Hi,
I recently got a quad of impecable DeForest 450´s with their boxes and everything and obviously want to make them sing, and as I like to be over the top I thought of making a universal parallel single ended tube amp so I can use any UX4 base DHT. This is because I also have many other triodes and don´t want to bother myself (and my pocket) by having to make a new amp each time I want to see them light up.
Anyway, to do so I thought of dividing the thing into 2 chassis, one containing just the HT supply for the output stage and in the other the amp itself with its own PS for the preamp as well as the filament supply for the output tubes. What I wanted to discuss (really ask for advice on) is the regulated power supply, the demands are an output voltage of 230 to 470 volts at 20-250mA.
Now the thing is I want to use DHT´s for the pass tubes and use all vintage tubes (Just for the mojo, if I wanted to be pragmatic i´d just use a 6336 and a 6SN7), so I drew the schematic in the attachment and simulated the thing in PSpice. It works like a charm, but we all know things are different in real life so here is my long list of doubts:
1) Is it better to use the centertap in the filament winding than to make a virtual one using the 2x20R resistors? Both for the rectifiers and the pass tubes. If so, why?
2) Should I add the dotted line caps in parallel with the 874´s? I read regulators are noisy, and I suppose these are the noisiest of them all by far being the oldest ones. I don´t want to risk having them oscillate if it´s not necessary.
3) The selectable 30K resistor (labelled Bleeder) is there to be used when the load is small so as to increase it allowing the chokes to be greater than the critical inductance (not sure I expressed my self correctly). Is there a more effective way of doing this? I also thought of making the regulator tubes resistor (the 40k one) switchable.
4) Will the thermistors be effective in protecting the filaments from inrush current? I´ve never used this techique so I am looking for opinions, if not what are the alternatives.
5) To roughly vary the rectified voltage output (I won´t risk brutally dropping 400 volts thorugh the 2a3´s/300b´s) I thought of just making selectable taps at the B+ transformer secondary. Thing is that I don´t know if the usual phenolic wafer rotary switches will widthstand the voltage. Should I tap the primary instead?
6) I know the LCLC filtering prior to the regulator is pretty overkill, but is it a TOTAL overkill? I mean, would just an LC filter suffice? Maybe an LCRC? I should also mention that the caps have to be PIO (Again, for the mojo) so I won´t just drop 50u ´lytics.
7) If there is anything that could cause major troubled I would greatly appreciate if you could point out. If something fails I want the rest of the components to be safe, would it be effective to add a couple of other fuses around?
Well, thanks for reading (If you bothered to do so
) and any kind of input is greatly appreciated.
I recently got a quad of impecable DeForest 450´s with their boxes and everything and obviously want to make them sing, and as I like to be over the top I thought of making a universal parallel single ended tube amp so I can use any UX4 base DHT. This is because I also have many other triodes and don´t want to bother myself (and my pocket) by having to make a new amp each time I want to see them light up.
Anyway, to do so I thought of dividing the thing into 2 chassis, one containing just the HT supply for the output stage and in the other the amp itself with its own PS for the preamp as well as the filament supply for the output tubes. What I wanted to discuss (really ask for advice on) is the regulated power supply, the demands are an output voltage of 230 to 470 volts at 20-250mA.
Now the thing is I want to use DHT´s for the pass tubes and use all vintage tubes (Just for the mojo, if I wanted to be pragmatic i´d just use a 6336 and a 6SN7), so I drew the schematic in the attachment and simulated the thing in PSpice. It works like a charm, but we all know things are different in real life so here is my long list of doubts:
1) Is it better to use the centertap in the filament winding than to make a virtual one using the 2x20R resistors? Both for the rectifiers and the pass tubes. If so, why?
2) Should I add the dotted line caps in parallel with the 874´s? I read regulators are noisy, and I suppose these are the noisiest of them all by far being the oldest ones. I don´t want to risk having them oscillate if it´s not necessary.
3) The selectable 30K resistor (labelled Bleeder) is there to be used when the load is small so as to increase it allowing the chokes to be greater than the critical inductance (not sure I expressed my self correctly). Is there a more effective way of doing this? I also thought of making the regulator tubes resistor (the 40k one) switchable.
4) Will the thermistors be effective in protecting the filaments from inrush current? I´ve never used this techique so I am looking for opinions, if not what are the alternatives.
5) To roughly vary the rectified voltage output (I won´t risk brutally dropping 400 volts thorugh the 2a3´s/300b´s) I thought of just making selectable taps at the B+ transformer secondary. Thing is that I don´t know if the usual phenolic wafer rotary switches will widthstand the voltage. Should I tap the primary instead?
6) I know the LCLC filtering prior to the regulator is pretty overkill, but is it a TOTAL overkill? I mean, would just an LC filter suffice? Maybe an LCRC? I should also mention that the caps have to be PIO (Again, for the mojo) so I won´t just drop 50u ´lytics.
7) If there is anything that could cause major troubled I would greatly appreciate if you could point out. If something fails I want the rest of the components to be safe, would it be effective to add a couple of other fuses around?
Well, thanks for reading (If you bothered to do so
) and any kind of input is greatly appreciated.Attachments
For the first question, the answer you can find on this link:
The Valve Wizard
Regarding with fourth question, is preferable to use a soft starter or bypassing the thermistor after it's job been done.
The Valve Wizard
Regarding with fourth question, is preferable to use a soft starter or bypassing the thermistor after it's job been done.
Try this
PSUD2
That will give you a fair idea of ps speed/ripple etc
Pete millett has a good design for a dc regulated filament heater he used in his "unnecessarily complex 300B amp"
Regulated DC filament supply
Low Voltage Regulator
I wouldnt be keen on having hv thru a rotary switch, there are several examples of hv regulated dc supplies out there and there a some on pete milletts site. There are also soft start systems out there where the hv is brought up gradually once the heater filaments have warmed up. Of course you could use a gz34 which is an indirectly heated rectifier. That gives you about 15 seconds before it conduts. The regulated dc heaters will avoid inruch currents. Of course true Valvisti will insist on ac for your heaters. I think the c tap is grounded for that but check that before you do it
Good luck
PSUD2
That will give you a fair idea of ps speed/ripple etc
Pete millett has a good design for a dc regulated filament heater he used in his "unnecessarily complex 300B amp"
Regulated DC filament supply
Low Voltage Regulator
I wouldnt be keen on having hv thru a rotary switch, there are several examples of hv regulated dc supplies out there and there a some on pete milletts site. There are also soft start systems out there where the hv is brought up gradually once the heater filaments have warmed up. Of course you could use a gz34 which is an indirectly heated rectifier. That gives you about 15 seconds before it conduts. The regulated dc heaters will avoid inruch currents. Of course true Valvisti will insist on ac for your heaters. I think the c tap is grounded for that but check that before you do it
Good luck
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.