I got Retro-Stereo's Spec-2 here for some badly needed attention. I was going to email him with what was done, but WTH, I'll post here.
Thing came in looking like the junkyard dog...dirty-looking but seemingly functional. Opening it up all I saw was crud, so off to the garage to blow out the biggest chunks. Then, the driver boards and heatsinks were pulled as an assembly (thank you Pioneer!), and the rest got a good cleaning and scrubbing.
The heatsinks were an odd color (faded gold-ish), so it seems that the anodized surface has been nearly stripped. A place in town does the plating for me, but unfortunately has a minimum charge or $50. Greg said to go for it, so the heatsinks were stripped of parts, and taken to the platers.
In the meantime, an examination of the boards revealed that this poor thing has obviously been in a shop more than it was in service. A ton of incorrect transistors and caps reside in the unit (some electrolytic caps in backwards!!), and way more than the fair-share of lifted traces. Also a good fist-full of burnt resistors. I told Greg this wins the award for the most farked-up unit that still runs. :yes: One positive note: still had all the original output devices (2SB600/2SD555, same as SX-1250)
One thing that concerned me was one of the biasing diodes (also called 'varistors' or 'thermal diodes' in some manuals) had a wire busted and somehow the previous tech managed to get the wire reconnected with a glob of solder. Not very confidence inspiring. You see these small 'horseshoe-shaped' devices on a lot of vintage gear mounted on the heatsinks with a single screw, and thin, fragile wires leading back to the driver board. Too much monkeying around with these and the wires break off flush with the diode itself leaving no way to repair it, and these are getting very very hard to find. Not only that, but on many amps a busted lead on one of these will instantly smoke the output devices and probably a lot more. The Spec-2 has a STV-4H diode in parallel with a STV-3H diode and a 470-ohm pot, so the loss of one diode might not blow the amp, but I see no reason to roll the dice. So here's a tip for you budding repair dudes: if you find you must manhandle a driver board/heatsink assembly with a thermal diode on the heatsink, loosten the screw and carefully move the diode someplace safe so you won't be putting any stress on it. Your output devices will thank you.
Anyway, I got lucky, and a guy on eBay had a couple of STV-4H's pulled from a Sansui something-or-other. Bought them, and sat down to order the rest of the parts I needed, mostly from Digikey. Also, we're replacing all the original 15,000µf 100V caps with 22,000µf 100V Nippon Chemi-Con caps
so I got them on the way.
One thing that I thought odd is the use of a crappy 4.7µf tantalum capacitor at the input of the amp. Even the Spec-4 uses a film cap here. I wanted to use a film cap, but once I got them in I could see that I was never going to fit a 4.7µf metal polypropylene cap in that space...the input connector is right there and they interfere with each other. Sooo...second best will have to do: stuck in a 4.7µf BlackGate 'N' cap.
I redressed the lifted traces the best I could, replaced all the electrolytic caps, and methodically went through and found all the burnt resistors. Another tip on old Pioneer's: don't neglect to check those resistors covered with the fabric weave...one tends to replace the ones that are obviously burnt, but the covered ones usually require desoldering to see if they have overheated. If they look burnt in any way, replace them even if they measure OK.
Lastly, all the oddball transistors get replaced.The Spec-2 is a little weird in that is uses a cascoded input differential feeding a voltage amp/biasing stage which is also a differential, but consisting of large TO-220 transistors. Pioneer feeds these transistors about 15mA of current, and with the huge supply voltage (93V with the switch in the 8-ohm position), these transistors end up dissipating a fair amout of heat, so they used transistors that are usually reserved for driver stages. These had some off-brand Chinese transistors stuck in there, so I put in some very fast Panasonic 2SA1535A's and 2SC3944A's (200MHz parts). The drivers were also some oddball mix, so I put in On-Semi MJE15032/33's (60MHz). All these transistors are much faster than the original Pioneer parts. As a matter of fact, the original Pioneer drivers and VAS transistors are 2SD610A's and 2SB630A's, which, although are very rugged devices, are only about 5MHz parts and in this case considerably slower than even the TO-3 output devices. Care must be taken when swapping out to much faster devices like this as you can introduce instability and the amp could even oscillate...the output needs to be checked with a 20KHz square wave when done to see if there is anything that needs attention.
