I will be building a PS for my ST-70. Use HEXFREDS to the following filter:
47 uF cap
ST70 choke
70 uF cap
ST70 choke
100 uF cap
no load output is about 510V.
Question is, do I want to put a resistor in the xfmr center tap return to drop voltage (figure somewhere near 100 ohm) or should I bias my EL34's a little lower?
This has got to be a question many have asked, but I couldn't find a thread on it.
Thanks.
47 uF cap
ST70 choke
70 uF cap
ST70 choke
100 uF cap
no load output is about 510V.
Question is, do I want to put a resistor in the xfmr center tap return to drop voltage (figure somewhere near 100 ohm) or should I bias my EL34's a little lower?
This has got to be a question many have asked, but I couldn't find a thread on it.
Thanks.
Put a pair of CL150 inrush current limiters between the ends of the power trafo's rectifier winding and the FREDs. The CL150s will slow the B+ rise down a little. That slight delay is enough time for the bias supply to rise and protect the EL34s against cathode stripping.
Connect the CT of the rectifier winding directly to ground. Take advantage of the extra B+ Volts. Use combination bias on the finals. 2X shared RC bias networks connect the cathodes of each pair of "finals" to ground. Size the bias resistor to drop the extra B+ Volts at the idle current you like. In addition to "burning" the extra B+ Volts, the RC bias networks compensate for minor matching errors and allow a single bias set pot. to work well with less than perfectly matched pairs.
Connect the CT of the rectifier winding directly to ground. Take advantage of the extra B+ Volts. Use combination bias on the finals. 2X shared RC bias networks connect the cathodes of each pair of "finals" to ground. Size the bias resistor to drop the extra B+ Volts at the idle current you like. In addition to "burning" the extra B+ Volts, the RC bias networks compensate for minor matching errors and allow a single bias set pot. to work well with less than perfectly matched pairs.
Maybe it does not matter here, but I have fixed bias (4 independent) for each output tube.
I like the specs on the CL150. Does it get fairly hot and reduce R or stay somewhat near the 5 ohm level?
Also, my controls consist of two independent timers driving relays, so I can delay turn on of B+ as long as I like, and control inrush with the second timer. So practically speaking, an NTC is not all that necessary here.
At the same time, you lost me on the combination bias. I expect to drop the 510V down to around 475V with the two chokes in series. So I basically need another 50-60V drop. Straight resistors concern me because now I just make regulation suffer. Series zeners sounds downright stupid.
What are you referring to?
I like the specs on the CL150. Does it get fairly hot and reduce R or stay somewhat near the 5 ohm level?
Also, my controls consist of two independent timers driving relays, so I can delay turn on of B+ as long as I like, and control inrush with the second timer. So practically speaking, an NTC is not all that necessary here.
At the same time, you lost me on the combination bias. I expect to drop the 510V down to around 475V with the two chokes in series. So I basically need another 50-60V drop. Straight resistors concern me because now I just make regulation suffer. Series zeners sounds downright stupid.
What are you referring to?
CL150s will not fully heat up in this situation. That's not a problem here, as you have more Volts than you need.
CL150s have 2 advantages over timers and relays. CL150s are inexpensive/simple and they keep the power trafo from going BOING 🙁 at turn on. The downside of NTC thermistors is that the protection they provide is gone, once they heat up. Don't turn the amp off and back on immediately.
Combination bias uses both cathode resistors of a fair size and an adjustable C- supply. Dropping some of the extra B+ Volts in cathode resistors of appreciable value does not impact B+ regulation and it stabilizes the operating point. Since you already have 4 bias pots. on hand, use a dedicated RC network between each EL34 cathode and ground.
BTW, if you keep the 1st filter cap. small, the power trafo will run cooler. Also, you can tweak the value of a small 1st filter cap. to hold the final B+ rail voltage down. Pile the energy storage up after the inductance.
Think about a small cap. after the FREDs. Follow with a LC section for each channel. You get better channel separation that way.
