SHFA2 Q amp diode question

[mb9513]

Hi all.  recapping an amp for a Q.  the diode pair CR101 is bad.  I had to replace the same part on the last shfa2 amp I did.  what diodes would I use to replace it? I remember asking the same question on the old forum but don't know how or if I can search those old archives.  Thanks Mark

[robnyc]

It is not critical there, 1N4002 or higher is fine. It will result in slightly higher AGC voltage, however, as long as the resulting voltage does not consistently exceed approx 25 volts on the G1 of the 6BJ6 w/loud record, the tube will be able to maintain control.

RobNYC

[Ron Rich]

Hi Mark,
I do not have the schematic handy, but if that's one of the dual diode,cement,encapsulated, type--use 4 silicone "signal"diodes ( 1n419, or eq.) Ron Rich

[robnyc]

Ron, CR101 is just AGC/squelch derivation -it doesn't pass any audio. I use "whatever" here in the mono and stereo versions.

Rob

[Ron Rich]

Rob,
Ain't that basically whut I rit ? I use 1N914's simply because I must have 10k, or so, of them--lol--Ron Rich

[robnyc]

Ron, I should-a been more clear in my response....What I meant was that you only need to use --one-- for each diode here are opposed to using 2 for each leg in the TSA-SHP when replacing the diodes that actually control the level.

BTW: I recently went through several TSA amps including a TSA-8 which uses those "stabistor" diodes. When I ungrounded the AGC voltage, one channel would be about 3 DB lower than the other. I've never thought it necessary to replace sil diodes here..diode test showed them all reading the same and drive was equal. That left nothing but the diodes themselves. I remembered you mentioning them as a defect point few months back. Replaced diodes and channels are now equal.

Live 'n learn.
Rob

[mb9513]

Thanks All! now if I could just magically get all those caps to jump in it and solder themselves.  Mark

[Ron Rich]

Hi Rob, and all,
I have played around with replacing those "stabistors"--found that if there was only 1 diode used, the AVC acted to quickly --especially at the start of a record that had a long lead-in -- I have used 3 of the small silicone's in each leg--Ron Rich

[robnyc]

Mark, while you are in diode mode --be sure to replace CR102--. This is the negative bias supply and preamp supply. If it fails the output tubes will conduct full-on and dissipate excessively. There is also the potential for output transformer damage.

Rob

[Ron Rich]

Ab-sol-lut-ley--ALWAYS replace this one AND, the limiting diode in the TSR/TSU. 1N4002, orbetter can be used for both. Ron Rich

[robnyc]

May 28, 2017 13:21:08 GMT Ron Rich said:

Hi Rob, and all,
I have played around with replacing those "stabistors"--found that if there was only 1 diode used, the AVC acted to quickly --especially at the start of a record that had a long lead-in -- I have used 3 of the small silicone's in each leg--Ron Rich

Ron, I found similar results in 1988 when rebuilding my HLPC (TSA-1). The original selenium package diodes that controlled the gain were no longer effective. Single silicon diodes reduced gain when conducting to a point where a full volume control setting was necessary for normal operation.  I ultimately used three sil diodes in each leg.

On the TSA-8's that I've been prepping for use I've used two sil diodes for each 'stabistor".   I do a full set of (probably unnecessary -but interesting) tests on the various sections of each amp that I've rebuilt.

One interesting thing is that for a given, measured AGC voltage at the time constant cap, the replacement diodes reduce gain less than the original stabistors. However, the AGC voltage for at the time constant cap is greater given the same input at the preamp.

Using 800Hz input at a level that produces 1 vdc at the TC: Stabistors reduce gain by 4-1/2 db but the two sil's (each leg) reduce it by only 2 db.    I do add partial decoupling to the emitter of the AGC drive transistors in all cases to increase the total potential gain reduction to a max of 19db.

Also interesting is the effect of the "scratch compensator".  As designed this never did much more than reduce the highs. But when the AGC drive is increased as above, the "medium" position results in:

Minus 5 db @10khz when .7 volts on the TC
Minus 2 db       "          "   1.3   "      "
and only minus 1/2 db when there is 1.4 vdc on the TC.     This illustrates the original intent of the circuit to take advantage of the varying impedance of the diodes as they compress more. problem was, it needed a lot more drive to the diodes for it to actually be a dynamic filter as intended.

Sorry for the long tome. I haven't worked on these amps since the mid-90's and it is interesting to go over my old notes and make some new ones.

Rob

[jukeboxmarty]

Lots of good information here, thanks guys!

So what's the bottom line? One, Two, or Three diodes for each leg?

Or do we just use what works?

[robnyc]

Marty, in your case one diode for each diode depicted. So, a total of two to replace Cr101 and one to replace CR102.

Rob