On Monday, October 13, 2003 at 1:30:37 PM UTC-5, N2EY wrote:
> In article <adJhb.611$zw4...@nwrdny01.gnilink.net>, Edward Knobloch
> <al...@nowhere.com> writes:
> >The 5R4 has a maximum piv voltage of 2800 volts,
> >so you would still need 3 1N4007's in series
> >in each leg for equivalent safety factor.
> Absolutely correct.
> >
> >The rated max supply voltage of 750V rms per plate means that
> >you should not use a transformer higher than
> >1500V rms center tapped in a full wave single phase
> >rectifier circuit using a 5R4.
> The 750 volt rating is for capacitor input. With choke input you can go a bit
> higher.
> >
> >The designers expected switching transients
> >when the current is interrupted in the HV transformer
> >and HV choke, and therefore included a max
> >supply voltage for guidance.
> Even with no transients, the peak voltages that the diodes have to withstand
> are quite high, compared to the RMS transformer voltages. Here's why:
> First off, the voltages in the transformer catalogs and tube manuals are RMS
> values, not peak values. 750 volts RMS equates to 1060 volts peak with sine
> waves. So under perfect conditions a 750-0-750 (1500VCT) transformer with
> fullwave center tap rectifier and capacitor input filter can have voltage
> output of 1060 volts.
> But there's more! In our 750-0-750 example, each transformer HV lead goes from
> 1060 volts (peak) positive to 1060 volts negative with respect to the center
> tap. So each diode is subjected to as much as 2120 volts when it is in the
> nonconducting (inverse) state. So you need *at least* three 1N4007s to replace
> one side of a 5R4. Four diodes won't hurt a thing.
> The tube manuals I have show voltages higher than 750 for capacitor input, but
> they are for the 5R4GYA.
> In any event please use at least three and preferably four 1N4007s or
> equivalents in a 5R4 replacement. Plate transformers and filter chokes can be
> expensive and hard to replace.
> 73 de Jim, N2EY

Jim, N2EY is correct in his analysis. You have to look at more than the output voltage - you need to actually know the circuit analysis to see the PIV on the diode. His method is exactly right. In many tube audio amps, two 1000 V PIV (like iN4007) series diodes in each string is enough. But above 200-300 V, in order to fully match the tube specs, like one needs a bit more than three. Never use one per side.

But that's not all. When you put diodes in series, the reverse voltages don't divide equally because each has a different leakage current. If those are not well matched, one should use "equalizing resistors" across each diode of about 5 to 10x the leakage current of the diodes.

Then there transients to consider, and good derating practices. All this is why the commercial SS replacement plug-ins so frequently fail. Most guys making them are not engineers and didn't "do the math".

I personally use two BY2000 2000 V PIV diodes.

Drew K3PA

Considering that this message was sent in 2003, I think it unlikely that
Jim or Edward will ever see your rely.

David

---
Drew Vonada-Smith wrote:
> On Monday, October 13, 2003 at 1:30:37 PM UTC-5, N2EY wrote:
>> In article <adJhb.611$zw4...@nwrdny01.gnilink.net>, Edward Knobloch
>> <al...@nowhere.com> writes:
>>> The 5R4 has a maximum piv voltage of 2800 volts,
>>> so you would still need 3 1N4007's in series
>>> in each leg for equivalent safety factor.
>> Absolutely correct.
>>> The rated max supply voltage of 750V rms per plate means that
>>> you should not use a transformer higher than
>>> 1500V rms center tapped in a full wave single phase
>>> rectifier circuit using a 5R4.
>> The 750 volt rating is for capacitor input. With choke input you can go a bit
>> higher.
>>> The designers expected switching transients
>>> when the current is interrupted in the HV transformer
>>> and HV choke, and therefore included a max
>>> supply voltage for guidance.
>> Even with no transients, the peak voltages that the diodes have to withstand
>> are quite high, compared to the RMS transformer voltages. Here's why:
>> First off, the voltages in the transformer catalogs and tube manuals are RMS
>> values, not peak values. 750 volts RMS equates to 1060 volts peak with sine
>> waves. So under perfect conditions a 750-0-750 (1500VCT) transformer with
>> fullwave center tap rectifier and capacitor input filter can have voltage
>> output of 1060 volts.
>> But there's more! In our 750-0-750 example, each transformer HV lead goes from
>> 1060 volts (peak) positive to 1060 volts negative with respect to the center
>> tap. So each diode is subjected to as much as 2120 volts when it is in the
>> nonconducting (inverse) state. So you need *at least* three 1N4007s to replace
>> one side of a 5R4. Four diodes won't hurt a thing.
>> The tube manuals I have show voltages higher than 750 for capacitor input, but
>> they are for the 5R4GYA.
>> In any event please use at least three and preferably four 1N4007s or
>> equivalents in a 5R4 replacement. Plate transformers and filter chokes can be
>> expensive and hard to replace.
>> 73 de Jim, N2EY
>
> Jim, N2EY is correct in his analysis. You have to look at more than the output voltage - you need to actually know the circuit analysis to see the PIV on the diode. His method is exactly right. In many tube audio amps, two 1000 V PIV (like iN4007) series diodes in each string is enough. But above 200-300 V, in order to fully match the tube specs, like one needs a bit more than three. Never use one per side.
>
> But that's not all. When you put diodes in series, the reverse voltages don't divide equally because each has a different leakage current. If those are not well matched, one should use "equalizing resistors" across each diode of about 5 to 10x the leakage current of the diodes.
>
> Then there transients to consider, and good derating practices. All this is why the commercial SS replacement plug-ins so frequently fail. Most guys making them are not engineers and didn't "do the math".
>
> I personally use two BY2000 2000 V PIV diodes.
>
> Drew K3PA
>

--
This email has been checked for viruses by Avast antivirus software.
www.avast.com

David Snowdon <do804@torfree.net> wrote:
>Considering that this message was sent in 2003, I think it unlikely that
>Jim or Edward will ever see your rely.

This is true, BUT, it's also good to remember microwave oven diodes!
Usually 10KV or so, but only around 100mA. Many of them have a bandgap
voltage as high as six or seven volts so clearly there is more than one
junction inside there. They are also handy to have in the parts bins.
--scott

--
"C'est un Nagra. C'est suisse, et tres, tres precis."