[Boatanchors] Diode replacements for 866 tubes, ratings needed?

[Jim Wiley]

Answers in-line, below


On 11/10/2014 4:16 PM, Rick Poole WA1RKT wrote:
> I'm rebuilding an old Heathkit HA-10 Warrior amp (4x811A, 1700 VDC on the plates).  I plan to replace the 866 rectifier tubes with solid state rectifiers.  I would prefer to replace them with plug-in solid state replacements but am having trouble finding them at a reasonable price and as I am coming up on three years of unemployment I really need to do this on the cheap.  So I figured on building in two strings of 1N5408 diodes, one string in each leg of the center-tapped plate transformer secondary.
>
> I have seriously limited (though not quite "none") experience in high-voltage power supply design, and have some questions...
>
> 1.  The Warrior power supply uses a swinging choke in the center tap lead of the power transformer (I must admit, I haven't seen that kind of "choke-input" filter before).  I'm told that choke-input filters tend to generate impressive back-EMF voltage spikes that the rectifiers need to be able to handle.  So, how many 1N5408 diodes (1 KV PIV, 3A) do I need for each leg for a full-wave rectifier?
As far as the number of rectifiers goes, I tend to use literally twice 
as many as good engineering specifies, to protect against power line 
surges and other issues.  Diodes are cheap (really cheap!) these days, 
so using a few extra won't hurt anyone's bankbook.  For this 
application,  I would probably use 8 to 10 per side


Putting the choke in the return lead of the power supply, in this case 
from the center tap of the HV transformer to ground removes the high 
voltage from the choke, thus requiring less insulation (and less cost) 
for the same result.

If you are worried about back-EMF, which could be caused by rapid 
transitions from full load to no load, as when using the amplifier for 
CW service, simply put a string of 4 or 5 1000V/3A diodes across the 
choke to absorb the energy.  The diodes would be connected in opposition 
to the normal rectifier diodes.
>
> 2.  Related to (1), does the choke in the center tap of the transformer, rather than at the input to the filter, provide any relief to that "impressive back-EMF voltage spikes" issue noted above?

No change
>
> 3.  Older Handbooks show a 0.01-uf cap and a several-hundred-Kohm resistor across each diode in the string.  I understand why the caps but not sure about why the resistors.  I've read lately that the resistors are not needed.  True or false?

Shunt capacitors and resistors are no longer needed.  They were used in 
days gone past because device uniformity was not very good. These days, 
manufacturing tolerances are so tight that simply connecting a suitable 
number of diodes in series will do the job. No "equalizing" capacitors 
or resistors are needed.
> 4.  One of those "older" Handbooks (1991) states that choke-input filters are not normally used with silicon rectifiers.  True or false, and if true, why?

Not true.  Choke input filters can be used with any type of rectifier.   
Choke input filters are the "easiest" on components, as they limit 
inrush (charging) current to filter capacitors.  Again, however,  modern 
diodes have so much reserve capacity for short-duration surges that 
capacitor input filters are no longer an issue.  Basically, even in HV 
supplies, the DC resistance of the transformer secondary is enough to 
limit inrush current to safe values.   Some constructors use so-called  
"soft-start" circuits in an abundance of caution, but they are not 
really necessary unless the supply transformer has an extraordinarily 
low DC resistance. Even then, installing diodes with a higher surge 
current rating will solve the issue.  Modern diodes have surge ratings 
into the hundreds of amperes, unlike older rectifiers such as 
mercury-vapor types that could be damaged by surges.  Most commercially 
available HV DC circuits don't bother with "step-start" circuitry.  
However, if it makes you feel good, do it.  It won't hurt anything.
>
> Thanks...
>
> Rick WA1RKT
>