[ARC5] Solid State 6AL5

[Tom Lee]

One should definitely take care when replacing vacuum tube rectifiers 
with solid-state ones. The much higher efficiency of the latter can lead 
to overvoltages, so if the B+ supply was designed with small margins to 
begin with, the solid-state rectifiers can produce a nice bang and let 
out all the magic smoke.

Someone earlier pointed out another important consideration: If the B+ 
comes up well before the tubes warm up, that can cause "cathode 
stripping" and accelerate the wearout of the tubes. So even if you solve 
the overvoltage problem (e.g., by adding zeners or resistors), there 
still remains the cathode stripping danger.

For AGC/AVC circuits, too, there can be many problems. I've analyzed a 
number of AVC loops, and a fair fraction actually have little right to 
work. To the extent that second-order effects seem to keep them from 
going unstable, any changes from the original design run the risk of 
making the loop fall off the cliff -- Murphy decrees that it can only go 
that way, despite the seeming 50/50 partitioning of outcomes.

Me, I like the warm glow of thermatrons, so I leave them in except in a 
very few circumstances (ratio detectors often benefit nicely from a 
switch to silicon, for example).

Cheers,
Tom

-- 
Prof. Thomas H. Lee
Allen Bldg., CIS-205
420 Via Palou Mall
Stanford University
Stanford, CA 94305-4070
http://www-smirc.stanford.edu
650-725-3383 (public fax; no confidential information, please)

On 2/7/2018 12:50 PM, Richard Knoppow wrote:
> FWIW, when I got one of my SP-600-JX's it had a solid state rectifier 
> in place of the 5R4. I removed it because I found all the voltages 
> were too high. This was not from high line voltage, I checked that, it 
> was the rectifier. When replaced with a standard 5R4 all were OK. This 
> is a molded case made of resin of some sort on a standard tube base. I 
> forgot about this until this thread and have no idea where to look for 
> it but it had a label and was not home made.
>     As fare as using solid state diodes for AVC, I think perhaps the 
> difference in minimum voltage and effects of contact potential might 
> require a change in the bias for AVC delay. I have not tested this. 
> You may be aware that it has been standard practice since about the 
> mid-1940s to put a dropping resistor in the filament of 6H6 and I 
> think also 6AL5 tubes when used as noise limiters. This affects the 
> contact potential and reduces the effects of hum from heater to 
> cathode leakage. Obviously it would have no effect on solid state 
> diodes. BTW, I have never found any technical paper about the filament 
> resistors but all receiver manufacturers began to add them about the 
> early to mid-1940. You would think the standard engineering texts 
> would have something. Maybe I missed it but if anyone knows please 
> tell me.
>
> On 2/7/2018 12:29 PM, J Mcvey via ARC5 wrote:
>> I'm curious...
>>
>>
>> On Wednesday, February 7, 2018 12:39 PM, John Watkins 
>> <jpwatkins9 at yahoo.com> wrote:
>>
>>
>> I have the metal case off of one of my mil 6AL5s, it is encased in an 
>> amber colored epoxy.  I could remove enough to see exactly what is in 
>> there and provide a few pictures if it would be of interest.
>>
>> John WD5ENU
>>
>
>