My newly built power supply is blowing fuses.

[Hextejas]

It also blew out a 24V AC wall wart.
As I first plugged the power supply into a toroidal transformer, it blew the 3.15 slow blow fuse.
So I thought that I would reduce the voltage by using the transformer and maybe get it to slightly work so that I could see what was wrong.
Nope, it blew out the wall wart.
So, here is the schematic and a picture.
Before all this, I put a thermistor in the input to the transformer and it stopped the fuse blowing for the toroid but apparently the power supply itself also has issues.
I dont know where to start.

I had a 2nd thought. The xformer is a 35-0-35 whereas the design called for 18-0-18. The output of the xformer is 38v.
Is that too much for the design of the power supply ? Should some of the components be beefed up ?

[Zero999]

You have 16.5mF on the output of each bridge rectifier, so this is exactly what you should expect to happen.

Presumably by thermistor, you're taking about a resistor with a negative temperature coefficient? If so, that's exactly what you should use.

[Kleinstein]

If the supply (rectifier and filter) is build for 18 V, the capacitors and maybe the diodes may not like twice the voltage.  For 36 V AC the capacitors should be good for 63 V at least, 50 V types would be too low.

The diodes, worst case see the peak to peak voltage, so something like nearly 3 times the RMS value. So 100 V diodes would be already too close.
The TO220 like case could be schottky diodes, and these tend to be low voltage only.

For a first test one would use a large (e.g. 1 K ) series resistors, just to check for a dead short and maybe give the capacitors time to reform.

[cbc02009]

It also blew out a 24V AC wall wart.
As I first plugged the power supply into a toroidal transformer, it blew the 3.15 slow blow fuse.
So I thought that I would reduce the voltage by using the transformer and maybe get it to slightly work so that I could see what was wrong.
Nope, it blew out the wall wart.
So, here is the schematic and a picture.
Before all this, I put a thermistor in the input to the transformer and it stopped the fuse blowing for the toroid but apparently the power supply itself also has issues.
I dont know where to start.

I had a 2nd thought. The xformer is a 35-0-35 whereas the design called for 18-0-18. The output of the xformer is 38v.
Is that too much for the design of the power supply ? Should some of the components be beefed up ?

If you're blowing fuses and power supplies, it's probably the current inrush to charge those capacitors. You have no resistance in series with those capacitors, so they're essentially charging at the input voltage divided by the ESR of the caps.

Actually, scratch that. It would only be the resistance of the wire since caps are essentially shorts initially at DC conditions.

[Hextejas]

Referring to the schematic, there are 4 wires coming from the toroid.  AC1, AC2, 0V1, and 0V2.
Would I put the 1K resistor on each of the AC lines only ?

Thanks

[Zero999]

Yes, put the 1k resistor on each  AC line. The capacitors should slowly charge up to close to the peak voltage, which will take a couple of minutes, with such large capacitors. You could try a lower value resistor, to speed it up, but it will needed to tolerate the full power, otherwise it will overheat. if there's a short.

The proper solution to this problem is to use a soft start circuit. If an NTC resistor isn't suitable, then you'll need a relay to bypass a current limiting resistor in series with the transformer's primary.

[Hextejas]

Thank you Zero. I decided to try and narrow down the problem by rail.
I started out with a 10k on one rail and connected it only.  The 10k 1/2 watt powered up ok so I started reducing it by half.
10k, 5k, 2k, 1k, 470 and things started heating up. This is still on 1 rail. I jumped right to a 25w, 25ohm resistor.
Well it still did not blow but the resistor sure got hot. And the xformer was making a lot of buzzing plus the output DC voltage was a lot lower than I had expected.
At that point I looked over my notes again re which xformer lead was connected to which power supply.
Yep, there was a pair switched. After a bit of further reading i got it sorted out and the xformer quieted down and did not blow any fuse.
I still needed the thermistor to handle the in-rush  current to the xformer.
So, bottom line was that I had 2 points of failure. The in-rush to the xformer and my screwing up the connection.

I Thank you again for helping me resolve this.

[wraper]

What the point in C5, C6, R3, R4? They do nothing useful in this circuit. Series resistors negate any effect 100n capacitors might give. Not to say C11, C12, C13, C14 will have barely any effect as well, if any at all. Wires between PSU and AMP will negate any effect of decoupling capacitors located in PSU. Considering capacitors used, looks like implementation of what people often call audiophoolery.

