Replacing bad output in vintage Yamaha amp, which outputs would result in the best end result?

I’m fixing a Yamaha CA-800 that had multiple issues. I’m trying to basically make it as good as it can possibly be. I replaced all of the 2SC458, 2SC1345, Sony 2SA706 and Sony 2SC1124 drivers on the amp driver boards. As well as all of the caps. One channel has bad outputs, they are 2SA663 and 2SC793. Sorry for the crappy image, I can’t seem to find a clean scan of the CA-800 circuit.

IMG_7876.jpeg


Here are the basic specs of the original outputs.

IMG_7875.jpeg

IMG_7871.jpeg


I can go with the same package and put in some new MJ21195G and MJ21196G and be done with the easy swap. They exceed in every way and have a close transition frequency of 4 MHz.

The other option would be using some TO3-P packages and using NJW0281G and NJW0302G, just bend the legs and fit them in and they’d be sitting towards the bottom of the heatsink.

IMG_7872.jpeg


Here are the specs of the TO3-P devices.

IMG_7873.jpeg

IMG_7869.jpeg



Again, they pretty much exceed the originals in every single aspect, but have a higher Ft of 30 MHz. How would the 30 MHz do in this circuit. Curious if they’d be any better, even slightly.

Thoughts?
 
I will stay on to3 mj 15002/3 15022/23 or mj 21194/95, finally this family and above all, no terminals that are too fast.
This amp doesn't have much interest but it would be a shame to make it oscillate, it's still a nice device.
Also look at the pinout if you stay in TO3.
 
  • Like
Reactions: 1 user
I will stay on to3 mj 15002/3 15022/23 or mj 21194/95, finally this family and above all, no terminals that are too fast.
This amp doesn't have much interest but it would be a shame to make it oscillate, it's still a nice device.
Also look at the pinout if you stay in TO3.

If I did go the TO3 route I would definitely be using that same family, as the 21195/21196 would really be the only option I’d use.

Regardless of whatever output device I use or have suggested in this thread, the pin out will not be an issue.

You think it’s possible the 30 MHz devices would cause oscillation issues?

Dan
 
  • Like
Reactions: 1 user
About 10 years ago, I went through this with my Marantz 1200. The OP's question is regarding changing the output transistors to transistors with a faster transition frequency. Short answer is that it made a great oscillator out of my amplifier with a similar amplifier circuit. On the Marantz unit, it did not oscillate only when the pre-driver transistors had a similar hFE and ft value.

After another couple years dealing with 50 year old pcb's and lifted traces, I had some Leach Amp PCBs made and changed out the crusty old Marantz power amp boards. It sounds as good as the original boards to my old man ears.

https://leachlegacy.ece.gatech.edu/lowtim/

You'll have lots of people poo-pooing all over this, telling you to build a Pass amplifier or some modern equivalent. But then you'll miss your wife/kids asking you why you spend so much time on that old amplifier, until you tell them a new, modern equivalent costs more than your car out in the driveway. And- you don't get to spread your o-scope, DC power supply, DMM, signal generator and all the impressive looking cables and connectors all over the covered dining room table. It took me a long time to learn how to use all this gear, let alone hook it all up.
 

Attachments

  • Leach Amp 1.JPG
    Leach Amp 1.JPG
    309.9 KB · Views: 36
  • Leach Amp Schematic.pdf
    121 KB · Views: 36
  • Screen Shot 2020-06-06 at 10.54.47 PM.png
    Screen Shot 2020-06-06 at 10.54.47 PM.png
    215.9 KB · Views: 35
  • Like
Reactions: 1 user
About 10 years ago, I went through this with my Marantz 1200. The OP's question is regarding changing the output transistors to transistors with a faster transition frequency. Short answer is that it made a great oscillator out of my amplifier with a similar amplifier circuit. On the Marantz unit, it did not oscillate only when the pre-driver transistors had a similar hFE and ft value.

After another couple years dealing with 50 year old pcb's and lifted traces, I had some Leach Amp PCBs made and changed out the crusty old Marantz power amp boards. It sounds as good as the original boards to my old man ears.

https://leachlegacy.ece.gatech.edu/lowtim/

You'll have lots of people poo-pooing all over this, telling you to build a Pass amplifier or some modern equivalent. But then you'll miss your wife/kids asking you why you spend so much time on that old amplifier, until you tell them a new, modern equivalent costs more than your car out in the driveway. And- you don't get to spread your o-scope, DC power supply, DMM, signal generator and all the impressive looking cables and connectors all over the covered dining room table. It took me a long time to learn how to use all this gear, let alone hook it all up.
I love your story and it’s good to see someone who has gone through something similar. So are you saying the amplifier behaved once you used drivers with similar specs? Currently the drivers are Toshiba TTA004b/TTC004b which have similar gain and an Ft of 100 MHz. Not sure how that combo would do.

Id have to check all of the specs, but I believe the MJE243/253 would easily fill in as drivers and their Ft is 40 MHz which is quite a bit closer to the spec of the 0281/0302 combo. Regardless I’ll follow through with finishing this as I’m repairing it for a friend who really loves this piece. Clearly I should just tell him to build the Wolverine or one of Pass’s designs, but yeah, cost might be an issue lol. I believe this has close family ties and he asked me to make it the best performing it can be, so I was just wondering if faster output devices might be able to squeeze a bit of performance from it.

Edit: yes, it looks like the MJE243/253 are far more than robust enough to be in there as drivers. So maybe they having an Ft of 40 MHz would be a better pick?? Then again, maybe the TO3 packages would just be better in the long run as any added benefit from the faster TO3-P wouldn’t be worth it. If they’re better I’d like to use them as I have about 350 each of the 0302/0281 where as I only have about 150 each of the MJ21195/96. Plenty of each, but if I have more of one….

Dan
 
Last edited:
Remember that these devices were made before Mass Flow Controllers and other modern devices were available. That's also why curve tracers were commonly used to match hFE and other transistor characteristics. Today, we can use components with the same lot number which assures the devices have similar operating characteristics.

My opinion (you know what is said about opinions,) is use the MJ21195G and MJ21196G. If all the other transistors are still good- and some of them on my Marantz amp only showed they were bad when I put them on the Tektronix Curve Tracer- then you should be okay with the MJ transistors.

I mentioned the original pre-driver transistors on the Marantz amp were difficult to find/buy, but it's also when I learned that the Ft value was important to consider so that you don't get oscillation.

You may also consider the MJ15004G (PNP) and MJ15001G (NPN) transistors for the outputs. These would use pre-drivers MJE15031G and MJE15030G. I'm pretty sure you can still buy these from Mouser/Digikey/Newark/your favorite source.

Dan