LM2904B: Problem when using LM2904BIDR to replace LM2904AVQDR

Manu Chang

Part Number: LM2904B
Other Parts Discussed in Thread: UCC28951, LM2904,

Hi,

My customer is facing a problem when use LM2904B to replace LM2904AV, as showed in the schematic below, customer is using LM2904 for a current and voltage sampling for a PSFB with using UCC28951.

When using LM2904B, they found that the Vadj is turned down when the input voltage is high, namely the current in COMP of UCC28951 is high. However, with LM2904A it doesn't occur.

Also, when customer changed the R406 and R431 in second figure from 100kohm to 50kohm, the Vadj value turned up a little bit. The normal voltage on Vadj is around 2V.

Customer would like to know, what differences between LM2904A and LM2904B that may cause that only LM2904B have this problem?

Thanks！

Manu

over 2 years ago

0 Clemens Ladisch over 2 years ago

Guru 284990 points

The B version has better noise filtering at its inputs.

How large is the switching noise at Vadj? Please show an oscilloscope trace.

0 Ron Michallick over 2 years ago in reply to Clemens Ladisch

TI__Guru**** 153026 points

Hello Manu,

Vadj is an input; what does Vadj is turned down mean?

Also check waveform for Isense.

0 Manu Chang over 2 years ago in reply to Ron Michallick

TI__Intellectual 1450 points

Hi Ron and Clemens，

pls check the waveform when facing the problem, The waveform is Vadj and DC_OUT. DC_OUT is going down mainly because Vadj goes down.

Vadj is falling down when the input voltage of PSFB is high, which means a higher current at the output pin COMP of LM2904B. The odd thing is that how could the output current have influence on input Vadj and make both the value smaller.

Hope that you can take a look and give some analysis on this scenario, thanks a lot

Best regards.

0 Clemens Ladisch over 2 years ago in reply to Manu Chang

Guru 284990 points

What PWM frequency are you using? Please show Vadj with a time base that allows to see the switching noise.

Does the PWM signal change in this case?

0 Ron Michallick over 2 years ago in reply to Clemens Ladisch

TI__Guru**** 153026 points

Manu,

The waveform is basically good because DC output voltage tracks Vadj. I did notice that the thickness of the VOUT waveform changes. Is this just a scope artifact or does the loop oscillate during the thick trace times?

Amplifiers

Amplifiers forum

LM2904B: Problem when using LM2904BIDR to replace LM2904AVQDR