Armstrong 421 output transistors

The output transistors on one channel of my son's 400 series receiver have blown. Unfortunately he temporarily used bare ended wires to connect the speakers pending receipt of the proper two pin plugs. This led to a stray thread of wire getting trapped and eventually shorting the pins in the socket.

Replacement germanium transistors seem to be unviable for all practical purposes and I was wondering whether the amp can be modified to use silicon transistors. Can anyone help please?
Have a look inside your 421.
There are 2 types of 'power amp' board. The 421 and earlier 521 amps used the A14 board which is all germanium.
The later 521 amps and some 421s that have been in for repair have the A15 silicon based power boards.
If you have A15 boards conversion to silicon outputs (2N3055s) is reasonably easy and covered on the net. Unfortunately the A14 germanium based boards do not lend themselves to conversion. (There are a few uncompleted threads, on several forums, where people try...)

So choice is if you have A14 power boards, replace the AL 102 transistors like for like or near equivalent, yes you can still find them. Or find a pair of A15 power boards and modify ALL the outputs to silicon devices.


  • Armstrong A15 PA Board v2.doc
    596.5 KB · Views: 54 of nebraska has germanium TO-36 stud mount power transistors for $4.50 ea. Located in Nebraska. May have others in the 4 pages of transistor listings, I didn't look.
Products I have received from them previously have been genuine. Mostly ex military stock I believe, they are near Offutt AFB.
Last edited:
Thank you for the replies.

It has the A14 board so. maybe, adapting it to silicon is impractical. Jon, the schematic is on the post by Alan, but it doesn't show the component values. I don't have the unit to hand to find out what they are but I can get it in a couple of days. The output transistors are AL102s. I have found an equivalents list which shows variations on AL100 and AL102s and 2n1906 TO3. They are all rather expensive at over £20 each and I wonder whether the receiver is worth the expense if simple replacement of the transistors is the only option.
Moderators if this is against the rules please delete ! I have read on the Uk vintage radio forum a no of articles converting the Armstrong 421 /521 to silicon there is mention of an all pnp version that will suit the A14 board
its simple enough to change the outputs and drivers
  • Like
Reactions: 1 user


2019-09-17 12:49 am
Not even a need to rewire the whole output section to NPN.

Seems like you could easily change to just PNP silicon devices like MJ2955

you would just need to adjust bias resistor values.

Something even easier would be to add parallel resistors to existing bias resistors.
which to me looks like what the factory schematic for A14 is doing to adjust idle current.
  • Like
Reactions: 1 user


2019-09-17 12:49 am
If I am not mistaking on A14 board
Resistor groups 281 , 279 and 280 set bias for TR25
Resistor groups 278, 277 and 276 set bias for TR27

as you can see resistor 280 and 276 have little dots next to them
these are parallel resistors to change bias

So technically you would only need to change resistor 280 and 276
to adjust bias for MJ2955 used as final transistors

  • Like
Reactions: 1 user


2019-09-17 12:49 am
hmm that is interesting, so it does have thermal tracking.

leave thermal tracking circuit and then bias can be adjusted on other side then instead.

you could change 278 , 271 to adjust bias.
and if needed more fine tune, again just add parallel resistors.
Excuse my extreme ignorance here but how does 271 affect the bias? Now for an even dafter question - would I increase or decrease the values of the resistors to get the increased bias for the Si transistors? Would variable resistors be a good idea? You can see I am an absolute novice regarding solid state circuits!


2007-10-10 5:21 pm
It is more customary to see this topology with an interstage driver transformer. R270 and 271 would be the secondary windings. C273 is a bootstrap - that may be what looks confusing.

Are the original Q24 and 27 Ge devices also, or were they able to obtain PNP silicon drivers back in those days?