The original transistors are currently installed. I put them back in after finding out the new ones didn't fix the problem.
I'd tried a KSC945C in Q1.
I'd tried a KSC945C in Q1.
I imagine you’ve already confirmed no drop across the output inductor.
It would be a tedious chore, but you could swap Q5 and Q6 across channels and see if the problem moves.
It would be a tedious chore, but you could swap Q5 and Q6 across channels and see if the problem moves.
That's just the bootstrapping capacitor discharging, it's normal. You replaced that?
So what was in there before, presumably with very poor gain if one channel was falling short on power?
Here is the factory Dyna Stereo 120 TIP modification. Note the higher output beta specs.
Q1 replace 40233 (100-250 beta) with BC108A (130-180 beta)
Q2 replace 2N3053 (100-200 beta) with 2N5320 (160-260 beta)
Q3 replace 2N3053 with TIP31C
Q4 replace 2N4037 with TIP32C
Q5, Q6 replace 2N3055 (17-25 beta) with 2N3772 (60-90 beta @ 1A)
D1 replace with 1N4733
D2,3 replace with 1N4004
C1 replace with 5 µF 10V tantalum
Add C13 68 pF from Q2C to Q2B, both channels
Add C14 0.01 µF from Q3C to GND, both channels
Add C15 27 pF from Q2C to Q1E, both channels
Add C16 .01 µF across xfmr secondary at rectifier bridge on PC-15
Replace C6 (30 µF, NP) with .47 µF mylar
Add 3.3 Ohm ½ watt emitter resistor to Q4, both channels
Add 1 k Ohm 1.2 watt resistor across the red/black output jacks
Q1 replace 40233 (100-250 beta) with BC108A (130-180 beta)
Q2 replace 2N3053 (100-200 beta) with 2N5320 (160-260 beta)
Q3 replace 2N3053 with TIP31C
Q4 replace 2N4037 with TIP32C
Q5, Q6 replace 2N3055 (17-25 beta) with 2N3772 (60-90 beta @ 1A)
D1 replace with 1N4733
D2,3 replace with 1N4004
C1 replace with 5 µF 10V tantalum
Add C13 68 pF from Q2C to Q2B, both channels
Add C14 0.01 µF from Q3C to GND, both channels
Add C15 27 pF from Q2C to Q1E, both channels
Add C16 .01 µF across xfmr secondary at rectifier bridge on PC-15
Replace C6 (30 µF, NP) with .47 µF mylar
Add 3.3 Ohm ½ watt emitter resistor to Q4, both channels
Add 1 k Ohm 1.2 watt resistor across the red/black output jacks
Old RCAs are usually good. Even the hometaxial 3055’s. RCA was the king of the single diffused process. Those Je ne sais quoi things were probably running out of beta. I’ll bet it was less than 17 at the 5A you could be drawing with an 8 ohm driver. At 4 ohms all bets are off. It might be a TO-3, but might not even be a 15A part. A 5 amp part might have a beta of 5 or 10 at 5 amps. I’ve had MJ15015’s (high voltage 3055’s) fall to the single digits at 7A and they’re supposed to be 15A. The parallel bank in an old QSC I was rebuilding with them just fell apart and I gave up and stuck in 15024’s the 2nd time around. But they still met spec……
I’ve been tempted to get some of the RCA1B01’s that electronicsurplus.com still has. Just tempted, not that I really need any more 2N3055-oids. Which is why I haven’t pulled that particular trigger. I still have a whole box full of 2N3772’s that were house numbered for TRW. Yes, they’re good.
When doing “the TIP mod” it’s just as common to go to the MJ15003 these days. One could only dream about them when the ST120 came out.
I’ve been tempted to get some of the RCA1B01’s that electronicsurplus.com still has. Just tempted, not that I really need any more 2N3055-oids. Which is why I haven’t pulled that particular trigger. I still have a whole box full of 2N3772’s that were house numbered for TRW. Yes, they’re good.
When doing “the TIP mod” it’s just as common to go to the MJ15003 these days. One could only dream about them when the ST120 came out.
