source for noise coupling into the pin.
So instead of decoupling the noise, this capacitor may just be coupling noise into the pin.
We recommend removing this capacitor and re-run the tests without the scope probe attached to the Rsynth pin.

One other thing is the ceramic capacitor dielectric. Are they rated for the operating temperature?

Have you looked at the mechanical aspects of low temperature operation such as solder connections, pcb conformal coating etc ?

Bulk capacitor value at low temp ?

Regards,

John