Hi guys,
I've seen power supplies with the filter stage split after the rectifier valve for each channel. I was wondering wether it would be possible to split this at the transformer secondary. That is, use a separate rectifier and filter stage for each channel drawing from one transformer.
I have a hammond 373ex that is rated at 345ma with a 5V 6A tap. I've recently been given a pair of matched rectifiers so was looking at conducting some experiments with the PS. I'm expecting about 40-60ma draw each channel. I know it's a bit OTT but its more for learning than anything else.
Hope everyone had a nice Christmas and all the best for the new year!
I've seen power supplies with the filter stage split after the rectifier valve for each channel. I was wondering wether it would be possible to split this at the transformer secondary. That is, use a separate rectifier and filter stage for each channel drawing from one transformer.
I have a hammond 373ex that is rated at 345ma with a 5V 6A tap. I've recently been given a pair of matched rectifiers so was looking at conducting some experiments with the PS. I'm expecting about 40-60ma draw each channel. I know it's a bit OTT but its more for learning than anything else.
Hope everyone had a nice Christmas and all the best for the new year!
Separate rectification from same HT winding works just fine, but if you are using 5V rectifiers like 5U4 or 5AR4, you'll need a separate 5V filament winding for each rectifier- remember that the filament is the cathode (B+ out) in a directly heated 5U4, and in case of indirect heated 5AR4, the cathode sleeve is internally connected to filament. I used this topology (2 rectifiers, one HT winding) in a preamp power supply with 6AX5GT rectifiers (cathode isolated from filament).
You could also look at using TV damper diodes if you have a suitable heater winding available - they are robust, cheap, and have excellent heater-cathode insulation, allowing use of common heater supply, BUT, they are half-wave - you'll need 4 individual tubes to get two separately rectified supplies from one centre-tapped HT winding.
You could also look at using TV damper diodes if you have a suitable heater winding available - they are robust, cheap, and have excellent heater-cathode insulation, allowing use of common heater supply, BUT, they are half-wave - you'll need 4 individual tubes to get two separately rectified supplies from one centre-tapped HT winding.
Hi,
Thanks for the response. Would you say then that it would defeat the purpose if I paralleled the 5v 6A tap (not the valves) to heat two GZ37? I'm looking to be able to use just what I have and I don't have any other 5v transformers. Thanks.
If you're stuck with only one winding, but it is capable of powering 2 x 3A rectifier filament (the GZ37 has a 3A heater), then your option is to simply parallel the two rectifiers, adding a small value of resistance in series with each anode (22-47 ohms) to ensure equal current sharing. You'll not be able to obtain separate HT from each rectifier, because as previously described, the filaments are internally connected to cathodes, so they're commoned together by the 5V winding (of which you only have one). I hope this makes sense, rgds
That makes sense. Finally (hopefully), if I implemented paralleled rectifiers I could then split the filter stage of each channel. Would there be any theoretical advantage in doing all of this to power a low powered SET amp?
series connected rectifer filaments
I have a 10V CT winding on my PT.
I've read that you are not suppose to connect directly heated rectifier filaments in series. Maybe this only holds true if you are taking the B+ from one side of the filament?
What about two 5V4GA with filaments in series? The cathode is connected internally to the heater, but I would take my B+ from the 10V center tap.
Would it matter how it is series connected? 10V to pin 2, jumper from pin 8 to pin 2 of the second filament, then the other side of the 10V winding to pin 8 of the second rectifier filament?
I have a 10V CT winding on my PT.
I've read that you are not suppose to connect directly heated rectifier filaments in series. Maybe this only holds true if you are taking the B+ from one side of the filament?
What about two 5V4GA with filaments in series? The cathode is connected internally to the heater, but I would take my B+ from the 10V center tap.
Would it matter how it is series connected? 10V to pin 2, jumper from pin 8 to pin 2 of the second filament, then the other side of the 10V winding to pin 8 of the second rectifier filament?
You want to connect the cathodes to the center tap and take your B+ from there. Piece of cake.
All good fortune,
Chris
So, I am connecting the cathode side filaments of both rectifiers together, with my 10V being on the opposing sides. Then I am taking my B+ from the center tap of my 10V winding?
Scott
Scott
Hi!
I don't see any advantage of using separate rectifiers from a common transformer, except if the rectifiers cannot handle the total current.
I always split the last LC stage between channels from a common supply. This decouples the channels from each other
Best regards
Thomas
I don't see any advantage of using separate rectifiers from a common transformer, except if the rectifiers cannot handle the total current.
I always split the last LC stage between channels from a common supply. This decouples the channels from each other
Best regards
Thomas
Thanks! Are there any cons in attempting such a configuration? Apart from cost, weight and space of course. Just as a perspective, I am comparitively new to the hobby. I've just spent the last couple of months observing different Q and now I'm starting to move on to observing what the PS is doing. The reason I'm keen on attempting this is that my phono/preamp is choke input (20H then 12H) followed by split RC filters for each channel. My power amp is a straight CLC. This is why I'm interested in another configuration to observe and test.
Hi!
To look at the difference of just splitting the filter or not in isolation, does not make a lot of sense. The effect it might or might not have also depends on the entire amp design. If your CLC ist set up such that interaction between channels is already minimal, the effect will be negligible. For example if you use large Cs. I prefer medium sized capacitances (20-30uF). There it makes more sense to split since the cap alone would not proveid much decoupling between channels. I also use the ultrapath concept which gives a defined signal path from B+ to cathode for each stage. With this a decoupling makes a lot of sense.
If you want to explore the differences a PSU makes, choke input filters are worth trying. IME the better alternative.
