Hi all...
I'm new to this forum, but far from new to DIY audio... (still have all my '70s vintage issues of Audio Amateur...)
Question... Recently I was given an old Dynaco ST-120. As expected, its non-functional and needs a complete re-build... and its an oldy... before the "TIP mod".
Now, I've brought several of these things back to life over the years, but its been at least 15 years since I've worked on my last one. So... what I'd like to know is what output transistors available today are people using to fix these?... I remember buying parts from an outfit in Ohio way back when that had all of Dynaco's old inventory, but they're loooong gone... and I also remember using 2N5630s (what Dyna put in their "400" if I remember correctly)... but those are now NLA (at least ones I would trust buying).
I've read of people using 2n3772s... but as with 2N3055s (which I've also seen people use) they don't seem to have the required ratings ( Dyna originally used a specially selected 3055s).
In reading on this forum and others, I've seen recommendations for MJ15003s, MJ802s and the previously mentioned 2n3772... but I'd love to hear input from anyone that's done a rebuild recently and hearing what they used and how they worked...
BTW... to ward off opinions as to whether this thing should be saved or not... Let me just say that I've built/owned/repaired several ST-120s over a 35 year span and know their pluses and minuses... this is just a nostalgia trip for me...
Thanks... jjw.
I'm new to this forum, but far from new to DIY audio... (still have all my '70s vintage issues of Audio Amateur...)
Question... Recently I was given an old Dynaco ST-120. As expected, its non-functional and needs a complete re-build... and its an oldy... before the "TIP mod".
Now, I've brought several of these things back to life over the years, but its been at least 15 years since I've worked on my last one. So... what I'd like to know is what output transistors available today are people using to fix these?... I remember buying parts from an outfit in Ohio way back when that had all of Dynaco's old inventory, but they're loooong gone... and I also remember using 2N5630s (what Dyna put in their "400" if I remember correctly)... but those are now NLA (at least ones I would trust buying).
I've read of people using 2n3772s... but as with 2N3055s (which I've also seen people use) they don't seem to have the required ratings ( Dyna originally used a specially selected 3055s).
In reading on this forum and others, I've seen recommendations for MJ15003s, MJ802s and the previously mentioned 2n3772... but I'd love to hear input from anyone that's done a rebuild recently and hearing what they used and how they worked...
BTW... to ward off opinions as to whether this thing should be saved or not... Let me just say that I've built/owned/repaired several ST-120s over a 35 year span and know their pluses and minuses... this is just a nostalgia trip for me...
Thanks... jjw.
The datasheet for the 2N5630 gives fT as 1MHz
For both MJ15003 and MJ802 it is 2MHz.
Close enough? Probably.
For both MJ15003 and MJ802 it is 2MHz.
Close enough? Probably.
If you don't want to gut it, try MJ15015. It has the Vceo and an actual SOA rating . MJ802 and MJ15003 would work but cost about 3 times as much. All these fast transistors could oscillate, so do the part of the TIP mod that involves putting 10-20 pf capacitors from b to e on the driver transistors.
Running mine at 20W more than 3 hours got the whole case and output capacitor so hot it melted the solder on the snap in terminal of the output capacitor. The resulting fireball under the lectern really impressed the junior choir. So I put PCAT fans on a bracket to blow on the two output transistor flanges. They are outside the cover and look like mouse ears. I put covers over them to keep the cat out of the blade. I power the fans with a 9v V wall transformer so they don't run full speed. Now I run my ST120 18 hours a day at home without heat or problem. Another guy looped the wires to the new snap in output cap through a little insulator board so they don't want to jump off the output capacitor. I recommend this. Snap in output caps are so much cheaper than screw terminal ones. Mine has all new electrolytic caps, except no 20 uf non polars.
Eliminating the cold crossover distortion with 6 transistors/channel really peps up the sound, along with the faster output transistors. djoffe came up with a design for that before he started selling LM3886 kits. I installed it and love the result. See this thread www.diyaudio.com/forums/solid-state/156627-dynaco-stereo-120-can-beautiful.html I'm using ZTX653 transistors from diodesinc instead of the 2N3904 on the input pair because when I got a little heat sink compound on the ring terminal the .5 ohm resistor is mounted on, it put 80V on the input transistors and blew them up. I built the two bias boards on 1.5"x3" fiber boards I drilled holes in and mounted them on angle brackets above the PC15's. They fit under the cover.