I also found a link to the Dynaco fill TIP mod doc on updatemydynaco website, wit a good description of additional cap placements. Unfortunately the file is prob too big to upload and save here, but it's available up until now at this link---> https://updatemydynaco.com/documents/TIPmod.pdf
TIP31C and 32C drivers will give you dull highs. My 1971 ST120 came with house number drivers Q3 Q4 that were probably 2n5320 5322. 3 times the Ft of TIP31/32 which is only 6 mhz. Ask me how I know? One pair survived the burnout, that side sounded better than the side with TIP. No 60636 or 2n3055 outputs survived when I bought the unit about 1985, so I started off with NTE60 which were probably white box MJ15003 due to the soa spec. Those outputs had 2 mhz Ft instead of 0.4 Mhz. A simulator posted on here that the original homotaxial output transistors were the cause of poor highs of the ST120. See the 1966 reviews. I did not monkey with Q1 and Q2 and the 1971 transistors sounded fine. Top octave solo Steinway piano on my SP2-XT speakers is my high freq test. My hearing goes to 14 khz.
The capacitor mods and the output zobel across the speaker jacks are oscillation preventatives. the Greg Dunn version of TIP mod has 1k series .1 uf on the output jacks. Also these capacitor mods instead of updatemydynaco version, which I did:
C1 replace with 5uF 10V tantalum
Add C13 68 pF from Q2C to Q2B, both channels
Add C14 0.01uF from Q3C to GND, both channels
Add C15 27 pF from Q2C to Q1E, both channels
Add C16 .01 uF across xfmr secondary at rectifier bridge on PC-15
Gregg dunn article was home.indy.net/~gregdunn/dynaco/components/st120/index.htm
I later replaced C2 with 120 pf ceramic cap to prevent AM sports talk radio from coming in the input from 12' RCA cables necessary to reach from disco mixer to amp.
One PC14 was had the lands damaged by TO220 driver with legs twisted around. So I replaced with Apex AX6 which sounds better cold due to better idle bias current control. Also fewer parts.
The original driver board PC14, I added the Djoffe 7 transistor bias control board. https://www.diyaudio.com/forums/solid-state/156627-dynaco-stereo-120-beautiful.html
Sound is great but the sense transistor has shorted twice leaving 200 ma idle bias current. Sounds good that way, but good thing I had PCAT PS fans on the boards.
PC15 the B+ regulator and max current limiter board, was limiting to 2 amps with a NTE181(MJ15003?) instead of a defined low gain homotaxial part. So I replaced it with a 69 v regulator board and a 10 amp fuse. To get the 6.75 amps the transformer will actually put out. 5 parallel TIP142 pass transistors with .5 ohm emitter resistors each, on a Pentium2 CPU heatsink. Bases driven by a stack of six 12 v zeners with series resistor to limit current.
Biggest reliability improvement is to replace the dynaco 1/8 angle heat sink on outputs with something like pentium2 cpu heatsinks. I only had one dead pentium, so I added a couple of TO3 finned sinks which would not fit under the bottom transistor, then added 2 PCAT PS fans like micky mouse ears blowing on both sides. Fan noise is one reason I had the amp behind the record rack near the speakers instead of near the mixer. Fans bolted to a piece of steel channel under the amp salvaged from a box fan. 9 v wall transformer drives the 12 vdc fans.
The capacitor mods and the output zobel across the speaker jacks are oscillation preventatives. the Greg Dunn version of TIP mod has 1k series .1 uf on the output jacks. Also these capacitor mods instead of updatemydynaco version, which I did:
C1 replace with 5uF 10V tantalum
Add C13 68 pF from Q2C to Q2B, both channels
Add C14 0.01uF from Q3C to GND, both channels
Add C15 27 pF from Q2C to Q1E, both channels
Add C16 .01 uF across xfmr secondary at rectifier bridge on PC-15
Gregg dunn article was home.indy.net/~gregdunn/dynaco/components/st120/index.htm
I later replaced C2 with 120 pf ceramic cap to prevent AM sports talk radio from coming in the input from 12' RCA cables necessary to reach from disco mixer to amp.
One PC14 was had the lands damaged by TO220 driver with legs twisted around. So I replaced with Apex AX6 which sounds better cold due to better idle bias current control. Also fewer parts.