[coppercone2]

What the point in C5, C6, R3, R4? They do nothing useful in this circuit. Series resistors negate any effect 100n capacitors might give. Not to say C11, C12, C13, C14 will have barely any effect as well, if any at all. Wires between PSU and AMP will negate any effect of decoupling capacitors located in PSU. Considering capacitors used, looks like implementation of what people often call audiophoolery.

Military schematics for mains filters (like Lambda style) actually use dampened capacitors, in addition to non dampened capacitors.

poor example on broken page

not the best example but look in the front end of the diagram on page 2. I have seen more elaborate ones I think
http://pdf1.alldatasheet.com/datasheet-pdf/view/870010/SYNQOR/MCOTS-F-28-T-HT.html

https://www.synqor.com/document-viewer?document=Input%20System%20Instability.pdf

[wraper]

What the point in C5, C6, R3, R4? They do nothing useful in this circuit. Series resistors negate any effect 100n capacitors might give. Not to say C11, C12, C13, C14 will have barely any effect as well, if any at all. Wires between PSU and AMP will negate any effect of decoupling capacitors located in PSU. Considering capacitors used, looks like implementation of what people often call audiophoolery.

Military schematics for mains filters (like Lambda style) actually use dampened capacitors, in addition to non dampened capacitors.

poor example on broken page

not the best example but look in the front end of the diagram on page 2. I have seen more elaborate ones I think
http://pdf1.alldatasheet.com/datasheet-pdf/view/870010/SYNQOR/MCOTS-F-28-T-HT.html

https://www.synqor.com/document-viewer?document=Input%20System%20Instability.pdf

That's resistor in series with bulk capacitor. Not 100n cap which already has barely any effect as is.
EDIT: Also placing 5W 1 ohm resistor in series with 100n cap is ridiculous in this case. Not to say it's almost guaranteed to be wire-wound = having high inductance.

[coppercone2]

maybe the designer had something in mind but fucked up.

the values used are indeed preposterous. If you guess at the inductance of the resistor being 25uH and the 100n cap and negate resistance its centered at 100KHz.

[Hextejas]

It is not my design. I borrowed it from the DIYAudio site.

[Hextejas]

If any of the critics would like to change it please do. I will go ahead and build it and see how it performs.
It needs to support a left and right pair of either 4 ohms or 8 ohms.
I would be curious to see if it makes a difference in the sound.

[wraper]

If any of the critics would like to change it please do.

You won't make it better. All modifications I'd do is removing some unnecessary parts and not use audio grade capacitors (especially when using not in signal path) as they generally are just usual good quality caps sold for a lot of money. But as they are already mounted, there is no point of removing anything as it won't make any difference.

[wraper]

Only things somewhat worth of removing are R1, R2 bleeding resistors. Amplifier will discharge capacitors by itself anyway and those resistors will cause some additional heating without adding any benefit.

[coppercone2]

i would just increase their value if your worried

[Jay_Diddy_B]

Hi,

To the original poster.

The schematic shows a transformer that is 0-18 and 0-18. This is two independent windings.
This is different than a 18-0-18 transformer.

The 0-18 transformer will give an output voltage of 18 x 1.41 = 25V

The maximum voltage on the diodes is also 25V

You need to determine if your transformer has two independent windings or is center tapped.

Regards,

Jay_Diddy_B

[Hextejas]

Hi,

To the original poster.

The schematic shows a transformer that is 0-18 and 0-18. This is two independent windings.
This is different than a 18-0-18 transformer.

The 0-18 transformer will give an output voltage of 18 x 1.41 = 25V

The maximum voltage on the diodes is also 25V

You need to determine if your transformer has two independent windings or is center tapped.

Regards,

Jay_Diddy_B

Jay, please explain more. The schematic has 18-0-18 printed above the symbol for the xformer so is there something in the wiring that says different ? You have me even more confused.

[wraper]

Jay, please explain more. The schematic has 18-0-18 printed above the symbol for the xformer so is there something in the wiring that says different ? You have me even more confused.

18-0-18 means single winding with center tap, not 2 independent windings. When using transformer with center tap, you should connect center tap to GND and use single rectifier. The only way using existing PCB with center tap transformer would be removing 4 diodes and connecting like I shown on the picture.

[info]

This looks like the power supply circuitbasics is using. Build an input current limiter circuit (search is your friend) or for testing, use a lightbulb in series with the transformer on the primary side

https://youtu.be/Gj5KlGOdUqA?t=896