Oh, and somebody did a simulation on a leak Delta 70 with similar geometry and found the distortion got lower with TIP3055 transistors instead of 2n3772, and even lower with something even faster. I'm using NTE60's actually (probably MJ802 equivalents, I bought them in 1985) and the sound is now good enough I can't tell the difference between this and the CS800s amp which has no capacitors in the sound path whatsoever. Both amps sound better than my recapped ST70 with new Output and rectifier tubes. One channel is still using the RCA TO5 driver transistors, so I don't think these are a limitation. Both channels have a 4.7 uf 50V aerovox gold ceramic input capacitor, which many people will say is stupid, but solved the problem of a frying pan noise I got with new tantalum input caps.
One further tip. I found a modern NTE60 for the power transistor on the current regulator PCB PC14 caused the output current to trip off at 2 amps total instead of the designed 6.5. Actually that is the way it was when I melted the output capacitor solder. To to replace that old low gain RCA transistor, you have to put a resistor between b & e until the trip point is correct. I'm using a 225 W 10 ohm adjustable ohmite resistor to set the trip point, just connecting it to the feed to the PC15's. PC14 is not really a useful safety device, but it does regulate the open circuit voltage to 80 from 100 the transformer puts out, so that part is useful. I installed two 5 amp fuses for the separate PC15's.
Running mine at 20W more than 3 hours got the whole case and output capacitor so hot it melted the solder on the snap in terminal of the output capacitor. The resulting fireball under the lectern really impressed the junior choir. So I put PCAT fans on a bracket to blow on the two output transistor flanges. They are outside the cover and look like mouse ears. I put covers over them to keep the cat out of the blade. I power the fans with a 9v V wall transformer so they don't run full speed. Now I run my ST120 18 hours a day at home without heat or problem. Another guy looped the wires to the new snap in output cap through a little insulator board so they don't want to jump off the output capacitor. I recommend this. Snap in output caps are so much cheaper than screw terminal ones. Mine has all new electrolytic caps, except no 20 uf non polars.
Eliminating the cold crossover distortion with 6 transistors/channel really peps up the sound, along with the faster output transistors. djoffe came up with a design for that before he started selling LM3886 kits. I installed it and love the result. See this thread www.diyaudio.com/forums/solid-state/156627-dynaco-stereo-120-can-beautiful.html I'm using ZTX653 transistors from diodesinc instead of the 2N3904 on the input pair because when I got a little heat sink compound on the ring terminal the .5 ohm resistor is mounted on, it put 80V on the input transistors and blew them up. I built the two bias boards on 1.5"x3" fiber boards I drilled holes in and mounted them on angle brackets above the PC15's. They fit under the cover.
Oh, and somebody did a simulation on a leak Delta 70 with similar geometry and found the distortion got lower with TIP3055 transistors instead of 2n3772, and even lower with something even faster. I'm using NTE60's actually (probably MJ802 equivalents, I bought them in 1985) and the sound is now good enough I can't tell the difference between this and the CS800s amp which has no capacitors in the sound path whatsoever. Both amps sound better than my recapped ST70 with new Output and rectifier tubes. One channel is still using the RCA TO5 driver transistors, so I don't think these are a limitation. Both channels have a 4.7 uf 50V aerovox gold ceramic input capacitor, which many people will say is stupid, but solved the problem of a frying pan noise I got with new tantalum input caps.
One further tip. I found a modern NTE60 for the power transistor on the current regulator PCB PC14 caused the output current to trip off at 2 amps total instead of the designed 6.5. Actually that is the way it was when I melted the output capacitor solder. To to replace that old low gain RCA transistor, you have to put a resistor between b & e until the trip point is correct. I'm using a 225 W 10 ohm adjustable ohmite resistor to set the trip point, just connecting it to the feed to the PC15's. PC14 is not really a useful safety device, but it does regulate the open circuit voltage to 80 from 100 the transformer puts out, so that part is useful. I installed two 5 amp fuses for the separate PC15's.
Last edited:
Thanks everyone for the input... this is a great forum. My primary concern is with stability when using alternate transistors... as its unlikely there would be any sonic difference.
The mods suggested are definitely interesting and food for thought, but I think I'll leave this one as stock as possible...
Thanks again... jjw
The mods suggested are definitely interesting and food for thought, but I think I'll leave this one as stock as possible...
