Hello guys. Im in the proces of making my SE EL84, but due to low voltage on the power transformer, I have to use solid state rectifiers.
Since this is SE amplifier, I need the most clean power supply, without any switching diode noise etc. But how to do that?
Every time I found solid state diodes to sound harsher and worse in highs than vacuum tube ones. So, how to use solid state diodes properly?
What precautions to use to make them sound good, what type of diodes should I use, should I bypass them, etc?
Thank much everyone for help.
Best regards, Michal
Since this is SE amplifier, I need the most clean power supply, without any switching diode noise etc. But how to do that?
Every time I found solid state diodes to sound harsher and worse in highs than vacuum tube ones. So, how to use solid state diodes properly?
What precautions to use to make them sound good, what type of diodes should I use, should I bypass them, etc?
Thank much everyone for help.
Best regards, Michal
and switch mode power supply too?
Absolutely not. Use sync rectification without the ugliness of smps.
If you insist on using semiconductor diodes i foretell a long, interesting and diverse thread with lots of illuminating advice, none of which works once the ears pass judgment
PN junction diodes exhibit a reverse recovery spike, AKA switching noise. UFnnnn diodes generate little noise and that noise is easily snubbed.
However, high PIV Schottky diodes are switching noise free.
Scan the archives for my posts about "hash" filtration. Large valued cap. I/P filters generate noise, even when the quietest diodes are employed.
However, high PIV Schottky diodes are switching noise free.
Scan the archives for my posts about "hash" filtration. Large valued cap. I/P filters generate noise, even when the quietest diodes are employed.
Do you listen to the diodes, or your own diy amps with ss diodes, to come to that conclusion? Or have you listened to any commercial amps?
I listened only to my DIY amps with solid state diodes.
My schematic for a power supply will be:
Solid state diode bridge with each diode bypassed 10nf ceramic
then 2uf PIO paralell with 100nf film cap paralell with say 200uf electrolyte
then 5H choke
then say 200uf electrolyte paralell with 100nf film cap
and then output transformers
then additional RC network (say 5k resistor and 10uF PIO cap) to the driver stages.
And I still found it to sound harsher and worse than solid state rectifier....
as the diodes I tried BYV26E and sic Cree C3D02060F.
The sic tent to sound liitle bit better but that could easily be my own imagination. But the difference between solid state and vacuum tube rectifier is like night and day.
Best regards, Michal
My schematic for a power supply will be:
Solid state diode bridge with each diode bypassed 10nf ceramic
then 2uf PIO paralell with 100nf film cap paralell with say 200uf electrolyte
then 5H choke
then say 200uf electrolyte paralell with 100nf film cap
and then output transformers
then additional RC network (say 5k resistor and 10uF PIO cap) to the driver stages.
And I still found it to sound harsher and worse than solid state rectifier....
as the diodes I tried BYV26E and sic Cree C3D02060F.
The sic tent to sound liitle bit better but that could easily be my own imagination. But the difference between solid state and vacuum tube rectifier is like night and day.
Best regards, Michal
MikyK,
Did you ever discover what really caused the harsh sound when you used solid stage rectifiers.
Was it possibly caused by the very hard high frequency transient caused by solid state rectifiers and a capacitor input filter. It could introduce a high frequency ground loop into the sensitive amplifier circuitry?
Also, the transient current load on the high voltage secondary can also reflect transient noise onto the filament secondary.
Even when I have to use a cap input filter, I never put snubber caps across the rectifier diodes.
Whenever possible, I use solid state rectifiers in combination with a Choke input filter.
The ripple is much smoother, low frequency, no high frequency components.
Less liable to produce ground loops.
Less liable to induce noise onto the filament secondaries.
Just my experience and opinions.
Did you ever discover what really caused the harsh sound when you used solid stage rectifiers.
Was it possibly caused by the very hard high frequency transient caused by solid state rectifiers and a capacitor input filter. It could introduce a high frequency ground loop into the sensitive amplifier circuitry?
Also, the transient current load on the high voltage secondary can also reflect transient noise onto the filament secondary.
Even when I have to use a cap input filter, I never put snubber caps across the rectifier diodes.
Whenever possible, I use solid state rectifiers in combination with a Choke input filter.
The ripple is much smoother, low frequency, no high frequency components.
Less liable to produce ground loops.
Less liable to induce noise onto the filament secondaries.
Just my experience and opinions.
Interesting point. I cant use choke input filter since the voltage drop will be too high and I cant do that.
Can I use small resistor (say 50-100 ohms) between rectifier and any filter that will be after it to smoothen the peaks a little bit?
