Wow, what a thread. Must be interesting but just too long for me to even try to get into. Can someone write down the conclusion and post it in a new seperate thread? Thnx.
That'll be hard. But in a few words, Salas has developed a high voltage shunt regulator which a number of people have implemented and have been really happy with. The latest schematic is somewhere in this thread, sadly, I don't know exactly where.
My personal summary of this thread:
1. According to the late Allen Wright, the power supply quality has 50% responsibility for the sound quality of an amplifier.
2. Shunt Regulators are the best possible regulators - wherever it's possible to use them (it's usually impossible to use them with reactive load, like output transformer or inter-stage transformer{s}). They are the best possible regulators both from the technical-engineering view and from listening reports.
3. Salas Shunt Regulators are easy to build, even with point-to-point wiring, and according to the testimonies of many users - give excellent results.
My bottom line:
The Salas HV Shunt Regulator is the most highly recommended solution to any tube pre-amp, including phono stages - and to the voltage stage(s) of any tube power amp.
1. According to the late Allen Wright, the power supply quality has 50% responsibility for the sound quality of an amplifier.
2. Shunt Regulators are the best possible regulators - wherever it's possible to use them (it's usually impossible to use them with reactive load, like output transformer or inter-stage transformer{s}). They are the best possible regulators both from the technical-engineering view and from listening reports.
3. Salas Shunt Regulators are easy to build, even with point-to-point wiring, and according to the testimonies of many users - give excellent results.
My bottom line:
The Salas HV Shunt Regulator is the most highly recommended solution to any tube pre-amp, including phono stages - and to the voltage stage(s) of any tube power amp.
I know at least two or three of us are using it with output transformer loads without issue. This is a SSHV single feed 5k:32 OPT WE417, 2H-60db, the bump at 60hz(under -100db) is from the power transformer emi into the opt:
Indeed, however it looks like a shunt regulator cannot be guarantied to work on all amps with output or inter-stage transformers without any issue.
Hi Salas
I tried to review this thread up to page 192 and cannot find the circuit that matched the above like you said is the latest version. Would you tell me D5 and D7 should be 1N4003 or 12V zener?
Because some latest circuit showed that D5 to D8 is all 12V zener including the redrawn circuit by Mr. Quang Hao.
I just P-to-P again last night but it didn't work. ha ha ha .......................😛
Thanks
Tim
They can be two zeners protecting a gate, or a zener and a reverse biased diode. Firstly take them all types out if in doubt, see if your circuit works without those protections.There could be some other priority to fix or not.
Hi Salas
Problem solved. A very careless mistake:the regulator did not grounded !!!!
A preliminary test is the audible hum, it is very very quite !!!!! The first version I still got an audible buzz......y............., but this time it is nothing can be audible !!
Thanks for your masterpiece !!!!!
Now is mid night in Hong Kong, I have to play it tomorrow.😀
Tim
Problem solved. A very careless mistake:the regulator did not grounded !!!!
A preliminary test is the audible hum, it is very very quite !!!!! The first version I still got an audible buzz......y............., but this time it is nothing can be audible !!
Thanks for your masterpiece !!!!!
Now is mid night in Hong Kong, I have to play it tomorrow.😀
Tim
Nice. Did you leave the diode-zener protection connected to gates as in the drawn circuit when it got grounded and worked?
The diode - zener protection was disconnected and found no problem, the regulator turned on up to this time seems no problem too and I proposed not to connect this protection.
P-to-P of transistors circuit is very interesting than tube....😀
I know it works without gate protection but its there in the case of a bad transient ever comes and punches the gate oxide insulation. Yes, I remember your last P2P directly on chassis pictures. Show us from the new P2P too when you have.
OK , protection will be added up later, when I get lazy enough😀
I found that the 120K connected to 2SK170 is very hot! My finger cannot stand for few seconds on it, maybe one second something.
The regulator current set at 42ma something, R1 is 47 ohms, raw DC input is 380V after load. R3 set to 79K by series connection of 10K + 22K + 47K.
Also use red LED for ref voltage of 2SK170, for nice look purpose and indeed it looks nice !!!
Except the Solen cap., all others are very cheap components.
This morning I play the amp. and feeling better than before. Don't know use cosmetic words to describe the change, but to me , it is very "stereo" and "attractive".
