Solution Finder Get Tech Support

TOPSwitch-Hx going into latched shut down.

Posted by: Prashanth on

Hi
I have constructed a 12V@10A power supply using TOP257YN device (schematic attached). The input voltage to this converter is 300Vdc from PFC stage.

The entire design was done with PI Expert software (V8).

I observed a problem, during testing in the following sequence:

1. Upon first power ON, 12V comes up normally and remains in regulation.
2. When the power supply is switched OFF, after about 5 seconds, the 12V drops to 2.8V (This is due to the input hold-up capacitors - not a problem).
3. When the power supplyis switched ON again, the output is not coming (no 12V). However, 300Vdc is present and it is stable and regulating (from PFC stage).
4. After switching OFF the power and waiting for more than 4 minutes, until the input voltage dropped below 100V (300V input does not have any bleeder), I turned ON the power again.
At this stage, sometimes 12V appeared and most of the time it did not appear.

Following observations were done:
1. Checked for drain node switching for auto-restart - NO auto-restart
2. Disabled the OVP sensing circuit by disconnecting the resistor (R3=5.1k) to V-pin. Still the problem existed.
3. Shorted the V-pin to source (disabled UV/OV feature). 12V comes up every time properly. However, during turn OFF, after 12V turns OFF, it momentarily turns ON two or three times, while the input drops from 300V to zero.

Schematic and layout snapshots are attached.

Request you to please provide suggestions on what could be the problem.

Thanks and Regards
S.Prashanth

Files

Attachment Size
12V-ckt_0.JPG 351.99 KB
pi-12vnot-up_0.pdf 269.59 KB

Comments

Submitted by VCastrellon on 05/16/2012

I saw your attached schematic and this is what I would do first. The circuit needs to be simplified.
FIRST: The total resistance made up of R9, R10, R18, and R19 changed it from 15K to about 75 to 85K. If you get the actual values form the software I think it is a nono from the program. Make sure when you increase the value of the total resistance, the DRAIN voltage does not go beyond 700v under any condition.

SECOND: Remove L1 from the circuit and place a short circuit instead.

THIRD: remove the RC network you have through the transformer's primary winding.

Try to test your circuit again and tell me if the problem still is there.

Submitted by Prashanth on 05/20/2012

Hi
I checked after modifying as per your suggestions. I still see this problem. Some more observations:

1. When the TOPSwitch is OFF (no 12V output), if I touch the V-pin with a multimeter probe or scope probe, 12V comes up and works properly.
2. When I left the probe connected to V-pin and repeated turned OFF and ON, 12V came up every time I turned ON.
3. When I remove the probe and turn OFF and turn ON, the problem re-appears.
4. When I power ON for the first time, 12V comes up. After I switch OFF and switch ON again within 5 seconds, the problem re-appears, that is no 12V.

Please let me know if there is any other check I could do.
Thanks and regards
S.Prashanth

Submitted by VCastrellon on 05/21/2012

make sure R66 and R64 are the right value and at the moment you test the supply, the input voltage is 300VDC. Make sure the trace connecting R64 and the V pin is a very short trace. In other words, R64 is very close to the V pin.
For some reason you are not getting enough current into the L pin or you are getting noise inside the pin

Hi. I checked R66 and R64 values and found them to be ok.
One more observation is:
1. This problem is seen on the boards that were subjected to dielectric strength test of 1500Vac.

Since the 300 V is from a PFC preregulator, in the PFC ckt, there is diode which is used to bypass the pfc choke initially at power ON. When we do dielectric (hi-pot) strength test, the line and neutral are shorted and 1500Vac is applied across this shorted line and neutral and Earth (chassis).
This bypass diode is the only link between shorted line and neutral and the TOPSwitch V pin. Also, underneath the TOPSwitch, in the inner layer, we have the chassis copper present.

Could this be causing any damage to the TOPSwitch?

We are seeing this issue with 7 boards consistently. There are 3 boards which are not yet subjected to hi-pot test. These 3 boards are turning ON properly and 12V comes up without any problem.

Request your advice and suggestions.

Thanks and regards
S.Prashanth

Submitted by VCastrellon on 05/29/2012

It is hard for me to see if any possible failure is happening when you run hi-pot test.
If this is the case, that some damaging is happening during this test, then it means that at some point of the circuit, there is current flowing between the circuit and the chassis.
Make sure HS4 does not have its legs connected to the circuit. If HS4 is connected to the top-switch is ok. But make sure it is not anywhere else connected to the circuit.
Connecting its legs to the circuit makes a path to flow currents under some circumstances like when running hi pot

Submitted by Prashanth on 05/29/2012

Hi
The heat sink legs are not connected to any circuit and are floating. It is connected to the Source (tab) of the TOPSwitch and the legs are soldered to free pads. There is a chassis layer of copper in the board beneath the TOPSwitch ckt. The gap between the TOPSwitch pins, the heat sink solder legs etc. and the chassis copper is 3mm minimum.

Is it possible that the input capacitors (on 300V dc input to TOPSwitch) can get damaged (leaky or partial damaged) due to the high voltage of 1500Vac with chassis layer just below?

Can TOPSwitch device withstand hipot test in normal flyback systems, where there is direct connection from line to the input of the flyback?

Regards
S.Prashanth

Submitted by VCastrellon on 05/31/2012

Topswitch can survive hi pot test.

What I think may happening is that there is coupling currents from the primary side of the supply to the chassis plate.

This is what I would do.

I will run the hi pot test with the V pin connected to primary return. After the test, re-install the V pin as the original circuit and verify if the supply is properly working.
Another thing I would do is to run the test adding an small cap to the V pin. This Cap will delay the V pin operation but would filter any noise injected to this Pin.

From this two test we can conclude if noise has been injected into the V pin and this noise is affecting the V pin.

Submitted by bartek_zet on 09/03/2013

Hello,

did you find a solution? I have simillar problem.

Submitted by Prashanth on 09/03/2013

Hi

The only way I found to resolve this problem was to short the V-pin to Source (that is disable the V-pin functionality). On my board, there is a copper layer beneath the TOPSwitch ,which is connected to chassis (Earth). I started seeing this problem AFTER doing a dielectric strength test (1500Vac for 1 minute between Line-Neutral shorted and Earth). So, according to me, the root cause may be the V-pin was subjected to Electric stress due to its proximity to chassis layer, which carried the 1500Vac and may have caused partial damage to the V-pin. Note that the V-pin is not exactly floating or open, since it is tied to +Vdc bus through a string of high resistance. Only other suggestion I have is to put a low capacitance TVS diode across V-pin to source, if you are expecting any transients on this pin. Please let me know if this helped.