Hello all,
A couple of days ago I was leafing through some old Audio Magazines and I found an interesting article that shows a very smart way to make a better power supply. The idea is originally from Andrew Nehan who came across an article in the Radiotron Designer Handbook and the RSGB Handbook, later, he published a paper about the addition of a capacitor in parallel with the secondary of transformers just before the bridge.
The addition of this capacitor can be used for single and double secondaries. For double secondaries the capacitor is connected over the whole secondary and have no contact with the center tap. This is a parallel tuned LC circuit. The mechanism on how this works is not known. It might reduce the hysteresis of the secondary. The addition of the correct capacitor is audible to unbelievable.
The best part of the tweak is that a capacitor value that is not optimum for the transformer will do no harm at all.
I have in my GainClones 160va 22+22v (double secondaries) transformers for each channel, I installed two 1500pf one on each secondary, and there was a big improvement. The quality of the capacitor is crucial but given the low value it won't cost you an arm. Polypropylene Pulse Capacitors are the best (Wima FKP-2). Place the capacitor as close as possible to the transformer. Experiment with close values, here is a starting list:
VA Volt Cap Value
50 15 600pF
50 2x15 1200pF
50 2x25 2200pF
100 25 2200pF
100 2x25 4400pF
Regards
A couple of days ago I was leafing through some old Audio Magazines and I found an interesting article that shows a very smart way to make a better power supply. The idea is originally from Andrew Nehan who came across an article in the Radiotron Designer Handbook and the RSGB Handbook, later, he published a paper about the addition of a capacitor in parallel with the secondary of transformers just before the bridge.
The addition of this capacitor can be used for single and double secondaries. For double secondaries the capacitor is connected over the whole secondary and have no contact with the center tap. This is a parallel tuned LC circuit. The mechanism on how this works is not known. It might reduce the hysteresis of the secondary. The addition of the correct capacitor is audible to unbelievable.
The best part of the tweak is that a capacitor value that is not optimum for the transformer will do no harm at all.
I have in my GainClones 160va 22+22v (double secondaries) transformers for each channel, I installed two 1500pf one on each secondary, and there was a big improvement. The quality of the capacitor is crucial but given the low value it won't cost you an arm. Polypropylene Pulse Capacitors are the best (Wima FKP-2). Place the capacitor as close as possible to the transformer. Experiment with close values, here is a starting list:
VA Volt Cap Value
50 15 600pF
50 2x15 1200pF
50 2x25 2200pF
100 25 2200pF
100 2x25 4400pF
Regards
konstantin said:
It is a snubber that dampens the diode ringing of your rectifier. Many a thread has been written about it in this forum.
Some people prefer one capacitor per diode, others think one on the AC side and one on the DC side is sufficient. There are variants with or without serial resistor, and there is a pdf-document circling the forum by Jim Hagerman from 1995, where he explains how to find the best component values.
Nuuk said:Interesting tweak and one that is easy to try. I'll get some caps and experiment here. Thanks for the heads up!
If I understand correctly, I will need 1500 pF caps for my 300VA 2 x 25v traffos.
If you have split secondaries (2x25) you will need one capacitor connected in parallel with the transformer secondaries, NOT connected to the center tap, for 300VA something in the range 10000 - 14000pF.
If you have dual secondaries (25+25) you will need one capacitor per secondary winding, divide the capacitor value by two.
Regards
Re: Re: Tuning a PS Transformer on a GainClone.
Maybe I'm wrong but isn't a snubber an RC circuit? A resistor in series with a capacitor tuned to a specific frequency?
The above mechanism is an LC circuit, a capacitor tuned to the inductance of the transformers secondary.
Regards
pacificblue said:
It is a snubber that dampens the diode ringing of your rectifier. Many a thread has been written about it in this forum.
Some people prefer one capacitor per diode, others think one on the AC side and one on the DC side is sufficient. There are variants with or without serial resistor, and there is a pdf-document circling the forum by Jim Hagerman from 1995, where he explains how to find the best component values.
Maybe I'm wrong but isn't a snubber an RC circuit? A resistor in series with a capacitor tuned to a specific frequency?
The above mechanism is an LC circuit, a capacitor tuned to the inductance of the transformers secondary.
Regards
I really appreciate this thread.
Correlating it to my experiments on the same topic yields this: When I don't like the results from adding caps on the ac side of the rectifier, then those caps are too large.
That is especially the case with the 100v polyester, internally an RC because of its very high ESR, and thus needs a smaller than anticipated size.
