So, how is the voltage doubler best implemented?
Playing with LTSpice, not real impressed with any
simulated doublers so far...
I don't want DC in the secondaries. I don't want
a boatload of caps. Don't want huge conduction
spike on the peaks. I've got BIG hammond choke,
the biggest they currently sell at PE to throw at
this problem. But doubler caps seem doomed to
defeat any PFC offered by a choke. What to do?
Can the spare primary winding be used to add to
the secondary instead? At 50% reduced VA rate?
And what of the bias winding?
Playing with LTSpice, not real impressed with any
simulated doublers so far...
I don't want DC in the secondaries. I don't want
a boatload of caps. Don't want huge conduction
spike on the peaks. I've got BIG hammond choke,
the biggest they currently sell at PE to throw at
this problem. But doubler caps seem doomed to
defeat any PFC offered by a choke. What to do?
Can the spare primary winding be used to add to
the secondary instead? At 50% reduced VA rate?
And what of the bias winding?
Last edited:
Hello
It is not just Spice. I have given up on doublers. I was not successful in getting rid of the hum in several attempts with a couple of different 120 to 120 isolation transformers.
DT
All just for fun!
It is not just Spice. I have given up on doublers. I was not successful in getting rid of the hum in several attempts with a couple of different 120 to 120 isolation transformers.
DT
All just for fun!
I was successful with them in guitar amps. Going to try now for hi-fi.
Edit: I mean voltage doublers.
Edit: I mean voltage doublers.
Last edited:
What I'm reading on transformer winding techniques
don't exactly sound promising that the alt primary is
well enough isolated to abuse as add-to secondary.
No way to know for sure, but tear one up, or hipot.
Again, what exactly you prefer for good doubler?
don't exactly sound promising that the alt primary is
well enough isolated to abuse as add-to secondary.
No way to know for sure, but tear one up, or hipot.
Again, what exactly you prefer for good doubler?
HF87 like many amps had voltage doubler - some hum in mine but its a mess with too high B+ I've heard of ok regulation with doublers
An externally hosted image should be here but it was not working when we last tested it.
Well, you can always do what the big dumb blonde one did. Use two transformers. They are cheap enough. I need 300 volts for the screen regulators and the input tubes. I need 600 volts for the output tube plates. I put a bridge rectifier on each secondary to make a 300 volt supply, then wired the two supplies in series. This gives me enough VA to run a 100 WPC version of the red board at full tilt with nothing smoking except the load resistors.
I did try a choke input voltage doubler once. Add a common mode choke in series with each diode. Since the current through each diode is equal, you can use an old power transformer that has 2 identical primary windings for a choke or you can buy the One Electron dual choke from AES. I didn't like the choke input deal since I lost too much voltage and it was a guitar amp.
If I build for stereo rather than monoblocks, stacking transformers is an option.
Only objection to doing so would be the pick up and move weight also doubles.
Unless the unified power were a third brick, and maybe thats the best answer?
Only objection to doing so would be the pick up and move weight also doubles.
Unless the unified power were a third brick, and maybe thats the best answer?
Last edited:
My "Kingfisher" amp uses a doubler power supply, and it's one of the most quiet amps I've done. Doublers were a standard feature in many Scott and Fisher products at the time and as the iron from this project came from a Fisher 500B (used a doubler), I did the same. Good diodes and a small loop area between transformer and filter caps are esssential.
Hello
I hope that it is not an over statement to say that voltage doublers present greater ripple that needs smoothing. I am looking forward to seeing some results posted here and seeing what level of filtering is required.
DT
All just for fun!
I hope that it is not an over statement to say that voltage doublers present greater ripple that needs smoothing. I am looking forward to seeing some results posted here and seeing what level of filtering is required.
DT
All just for fun!
Certainly a fair assessment. From playing with PSUD, it seems that a voltage doubler supply might require 4x the capacitance of a FWCT supply in order to achieve a similar level of residual ripple. If you can get satisfactory results with a 220uF cap on a FWCT supply, you'd likely need two 470uF caps for your doubler.
I see the advantage of the doubler is that the voltage rating of the caps need only be half of what you'd require for the "conventional" supply. That can equate to a big cost savings. The PIV across the rectifier diodes is also half, which might open up all kinds of possibilities.
