I have a pretty high voltage tranny that has been sitting around... around 850V output after a bridge rectifier, and I want to drop it to a more sensable level... as in under 600v. Typically, I would just rig up a mosfet after the first bank of filter caps and choke with a voltage divider at the gate for such a thing, but I would like to avoid having to use 1KV capacitor bank/ choke at the input for money and space constraints, so would it be prudent to place a mosfet (1200v) voltage dropper right after the rectifier bridge and a small series resistor? Maybe a zener string-to-mosfet arrangement would work better?
If need be, I can use a few oil can input capacitors in series to get the proper voltage rating, but they are big suckers.. and that means big sucker*2=layout problem.
If need be, I can use a few oil can input capacitors in series to get the proper voltage rating, but they are big suckers.. and that means big sucker*2=layout problem.
Nice idea but I can see practical problems.
To drop any significant voltage at a reasonable current will need the Mosfet to be on a serious heatsink. Isolating the Mosfet from the heatsink will be a real issue at these voltages. Otherwise face large areas of metal at 900V, which will still need to be isolated from the case.
I personally think the practical issues will make it a none starter.
Shoog
To drop any significant voltage at a reasonable current will need the Mosfet to be on a serious heatsink. Isolating the Mosfet from the heatsink will be a real issue at these voltages. Otherwise face large areas of metal at 900V, which will still need to be isolated from the case.
I personally think the practical issues will make it a none starter.
Shoog
My experiments with using mosfets to drop (or regulate) voltage at that voltage level has usually resulted in dead mosfets. I am not saying that it can't be done (I am sure that it can), I just want to remind you of one nasty consequence. When a mosfet fails, it usually fails to a short, this will put the full voltage on all of the parts downstream. They will usually voice their displeasure with smoke!
The first thing to try is a choke input filter. If you weren't already planning this, it is a good way to lose 150 or so volts.
It is possible to build a regulator using vacuum tubes that would lose a few hundred volts and offer regulation and low source impedance.
The first thing to try is a choke input filter. If you weren't already planning this, it is a good way to lose 150 or so volts.
It is possible to build a regulator using vacuum tubes that would lose a few hundred volts and offer regulation and low source impedance.
Have you tried the solid state group?
At any rate, you can use caps in series in parallel with 220K or 470K resistors for filtering 1000 volts. The resistors make sure each cap gets no more than its rated voltage. That is, if you use three 350 volt caps with three 220K resistors across, each cap will get a relatively fixed 333 volts across.
As for dropping, a high voltage MOSFET can be used as a regulator with the mentioned Zener string biasing it.
Place the MOSFET after the first filter cap bank with the drain to the output of the rectifiers. I think you'ld need an N-channel. The output will be taken from the source, since we want it to be a constant voltage source. Connect a 47K 15-watt resistor from the B+ to the gate, and the Zener string (I think you can get 300 volters at 15 watt, or use 4 150 volters) from the gate to ground, with a cap across for filtering. If it is done correctly, you should not need a choke or much more filtering after. Kinda 'spensive, Lucy.
Better yet, use a BJT to bias the MOSFET, and use a smaller voltage/wattage Zener to fix it. Maybe a little cheaper than 15 watt Zeners (that's for 5 milliamp bias current). Then again, the power MOSFET you need will likely cost 20 bucks or more.
You could also use a 40 watt resistor to drop the 250 volts. What a waste, though... either way... and no regulation. All that power going to heat.
OR
If the 850 volt transformer has a center tap, use the standard two diode full wave rectifier with a 120 ohm resistor from the center tap to ground for current limiting (so you don't burn out the secondary of the trannie for too much current draw) and get about 590 volts (close to peak, since it would be 425 RMS).
OR...
Buy a 600 (300-0-300) volt power transformer for 55 bucks and call it a day! With a bridge rectifier and no use of the center tap, you'll have your 600 volts. Sell the 850 volt one on eBay.
Gabe
At any rate, you can use caps in series in parallel with 220K or 470K resistors for filtering 1000 volts. The resistors make sure each cap gets no more than its rated voltage. That is, if you use three 350 volt caps with three 220K resistors across, each cap will get a relatively fixed 333 volts across.
As for dropping, a high voltage MOSFET can be used as a regulator with the mentioned Zener string biasing it.
Place the MOSFET after the first filter cap bank with the drain to the output of the rectifiers. I think you'ld need an N-channel. The output will be taken from the source, since we want it to be a constant voltage source. Connect a 47K 15-watt resistor from the B+ to the gate, and the Zener string (I think you can get 300 volters at 15 watt, or use 4 150 volters) from the gate to ground, with a cap across for filtering. If it is done correctly, you should not need a choke or much more filtering after. Kinda 'spensive, Lucy.
Better yet, use a BJT to bias the MOSFET, and use a smaller voltage/wattage Zener to fix it. Maybe a little cheaper than 15 watt Zeners (that's for 5 milliamp bias current). Then again, the power MOSFET you need will likely cost 20 bucks or more.
You could also use a 40 watt resistor to drop the 250 volts. What a waste, though... either way... and no regulation. All that power going to heat.
OR
If the 850 volt transformer has a center tap, use the standard two diode full wave rectifier with a 120 ohm resistor from the center tap to ground for current limiting (so you don't burn out the secondary of the trannie for too much current draw) and get about 590 volts (close to peak, since it would be 425 RMS).
OR...
Buy a 600 (300-0-300) volt power transformer for 55 bucks and call it a day! With a bridge rectifier and no use of the center tap, you'll have your 600 volts. Sell the 850 volt one on eBay.
