Strange Voltage Drop 2

[Kepco]

Hi Guys,
Here is a circuit that is two part. Top part is 48V ( Telecom -48V) monitoring circuit. This circuit converts 48V to a low isolated voltage (~1.6V) 48V_MNTR. Also the 12V that power U26 is from same 48V going through a DC to DC (48V,4.5A to 12V) converter. Bottom part is a 15V monitoring circuit. This circuit converts 15V to a low isolated voltage ( ~1.6V) 15V_MNTR. The problem is when both 48V and 15V are present the U626 output voltage drops from 5V(5V_ISOL) to 3.8V. But if only 15V is present then 5V is stable. U626,U27,U6475,R6325 and R6326 were replaced with no change in circuit behavior. Any ideas what is going on? U626 is getting loaded when VDD2 on U27 becomes active it seems. U626 can handle 80mA max and U27 on VDD1 side needs 16mA max so not sure why there is overloading. 15V is used to power only the 15V monitoring circuit and does not go any where else and its value is stable during the test. 15V and 48V each are completely isolated sources as well.
Thanks

 

[VEBill]

With the disclaimer that I'm not familiar with some of those ICs, it seems to be a very logical design. Perhaps it's a build fault (as opposed to a design error).

 

[Skogsgurra]

With that nick (Kepco) one expects lot of expertise. So this has to be a tough one.
If this is a circuit that usually works and nothing has been changed, well...
If this is a newly developed circuit, then two questions arise:
1. Did you look at all voltages with a scope? Is there any HF present? That usually explains fun behaviour.
2. What does the simulation say? Does it work? Or same fun behaviour?

Gunnar Englund

http://www.gke.org

--------------------------------------
Half full - Half empty? I don't mind. It's what in it that counts.

 

[iop995]

It's not any overloading; I think when 48V line is present, it influence 15V RTN potential. Check if voltage between +15V and 15V RTB change when power-on 48V line. Also, may be influence of U27 in output voltage of U626 when 48V came on (VDD2 of U27 is powered); reduce R6325 and R6326 by minim a order.

 

[Kepco]

"reduce R6325 and R6326 by minim a order. "

This was it. Changing the divider current to 1mA did the trick. Original value was 1.5uA then changed to 5uA with not difference. So new values of U6325=3.3K and U6326=1K provides current of 1.2mA and it worked.

Thanks a lot iop995 and all who took time to comment on the issue.