diyAudio

diyAudio (http://www.diyaudio.com/forums/)
-   Solid State (http://www.diyaudio.com/forums/solid-state/)
-   -   PS Audio 200c repair and upgrade (http://www.diyaudio.com/forums/solid-state/136822-ps-audio-200c-repair-upgrade.html)

SQLGuy 18th January 2009 05:52 PM

PS Audio 200c repair and upgrade
 
Hi all,

I am in the process of repairing a PS Audio 200c. It's previous owner blew out one channel by accidentally shorting the outputs, internally disconnected the rails from that channel, and then later blew the other channel due to a bad biwire.

I haven't finished pulling everything yet. All the outputs I've found so far are shorted, and it looks like at least one of the driver pairs are shorted as well. This amp originally used NEC outputs, 2SD555's and 2SB600's. One channel had been repaired in the past with NSC(?) versions of these numbers. Anyway, since these devices are a bit scarce and highly counterfeited, I looked into subs. In the later version of this amp, the 200cx, PS Audio used MJ21193's and MJ21194's. I checked the datasheets and these looked like a pretty good sub; Paul McGowan was also kind enough to confirm that they should work fine.

So, at this point, I'm swapping out the blown outputs, drivers, and one burned resistor for sure. I am also planning to replace the on-board electrolytics. They are mostly Matsushitas, which I have found to be very long-lived and heat-resistant, but these are the 85 degree variety and they are over 20 years old as well. The stock filter caps actually still read about 17000uF on my Fluke (they're rated at 15000uF) but I'll probably upgrade/replace those once the rest of the amp is up and running. I also saw a post on audioreview.com where someone recommended replacing the original bridge rectifiers with FRED rectifiers. I would appreciate your thoughts on that suggestion.

Lastly, I have some questions about fuses for this amp. I know that the stock output fuses are supposed to be 8A (speaker protection, not rail). This seems a bit odd in that the amp is rated 400W continuous into 4 Ohms, and 800W continuous into 2 Ohms. Perhaps the manufacturer expected larger fuses to be installed when driving lower impedance loads? Also, the line fuse was a 5A slo-blo. This doesn't seem a likely value and I wonder if that's just something the PO swapped in at one point. I would have expected more like 12A+. Any ideas what it's supposed to be?

One other note: I saw another thread somewhere where someone repairing one of these mentioned that he didn't see any heat sink grease under the outputs. That degraded into a bit of back and forth about clear silicone grease, evaporation, new micas, etc. All of this was sort of out of place on that thread because these amps use copper bus bars to couple rail current from the filter caps to the output transistors. The cases are supposed to be connected to the copper, and the copper is, in turn, insulated from the rest of the heat sink. One question, though: is there any electrically conductive heat sink compound that could and/or should be applied here?

Thanks,
Paul

SQLGuy 19th January 2009 12:18 AM

A bit of an update: I think I've pulled all the blown parts now. Ironically, one of the NSC 2SB600's actually survived. Every other output was blown. The two PNP drivers were blown, but the two NPN ones survived - I think I'm going to swap all four anyway, just to be safe... they're only about $1 a piece.

On one side the bias transistors are OK, but on the other they're blown as well. These are the annoying ones (MPSU95 and MPSU45).

Anything else I should be looking at further back towards the input side of things?

Thanks,
Paul

SQLGuy 19th January 2009 01:36 AM

One more update: parts are ordered! Most everything is coming from Mouser, except for the MPSU's which are from an eBay store.

One thing I noticed that I thought some of you might find interesting... these NSC 2SB600's have a base plate about half as thick as the original NEC 2SB600's. The NEC's also weigh (not surprisingly) about 30% more.

On the other hand, all the NEC's are dead, and one NSC lived to tell the tale... so either they are at least as good or they got lucky :)

Cheers,
Paul

SQLGuy 21st January 2009 04:50 PM

2 Attachment(s)
Well, this is a bit of a shameless bump, but I thought a picture might help get attention.

While you're looking at this, any thoughts on the FRED rectifier suggestion, output transistor sub, or heat sink compound questions from above?

SQLGuy 28th January 2009 04:17 AM

2 Attachment(s)
Got the Variac today and fired up the 200c. Looks like I didn't forget anything too important, as she started up just fine and responded well to adjustments of DC offset.

