I can understand using a shunt reg where there are comparatively small currents flowing, such as regulating pentode screen grids or a preamp's HT, but a whole stereo power amp?
I agree series regs can sound good. I will resist the temptation to wittier too much about my views in your thread, but if you have a output stage passing 50ma, what’s so hard about having a shunt valve also passing 50ma? If I call it a push pull output stage you would be happy using two valves passing 50ma.
Second, don't fall into considering the load being passive, a loudspeaker will push back, that back current will create a voltage across the tx primary. Series regs are very bad at dealing with over voltage, shunt regs are very good at that.
Whenever an honest man discovers that he's mistaken, he will either cease to be mistaken or he will cease to be honest.
Nick,
I would welcome you wittering about your views in this thread.
I discovered, reading back in the archive, that I'm waaay behind the curve these days. I suspect that things have moved on significantly. I mean, why use valves in a reg anyway, why not HV transistors? I expect that properly implemented, they would p*ss on any valve reg, but sadly I know next to nothing about solid state, so wouldn't know where to start with an SS reg.
It would be helpful if I could be pointed to something.
Sgt. Baker started talkin’ with a Bullhorn in his hand.
If you're going SS you could start with an MJ "Maida regulator" (based on LM317 or LT3080) but he says that the "Statistical regulator" is quieter and more robust.
Cheers Nick,
It's a new world of PCB design and all that stuff.
I'm sure if I tried lashing any solid state reg up it would oscillate itself into meltdown.
What are the things you need to consider when laying out something like this.
Sgt. Baker started talkin’ with a Bullhorn in his hand.
Iirc your mj statistical reg board was capable of 20ma or so, so presumably then if the zeners were a different higher current capable type, and the number of them altered to give the correct v, and the 3 pin regs scaled appropriately then the concept would be scaleable to ht voltages and currents?
The difference that board made to my nuvistor phono stage was not subtle, to a power amp, that kind of regulation could make a big difference
Yes,
The statistical reg would be just the ticket if beefed up somehow and several could be used to isolate each individual valve, providing said valves with their own regulated supplies, sharing the dissipation requirements instead of concentrating it all in one place and also reducing unwanted interaction between stages/channels.
Last edited by Cressy Snr on Sat Aug 19, 2017 12:16 pm, edited 2 times in total.
Sgt. Baker started talkin’ with a Bullhorn in his hand.
pre65 wrote: ↑Sat Aug 19, 2017 11:58 am
Steve, I thought you were going to recapture something wonderful from the past ?
If you do things in a vastly different way, will the objective be reached ?
That's the dilemma I'm facing, but equally, I want to be open-minded enough to consider the ways that these things have been taken forward over the five years since I did this amp. For example Steve S has done wonders with the power supply for his PX25 setup by using individual supplies for each valve. The differences between this and his old amp are not subtle, and on his new baffles in his house, they sound stunning. In fact I was shocked at how well my Flatback speakers sounded on his amp too. It made them sing and caused me to reconsider what it is I actually want from my system. That and a few emails exchanged with Simon, concerning speaker design sort of sealed the decision to re-investigate my old SE amps.
Sgt. Baker started talkin’ with a Bullhorn in his hand.
Ant wrote: ↑Sat Aug 19, 2017 11:55 am
Iirc your mj statistical reg board was capable of 20ma or so, so presumably then if the zeners were a different higher current capable type, and the number of them altered to give the correct v, and the 3 pin regs scaled appropriately then the concept would be scaleable to ht voltages and currents?
The difference that board made to my nuvistor phono stage was not subtle, to a power amp, that kind of regulation could make a big difference
MJ regards the statistical regulator as a general HT supply - it would have to be up-rated for general HT use and I would definitely make changes a second time round.
For a start, I wouldn't be so anal about making it small - the thermal dissipation from even that board was quite high - the zeners get hot - space is needed. And airflow.
I could re-lay the board for higher currents - I'd also downrate one of the FETs and probably move to the "improved" constant current source (only a couple of extra passives).
If you want some boards to play with, I have a few left and I'll be back in the UK next week for a couple of weeks so I could stick them in the post.
In the phono I was running it flat out and it wasnt too bad for heat, the small size made it abit fiddly for assembly in terms of the zeners but other than that, for a preamp it was spot on
A physically larger board for a power amp wouldnt be an issue.
Ive got another board, you sent me 2.
It'd be interesting to see what making the nuvistor board dual mono did to it, it all needs redoing anyway
Last edited by Ant on Sat Aug 19, 2017 12:37 pm, edited 1 time in total.
Ok, question to Nick and Nick . If the signal return path is mainly via the shunt, which is SS, how do you minimise potential SS colourations which I suspect would be most apparent in the HF ( which was wonderful on Steve's original amps ) ? I'm thinking about the usual non-linear capacitance effects in Mosfets for example. What's that Q1 thing in the circuit above ?
"Once you find out ... the Circumstances ; then you can go out"
Well, off the top of my head, the non linear capacitance thing is a issue when the voltage varies. The voltage across the shunt wont be carrying much anyway, so the varying capacitance is less of a issue. And probably more important, the variable capacitance is more relevant between the gate and the source, the signal path is not going that route. And finally you will probably have a nice cap of your choice across the shunt to prevent HF instability anyway.
Whenever an honest man discovers that he's mistaken, he will either cease to be mistaken or he will cease to be honest.