CL150s have 2 advantages over timers and relays. CL150s are inexpensive/simple and they keep the power trafo from going BOING 🙁 at turn on. The downside of NTC thermistors is that the protection they provide is gone, once they heat up. Don't turn the amp off and back on immediately.
Combination bias uses both cathode resistors of a fair size and an adjustable C- supply. Dropping some of the extra B+ Volts in cathode resistors of appreciable value does not impact B+ regulation and it stabilizes the operating point. Since you already have 4 bias pots. on hand, use a dedicated RC network between each EL34 cathode and ground.
BTW, if you keep the 1st filter cap. small, the power trafo will run cooler. Also, you can tweak the value of a small 1st filter cap. to hold the final B+ rail voltage down. Pile the energy storage up after the inductance.
Think about a small cap. after the FREDs. Follow with a LC section for each channel. You get better channel separation that way.
Now I think I follow you. 1K resistor between cathode and B- equals 50V drop, effectively lowering my B+ supply. Cap in parallel to bypass audio.
That's gotta be a big cap, I would think, to be on the output stage. Isn't that why the original design kept a low v-drop across the cathode resistor, effectively eliminating the need for the cap?
That's gotta be a big cap, I would think, to be on the output stage. Isn't that why the original design kept a low v-drop across the cathode resistor, effectively eliminating the need for the cap?
One last thought: I'm not using any of my filament windings on the pwr tranny. What about bucking the secondary?
I think you mean wiring 1 or more of the filament windings in the bucking configuration. You wire the filament winding(s) in series with the PRIMARY.
You have considerable flexibility in adjusting the B+ rail voltage. Resort to combination bias is probably not going to be needed.
You have considerable flexibility in adjusting the B+ rail voltage. Resort to combination bias is probably not going to be needed.
I guess it's a matter of nomenclature.
In my industry, the primary and secondary do not determine high and low voltage, but rather direction of positive power flow.
Therefore the primary of ST70 is 120V, secondary is 720VCT.
In the end, I could either buck the 720V 'secondary' with one or more filament windings (just reverse polarity), or I could add additional turns to the 120V 'primary' (keep polarity in phase).
Come to think of it, I kind of like the idea of adding more turns to the primary; that way, the 720V winding simply puts out less voltage.
Tranny runs at lower VA for a given current, instead of wasting watts by dropping B+ with resistors. Fewer concerns about regulation.
Not bad at all.
<|:^)
In my industry, the primary and secondary do not determine high and low voltage, but rather direction of positive power flow.
Therefore the primary of ST70 is 120V, secondary is 720VCT.
In the end, I could either buck the 720V 'secondary' with one or more filament windings (just reverse polarity), or I could add additional turns to the 120V 'primary' (keep polarity in phase).
Come to think of it, I kind of like the idea of adding more turns to the primary; that way, the 720V winding simply puts out less voltage.
Tranny runs at lower VA for a given current, instead of wasting watts by dropping B+ with resistors. Fewer concerns about regulation.
Not bad at all.
<|:^)
My suggestion is to reduce the cathode resistors in the output tubes to as close to nil as you can manage!
Lower = cleaner bass each and every time it has been tried by anyone I know in a ST-70.
The purpose of those cathode resistors is/was to make it easy to bias the tubes with an old time Simpson or VTVM, and to act as fuses in case the tubes runaway or short.
Size the new lowZ resistors appropriately to maintain the "fuse" capability if you want that...
One could consider a choke input filter, given the higher secondary voltage??
The problem with import EL34s is that they will not handle the screen current that NOS tubes would have, and will flame out faster if pushed harder, which is what will happen with a higher B+. Until they fry they will sound ok fine...
One can consider means to decrease the screen current, but then you get less max output! (dog chases tail?)
Then too, I think they sound best rigged as triodes, not as UL pentodes or pure pentodes... so the power drops some in that case too.
After having worked with them for a long time, and having at least one audio friend who went seriously bonkers into them (now quite a bit more than a decade back...) the overall conclusion was that it is better to have them sound good than to put out more power... since you can't seem to get both.
