I'm trying to increase the B+ up to closer to 405VDCin this PS. As I reduce the choke size I go from
Figure A to Figure B. I understand there would be an increase in ripple with the smaller choke but the irregularity and "noise" increase seems out of proportion to the change. Things look really ugly if I lengthen the window. Figure C.
I seem to be able to get either much lower or much higher but can't find the sweet spot.
basic help would be appreciated.
Thanks
Figure A
Figure B
Figure C
Figure A to Figure B. I understand there would be an increase in ripple with the smaller choke but the irregularity and "noise" increase seems out of proportion to the change. Things look really ugly if I lengthen the window. Figure C.
I seem to be able to get either much lower or much higher but can't find the sweet spot.
basic help would be appreciated.
Thanks
Figure A
Figure B
Figure C
Last edited:
Figure C, Psd2 seems upset even the "noise" is only about 0.03%, a small cap of a few nF direct after the rectifier and /or some resistive load migth help.
LCLC circuits are tricky, need proper damping.
And b.t.w., with so much voltage and little current draw, first of all, put the bigger choke at the input
LCLC circuits are tricky, need proper damping.
And b.t.w., with so much voltage and little current draw, first of all, put the bigger choke at the input
Last edited:
If I remember correctly most valve rectifiers have a 47uf just after the rectifier before the choke.
The spec for rectifier will tell you maximum cap value.
The spec for rectifier will tell you maximum cap value.
Figure A looks ok, you have to reduce the resistance of the chokes if you want 405Vdc out.
Also, will propably need a bleeder resistor unless your caps can take 700V or more,, but getting 405Vdc with additional load from bleeder may be unobtainium.
Also, will propably need a bleeder resistor unless your caps can take 700V or more,, but getting 405Vdc with additional load from bleeder may be unobtainium.
Last edited:
Nigel, he has a 495V CT transformer and wants 405V dc. With C input, that would give almost 700V.
With L input he could get close to 400V if the total resistance of the chokes can be kept low enough (not much left for bleeder)
With L input he could get close to 400V if the total resistance of the chokes can be kept low enough (not much left for bleeder)
Last edited:
Yes the bigger the xl of the input choke the less ripple current the choke has to carry. If you go to small on the xl or inductance the rail voltage will ramp up way beyong your target voltage as already pointed out.choke input power supplies are tricky to impliment.must follow rules ,laws of physics.
Usually, but these rectifiers require a choke first.
I do have a voltage divider for elevating the filaments of driver and follower that acts as a bleeder.
So that resonance that shows up is obviously not desirable?
Any suggestions as to appropriate chokes?
So that resonance that shows up is obviously not desirable?
Any suggestions as to appropriate chokes?
What are your rquirements? ("Sweet spot" does not mean a thing)
MIN and MAX DC volts and currents?
(rememeber, at start up a cold tube does not take any current before it heats up)
What are the parts you alleady have and what are there ratings?
MIN and MAX DC volts and currents?
(rememeber, at start up a cold tube does not take any current before it heats up)
What are the parts you alleady have and what are there ratings?
My goal is 405B+ ( I now have 380V B+ )
Total draw is 106mA
I have 450-0-450 PT
8H @150Ω chokes x 2
Caps are 100uf rated at 600V. DC Link.
Not against buying other parts - not really budget limited - just trying to get to the B+ of the design.
I see I can get to 437 but the PSUD sim looks very ugly as in Fig C.
Thanks for the help!
Total draw is 106mA
I have 450-0-450 PT
8H @150Ω chokes x 2
Caps are 100uf rated at 600V. DC Link.
Not against buying other parts - not really budget limited - just trying to get to the B+ of the design.
I see I can get to 437 but the PSUD sim looks very ugly as in Fig C.
Thanks for the help!
Be aware that tubes have a max Ua0 rating, even at start up a tube should never see a higher voltage than that.
What tubes do you want to use?
MIN current?
The input choke inductance must be selected big enough to prevent voltage rise with minimum load (resistor load ???).
(be aware, that at start up a cold tube does not take any current before it heats up, so tube current does not count)
Need complete transformer rating (heater included), primary and secundary resistance, measure secundary from CT,
450-0-450 V transformer would need rectifier and chokes with zero voltage loss, not possible, you will need a higher transformer voltage to get 405Vdc
But I will look into it tomorrow, bit late here 😉
What tubes do you want to use?
MIN current?
The input choke inductance must be selected big enough to prevent voltage rise with minimum load (resistor load ???).
(be aware, that at start up a cold tube does not take any current before it heats up, so tube current does not count)
Need complete transformer rating (heater included), primary and secundary resistance, measure secundary from CT,
450-0-450 V transformer would need rectifier and chokes with zero voltage loss, not possible, you will need a higher transformer voltage to get 405Vdc
But I will look into it tomorrow, bit late here 😉
Last edited:
In my experience with PSUD2, I've found that massive damping is required to get the behavior looking okay for LC input. I think you're SOL with the given constraints here.
Is there a reason this needs to be LCLC? A lot of people seem to think that's the best way to do it, but I've never understood that. Ultimately it should be about transient response, lack of ringing, and lack of ripple, right? Many paths to that goal.
Personally, I would ditch the LCLC, combine the two Ls into one:
But you're only going to have ~385V B+ with ~125mV ripple. Neither of those aspects of this attempt are very attractive, I'd guess.