Doublers yet again- when I finally get around to doing my GM70 monoblocks, I'm strongly considering using a ZVS flyback supply operating from doubled mains with a choke for partial power factor correction/peak smoothing. This approach was used for a while in the better PC power supplies before active PFC became pretty much ubiquitous. The chokes were wound on a small, chunky stack of lams for compactness. This approach avoids the extra complexity and EMI burden of an active PFC stage, but offers some of the benefits at the cost of a somewhat lower doubled B+. I'm a switching power supply designer by trade (30+ years experience), but in my hobby side, I'd rather concentrate on the amp than the power supply, so anything that makes the supply less complex is welcome.
On the subject of the doubled supply for the "Kingfisher", I used a pair of ~390uf (maybe more), 250Vcaps (midget computer electrolytics) to double my B+. Modern electrolytics in the 200-250V range are extremely compact, so adding extra capacitance to meet the ripple requirements is no big burden. Even the 400-450V caps aren't all that big, and all are pretty plentiful either new or as surplus.
On the subject of the doubled supply for the "Kingfisher", I used a pair of ~390uf (maybe more), 250Vcaps (midget computer electrolytics) to double my B+. Modern electrolytics in the 200-250V range are extremely compact, so adding extra capacitance to meet the ripple requirements is no big burden. Even the 400-450V caps aren't all that big, and all are pretty plentiful either new or as surplus.
Last edited:
I don't see it happing with one choke. Caps required to double cause all
the conduction to happen at the peaks regardless choke filter afterward.
Maybe one choke per diode, how big chokes are we talking here?
the conduction to happen at the peaks regardless choke filter afterward.
Maybe one choke per diode, how big chokes are we talking here?
It's one choke, located at the cap center tap (for 110V doubler). The computer supplies used two sets of windings, switched in for 115-230V operation.
I just ran the Orcad sim for the classic voltage doubler with inductor in series with the center point of the two caps. A few tens of mH brings down the peak current at the cost of about 40V of B+. A short stack of lams or a largish powdered iron core will get you this without a whole lot of trouble.
Leakage inductance of the power transformer doesn't already amount to that?
My sim doesn't account for any such leakage yet.
Something probably ought be done. In my sim, peak secondary currents spike
up to 1.6A! (100uF caps in doubler). Is that a big deal or completely normal?
And we aren't talking inrush here.. This is after stabilizing out at 400V 120mA.
My sim doesn't account for any such leakage yet.
Something probably ought be done. In my sim, peak secondary currents spike
up to 1.6A! (100uF caps in doubler). Is that a big deal or completely normal?
And we aren't talking inrush here.. This is after stabilizing out at 400V 120mA.
Last edited:
Short pulses of 1.6 amps. That's what makes a Hammond buzz.
You can reduce it a little by using smaller caps in the doubler, followed by one or two chokes (std or common mode made from an old power transformer) followed by the usual BFC.
I run a 480 volt industrial transformer into a v doubler made with 5AR4's using a single choke in my 845SE amp. I get 1050 to 1100 volts depending on how hot I set the bias. It is about 5 years old and still has the original tubes. I don't have the cap values here at work. The first set are small and the set after the choke are not!
You can reduce it a little by using smaller caps in the doubler, followed by one or two chokes (std or common mode made from an old power transformer) followed by the usual BFC.
I run a 480 volt industrial transformer into a v doubler made with 5AR4's using a single choke in my 845SE amp. I get 1050 to 1100 volts depending on how hot I set the bias. It is about 5 years old and still has the original tubes. I don't have the cap values here at work. The first set are small and the set after the choke are not!
I believe one can take say an ICT and build either a FWB cap input with the secondaries in series or a doubler with the secondaries in parallel, and after scaling component values,one will obtain approximately the same performance, same ripple, regulation, same peak primary current peaks. The doubler does take more capacitance (4x total) at 1/2 the working voltage, which nets out to the same storage capacity. One big difference is the secondary current peaks are 2X with the doubler, which will increase the voltage drop across the rectifier even though there are fewer diodes in series for each conduction cycle. I think in both cases the rectifiers see approximately the DC output voltage when reversed. In reality the doubler is slightly less efficient, has a little worse regulation but it's not a significant difference in my experience.