Gabe
Good info all around guys. I had not taken the heatsink insulation voltage rating tinto account... that WOULD be a problem. I used the mosfet with a pot as the voltage divider-to-gate in my DIY bench supply, and it worked flawlessly well... but at 500v, which is much more managable than 800+v.
I had thought choke input, but a big, high curent choke that would take that voltage rating is more expensive than getting a bigger chassis built and running a fancy schmancy power supply into it... plus I have killed silicon diode bridges with choke inputs in the past, so that worries me too at those voltages.
Another thought that I had was to run a hybrid tube/ss recto bridge, and then add a little series resistance in a CRC filter to drop the voltage down a bit, then use a mosfet rectifier at safer voltages. That might work, bt the 'squish' factor on transients would be pretty high since I was planning AB1 ultralinear... which might not sound too bad, but it would definately be distinctive.
I had thought choke input, but a big, high curent choke that would take that voltage rating is more expensive than getting a bigger chassis built and running a fancy schmancy power supply into it... plus I have killed silicon diode bridges with choke inputs in the past, so that worries me too at those voltages.
Another thought that I had was to run a hybrid tube/ss recto bridge, and then add a little series resistance in a CRC filter to drop the voltage down a bit, then use a mosfet rectifier at safer voltages. That might work, bt the 'squish' factor on transients would be pretty high since I was planning AB1 ultralinear... which might not sound too bad, but it would definately be distinctive.
Gabevee said:
PFFFT! Yes... they deal with low voltage, high current usually... not much help.
If the 850 volt transformer has a center tap, use the standard two diode full wave rectifier with a 120 ohm resistor from the center tap to ground for current limiting (so you don't burn out the secondary of the trannie for too much current draw) and get about 590 volts (close to peak, since it would be 425 RMS). [/QUOTE]
No center tap... unfortunately, then I could just reference it to a negative voltag and solve the whole problem! it is a nice big toroid that I have had sitting around as a paperweight.
That makes me wonder... if I run a full wave bridge, then the junction of the two diodes-to-ground forms the "center tap" there. Couldn't I just reference that to a negative voltage... like a zener or mosfet... or a simple resistor, and get the same result? or am I missing something...
Hmm... Sounds interesting. Use a negative voltage there to make the MOSFET into a virtual dropping resistor. Oh, that is what a regulator is (dummy me!).
Use a zener from the negative to ground... or would one need a resistor in parallel... to get a fixed negative voltage, and a pot to adjust the mosfet from that reference. Cool!
Gabe
Gabevee said:
i ALWAYS forget to mention that! it is a push-pull class AB1 Ultralinear KT88 output stage. Plates will be around 550-600V, the OT load around 4.5K, with an expected output at the speaker just under 100W-ish.
I bought tht toroid for another project that never got finished, and I figured that i could use it for this if I applied a few 'band-aids' to get the voltage down.
Gabevee said:
Use a zener from the negative to ground... or would one need a resistor in parallel... to get a fixed negative voltage, and a pot to adjust the mosfet from that reference. Cool!
Gabe
I'm thinking that a zener bypassed with a cap would work best there, so that it wouldn't vary the voltage with the current draw... but power rating might be an issue, so maybe a mosfet with a zener string on the gate, just like a series regulator, set up to drop whatever voltage is needed. A simple resistor might be best for class A, where curent is pretty steady.
tubelab.com said:
Cool. I just took an old Hammond 5H 150mA 105r choke and tied it between the bridge and ground, and I'll be darned! It dropped the voltage by about 20%! I was expecting 10%, but more is better. Thanx for the great tip. I'll definately store that one away in my bag of dirty tricks. I wonder why that is never used more often?
Gyrator may be the go.
SUGGESTION: Use Choke Input Power Supply Configuration BUT substitute a MOSFET Gyrator for the Choke. A netsearch on "Gyrator" came up with several MOSFET and BJT Gyrator designs which would be suitable. About to try this for myself with a C-L(Gyrator)-C arrangement to ditch about 50V on a 480V supply.
Cheers,
Ian
SUGGESTION: Use Choke Input Power Supply Configuration BUT substitute a MOSFET Gyrator for the Choke. A netsearch on "Gyrator" came up with several MOSFET and BJT Gyrator designs which would be suitable. About to try this for myself with a C-L(Gyrator)-C arrangement to ditch about 50V on a 480V supply.
Cheers,
Ian
Alethian,
Similar (same principle) but a power version - I just did a quick search and here is an example of what I mean - a picture is worth 1K words. Its in German BUT the schematic and component values are readable. This is a failrly common technique. It doen't behave exactly like a choke in that there is no stored magnetic energy BUT it does emulate choke behaviour with regard to its "resisting current change" behavior.
http://www.loetstelle.net/projekte/gyrator/gyrator.php
Cheers,
Ian
Edit: Afterthought - watch the MOSFET voltage rating - you could put it in the ground side by using a P Channel MOSFET.
Similar (same principle) but a power version - I just did a quick search and here is an example of what I mean - a picture is worth 1K words. Its in German BUT the schematic and component values are readable. This is a failrly common technique. It doen't behave exactly like a choke in that there is no stored magnetic energy BUT it does emulate choke behaviour with regard to its "resisting current change" behavior.
http://www.loetstelle.net/projekte/gyrator/gyrator.php
Cheers,
Ian
Edit: Afterthought - watch the MOSFET voltage rating - you could put it in the ground side by using a P Channel MOSFET.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.