Bridge rectifier using ultrafast diodes

[The Slab Bacon]

Gerry,
         by the looks of that schematic, the 540K resistors are more for equalizing the voltage across all of the caps. As far as them being "bleeders" goes they are too high of a value to be anything but safety bleeders. They are too high in value to actually load down the power supply for any kind of regulation.


Bob,
      If you are only going for 2kV you might want to reduce the amount of capacitors and/or decrease the value of the equalizing resistors across them to inprove the equalizing effect.

[The Slab Bacon]

Gravy goes on turkey

Good man! ! ! ! 

[WA1GFZ]

Check the voltage rating of a common resistor and you find it around 300V.
So when you put it across a 1kv diode it could fail in time.

[KA2DZT]

I think most resistors can handle more that 300V  probably closer to 450V, 1/2 watt and above.  I use 470K 1watt MOX resistors across the diodes, never had a problem.

Those 500K resistors across the caps probably due very little to keep the voltage equal.  The resistance in the caps on start-up will be much lower and can be very unequal for each cap.  At that point the 500K resistors are doing nothing.  The resistors need to be much lower in value.  Probably more like 20K for each.  This will give about 120-160K bleeder, depending on how many caps are used.

In addition,  I always like to use an input choke on any HV supply.

Fred

[KD6VXI]

Ditch the equalizing caps and resistors.  Put a single cap across the entire string (for each side of the FWB, if that's the method you're using), so 4 total.

If you're using modern day rectumfires then you don't need them.  A single cap across the entire string will serve to kill transients..  Just need to use one properly sized.

MOF resistors will take more than 300 volts, but it's not recommended.  I've run them to 1Kv before...  Typically I try to keep them at 250 volts each.  I can't remember the brand off the top of my head, but one MFG of MOF's does rate them to 500 volts for their standard line, and they have some either 1kv or 1.5 kv rated ones, too.  I used them when I replaced the metering string in my SB220.

It would have been less expensive to go the route of the 300 volt resistors, and use more.  The higher voltage ones where not on the inexpensive side of a resistor, but.... 

--Shane
KD6VXI

[KM1H]

Anybody still using the resistor capacitor crap is living in the 70's. Even the ARRL dropped them around the early 90's in the HB.

There have been discussions on other forums by those in the industry showing documents stating they can even cause failures.

IMO the 1N4007 should be retired or relegated to LV apps and the 1N5408 take over HV apps until the 6A10 is required for tubes with handles.

Carl

[Opcom]

The eight resistors  - sure they will bleed off the charge after a while. The comment about the V rating is very important also and people overlook it.

It's a safety advantage to add a separate bleeder that will for sure drain off the charge in a few seconds. I run a 34K bleeder on a 3KV supply. It's a lot of power but the supply has 100mA to spare and when the HV is off, the charge is definitely gone in 10 seconds.

I chose the heavy bleed because the transformer is a pole pig with a 120 Ohm HV side and there is an input cap of 8uF. The power supply makes 3000V under load. If there was no bleed, the no-load voltage would rise to 4KV or a little more instead of 3500V. So it reduces stress on old components as well.

Bleeders help regulation, the point is that how well a bleeder helps regulation depends on the characteristics of the transformer. If the transformer's impedance is higher, like the older transformers, a heavy bleed will help keep the voltage to a reasonable level when the load is light.

[Opcom]

Anybody still using the resistor capacitor crap is living in the 70's. Even the ARRL dropped them around the early 90's in the HB.

There have been discussions on other forums by those in the industry showing documents stating they can even cause failures.

IMO the 1N4007 should be retired or relegated to LV apps and the 1N5408 take over HV apps until the 6A10 is required for tubes with handles.

Carl

Can you point out this information?

[W0BTU]

There have been discussions on other forums by those in the industry showing documents stating they can even cause failures.

I remember that discussion. As unbelievable as this may seem, it sure seemed to be true. This may be the thread: http://forums.qrz.com/showthread.php?299447-ballancing-resisitors-in-bridge-rectifier

[VK7ZL]

Thanks to everyone for the informed comments on the subject.

The transformer is 750V/side so the DC output will be about 2Kv and I will go with the following:
2 x 1N5408's in each string (giving a PIV of 4Kv for each half of the rectifier) with one 1.5nf, 4Kv, cap. across each pair of diodes and a fuse in the secondary in case of a shorted diode.

For the capacitor bank:
6 x 220uF, 450V caps with 2 x 270K, 1W res. in series across each cap. Total current flow is 0.00068A so the voltage across each resistor should be 183V - well within spec.

Any final comments welcomed.

Bob

[WD5JKO]

Thanks to everyone for the informed comments on the subject.

The transformer is 750V/side so the DC output will be about 2Kv and I will go with the following:
2 x 1N5408's in each string (giving a PIV of 4Kv for each half of the rectifier) with one 1.5nf, 4Kv, cap. across each pair of diodes and a fuse in the secondary in case of a shorted diode.

For the capacitor bank:
6 x 220uF, 450V caps with 2 x 270K, 1W res. in series across each cap. Total current flow is 0.00068A so the voltage across each resistor should be 183V - well within spec.

Any final comments welcomed.
Bob

   High voltage fuses can be tricky; I wonder what type of fuse you are going to use? What about a varistor across the primary? Will the supply be left on, or keyed for PTT? How do you limit the turn on surge current? If the transformer is 1500v CT, is that CT rated to float? Also if 1500v CT, the current rating must be set to 1/2 when using a FWB rectifier. You cannot run 2KW from a 1KW transformer by going from FW-CT to FWB. Well not for long that is.

   The following thread discusses the filter cap. I added some comments about helping series electrolytic's survive real world conditions.

http://amfone.net/Amforum/index.php?topic=17729.msg122355;topicseen#msg122355

Good Luck!
Jim
WD5JKO

[The Slab Bacon]

Bob,
      be very wary of using a transformer with a center tap in a FWB situation, with the center tap floating. Many of them are made with the CT close to the core figgering that it will be grounded anyway. there is often very little insulation between the CT and the core. They WILL often break down and short the CT to the core. They must be rated for FWB service. I know this for a fact because I didn't listen to others that told me not to do it that way and zorched a few really nice transformers  

[WQ9E]

Great advice by SB and even putting a filter choke between center tap and ground can cause the same transformer failure since center tap will be raised to the ripple voltage above ground.  Bill Orr discussed these issues in the "West Coast" handbooks.

[The Slab Bacon]

Great advice by SB and even putting a filter choke between center tap and ground can cause the same transformer failure since center tap will be raised to the ripple voltage above ground.  Bill Orr discussed these issues in the "West Coast" handbooks.

Roger,
         Been there, done that, learned the hard way! ! !      

The Ripple, as well as straight AC (or modulated DC) seems to have a much greater probability to "punch" through the insulation than straight DC.

[KA2DZT]

I never put chokes in the negative lead.  So far I don't like much of what the OP is planning.  I posted a number of suggestion but maybe he is not reading them.

Jim,  I read your notes on the series caps, it's very good.  I only use oil caps.

Frank, you're right about the CT issues with FWB circuits.  Many of the common plate xfmrs can not be used with FWB circuits.  The Stancor 80 series in one series that can be used with a FWB.

I've found that most small lower voltage xfmrs (say under 700 volts CT)  can be run with a FWB without having them fail. Just about any amount of some insulation near the core will handle 400-500 volts.

I used two old (NOS) RCA (1950's) TV xfmrs with 735vct 300ma windings to make 1400 volts at 300ma.  I put FWB rectifiers on both xfmrs.  I then connected the output from the first FWB diodes directly to the negative terminal of the second xfmr's FWB diodes.  The second xfmr's CT was running over 1000 volts above ground.  The PS worked perfect,  I ran it for days pulling near full current.  The two xfmrs hardly got warm and nothing shorted.  I did open the second xfmr and added some insulation about the filament lead wires (filaments not used) and added some plastic strip insulation where the leads past through the metal end bells.

Fred

 

[The Slab Bacon]

Here is another one to ponder as well. I am a big advocate of using oil caps especially in HV power supplies. I am much more comfortable with them in lieu of stacked aluminum electrolytics. (anything over 400-450v) I also have a goodly supply of them on hand.

I have also always pondered what the breakdown voltage is on them (from the cap guts to the metal case) I would think that it would be considerably more than the WVDC rating of the cap.

Here is another one  I bought a half a dozen 100uF @ 3Kv oil caps at a fester last year for $2 each. For the ratings, the cans weren't real big. They supposedly came out HID lighting. All of them measured up way low in capacitance (60-80uF). In an oil cap, how can this happen?? they passed a DC HI-POT test at 3500v. Do the plates get burned off inside?  One of then showed no capacitance at all. Since they are all labled "NO PCB" I might have to take a can opener to the dead one.   

[KA2DZT]

Frank

As for the dead cap, a lot of those newer oil caps read "protected" on the label.  They must have some sort of fuse inside them.  Also I think they have a bleed resistor.

I also wondered about the breakdown voltage to the can.  Doesn't matter unless you start running them in series.  Something I never do.

As for the low capacitance,  I wonder if they need to run at near full voltage for a period of time to maybe come up to full capacitance.

Fred

[The Slab Bacon]

As for the low capacitance,  I wonder if they need to run at near full voltage for a period of time to maybe come up to full capacitance.

Phred,
         That could be an interesting scenario Hmmmmmm.............

But, I have never heard of oil capz needing to be "re-formed". Keeping in mind that many people store them with shorting jumperz on them. I would imagine they are just plates, a simple dialectric and the whole shebang submerged in mineral oil. I guess the can opener will tell the whole tale.

[Tom WA3KLR]

Hi Frank,

What capacitance meter are you using?

[The Slab Bacon]

Hi Frank,
What capacitance meter are you using?

Tom,
        It is one of those cheap chineese digital LCR meters. But it also agrees with my Wavetech DMM and another DMM that I also have at the bench. All 3 read pretty much the same on the capacitance readings.

[WA1GFZ]

These little meters excite the cap with a low level AC signal. Maybe leakage  effects the reading. The meter reads reactance and converts to C.

[VK7ZL]

I never put chokes in the negative lead.  So far I don't like much of what the OP is planning.  I posted a number of suggestion but maybe he is not reading them.

Jim,  I read your notes on the series caps, it's very good.  I only use oil caps.

Frank, you're right about the CT issues with FWB circuits.  Many of the common plate xfmrs can not be used with FWB circuits.  The Stancor 80 series in one series that can be used with a FWB.

I've found that most small lower voltage xfmrs (say under 700 volts CT)  can be run with a FWB without having them fail. Just about any amount of some insulation near the core will handle 400-500 volts.

Fred
 

Fred
I did read your comments re the diodes and I will use 4 as per the original circuit. I also read Jim's notes series caps but unfortunately I have to go with resistor values that are commonly available off the shelf. The lowest I can go is 6 x 220K, 1W with a dissipation of 0.6W / resistor @ ~2200V.

Re the CT and FWB.
I don't have any specs on the transformer other than 240V primary and 750V / side secondary with a very heavy 6V winding and a lighter 14V. The iron core measures 6" x 5" x 2.5" with a total weight of 21lbs.

The amplifier will only be running 400W (our legal limit) in SSB mode.

Bob

[KA2DZT]

Bob,

I see you're using a 1500vct xfmr.  I thought you were trying to use MO HV xfmrs. for the plate supply.  I don't like MO xfmrs for plate supplies.

Using the 1500vct xfmr with a FWB rectifier should work.  As mentioned,  the only concern is the CT running above ground by 750 volts.  The only way you'll know is after you build the supply.  If it doesn't short (CT) to ground then you'll be OK.  I've run many xfmrs in this voltage range with FWB rectifiers and never had any short.

My other concern is the amount of diodes you're using in the bridge circuit.  Two diodes in each of the four legs gives you a total of 4KV PIV.  I would suggest that you use at least three diodes in each leg giving a total of 6KV PIV.  4KV PIV, you're cutting it too close.  I run seven diodes in each of two legs of a FW rectifier circuit with 1500/ct/1500 volts (CT grounded).

Remember, the diodes are rated at 1000v PEAK IV.  Your 1500v RMS is over 2100 volts peak.  I use eight diodes in a FWB circuit for supplies in the 700-1000 volt range.

Many have said that you don't need to use the resistor and cap across each diode, I still use them.  If you decide to use them, the resistor can be ONE 470-510Kohm 1W across each diode.  The caps are .01ufd 1KV disc ceramics.

Fred

[Tom WA3KLR]

For series strings of diodes, I always base the number of diodes on being very conservative plus assume that one diode in the string has failed shorted, and so the N-1 number of diodes is still a conservative PIV; no propagation from a single diode failure.  You don't want to go overboard either, but diodes are cheap.

I'm for no parallel caps and resistors.  They just lower the MTBF of the diode string.  (We had to do that in the 1960's and this is 50 years later.) I don't think any high voltage power supply manufacturer today adds caps or resistors across the diodes.

By the way, is Australia 50 Hertz power?  (Even more time for regular diodes to switch.)

[KA2DZT]

Tom,

What is the MTBF??  Maybe you could expand on that a bit.

Thanks,
Fred