DSP-167DP LOW VCC at IC901

[TJM2005]

Did you check D902 and/or D902B yet?

I just double checked them, both are reading about .47 and are not showing shorted.
I also swapped out IC901 with another one I had, and it didn't make a difference.

[Agent24]

Sounds very strange the Vcc pin is only 1.1v, it should be much higher for sure.
That IC won't try to start until the Vcc pin is at least 16v, before that it will be in UVLO state.

There are various other protections built into the IC which will shut down the gate drive if it detects issues,
But from what I can gather, it only affects drive to the MOSFET, not the startup power to the Vcc pin.

So even if say the feedback pin is shorted to ground, I believe the Vcc pin would still be at a normal voltage,
But the IC just won't be sending drive to the MOSFET.


Normally a low Vcc pin reading would indicate a problem like open circuit or shorted/leaky capacitor C904B,
Or the diode D902 being shorted.
Or in your case, the current sense resistors being open and breaking the circuit - but you already replaced these.

Unless you got very unlucky and have two DOA ICs!

Are you measuring from the hot side ground?
What is the resistance from Vcc pin to Ground pin?
What is the resistance from BNO and COMP pins to Ground pin?

Did you check or change the optoisolator?


Also, can we get photos of your board, both sides, and showing the worked areas?

[TJM2005]

• Are you measuring from the hot side ground?
  • Yes, scope ground on the - pin of the main cap
• Did you check or change the optoisolator?
  • No, this would be PC902?
  • I have one and could try to swap it out.
• I found this near the bottom of the datasheet, everything else calls out stopping the gate driver:
  • Open/Short fault Protection
    Some crucial protections are integrated in the LD7913J to
    prevent the power supply from damage in open/short fault
    condition. The LD7913J will shut down to protect the power
    circuit under following open or short conditions:
    1. COMP pin short to ground
    2. S pin floats
• Also, can we get photos of your board, both sides, and showing the worked areas?
  • Attached, I tried to make them useful. If you need to see something else I can grab move.

Resistance to Hot Ground / Pin 3:

• S: 2.5Ω
• BNO: 156KΩ
• GND: 0Ω
• COMP: OL - 20MΩ I could only catch it for a split second
• DRAIN: 33kΩ-200KΩ maybe more I didn't wait for it to stop charging.
• VCC: 1.76KΩ

[Agent24]

I find it interesting you have 1.76k to ground on Vcc Pin.

I see that there is a 2k SMD resistor (R903A) that goes in series to ground with ZD907.
I guess this is to clamp the Vcc voltage.

I wonder if ZD907 is okay? If it were shorted, that may cause the problem.

It looks like the Vcc supply of IC901 also feeds a pin on IC902. Can you confirm this?
If it does, then it's possible a fault with IC902 or its associated circuit could be pulling the Vcc pin low.

[Agent24]

Also I would definitely recommend cleaning the flux around IC901 given it runs from the high voltage supply.

Sometimes flux can breakdown with high voltage and cause leakage or even arcing.

[TJM2005]

• ZD907 / .74 no shorts
• The blue line should be the connection from the comp pin to PC902
  • It is the only pin on the hot side with continuity to the comp pin, amongst PC901 - PC903.
• I pulled PC902 from the board and powered it on, I am still getting 1.6V at VCC.
  • It doesn't seem to be bad, I posted a couple of pictures of the old one vs a new one.
• I gave the flux a good cleaning, I forgot to clean it after I swapped the chip this afternoon.

OLD Octo:

New LTV-816 https://www.mouser.com/ProductDetail/859-LTV-816

[Agent24]

Hmm. I really thought ZD907 might have been it.

Well, the next thing which might be a problem I can think of, is IC902, since it seems also connected to the Vcc pin of IC901.
It probably receives power from the same place.

Can you identify IC902? Or post a clear photo of the markings?

[R_J]

The ic should be able to charge the vcc capacitor before anything else would be detected, it should be able to charge the cap with about 1ma of current. If you can isolate the vcc pin from the rest of the circuit and only connect a 10~47µf cap from H-ground to only that pin and see if the ic can charge the cap, or try removing one end of ZD907 (it may be zenering at 1 volt?) and try to isolate ic902 vcc pin if you can (careful not to break pin) I would also remove one end of D902B in case it is leaky and loading down the vcc line.
It is possible the ic is defective
IC902 is likely a ST9S101-t6G or IN1M101 (marked S101)

[TJM2005]

• Can you identify IC902? Or post a clear photo of the markings?
  • Sorry I thought you was talking about the PC902, I now see IC902 on the bottom side.
  • looks like an S101 /408A
  • 

• If you can isolate the vcc pin from the rest of the circuit and only connect a 10~47µf cap from H-ground to only that pin and see if the ic can charge the cap.
  • I have extras 22uf 50v caps
• Or try removing one end of ZD907 (it may be zenering at 1 volt?) and try to isolate IC902 VCC pin if you can (careful not to break pin)
  • Is this the correct pinout? VCC would be under the letter A of 408A on the chip?
  • 
• I would also remove one end of D902B in case it is leaky and loading down the VCC line.
• It is possible the IC is defective
  • I swapped it this afternoon, both are doing the same thing, although nothing to say that both are no good haha.

What are your thoughts on best next steps? Lift the IC901 VCC pin and install the spare cap between it and hot ground to verify the IC is good?
If this test good then try lifting ZD907 and D902B along with IC902 VCC, see if the voltage increases then add a leg back one by one?

Thanks again for all of the help, I am trying to learn as I go, but all of this complex circuitry working together is hard to wrap my head around.

[Agent24]

Is this the correct pinout? VCC would be under the letter A of 408A on the chip?

Looks like it to me. You should get continuity between that pin, and Pin 5 (Vcc) of IC901.

What are your thoughts on best next steps? Lift the IC901 VCC pin and install the spare cap between it and hot ground to verify the IC is good?

That does sound like a way to check it with everything else disconnected.
You could desolder the chip, lift Pin 5, resolder it, then attach the cap on the top side of the IC between pins 5 and 3.

Of course, if this does work, the IC may start up and begin oscillating, then shutdown again when the capacitor runs out of charge. So you won't see a stable voltage on Vcc, but probably a kind of sawtooth wave, but it should be higher than 1.x volts that you have now.