Where did you get the 495V from? When I use PSUD2, I put the parameters into the TX dialog screen and it converts to RMS:
I'm guessing you measured the 225 Ohm DCR for the TX.
Using RMS values we're even lower:
Somebody else might know better, but to me it looks like you should:
Is there a reason this needs to be LCLC? A lot of people seem to think that's the best way to do it, but I've never understood that. Ultimately it should be about transient response, lack of ringing, and lack of ripple, right? Many paths to that goal.
Personally, I would ditch the LCLC, combine the two Ls into one:
But you're only going to have ~385V B+ with ~125mV ripple. Neither of those aspects of this attempt are very attractive, I'd guess.
Where did you get the 495V from? When I use PSUD2, I put the parameters into the TX dialog screen and it converts to RMS:
I'm guessing you measured the 225 Ohm DCR for the TX.
Using RMS values we're even lower:
Somebody else might know better, but to me it looks like you should:
- Get a different TX so you can get the target B+ with adequate filtering,
- And/Or Use a different rectifier,
- And/Or redesign to use lower B+
The reason for the L input is that's what the 866A rectifiers need. No C input. No, I haven't measured the DCR of the PT. I just installed them in place of the Edcors that died an early death(one did). The new transformers, which were supposedly the same specifications, resulted in the lower B+. They were accurately wound and the Edcors were 40V higher. I really would like to avoid buying ANOTHER pair of transformers.
WntrMute2, you could gain a few volts (may be 15 or so) by replacing the rectifier tube with avalanche diodes costing only a few cent, are you open for that?
There could also be some other option, but i dont know enough about TX and your load, can you post the datasheet of the transformer and the circuit diagram of your load?
There could also be some other option, but i dont know enough about TX and your load, can you post the datasheet of the transformer and the circuit diagram of your load?
I'm afraid I'm married to the 866A rectifier tubes as I went to a lot of trouble to use them and I have a bunch on the shelf. Data sheets from Heyboer and Edcor are non-existent.
Edcor unloaded 981V (surviving one)
Heyboer unloaded 932V - so there is the problem in a nutshell. I don't want to use another 2 Edcors as it is mechanically noisy. Does this work?
Edcor unloaded 981V (surviving one)
Heyboer unloaded 932V - so there is the problem in a nutshell. I don't want to use another 2 Edcors as it is mechanically noisy. Does this work?
When I model that in PSUD2, it looks like the voltage never really settles anywhere.
5, 10, 15, 20 seconds in, zooming in on a 50 ms span--it's oscillating and all over the place.
I didn't use 225 Ohm for the TX, though. You really ought to measure the secondaries like Gorgon suggested so you can accurately simulate in PSUD.
You should get a higher voltage transformer, design a power supply for it, and enjoy. I think you're going to end up with lousy performance if you do anything else (given that you need to use this rectifier and a choke-input design). I don't know why you would go through all this trouble only to NOT create the best possible power supply within your budget. But maybe someone here knows how to get blood from this stone.
If the Edcors are noisy, you could try Hammond. I've used plenty of Antek toroidal transformers and have always been happy with them... but I've never run choke input or such highly-rated secondary voltages.
5, 10, 15, 20 seconds in, zooming in on a 50 ms span--it's oscillating and all over the place.
I didn't use 225 Ohm for the TX, though. You really ought to measure the secondaries like Gorgon suggested so you can accurately simulate in PSUD.
You should get a higher voltage transformer, design a power supply for it, and enjoy. I think you're going to end up with lousy performance if you do anything else (given that you need to use this rectifier and a choke-input design). I don't know why you would go through all this trouble only to NOT create the best possible power supply within your budget. But maybe someone here knows how to get blood from this stone.
If the Edcors are noisy, you could try Hammond. I've used plenty of Antek toroidal transformers and have always been happy with them... but I've never run choke input or such highly-rated secondary voltages.
WntrMute, my version of PSD2 givs a bit different results, but it shows clearly why the voltage is higher than it should be for a pure LC input.
Look a I(L), choke current goes negativ! Look at V(L1), see?
Real LC loaded with the 106mA constant current sink need a L1 with 4.3Hy minimum, but then your voltage would drop to 370V or so.
What you now have, is something in between C and L input
(actually, the so called C input is a RLC circuit with minimal R and L but large C).
Your circuit will work, sort of, but the voltage is highly depended on the actual inductance of L1 and your load current variations.
B.t.w., why now 0.5 ohm esr for the caps, earlier you used the 100 times smaller, correct value for DC link caps.
And what do you try to acheave with that 0.5ohm 100uF RC filter?
Still missing part of the info i asked for...
Look a I(L), choke current goes negativ! Look at V(L1), see?
Real LC loaded with the 106mA constant current sink need a L1 with 4.3Hy minimum, but then your voltage would drop to 370V or so.
What you now have, is something in between C and L input
(actually, the so called C input is a RLC circuit with minimal R and L but large C).
Your circuit will work, sort of, but the voltage is highly depended on the actual inductance of L1 and your load current variations.
B.t.w., why now 0.5 ohm esr for the caps, earlier you used the 100 times smaller, correct value for DC link caps.
And what do you try to acheave with that 0.5ohm 100uF RC filter?
Still missing part of the info i asked for...
Last edited: