Yamaha P-2200 measurements

Yamaha P-2200 uncased

Yamaha P-2200 PA power amplifier

Measurement results for Yamaha P-2200

Measurement results for Yamaha P-2200

PDF version

After many hours of work, I finally got my Yamaha P-2200 into a state worthy of being tested. The left channel is 100 % original, save for new electrolytic capacitors. It should be fairly representative of a completely original unit. The right channel has received major repairs. The high vol = FS noise floor on the right channel is not due to the repairs done to the amplifier, but rather the factory wiring of the amplifier; it’s mostly made up of 50 Hz hum. There’s probably a signal wire going too close to the transformer, somewhere.

All in all, the Yamaha P-2200 is a very well-performing amplifier, especially for its age and the fact that it’s Yamaha’s first ever dive into the world of PA amps. With plenty of passively-cooled power, tank-like build quality, an impressive noise floor and a good damping factor to boot, Yamaha made sure to make their PA power amp début one for the ages!

Just for fun, four audio gear gifs!

These files are huge (about 3 MB each), so give it some time to load!

JVC SEA-M9 in spectrum analyser mode

JVC SEA-M9 in spectrum analyser mode

The JVC SEA-M9 in spectrum analyser mode, filmed during an earthquake.

The JVC SEA-M9 in spectrum analyser mode, filmed during an earthquake.

The dancing meters of the Yamaha P-2200

The dancing meters of the Yamaha P-2200

Freshly refoamed Acoustic Research AR-7 woofer doing its thing.

Freshly re-foamed Acoustic Research AR-7 woofer doing its thing.

Substituting the SFC6120 double transistor in the Tandberg TR-1055

Tandberg TR-1055

Tandberg TR-1055

This nice-looking unit came into the shop with a noisy left channel as well as an intermittent DC offset. After seemingly repairing it by replacing the output transistors, the customer returned it complaining about the same issue arising after a few hours of use. After some further troubleshooting, the problem was found to be a Motorola-branded double transistor in the power amplifier.

The SFX6120 double transistor

The SFC6120 double transistor in the negative feedback circuit

Neither the part nor a datasheet for it was anywhere to be found, so a substitute had to be manufactured. I settled for a matched pair of the common KSC1845 to do the job. Gain matching is important, as an unmatched pair will result in a DC offset on the output of the amplifier.

Installing the transistors is easy, as the pin-out for the SFC6120 is printed on the circuit board.

Transistor 1 installed

Transistor 1 installed

Both replacement transistors installed

Both replacement transistors installed

However, my KSC1845s had roughly twice the gain of the SFC6120 (380 vs. 160), which resulted in a considerable increase of the amplifier’s gain. To counteract this, feedback resistor R712 was decreased from 10 kOhm down to 3,6 kOhm.

R712, the blue resistor, determines the gain of the amplifier module

R712, the blue resistor, determines the gain of the amplifier module

Since the modification altered the gain of the amplifier, I decided to perform it on both channels to ensure proper matching and guard against future SFC6120 failures. It is important to ensure thermal coupling between the two transistors, in order to guard against DC offset when the amplifier warms up. That’s probably why Tandberg decided to use a double transistor in the first place.

Somewhat unexpectedly, the THD+N of the amplifier decreased from 0,08 % at rated output into 4 Ohm, to a mere 0,033 % after the modification. (Measured with my HP 339A at 1 kHz)

After many hours of heavy load testing into 4 Ohm, I think this unit is ready to go back to the customer again – and hopefully not return!

The finished pair of output modules

The finished pair of output modules. Note the thermal goop on the KSC1845s. The two modules have been (partially) recapped at different occasions.

An untouched Tandberg TR-1055 output module

An untouched Tandberg TR-1055 output module (For reference)

JVC SEA-M9 in the house!

Who would think that such an unusual unit would appear on Åland? It’s been kept very well – it’s damn near mint aside from the little scratches on the top. The original service manual was a nice bonus, since it doesn’t seem to be available anywhere on the net.

The Yamaha P-2200 restoration begins

The fun part about giving this guy a first look-through was that it did actually turn on, and it did actually play on both channels … barely. The right channel was heavily distorted and  about half of the time when the power switch was flicked, it’d only flicker and buzz for a while before setting the breaker off.

The turn-on issue was caused by a severely worn-down power switch. The contact surfaces had literally vaporised and disappeared from switching that 1,3 kVA transformer for thirty years. There’s no remedy able to get that into shape for switching big loads again. At least not without giving it a helping hand.

A helping hand in the form of a 230 V relay. I even went the extra step and added a capacitor across the coil in order to make the life of the switch easier.

Testing the relay circuit (animated GIF – clickit!)

While I wouldn’t trust a transformer to this, it should switch the relay for many years to come.

The hot-glued, PVC-taped relay board is simply placed loosely behind the front panel.

It workes very well, however there should be two relays used if you’re to be picky; the original switch is configured to turn the two primary windings on at slightly different times. This is in order to reduce the start-up current somewhat.

(I was later informed that switches such as these, commonly found in AT computer chassis’ would probably work as replacements.)

With the power switch issue out of the way, let’s move on to some intensive maintenance of the actual electronics. Starting with a re-cap.

Two 470 µF capacitors for the size of a thirty-year-old 220 µF

Old and nasty.

Worth noting is that C101 and C201 (100 µF/10 V) were actually completely dried-out; they didn’t even register on the meter. It’s odd that the amplifier worked at all without them, as they seem to be in the signal path.

It’s hard to imagine that these capacitors are both more reliable and more powerful than the old ones.

While new caps are always fresh and nice, it didn’t remedy the right-channel bias issue. The cause for that was two-fold and somewhat unexpected. For starters, the bias adjustment potentiometer looked like this on the inside:

The less obvious issue, however, was that a small (0,47 µF) tantalum capacitor, C113, was shorted. It’s connected across the base and collector of one of the transistors in the bias circuit, as well as across the bias adjustment potentiometer. Short-circuiting C113 is equivalent to turning the bias down to zero.

The affected channel had been serviced before and a lot has been replaced. I’d wager that C113 failed at that time and that the service technician didn’t bother actually listening to the amplifier before giving it back to the customer. Nevertheless, replacing C113 (with a 2,2 µF electrolytic; C213 was also replaced to avoid any channel imbalance) and VR101 (the bias pot) remedied the problem. The channel still has a slight DC offset of 20 mV or so, but I’ll live with that for the moment; there’s still a few transistors that should be renewed.

The same service technician also installed MJ15015 and MJ15016 transistors for the output of the channel – transistors with a maximum VCEO of 120 V! The power supply rails in the P-2200 add up to 160 V; 120 V transistors have no place in such an amplifier. It may work, but any safety margins intended by the designer are lost. If the output terminals are shorted or over-loaded, these transistors will be up in smoke before you can even think “blown outputs”! The proper MJ series transistors to use would be MJ15022 and MJ15023.

For shame.

Anyhow, with that bitter-sweet story out of the way, only the cleaning-up remains. I’ll spare you the narration, enjoy!

And that’s it! The only things remaining to do is to give the incompetently serviced channel another overview in the future, replacing driver transistors and of course outputs. One of the thermal indication thermostats has also failed, so I might replace that some time in the future. That’s a minor issue though.

Thank you for visiting, I hope you enjoyed reading about this project as much as I enjoy listening to it!