In 2019, I bought a Yamaha P2150 from UCSD’s surplus department. It was my first post in this blog! I don’t think I did much in-depth testing of this amp at the surplus warehouse before springing for it, and I already had my doubts that this might not be in good condition. Usually, this surplus department is the place that broken audio gear is put on sale to be snatched up (by people like me, I guess). If it indeed was broken, this would be a good project to restore it back to working condition.
Prologue: Ode to Yamaha
I’ve always respected Yamaha for their work in audio over the past few decades, despite a couple design mistakes1. What I loved specifically was that many of their old user manuals from the 70’s and 80’s include comprehensive guides that explain many of the fundamentals behind properly setting up that particular piece of equipment in a system. These writings were the work of Gary D. Davis, a writer in Santa Monica that Yamaha had commissioned, in hopes that releasing informative guides with their products would help them emerge as a leader in a field in which they had no reputation. About ten years after beginning to write these manuals, he went on to co-author Yamaha’s Sound Reinforcement Handbook in 1989, which has since been widely regarded as the “Bible of Live Sound.”
Chapter 1: Condition
My amp came in a very dusty state. And it wasn’t the type of fluff that accumulates on your fireplace mantle after a couple weeks – it was really caked onto there. It looks like nobody has touched the front panel or the power button for at least a decade. It also came in a plastic SKB rack case without any lids.
Also, this amp is a P2150C, which is the commercial version of the P2150. The only difference is just that the volume knobs are smaller and slightly recessed. In the pecking order of flashy aesthetic amps, I guess that’s a step down from big fancy knobs, which is in turn a step down from huge lighted VU meters.That’s no matter, I’m personally a fan of the minimally designed amps.
This amp comes with a fan that runs at a couple speeds depending on its internal temperature. I opened up the grille to find that it even comes with a nice little particle filter! The only catch is that these things work as long as they’re not deteriorating themselves… Guess what condition I found this filter.
The first couple times I turned on the amp, it seemed to power up fine: Protection light stayed on for a couple seconds as expected, and switched off once everything was charged up and ready to go. However, one channel was only sending a very faint crackly signal through to the output even at full volume, and the other channel was not passing signal at all.
I put this amp away for a few years and took it back out this year to finally have a swing at it. Upon turning it on, it wouldn’t exit protection mode. I tried probing around and noticed that the input stage was working fine (the part that unbalances the balanced signal), so the problems were in the power amp. Unfortunately, I ended up bridging the collector and base of one of the transistors in an early balancing stage of the power amp and ended up letting some magic smoke :(. Given that I did not want to figure out the entire signal path of this amp and risk blowing up more parts with my butter fingers, it’s probably worth my time to do a partial rebuild and swap out all of the transistors.
Chapter 2: Repair
Note: If you are reading this to rebuild a Yamaha P1150/P1250/P2150/P2250 amp yourself, do not follow these quantities! Spoiler alert, not all of the replacements were correct.
I replaced the following BJT’s (the following quantities are per channel):
- 2SA970 $\rightarrow$ ZTX795A (5x PNP)
- 2SC2240 $\rightarrow$ ZTX694B (2x NPN)
- 2SA1360 $\rightarrow$ KSA940 (3x PNP)
- 2SA1015 $\rightarrow$ KSA1281 (1x PNP)
- 2SC1815 $\rightarrow$ ZTX968 (2x PNP)
- 2SC3423 $\rightarrow$ KSC2073 (3x NPN)
- 2SC2238 $\rightarrow$ KSC2298 (1x NPN)
- 2SA968 $\rightarrow$ KSA1220A (1x PNP)
Mystery 1: The case of the incompatible pinouts
First off, the pinouts of almost every replacement transistor was not the same as the original part that it was replacing. Pins 2 and 3 of the small “D” shaped transistors had to be swapped so that it matched the 1-emitter 2-collector 3-base convention. I also had to take care to slide a couple sleevs of heat shrink over the crossed legs to make sure they weren’t bridging. It’s a little nerve-wracking soldering them in place, hoping that the temperature doesn’t melt the legs through the layers of heat shrink.
The larger transistor pinouts were just reversed, so I just had to install them backwards.
Mystery 2: The case of the asymmetrical solder joints
Aside from the main output transistors and thermistors being attached to the main heatsinks, there is also an NPN that is screwed onto the main sink (Channel A: Q214, Channel B: Q314). Its legs are attached to a piece of PCB that is intentionally broken off from the main power amp board upon assembly, and a set of wires connects it back to that board. I noticed that the transistor on one side did not have a solder joint connecting the collector leg to the wire. On the other channel, the collector was connected to the wire.
Given that these solder pads were not connected in the board, it looks like these connections from leg to wire were made by hand with globs of solder. Maybe the person responsible for making this connection just made a mistake and overlooked this one. I ended up soldering that leg to the wire to make it match the others.
Mystery 3: The case of the over-specced transistors
Okay, this one was my fault. I was trying to find a transistor to replace the 2SC2238 NPN, and I could not find one with the exact same specs. I did find one with the same maximum voltage limits but a higher maximum current: the Toshiba TTC-0002. I thought it wouldn’t hurt to over-spec this part, especially with current rating.
Turns out the difference between a 1.5A rated transistor and a 35A rated transistor is quite large, physically.
In hindsight, it probably didn’t make sense to use that power transistor when my home outlet wasn’t even rated for half of its rated current.
Mystery 4: The case of the DC output
I fired up the amp and braced myself for any arcs or smokes to occur… and nothing catastrophic happened! I was reading a healthy idle current of 0.4A, so I started to set the bias pots.
After watching each bias signal stabilize to 12mV, however, the protection light stayed on. And I noticed that the signal clip lights were also constantly lit even with no signal. The clip lights turn on once a threshold level of THD is detected and its current limiting is active, and the protection light turns on when there is a DC offset detected across the outputs and the relays disconnect the outputs from the speaker terminals.
I measured a stable 1.8V across the terminals of both outputs. How weird, the amp seemed to be idling well (i.e. no smoke or runaway current), but the DC voltage seems to be the problem.
My mind immediately went to the asymmetrical solder joints. Maybe that middle leg was supposed to be disconnected. I powered off and disconnected the collector leg on one channel and turned it back on. Protection light stayed on and DC was still present on both channels, but one clip light turned off! Maybe this is a good direction.
I powered off again and disconnected the collector leg on the other channel. The 1.8V offset persisted, and this time, the amp started drawing upwards of 5A of current. I gave it the benefit of the doubt, and decided after about 15 seconds that the current draw was not normal. In hindsight, 5A was probably a red flag, considering that meant I was drawing at least 600W of power in “idle” when this amp is rated for 330W max.
I decided to reconnect the transistor legs on both channels. Thankfully, I was able to get back to the stable current draw after re-soldering.
I then thought about my massive Toshiba transistor replacement. I had another look at the BOM and realized that Q219 and Q220 were supposed to be a complementary pair. I was able to replace Q219 with a similarly-looking part, but Q220 was replaced with that monster of a chip. Maybe there some imbalance between these two was causing the DC offset. I installed the original parts back into the amp, and the DC problem persisted.
Until next time…
Like all good crime mysteries, we are left with a cliffhanger. Every little mystery is not solved in full, and the clues point to yet another unsolved mystery.
I had a look at the rest of the transistors in my spreadsheet and noticed that quite a few of the original transistors were installed as complementary pairs, and I did not replace them with their contemporary complementary counterparts (try saying that three times fast).
When will this case be solved? Tune in next Mouser order…