Oh btw, the more C you can put in the spot that feed the outputs, the better... which then leads to wanting to put in a heftier power transformer... (dog chases bone?) 😀
As always, ymmv.
_-_-bear
Lower = cleaner bass each and every time it has been tried by anyone I know in a ST-70.
The purpose of those cathode resistors is/was to make it easy to bias the tubes with an old time Simpson or VTVM, and to act as fuses in case the tubes runaway or short.
Size the new lowZ resistors appropriately to maintain the "fuse" capability if you want that...
One could consider a choke input filter, given the higher secondary voltage??
The problem with import EL34s is that they will not handle the screen current that NOS tubes would have, and will flame out faster if pushed harder, which is what will happen with a higher B+. Until they fry they will sound ok fine...
One can consider means to decrease the screen current, but then you get less max output! (dog chases tail?)
Then too, I think they sound best rigged as triodes, not as UL pentodes or pure pentodes... so the power drops some in that case too.
After having worked with them for a long time, and having at least one audio friend who went seriously bonkers into them (now quite a bit more than a decade back...) the overall conclusion was that it is better to have them sound good than to put out more power... since you can't seem to get both.
Oh btw, the more C you can put in the spot that feed the outputs, the better... which then leads to wanting to put in a heftier power transformer... (dog chases bone?) 😀
As always, ymmv.
_-_-bear
That's very good advice, bear. I was on fence about adding more cathode resistance.
In the end, I'll add turns to the primary. Did a quick test last night, and output was around 460 at the first cap bank. I can deal with that, especially with two chokes in series.
And I will take your advice about some additional, local capacitance at the output stage. Maybe an ASC PP or something. I don't think I'll bother with PIO caps until I do my SET project.
Thanks, everyone !
In the end, I'll add turns to the primary. Did a quick test last night, and output was around 460 at the first cap bank. I can deal with that, especially with two chokes in series.
And I will take your advice about some additional, local capacitance at the output stage. Maybe an ASC PP or something. I don't think I'll bother with PIO caps until I do my SET project.
Thanks, everyone !
What is the actual B+??
Without tubes? (don't blow the filter caps!!)
With tubes biased up??
That would be how I decided what to do...
if the choke input would give you enough B+ then I think I'd give that some consideration...
_-_-bear
Without tubes? (don't blow the filter caps!!)
With tubes biased up??
That would be how I decided what to do...
if the choke input would give you enough B+ then I think I'd give that some consideration...
_-_-bear
Bear makes an interesting point about a choke I/P filter. The ST70 power trafo's rectifier winding is 720 VCT. That would become about 324 VDC, with choke I/P filtration. Of course, 324 V. is way too low for B+, but the "flapping" filament windings could be configured to boost the voltage the rectifier winding produces. Wire them in series with the primary.
Choke I/P filters are inherently well regulated and the continuous draw from the rectifier winding leads to a cool power trafo. 🙂
Choke I/P filters are inherently well regulated and the continuous draw from the rectifier winding leads to a cool power trafo. 🙂
Unfortunately, bear, I do not have the rest of the circuit built. I am building as I go, and only loading the PS with resistors to simulate a load (and approximate at best).
So, with no tubes, the first cap bank runs at 510V. Loaded, we are close to 470V. I can bring the loaded results down closer to 440 by adding turns to the 120V winding.
When I actually get some load tests (probably Thursday before I can continue building), I will advise as to the success.
With all the ST70 mods out there, I'm a little surprised no one has advised on their successes in swapping the GZ34 out with SS. Simple voltage drop tells us we have a higher B+. No doubt someone has done this without significant circuit modifications, and have probably not had the luxury of free filament windings to play with?
So, with no tubes, the first cap bank runs at 510V. Loaded, we are close to 470V. I can bring the loaded results down closer to 440 by adding turns to the 120V winding.
When I actually get some load tests (probably Thursday before I can continue building), I will advise as to the success.
With all the ST70 mods out there, I'm a little surprised no one has advised on their successes in swapping the GZ34 out with SS. Simple voltage drop tells us we have a higher B+. No doubt someone has done this without significant circuit modifications, and have probably not had the luxury of free filament windings to play with?