Because of that, I would guess that monoceramic / multilayer ceramic would be ideal for this application, wheras other types would need to be smaller in capacitance value.
Observations:
Amplifier heatsink heat decreased in greater proportion than represented by the tiny voltage drop--both are results from caps on the AC side. This is my use of it.
Thanks again!!
Correlating it to my experiments on the same topic yields this: When I don't like the results from adding caps on the ac side of the rectifier, then those caps are too large.
That is especially the case with the 100v polyester, internally an RC because of its very high ESR, and thus needs a smaller than anticipated size.
Because of that, I would guess that monoceramic / multilayer ceramic would be ideal for this application, wheras other types would need to be smaller in capacitance value.
Observations:
Amplifier heatsink heat decreased in greater proportion than represented by the tiny voltage drop--both are results from caps on the AC side. This is my use of it.
Thanks again!!
Re: Re: Re: Tuning a PS Transformer on a GainClone.
BTW
Absolutely right. In this case R is very small, composed only of the transformers copper resistance, resistance of leads and PCB traces and capacitors ESR.konstantin said:
The pdf-document I mentioned above is explaining just that. The snubber to dampen the diode ringing must be tuned to the complete network of transformer inductance, transformer capacitance, transformer resistance and diode capacitance. As opposed to the routine application of a 100nF (or so) capacitor across the rectifier that is widely used.konstantin said:
BTW
A snubber must not necessarily be an RC circuit. A snubber can also be a diode, a varistor, etc. Anything that suppresses transients in a given application efficiently.
konstantin said:
Could you make an incremental application chart?
I have an Excel spreadsheet attached with common transformer varities related to incremental VA selections, but the capacitor values are blank, as in fill in the blank.
If that were filled in, I'd gladly turn the result into photographs and post. The data would answer most questions with "at a glance" ease.
(attached chart is an excel spreadsheet within a zip file)
Konstantin, I think a few folks will be confused by what you mean by split secondaries and dual (or twin) secondaries. I presume that this is what you mean.
An externally hosted image should be here but it was not working when we last tested it.
Oh yes, that would work so much better than a regular Internet search.
Okay guys, I filled out that Excel spreadsheet for both toroid and EI core transformers at a variety of voltages and amperages.
My math skills aren't sharp and the starting data was very skimpy; however, this, now-full, chart should result in workable values for ceramic and polypropylene capacitors.
If anyone would like to contribute a more accurate version, that would be great! Until then. . .
Okay guys, I filled out that Excel spreadsheet for both toroid and EI core transformers at a variety of voltages and amperages.
My math skills aren't sharp and the starting data was very skimpy; however, this, now-full, chart should result in workable values for ceramic and polypropylene capacitors.
If anyone would like to contribute a more accurate version, that would be great! Until then. . .
ratza said:
There are 7 known values for transformers, and the rest is an estimate, following the progression in the example at the first post.
There was no diode information given. As far as I known, a capacitor on the AC side will remove some of the ac signal, resulting in a slight voltage decrease.
Currently, I'm working on getting the EI core data a bit more accurate (its currently about 8% too high for "dual bobbin" types and leaves no "room" for ESR errata), although the result will still be an estimate.
After that's done, I'll try to do up a chart with "real" physical capacitor values.
danielwritesbac said:
Could you make an incremental application chart?
I have an Excel spreadsheet attached with common transformer varities related to incremental VA selections, but the capacitor values are blank, as in fill in the blank.
If that were filled in, I'd gladly turn the result into photographs and post. The data would answer most questions with "at a glance" ease.
(attached chart is an excel spreadsheet within a zip file)
Nuuk said:
Actually you will need 2 x 6800pF.
The really bad thing is that in the article I have on hand there is not a formula given to calculate the exact capacitor value but a board with fix VA-Volt values and capacitors.
What is at first apparent is that for a doubling of the transformers VA doubling there is a doubling on the capacitor value, the same applies to the Voltage rating. Here is an example:
VA Voltage Capacitor
50 2x25 2200pF
50 25+25 2 x 1100pF
100 2x25 4400pF
100 25+25 2 x 2200pF
250 2x25 11000pF
250 25+25 2 x 5600pF
500 2x25 22000pF
500 25+25 2 x 11000pF
I hope you get the idea of the calculation, these are correct values so for something in between you will probably need to experiment a little with close capacitors values, all I can say is that this little cap there makes a very positive surprise and worth to try.
The capacitor quality is an issue, I tried some ceramics I had on my parts box and there was no difference , the article specifies Wima FKP, these are Film and Foil for Pulse applications capacitors, they really make a difference.
Regards
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.