I found that the 120K connected to 2SK170 is very hot! My finger cannot stand for few seconds on it, maybe one second something.
The regulator current set at 42ma something, R1 is 47 ohms, raw DC input is 380V after load. R3 set to 79K by series connection of 10K + 22K + 47K.
Also use red LED for ref voltage of 2SK170, for nice look purpose and indeed it looks nice !!!
Except the Solen cap., all others are very cheap components.
This morning I play the amp. and feeling better than before. Don't know use cosmetic words to describe the change, but to me , it is very "stereo" and "attractive".
Attachments
The brown color cheap metal film cap is for the by pass purpose of the lift up voltage for cathode follower filament.
Star ground used for the regulator , totally six ground points for the whole circuit if zener protection not included. That's the point that was missing to connect a wire to the raw DC ground point !!
Star ground used for the regulator , totally six ground points for the whole circuit if zener protection not included. That's the point that was missing to connect a wire to the raw DC ground point !!
Nice spider's web, traditional. Big gate resistors and long legs here, its not like in tight PCB with tiny gate stoppers, better avoid adding more components around the gate paths since it works reliably for you now, I am afraid it might get extra ghost inductance. There is ~1W on that 120K if you use it for 350Vout for example. Since you find that one does not take it well, use two 68k in series like that one, or one 120k double fat. Its good you could find it better by changing version, so it was not work for no reason. Regards.
Current Update?
Hey Salas,
Just wondering where this reg is currently at? I have Nov 09 dated schematics I am looking to use for +/-200v, is there an updated version i can use? P2P is ok, or I can get vero board?
Want to use these in a dac for the B+ on the output tubes, would appreciate a looksee if there is an update about? Trying to keep up to speed on this forum is a little hard sometimes, not sure how you do it!
Thanks,
Drew.
Hey Salas,
Just wondering where this reg is currently at? I have Nov 09 dated schematics I am looking to use for +/-200v, is there an updated version i can use? P2P is ok, or I can get vero board?
Want to use these in a dac for the B+ on the output tubes, would appreciate a looksee if there is an update about? Trying to keep up to speed on this forum is a little hard sometimes, not sure how you do it!
Thanks,
Drew.
Hi, I had given you that link before I think. I have changed nothing yet. P2P or vero are OK. At your voltage skip the zeners.
http://www.diyaudio.com/forums/powe...tic-mosfet-hv-shunt-regs-182.html#post2604016
http://www.diyaudio.com/forums/powe...tic-mosfet-hv-shunt-regs-182.html#post2604016
Technically if you want the achieve the lowest output impedance that the design might allow you should use force/sense connections. However, we started talking about the Kelvin connection only when I introduced the bjt buffer before the shunt mostfet on the low voltage circuit, where the design promised low output impedance well beyond 20kHz. In that case, there was no way to achieve it without the Kelvin connection. If the current Salas HV schematic doesn't have the buffer, then you don't worry about low output impedance at 20kHz anyway. The HV shunt reg that I built recently has both the bjt buffer and the force/sense wires, and low output impedance in a wide frequency range.
Hey Salas, I spent some time on my HV shunt reg recently and ran into an issue which might be an issue for you guys too.
I'm talking about the 2SK170 (Q4) and MJE350 (Q2), because I'm using this pair exactly in my reg too. I've had a number of them go puff because of spikes at turn on/off, although you'd think that the E-B junction of the MJE350 would not allow a high voltage across the 2SK170. A 12V zener across the 2SK170 fixed the issue entirely for me and I've been using the regulator trouble free ever since. The rest of the reg is different than yours, different CCS, and bjt buffer on the shunt mosfet, to get low Zout beyond 15kHz.
I'm talking about the 2SK170 (Q4) and MJE350 (Q2), because I'm using this pair exactly in my reg too. I've had a number of them go puff because of spikes at turn on/off, although you'd think that the E-B junction of the MJE350 would not allow a high voltage across the 2SK170. A 12V zener across the 2SK170 fixed the issue entirely for me and I've been using the regulator trouble free ever since. The rest of the reg is different than yours, different CCS, and bjt buffer on the shunt mosfet, to get low Zout beyond 15kHz.