Need help with a F5 PSU mistake

[peranders]

anlive, you have gotten bad advice here. You should mirror the negative part. The resistors must be at the bottom but the PSU will still work but not as intended.

Power Amp PSU Board (for e.g. LYNX Amp)
Send me a message if you want to take a look at the schematics, too large to be uploaded.

[anilva]

Originally Posted by peranders

anlive, you have gotten bad advice here. You should mirror the negative part. The resistors must be at the bottom but the PSU will still work but not as intended.

Power Amp PSU Board (for e.g. LYNX Amp)
Send me a message if you want to take a look at the schematics, too large to be uploaded.

Thanks. I know what you mean. NPs schematic also has a mirrored CCRCC network for each leg of the PSU.

Having committed the error in my PCB, I am trying to find out the impact if at all there is any. The cost of correcting the PCB is cutting the tracks and hardwiring them right or desoldering all the electrolytic.

I buy the AC hum theory and that is should not impact in which leg the RC network is. What is the logic in your view?

Cheers.

[peranders]

You PS will get unsymmetrical with this equivalent circuit. Notice that the negative half will be unfiltered plus two caps doing nothing.

[anilva]

Originally Posted by peranders

You PS will get unsymmetrical with this equivalent circuit. Notice that the negative half will be unfiltered plus two caps doing nothing.

Please refer to the schematic posted by me earlier. They are two separate but identical DC power supplies employing RC filter but connected at the outputs with -ve of one supply and +ve of another supply to make a 0v reference for a +,0,- power supply.

I cannot figure out in what way the RC filter of the bottom rail becomes ineffective. My knowledge is definitely limited and would like to be corrected. Your schematic looks very different from mine.

Cheers.

[cnpope]

As long as you do exactly what is shown in drazen's schematic, then everything will be fine. This means modifying your existing schematic, to break the link between W2 and W3 that is shown in your schematic. i.e., the two supplies should be connected only by the link between W6 and W7.

If you stick with your schematic in unmodified form, then peranders's equivalent schematic is exactly right, and, as he says, this is not what you want.

Chris

[AndrewT]

You must break/cut the connection at the input just as D has shown.

The two supplies are isolated. They will both filter and give DC with a tiny rounded sawtooth superimposed. The +ve sawtooth should be rounded, I don't know if the -ve will be rounded, or inverted & rounded.

In my understanding, connecting at the output does not change the DC quality of either supply.

I have never done it this way, but you can easily measure the finished PSU (no amplifier/s) to check if both polarities have the same ripple. Do this using a bulb tester first and then direct on line.

[anilva]

Originally Posted by cnpope

As long as you do exactly what is shown in drazen's schematic, then everything will be fine. This means modifying your existing schematic, to break the link between W2 and W3 that is shown in your schematic. i.e., the two supplies should be connected only by the link between W6 and W7.

If you stick with your schematic in unmodified form, then peranders's equivalent schematic is exactly right, and, as he says, this is not what you want.

Chris

Originally Posted by AndrewT

You must break/cut the connection at the input just as D has shown.

The two supplies are isolated. They will both filter and give DC with a tiny rounded sawtooth superimposed. The +ve sawtooth should be rounded, I don't know if the -ve will be rounded, or inverted & rounded.

In my understanding, connecting at the output does not change the DC quality of either supply.

I have never done it this way, but you can easily measure the finished PSU (no amplifier/s) to check if both polarities have the same ripple. Do this using a bulb tester first and then direct on line.

Thanks everyone. My mistake. I have not connected W2 & W3 together as I have indicated in my original schematic but followed drazen's schematic.

I have checked AC ripple at the outputs and my Fluke DMM cannot detect any AC outputs.

Finally connected to the amp and works well. No hum.

Cheers.

[anilva]

Originally Posted by anilva

Thanks everyone. My mistake. I have not connected W2 & W3 together as I have indicated in my original schematic but followed drazen's schematic.

I have checked AC ripple at the outputs and my Fluke DMM cannot detect any AC outputs.

Finally connected to the amp and works well. No hum.

Cheers.

Correction: I measured AC ripple at the PS output. Both rails measure around 35mV of AC, whereas at bridge rectifier, the ripple measures 450mV of AC. There seems to be nearly 1/10th reduction of ripple across the RC network.

The caps are hot (my guess around 35-40 degrees centigrade) and the resistors are quite hot 50-60 degrees centigrade, when the power is being drawn by the amp.

[AndrewT]

Improve the ventilation around the capacitors. You must do everything you can to keep them cool. Even adding extra C if necessary. Try replacing the first 22mF C with ten 2m2F Cs in parallel. Use the 22mF as an extra final C

[AndrewT]

Originally Posted by anilva

I have checked AC ripple at the outputs and my Fluke DMM cannot detect any AC outputs.

I was about to tell you to buy a new (cheap) DMM, but post18 explained.