Thanks again... jjw
The original transistor were SOOOO slow that it is hard to not improve the sound by replacing them. They don't make anything that slow anymore. I've got one survivor 1970 TO3 transistor, I can't measure the ft but the gain is pitiful. It is acting as the DC regulator in PC14. Since modern O.T.s have ft of 2-4 mhz, I suggest the TIP mod with the resistor across the output jacks, and the 20 pf caps across driver b-e, even if you don't change the driver transistors to TIP's. I did those mods. RF oscillation is a nasty thing that heats things up, and these sure don't need any extra heat source as bad as the heat sinks are. I left the RCA to5 transistors on one side. The side with the avalanced driver transistors I used NTE49 & 50 , and I had to use an NTE 249 for the input transistor. Because all the old stuff was blown up on one channel. All except the 5.1V zener that everybody says blows every time. If I had to do it again I would use MPS8099 for the little transistor and TIP31C and TIP32C for the drivers. The pinout is different on the TO220's than TO5, you have to twist the legs around.
Last edited:
"The original transistor were SOOOO slow that it is hard to not improve the sound by replacing them."
Hence my concern for stability and wondering what people were using successfully these days... the last thing I want to do is go through the process and then wind up with a smoldering 60 wpc oscillator... ;-)
As stated earlier, the amp I'm working with now is an oldie... before the "TIP" update... so I'll be bringing it up to the last factory configuration. Also, there is evidence of previous (attempted) repair demonstrating a lack of experience with this designs quirks...
...if I had a quarter for every ST-120 I've seen like this....
Thanks again... JJW
Hence my concern for stability and wondering what people were using successfully these days... the last thing I want to do is go through the process and then wind up with a smoldering 60 wpc oscillator... ;-)
As stated earlier, the amp I'm working with now is an oldie... before the "TIP" update... so I'll be bringing it up to the last factory configuration. Also, there is evidence of previous (attempted) repair demonstrating a lack of experience with this designs quirks...
...if I had a quarter for every ST-120 I've seen like this....
Thanks again... JJW
The ON Semi MJ15015 datasheet I have shows Ft between 800 khz and 6 mhz. The parts, if they perform at 5.9 mhz, are very fast, and should be protected by the driver b-e capacitor and the output resistor of the TIP mod.
The 2N3772 are also selected as it says in my manual:
Q5 2N3772 571844 40-90 BETA @ 1A
Q6 2N3772 571844 RED DOT 60-90 BETA @1A
Now that is a difficult task. Better to look for alternatives.
BTW my 120 is 40 years old and origginal (all caps should be replaced) still works and had no problems, But I want to part it out for a project. So the amps and power supply are for sale as of yesterday. All I want from the parts are the big heatsinks. If you interested in buying mine for a reasonable price PM me. I want your heatsinks (3 of them) so I dont have to take mine apart (better for you). You will get the manual with the latest parts list.
I did a rebuilt 3 weeks ago,,I used 2n3773 onsemi transistors I,ve had no problem with them. Evette
I have recently refurbished (3) ST120's and (2) St80's. I have (4) left to go. I am running out of the 2N3772 Transistor pairs. Can someone please tell me why one is marked with a red dot and has to go in it's respective position on the heat sink ? I read the TIP Mod and everything else I could get my hands on to explain what 40-90 BETA @ 1 Amp and 60-90 BETA @ 1 Amp actually means ? Except that one takes over for the other during some sort of change over. Also, why not just use two MJ15003's or two NTE 130's, which more resembles the 571844 that have a DC current gain HFE 20-70 @ Ic=4 Amps ? Did HFE replace the Beta term ? Will the MJ15003 be O.K. to run the amps with no problems. By the way, sellers are offering thease as replacements already.
Beta and hFE are pretty much the same number in different context.
Beta/hFE vary a LOT with current, especially in the day the ST120 was designed. I suspect the "1 Amp" number is applicable to the original 2N3772 parts (which varied over the years).
So while I could show you how to do the specified test, probably best to see what parts others have used with good result.
FWIW: I used to get paid a salary for (among other duties) repairing ST120/80 amps. We did a lot of fun stuff with them, but in retrospect I think they are crude. Do not let you dissuade you from fixing them.
Beta/hFE vary a LOT with current, especially in the day the ST120 was designed. I suspect the "1 Amp" number is applicable to the original 2N3772 parts (which varied over the years).
So while I could show you how to do the specified test, probably best to see what parts others have used with good result.
FWIW: I used to get paid a salary for (among other duties) repairing ST120/80 amps. We did a lot of fun stuff with them, but in retrospect I think they are crude. Do not let you dissuade you from fixing them.