After this resistor I can do all bypasses to all caps, the choke etc.
Best regards, Michal
Can I use small resistor (say 50-100 ohms) between rectifier and any filter that will be after it to smoothen the peaks a little bit?
After this resistor I can do all bypasses to all caps, the choke etc.
Best regards, Michal
Don't. This isn't the proper way to snubber.
Reduce the input cap value, maybe to 22-47uF. You may increase the second cap value if you want less ripple.
And if your electrolytic caps are even remotely decent, don't bypass them by anything - that causes more problems than it solves.
Be careful of the choke's magnetic field.
The same goes for the power transformer.
Keep some space between the output transformers, to the choke and power tranformer.
Keep the choke and power transformer laminations at right angles, versus the output transformers laminations.
Do not use a magnetic steel chassis, the choke and power transformer fields will be magnetically conducted to the output transformers.
Do not ask how I know about magnetic steel chassis.
A series resistance from the solid state diodes to your capacitor input filter will tend to soften the transient currents, but the more resistance you use, the more it will be like using a tube rectifier (it will lower the B+ voltage).
As expensive as it is, perhaps a new power transformer with more B+ secondary voltage, for either of the following:
Tube Rectifier and a Cap input filter
Solid state rectifier and a choke input filter
Or, with even more B+ secondary voltage, a tube rectifier and a choke input filter.
The same goes for the power transformer.
Keep some space between the output transformers, to the choke and power tranformer.
Keep the choke and power transformer laminations at right angles, versus the output transformers laminations.
Do not use a magnetic steel chassis, the choke and power transformer fields will be magnetically conducted to the output transformers.
Do not ask how I know about magnetic steel chassis.
A series resistance from the solid state diodes to your capacitor input filter will tend to soften the transient currents, but the more resistance you use, the more it will be like using a tube rectifier (it will lower the B+ voltage).
As expensive as it is, perhaps a new power transformer with more B+ secondary voltage, for either of the following:
Tube Rectifier and a Cap input filter
Solid state rectifier and a choke input filter
Or, with even more B+ secondary voltage, a tube rectifier and a choke input filter.
Last edited:
Thanks for the magnetic field advice too.
Im using solid aluminium plate, choke and power transformer will be mounted from the bottom and output transformers from the top.
Output transformers will also have its own steel covers with thickness of I dont know, maybe one milimeter.
I will start by making the simplest power supply possible without any bypasses and then move my way up, listening to it and trying to find out which will sound the best.
I dont have the best electrolytic capacitors. Only some Philips LL 40/085/56 and Nichicon CE.
I still think about buying new KEMETs. But they are pretty expensive.
best regards, Michal
Im using solid aluminium plate, choke and power transformer will be mounted from the bottom and output transformers from the top.
Output transformers will also have its own steel covers with thickness of I dont know, maybe one milimeter.
I will start by making the simplest power supply possible without any bypasses and then move my way up, listening to it and trying to find out which will sound the best.
I dont have the best electrolytic capacitors. Only some Philips LL 40/085/56 and Nichicon CE.
I still think about buying new KEMETs. But they are pretty expensive.
best regards, Michal
To the power transformer. I have one winding 200V 0.3A and 30V 0.1A which was used as fixed bias supply in its old purpose.
I can tie these together and get 230V 0.1A, but now the current capability will drop significantly. With this I will be able to put out only roughly 0.1 * 0.71 = 71mA of current. Im I right?
This thread over on the power supply forum is a great source of info on snubbers. Simple, no-math transformer snubber using Quasimodo test-jig
There is another thread where people report values for various transformers.
There is another thread where people report values for various transformers.
I only have my limited experience to go by, but I made a tube headphone amp with UF4007 rectifiers then LCRC (choke input) PSU that I think sounds quite good, nice and smooth. I figure I should be able to hear problems with headphones, since there's no room sound to obscure things.
You may be right about choke input supplies. They might actually make a difference in subjective sound quality.
--
+1 on that great thread. I use the quasimodo technique described there with this Q tester works great. Random snubbers are hit or miss, but a snubber that is RC tuned exactly for your transformer and diode combo is provable on the scope. And its so easy, attach a trimmer, watch the scope for total ring suppression, replace the trimmer with a resistor value of your trimmer.
https://www.ebay.com/sch/i.html?_from=R40&_trksid=p2047675.m570.l1313&_nkw=anatek+q+tester&_sacat=0
Hey Eli -- Are you thinking this is for use when the B+ and heater windings are on the same transformer? Or also for when you have separate transformers for the plate and heater supplies?