Best regards
Thomas
To look at the difference of just splitting the filter or not in isolation, does not make a lot of sense. The effect it might or might not have also depends on the entire amp design. If your CLC ist set up such that interaction between channels is already minimal, the effect will be negligible. For example if you use large Cs. I prefer medium sized capacitances (20-30uF). There it makes more sense to split since the cap alone would not proveid much decoupling between channels. I also use the ultrapath concept which gives a defined signal path from B+ to cathode for each stage. With this a decoupling makes a lot of sense.
If you want to explore the differences a PSU makes, choke input filters are worth trying. IME the better alternative.
Best regards
Thomas
I am havingthe following situation: My transformer has only one Vac winding. For everything. The amp is a classical El34-pp in Ul.
I had very good results in using a Gz34-Ss-Hybrid-Bridge as it does not have a CT. The GZ 34 works into a 30uf Polyprop, than a 8H-choke, than 400uF Polyprop from here 360V to the Powertubes and 40H-50uF for each driver tube.
I would like to improve this setup further. On the one hand, the GZ34 needs to deliver 245ma currently and is a bit at its limits. On the other hand, I want to do transformer-coupling and provide the current for the driver stage (6n6p in differential mode, each triode takes 20mA, 80mA alone for the drivers) through the IT and before that through a CCS and shunt-regulated by a OC3VR150(basically the Karna setup of Lynn Olsen for the driver). This requires to have a 150V supply, the 360 V is currently a bit high and I need to destroy more than 200V...
My question:
Is it possbile to do the following:
- One Hybrid-Bridge stays as it is, so producing only for the power tubes 360V with a chain of 30uF-8H-400uF.
- a new Hybrid-bridge connected at the same HV-Winding, but working directly into a 8H or more choke-input-setup like 10H-70uF-40H-50uF for the drivers, delivering something around 250V initially which will be dropped by another 100V through another RC, so that we have a shunt regulated 150 V at the Anode of the driver tube.
So, I want isolation between drivers and power stage as much as possible, highest quality driver voltage with choke-input and shunt reg for the driver and from the beginning not produce a much to high voltage for what the driver in reality needs...all from one winding ?
What I am not sire about: When I try to imagine how the current flows: When using two bridges are the diodes of the one bridges not parallel to the diodes of the other bridge ? So, how can they work if they produce two different voltages ?
Or is the only option to let them produce the same voltage, so channel separation, but not driver/power-tube separation possible ?
I had very good results in using a Gz34-Ss-Hybrid-Bridge as it does not have a CT. The GZ 34 works into a 30uf Polyprop, than a 8H-choke, than 400uF Polyprop from here 360V to the Powertubes and 40H-50uF for each driver tube.
I would like to improve this setup further. On the one hand, the GZ34 needs to deliver 245ma currently and is a bit at its limits. On the other hand, I want to do transformer-coupling and provide the current for the driver stage (6n6p in differential mode, each triode takes 20mA, 80mA alone for the drivers) through the IT and before that through a CCS and shunt-regulated by a OC3VR150(basically the Karna setup of Lynn Olsen for the driver). This requires to have a 150V supply, the 360 V is currently a bit high and I need to destroy more than 200V...
My question:
Is it possbile to do the following:
- One Hybrid-Bridge stays as it is, so producing only for the power tubes 360V with a chain of 30uF-8H-400uF.
- a new Hybrid-bridge connected at the same HV-Winding, but working directly into a 8H or more choke-input-setup like 10H-70uF-40H-50uF for the drivers, delivering something around 250V initially which will be dropped by another 100V through another RC, so that we have a shunt regulated 150 V at the Anode of the driver tube.
So, I want isolation between drivers and power stage as much as possible, highest quality driver voltage with choke-input and shunt reg for the driver and from the beginning not produce a much to high voltage for what the driver in reality needs...all from one winding ?
What I am not sire about: When I try to imagine how the current flows: When using two bridges are the diodes of the one bridges not parallel to the diodes of the other bridge ? So, how can they work if they produce two different voltages ?
Or is the only option to let them produce the same voltage, so channel separation, but not driver/power-tube separation possible ?
You can run a choke input PSU and a capacitor input PSU from the same secondary and rectifiers. What you need to do is attach the choke input PSU directly to the rectifier output (i.e. where your current cap input PSU is). Then connect the cap input PSU to this same point via a single rectifier diode. This ensures that each PSU sees the rectifier arrangement it needs. The two PSUs may have to share a ground bus for lowest hum/buzz but as they are feeding the same audio circuit they would need just one ground anyway.
Thx a lot for the answers. Df96, If I understand you looking at the normal Hybrid-Schematic:
When we can assume that a second gz34 or any other double driode would be connected to the same cathodes of the solid-state rectifiers: so two gz34 with their anodes conected in parallel (not within the envelope, but both connected like shown above), but one gz34 with its connected cathodes works into the pi-filter and for the output tubes and the other would work into the choke-input filter and the driver stage: That would work and is your suggestion ?
I saw somewhere a schematic for that, but cant find it anymore. But basically one winding, one set of solid state diodes with virtual centertap/ground connected to ac and two gz 34 connected to the ac/ss-diodes but with cathodes working into different filters.
An externally hosted image should be here but it was not working when we last tested it.
When we can assume that a second gz34 or any other double driode would be connected to the same cathodes of the solid-state rectifiers: so two gz34 with their anodes conected in parallel (not within the envelope, but both connected like shown above), but one gz34 with its connected cathodes works into the pi-filter and for the output tubes and the other would work into the choke-input filter and the driver stage: That would work and is your suggestion ?
I saw somewhere a schematic for that, but cant find it anymore. But basically one winding, one set of solid state diodes with virtual centertap/ground connected to ac and two gz 34 connected to the ac/ss-diodes but with cathodes working into different filters.
Last edited:
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.