Red Dot On Transistors
Crude and simple is what a lot of people were satisfied with. So to help me with this perplexing red dot issue, you're suggesting I pay a salary for the answer or see what others are doing ? Well, I have read a lot of what others are doing up to and including changing the amp. boards, while partially retaining some of the original design. Is there anyone on here that can answer the original question ? Anyone ? Using the Peak Meter, I can get all the info I need about the transistor but what good is a lot of information if the application of a modification does not explain it's purpose. Nobody, and I mean anybody in all the articles anywhere, explain why one transistor is discriminated. Also, I have taken so many of these apart and not one had a 1,000 ohm resistor across the speaker output jacks as required. What's weird is that all have had various stages of the TIP Mod put in, but never the complete Mod. Others that do use them, up to 3 watts, are also installing further variations of their own modifications. I'm looking to just put it back to factory specification, which includes only the TIP Mod. I know others were issued but crude and simple is good for now. Can anyone help here ?:confused
Crude and simple is what a lot of people were satisfied with. So to help me with this perplexing red dot issue, you're suggesting I pay a salary for the answer or see what others are doing ? Well, I have read a lot of what others are doing up to and including changing the amp. boards, while partially retaining some of the original design. Is there anyone on here that can answer the original question ? Anyone ? Using the Peak Meter, I can get all the info I need about the transistor but what good is a lot of information if the application of a modification does not explain it's purpose. Nobody, and I mean anybody in all the articles anywhere, explain why one transistor is discriminated. Also, I have taken so many of these apart and not one had a 1,000 ohm resistor across the speaker output jacks as required. What's weird is that all have had various stages of the TIP Mod put in, but never the complete Mod. Others that do use them, up to 3 watts, are also installing further variations of their own modifications. I'm looking to just put it back to factory specification, which includes only the TIP Mod. I know others were issued but crude and simple is good for now. Can anyone help here ?:confused
I wouldn't worry about the red dot. In 1966 TO3 transistor had a hard time holding off 65 vdc. Dynaco bought specially selected ones out of the 2n3055 process. Now holding off that voltage is a snap. None of my 1970 burnt out transistors had any red dots. They did have rca 6 digit part numbers. If you do have ancient red dot transistors, do a Iceo test at 65 v or more, and see if the red dot ones leak less current C-E. Put a 1 k resistor in series with C to limit the breakdown current and not damage the part.
IMHO the snubber (1000 ohm series capacitor) on the output jacks is to prevent oscillation triggered by FM police radios, business band, CB, Ham, ete. Those 11 turns around the output caps weren't much of a zobel, and they were far away from the output jacks, too. I could see the output jacks radiating RF to the input transistor pretty easily, it was only 3" away from the jack on the left side.
Another trick to stop fast Ft transistors from oscillating is use a 10 ohm 1 watt carbon resistor instead of a wire between driver & base. I did that to provide a cheap "fuse" in case the output transistors overheated again, but Mooly does that to his replacement transistors in his Leak delta 70's. I think the dynaco ST-120 was copied from the leak; but who knows, the schematics didn't have copyright dates. Or the Armstrong 621, also british.
Faster output transistors than 200 khz cleans up some of the distortion 1966 reviewers complained about. Letting them warm up by playing loud helps too. The idle bias current capacitor does eventually charge up and damp down the cold quiet crossover distortion problem.
I put the djoffe closed feedback idle bias mod board in mine, but it tends to blow the sense transistor and run OT idle current up to 200 ma or more. Dynaco Stereo 120...can be beautiful - diyAudio
soa violation of the sense transistor maybe? MPSW06 might help, the 1 watt version of MPSA06. With my twin fans on the heatsinks my unit withstood that bias current for some considerable time.
The last burnup damaged a PC14 badly so instead of building it point to point again, I built an Apex AX6 for that channel. Sounds exactly the same as the PC14 with djoffe bias control except a small volume imbalance. Both channels will put out 70 W peaks on loud R&R tracks with heavy bass. On my 8 ohm SP2-XT speakers.
I'm listening to the unit tonight. My output transistors are NTE60, which the datasheet said was the same SOA as an MJ802. Which they probably were. I bought them about 1986. I wouldn't hesitate to use MJ15015, MJ15003, or MJ21194 either. 2n3773 cost more now, we really don't need the heavy bond wire of those. I wouldn't obsess about the gain either, especially if using On semi G suffix parts which have high gain anyway. The wattage above (1 second at a time on an analog meter, not a o-scope peak) indicates my RCA TO5 transistors are having no trouble driving the output transistors beyond the specification. I listened to the amp the rest of the evening after that max power test , and didn't blow up anything either.
IMHO the snubber (1000 ohm series capacitor) on the output jacks is to prevent oscillation triggered by FM police radios, business band, CB, Ham, ete. Those 11 turns around the output caps weren't much of a zobel, and they were far away from the output jacks, too. I could see the output jacks radiating RF to the input transistor pretty easily, it was only 3" away from the jack on the left side.