I've been building doublers and cap input bridge supplies using machine tool control transformers and have not had any noticeable mechanical buzz from the transformers up to at least 40% of their VA rating.
I've been building doublers and cap input bridge supplies using machine tool control transformers and have not had any noticeable mechanical buzz from the transformers up to at least 40% of their VA rating.
So Wrenchone
You are the guy responsible for all the harmonic laden Power Factor shifted building AC making all the K rated building transformers run hot. Lots of harmonics on those neutral conductors. It is not my VFD’s it is your computer and lighting power supplies that does it.
Nice Ridgid pipe wrench, nice rust.
More seriously thanks for bringing up the value of a choke for Power Factor Correction. It is usually capacitors switched in by the utility to correct for motor loads.
DT
Just for fun!
You are the guy responsible for all the harmonic laden Power Factor shifted building AC making all the K rated building transformers run hot. Lots of harmonics on those neutral conductors. It is not my VFD’s it is your computer and lighting power supplies that does it.
Nice Ridgid pipe wrench, nice rust.
More seriously thanks for bringing up the value of a choke for Power Factor Correction. It is usually capacitors switched in by the utility to correct for motor loads.
DT
Just for fun!
Putting an inductor in series with the center cap feed point (from xfmr) for a doubler causes a loss of regulation since AC current occurs there (omega L in series). Either have to keep the inductor small (mH) or use a split inductor and put each half in one of the DC diode pathes. I don't believe a common mode inductor made from an un-gapped xfmr will work, since it doesn't store any energy that way.
A trick you might try (recommend trying first on simulator or PSUD) would be to put an inductor and a cap (AC rated) in series at the center feed point. Make them resonant at 60 Hz. That way only smooth 60 Hz current can pass thru, but with low series reactance to hold the regulation better. You likely have to tweek the AC cap value a little (for 60 Hz resonance) to allow for the other electrolytics in series with it. But they will be much larger, and have only a small effect on the series equivalent.
A trick you might try (recommend trying first on simulator or PSUD) would be to put an inductor and a cap (AC rated) in series at the center feed point. Make them resonant at 60 Hz. That way only smooth 60 Hz current can pass thru, but with low series reactance to hold the regulation better. You likely have to tweek the AC cap value a little (for 60 Hz resonance) to allow for the other electrolytics in series with it. But they will be much larger, and have only a small effect on the series equivalent.
Last edited:
The PC power supply designers used just an inductor without the extra capacitor (cheapskates, but every penny and fraction thereof counts in those designs). The loss of regulation was relatively unimportant, as the SMPS circuit following the filter takes care of that. The SMPS transformer is rewound with hotter turns ratio to make up for the loss in voltage introduced by the inductor. The designers were more concerned with controlling AC harmonic content to meet IEC specs rather than attaining a really high power factor. This takes a lot less finesse.
For my nefarious purposes, this is probably the approach I'll use, as I'd be feeding the doubler from the AC line to power an SMPS. Dirty (actually, less dirty than a straight capacitor input filter), simple, uggish. I'm concerned with reducting the peak current each cycle to charge the input caps, as well as stretching out the rectifier conduction angle. Harmonic reduction is a cherry on top of that sundae.
Those wanting to try passive PFC on a high voltage transformer secondary will need at least a few hundred millihenries, albeit at a lower current rating than needed at the AC input. The resonating capacitor would come in handy. I simulated the simple inductor-only approach last night, - I'll try throwing in a cap tonight to see what happens.
For my nefarious purposes, this is probably the approach I'll use, as I'd be feeding the doubler from the AC line to power an SMPS. Dirty (actually, less dirty than a straight capacitor input filter), simple, uggish. I'm concerned with reducting the peak current each cycle to charge the input caps, as well as stretching out the rectifier conduction angle. Harmonic reduction is a cherry on top of that sundae.
Those wanting to try passive PFC on a high voltage transformer secondary will need at least a few hundred millihenries, albeit at a lower current rating than needed at the AC input. The resonating capacitor would come in handy. I simulated the simple inductor-only approach last night, - I'll try throwing in a cap tonight to see what happens.
- Status
- Not open for further replies.