Have you seen the ST-70 rebuild by Joe Curcio that appeared in the very first issue of Glass Audio?
Finally, some good news !!
I am not sure, SY, as I have not seen the first issue of Glass Audio. However, I will do some searching to try and find it.
I can say this much; I am very familiar with Joe Curcio's current design, and have used some of his ideas (in combination with ideas from others) to build my own little custy.
This is part of what stumps me; Curcio's power supply upgrade does not appear to address the higher B+ voltage one can expect to find. No doubt Curcio is head and shoulders above me, so what's the deal? Does he simply run his output tubes that much hotter? There was no indication he fundamentally changed anything in the output stage (cathode resistor split between tubes, but still same 50mA per tube).
Anyways, maybe some good news for everyone. I did some more experimentation, and may have a slick little solution (IF one is concerned about hotter B+). I tried a number of iterations using various filament windings, and this method seems to work best:
Since you are using SS rectification, you no longer need the 5V filament winding off the pwr tranny. Wire it in series with the 120V primary winding, with polarity such that it adds turns to the 'new primary' winding.
Connect 120VAC up to the full 'new winding' (i.e. one of the 'old' 120V winding leads will not be directly excited, merely tied to the 5V filament winding).
The net result will be a 720VCT secondary that is LESS than nominal. I did some load tests, and came up with the following results. Load resistances of 1500 and 3000 ohms were used (all I had available for a load). Measurements were taken such that I could compare to the ST70 manual, which specifies 435VDC at the output of the rectifier tube. My assumption is that later stages will be appropriate if I can be relatively close to this 435V target.
No load: output = 490V
Load of 3000 ohms: output = 442V
Load of 1500 ohms: output = 422V
So yeah, it's only a resistive load bank, but at least I'm pulling amps in the general vicinity of what I could expect.
Granted, I could be missing something very important here, but I think this method is slicker than snot on a doorknob. No additional parts are needed, and I use the 5V filament winding that I wasn't using anyway (but at very low current).
What do y'all think?
I am not sure, SY, as I have not seen the first issue of Glass Audio. However, I will do some searching to try and find it.
I can say this much; I am very familiar with Joe Curcio's current design, and have used some of his ideas (in combination with ideas from others) to build my own little custy.
This is part of what stumps me; Curcio's power supply upgrade does not appear to address the higher B+ voltage one can expect to find. No doubt Curcio is head and shoulders above me, so what's the deal? Does he simply run his output tubes that much hotter? There was no indication he fundamentally changed anything in the output stage (cathode resistor split between tubes, but still same 50mA per tube).
Anyways, maybe some good news for everyone. I did some more experimentation, and may have a slick little solution (IF one is concerned about hotter B+). I tried a number of iterations using various filament windings, and this method seems to work best:
Since you are using SS rectification, you no longer need the 5V filament winding off the pwr tranny. Wire it in series with the 120V primary winding, with polarity such that it adds turns to the 'new primary' winding.
Connect 120VAC up to the full 'new winding' (i.e. one of the 'old' 120V winding leads will not be directly excited, merely tied to the 5V filament winding).
The net result will be a 720VCT secondary that is LESS than nominal. I did some load tests, and came up with the following results. Load resistances of 1500 and 3000 ohms were used (all I had available for a load). Measurements were taken such that I could compare to the ST70 manual, which specifies 435VDC at the output of the rectifier tube. My assumption is that later stages will be appropriate if I can be relatively close to this 435V target.
No load: output = 490V
Load of 3000 ohms: output = 442V
Load of 1500 ohms: output = 422V
So yeah, it's only a resistive load bank, but at least I'm pulling amps in the general vicinity of what I could expect.
Granted, I could be missing something very important here, but I think this method is slicker than snot on a doorknob. No additional parts are needed, and I use the 5V filament winding that I wasn't using anyway (but at very low current).
What do y'all think?
- Status
- Not open for further replies.