The original driver board PC14, I added the Djoffe 7 transistor bias control board. https://www.diyaudio.com/forums/solid-state/156627-dynaco-stereo-120-beautiful.html
Sound is great but the sense transistor has shorted twice leaving 200 ma idle bias current. Sounds good that way, but good thing I had PCAT PS fans on the boards.
PC15 the B+ regulator and max current limiter board, was limiting to 2 amps with a NTE181(MJ15003?) instead of a defined low gain homotaxial part. So I replaced it with a 69 v regulator board and a 10 amp fuse. To get the 6.75 amps the transformer will actually put out. 5 parallel TIP142 pass transistors with .5 ohm emitter resistors each, on a Pentium2 CPU heatsink. Bases driven by a stack of six 12 v zeners with series resistor to limit current.
Biggest reliability improvement is to replace the dynaco 1/8 angle heat sink on outputs with something like pentium2 cpu heatsinks. I only had one dead pentium, so I added a couple of TO3 finned sinks which would not fit under the bottom transistor, then added 2 PCAT PS fans like micky mouse ears blowing on both sides. Fan noise is one reason I had the amp behind the record rack near the speakers instead of near the mixer. Fans bolted to a piece of steel channel under the amp salvaged from a box fan. 9 v wall transformer drives the 12 vdc fans.
Last edited:
Good info, and valid point about the TIP mod. A more suitable TO-220 sub may be a Onsemi MJE15028/29 combo, having a rated fT of 30Mhz. It's been noted by at least one amp designer that, a higher fT rated output driver does help to control the CAS final outputs better. The stock of certified source 2N originals is scarce, there's some around from non-certified parts houses according to Octopart, but, I would think that using the Onsemi MJ's which are still in production would be a better modern choice than the TIP's.
Last edited:
I replaced TIP31c/32c on the AX6 board with MJ15028/29. True did solve the dull highs even with epitaxial output transistors (NTE181) problem. See https://www.diyaudio.com/forums/solid-state/236256-retro-amp-50w-single-supply-42.html for a picture of my point to point AX6 board in the ST120 chassis.
The TO220 parts don't fit the PC14 which is designed for EBC legs. I ruined one PC14 board with TO220 parts. When the leg broke from twisting, it pulled the land off the board. 2N5320/22 in TO39 were available from ST a few years ago and I wished I had access to Newark stock in 1986 when I was repairing the ST120 the first time. Before debit cards were invented individuals without a tax # could not buy from Newark. Digikey was a surplus house then with a mail flyer full of assembler leftovers.
The TO220 parts don't fit the PC14 which is designed for EBC legs. I ruined one PC14 board with TO220 parts. When the leg broke from twisting, it pulled the land off the board. 2N5320/22 in TO39 were available from ST a few years ago and I wished I had access to Newark stock in 1986 when I was repairing the ST120 the first time. Before debit cards were invented individuals without a tax # could not buy from Newark. Digikey was a surplus house then with a mail flyer full of assembler leftovers.
"The TO220 parts don't fit the PC14 which is designed for EBC legs. I ruined one PC14 board with TO220 parts. When the leg broke from twisting, it pulled the land off the board. 2N5320/22 in TO39 were available from ST a few years ago and I wished I had access to Newark stock in 1986 when I was repairing the ST120 the first time. Before debit cards were invented individuals without a tax # could not buy from Newark. Digikey was a surplus house then with a mail flyer full of assembler leftovers."
Yes, the "twisty leg" install of the TO-220's is an issue for sure. I have gone as far as designing new PC14 PCB's using EasyEDA, and making a coherent TO-220 footprint in the mix, but, I have not pulled the trigger on making an order as of yet. I do have one ST-120 to play with once I do, and I appreciate the info on the TIP / MJ Q3-4 issue.
Yes, the "twisty leg" install of the TO-220's is an issue for sure. I have gone as far as designing new PC14 PCB's using EasyEDA, and making a coherent TO-220 footprint in the mix, but, I have not pulled the trigger on making an order as of yet. I do have one ST-120 to play with once I do, and I appreciate the info on the TIP / MJ Q3-4 issue.