Another trick to stop fast Ft transistors from oscillating is use a 10 ohm 1 watt carbon resistor instead of a wire between driver & base. I did that to provide a cheap "fuse" in case the output transistors overheated again, but Mooly does that to his replacement transistors in his Leak delta 70's. I think the dynaco ST-120 was copied from the leak; but who knows, the schematics didn't have copyright dates. Or the Armstrong 621, also british.
Faster output transistors than 200 khz cleans up some of the distortion 1966 reviewers complained about. Letting them warm up by playing loud helps too. The idle bias current capacitor does eventually charge up and damp down the cold quiet crossover distortion problem.
I put the djoffe closed feedback idle bias mod board in mine, but it tends to blow the sense transistor and run OT idle current up to 200 ma or more. Dynaco Stereo 120...can be beautiful - diyAudio
soa violation of the sense transistor maybe? MPSW06 might help, the 1 watt version of MPSA06. With my twin fans on the heatsinks my unit withstood that bias current for some considerable time.
The last burnup damaged a PC14 badly so instead of building it point to point again, I built an Apex AX6 for that channel. Sounds exactly the same as the PC14 with djoffe bias control except a small volume imbalance. Both channels will put out 70 W peaks on loud R&R tracks with heavy bass. On my 8 ohm SP2-XT speakers.
I'm listening to the unit tonight. My output transistors are NTE60, which the datasheet said was the same SOA as an MJ802. Which they probably were. I bought them about 1986. I wouldn't hesitate to use MJ15015, MJ15003, or MJ21194 either. 2n3773 cost more now, we really don't need the heavy bond wire of those. I wouldn't obsess about the gain either, especially if using On semi G suffix parts which have high gain anyway. The wattage above (1 second at a time on an analog meter, not a o-scope peak) indicates my RCA TO5 transistors are having no trouble driving the output transistors beyond the specification. I listened to the amp the rest of the evening after that max power test , and didn't blow up anything either.
Last edited:
Thanks for your help ! While at college, my Organic Chemistry professor sent us home with a take home test the answers to which he said, could not be found in any publications, textbooks etc.. After spending all night in the Weslyan Science library, I went home with the best case scenario for ten answers to the test. I did manage to get them all correct. There was one that took me a very long time. This thing with the dot started to stick in my craw, like the answer to that question, to the point of which, I looked up everything there was and could not find anything about it. It was just the point that a statement was made by Dynaco and I wanted to understand why but couldn't.
Anyway, I agree that there are never any dots on the old units I refurbished and I have two more on my bench that don't have them. I have a room full of refurbished analog stereo receivers all covered up. I just recently (last 2 yrs.) started with just analog audio amp products. Anyway, thanks.
Anyway, I agree that there are never any dots on the old units I refurbished and I have two more on my bench that don't have them. I have a room full of refurbished analog stereo receivers all covered up. I just recently (last 2 yrs.) started with just analog audio amp products. Anyway, thanks.
It occured to me that dynaco ST-120 has rather reviled feedback to the resistor of the emitter of the input transistor. It is possible that the gain of the feedback in the lower voltage part of the curve (below centerline) is lower than the gain in the upper part of the curve (above centerline). So maybe they selected higher gain output transistors for the bottom one and lower gain output transistors as the upper one. I don't know. My replacement transistors were NTE matched pairs, NTE60MP, so I missed out on this trick to maybe lower distortion at loud volumes. I can't do sims to test this theory on lubuntu (linux) PC op system, and I don't have a distortion analyzer, or a smart phone to run sims either.
The biggest problem of the ST-120 is IMHO crossover distortion at lower volumes due to the weird idle bias current configuration. Since I listen most at 1.5W average with occasional 50 W peaks on classical music, low volume distortion is definitely my empahsis. My ST-120 + djoffe bias control sounds very like my Peavey CS800s which is rated at .02% HD full power, and has no speaker cap. The ST-120 is also miles ahead of my fully maintained dynaco ST70 in clean sound at 1.5 W.
The biggest problem of the ST-120 is IMHO crossover distortion at lower volumes due to the weird idle bias current configuration. Since I listen most at 1.5W average with occasional 50 W peaks on classical music, low volume distortion is definitely my empahsis. My ST-120 + djoffe bias control sounds very like my Peavey CS800s which is rated at .02% HD full power, and has no speaker cap. The ST-120 is also miles ahead of my fully maintained dynaco ST70 in clean sound at 1.5 W.