I also did a bit of quick testing with music and both channels sound OK.

In the process of setting bias now. I'm guessing about 100mA per device based on the published idle of 100W or so. Those heat sinks get decently warm at that bias. Still checking and waiting to see how thermally stable she is.

Two notes:

1. The bias pot might be opposite of what you'd expect - CW decreases bias, CCW increases it

2. That contactor (giant relay to provide on/off for the mains current) has to go - constant low level buzz!


Cheers,
Paul

SQLGuy 31st January 2009 02:32 AM

Pretty much done
 
2 Attachment(s)
Well, I did a bit more testing, and a bit of test listening. The amp drove my test speakers (Magnepan MMG's) very nicely indeed. I was surprised at how much bass it got out of them. During listening tests, at pretty good volume, I was constantly checking the bias and DC offset. I also tweaked the bias a bit when the amp was well warmed up, then checked it again after idling to make sure it was staying stable. It seems to rise a bit when driven hard, and the amp is definitely a bit underbiased when cold, but after about 20 minutes of idle it stabilizes and drifts back and forth within a 5mA per device window.

I was surprised, though, at the variation I see between these MJ21193's/94's that were part of the same manufacturing run. The biggest mismatch is among two of the NPN's in the right channel. With one biased to 120mA, its neighbor will only be pull 95mA or so.

The bench testing included some sine waves from the function generator into 8 Ohm loads to check for any crossover notching on the 'scope. Looked good there.

I also tried an about 100W 20Hz square wave with both channels driven into 4 Ohm loads. Main thing I was looking for there was whether I'd see any ripple on the tops and bottoms of the square wave... it was razor flat. I don't know if this a good test of the power supply caps, but I found it reassuring.

Lastly, I got a good scare, and a good indication of the current capability of the amp, when I first went to check the square wave output in the left channel. I had made load adaptors by wiring together two 8 Ohm 50W resistors with steel hookup wire about as thick as the leads of a 1W resistor; these I then attached directly to Pomona banana's for easy plugging in. Well, when I went to hook up the ground of my scope, I mistakenly attached it to the + out. I have two scopes, one with ground defeated and one that's grounded... this was the grounded one. The output from the amp instantly vaporized the steel wire between the plug and the resistors, but didn't seem to adversely affect the amp at all! Follow-up testing of that channel still looks fine.

Here's a pic of the amp in its next install location, behind the Tympani IV's.

SQLGuy 31st January 2009 05:12 AM

Not ready yet
 
I installed the amp in my main system this evening, and, while it sounds decent, it's very noisy (idle background noise, not distortion). Part of that noise is because it's more sensitive than the Adcom GFA-555ii that it's supposed to be replacing, but a lot of the noise is coming from the amp itself.

Even with inputs disconnected, I am getting about 7mV A/C of noise from the left channel, which is pretty offensive through the Magnepan bass panels. The right channel is a bit better, at about 5mV. The Adcom, by comparison, is dead silent at idle. For now, I've put the Adcom back.

At idle, my meter shows about 50mV of ripple between supply rails and ground.

These are still the original filter caps, though all the other electrolytics are new. What do you guys think? Is this likely just tired filter caps or should I be looking elsewhere first?

SQLGuy 31st January 2009 04:28 PM

2 Attachment(s)
FWIW, here's a close schematic. Differences are that mine has no base resistor on the outputs, and the resistor between the drivers is 39 Ohms instead of 83.2 Ohms.

Paul

SQLGuy 1st February 2009 01:05 AM

2 Attachment(s)
Here's what the output noise looks like. This is at 5mV/div. By the way, the spikes correspond to 60Hz (5ms/div).

Both this and the next shot were taken with the amp somewhat warmed up, running at idle, and with the input shorted.

SQLGuy 1st February 2009 01:07 AM

2 Attachment(s)
And this is the ripple seen on the rails. This is at 50mV/div. All four rails look more or less like this.

I would really appreciate some feedback from the group as to whether this looks unreasonably noisy for a 75V rail.

Thanks,
Paul


All times are GMT. The time now is 12:03 PM.


vBulletin Optimisation provided by vB Optimise (Pro) - vBulletin Mods & Addons Copyright © 2014 DragonByte Technologies Ltd.
Copyright 1999-2014 diyAudio


Content Relevant URLs by vBSEO 3.3.2