A problem with a unit that has been operated on as many times as this one is fragile wires. Pioneer (and about everybody else in this era) used solid-core wires wirewrapped to posts, and these will only take so much flexing before they break. I fought this the whole time I had this amp apart, and had to re-do a good dozen broken connections (I usually remove the original wirewrap, solder and heatshrink a short extention to the busted wire, and wrap 2 or 3 times around the post and solder it). Sucks, but what can you do?
The 470 ohm cheapo bias pot was replaced with a Bourns 25-turn trimpot, and the current limit pots were replaced with sealed single-turn pots, as were the meter adjust pots (you can see them in the pics).
The original Pioneer output posts were trashed, so I ordered some 5-way gold posts from PE and popped 'em in. Easy on the Spec-2.
Picked up the heatsinks from the platers, and put the drivers and outputs back on 'em. Installed the new 22,000µf caps (a lot easier typed than done), and then put in the driver/heatsink assemblies in. Fired it up the first time with the 100W lamp in series, and made sure that bias was at a minimum. Here's where I discovered that one of the driver boards was not providing any idling current for the outputs. Crap...I must have missed something.
Seems that the biasing is low to the tune of about a diode drop. Methinks a shorted base-emitter junction, so if that's right, it shouldn't pass a signal. I feed it a sine wave, and it looks fine at the output. Very wierd. Long story short, the last guy to fix it replaced one of the STV-4H thermal diodes with an STV-3H diode. There's where my missing diode drop went. Hard to see WTF, since they look identical, and they aren't marked in a fashion that tells you what they are. Couple more hours down the drain... Luckily I bought two from the eBay guy, but I was hoping to have a spare since they are so rare.
OK, so I get it back together again, and now it biases fine. I set the bias, check the offset (-22mV left, -46mV right, and non-adjustable on the Spec-2
), and connect the dummy load set for 4 ohms to adjust the current limit. I get it set for both channels, reconfigure for 8 ohms, and connect the function gen for a square wave to check stability. Bottom line, 20KHz wave look really really good. :thmbsp: No problems, no instability.
So I calibrate the meters at 250W, and check distortion. Measures 0.0077% at 250W, which is getting close to my measurement limit. Then I push the amp to clipping, and I am really suprised at what I read: 340WPC into 8 ohms!! Not too shabby... :thmbsp: No, I can't do 4 ohms with this much power without investing in another $150-worth of dummy loads. Let's just say that if it was around 400W I wouldn't be suprised.
The last thing to do on the amp I'll be working on next week...Retro wants to add an offset servo to each channel since the Spec-2 has no adjustment. I blew it when I ordered the parts and forgot some stuff I needed to do this, so I'll have to place another order and get the rest of what I need on the way. I'll take some more pics of the servo install and post 'em in this thread.
Thing came in looking like the junkyard dog...dirty-looking but seemingly functional. Opening it up all I saw was crud, so off to the garage to blow out the biggest chunks. Then, the driver boards and heatsinks were pulled as an assembly (thank you Pioneer!), and the rest got a good cleaning and scrubbing.
The heatsinks were an odd color (faded gold-ish), so it seems that the anodized surface has been nearly stripped. A place in town does the plating for me, but unfortunately has a minimum charge or $50. Greg said to go for it, so the heatsinks were stripped of parts, and taken to the platers.
In the meantime, an examination of the boards revealed that this poor thing has obviously been in a shop more than it was in service. A ton of incorrect transistors and caps reside in the unit (some electrolytic caps in backwards!!), and way more than the fair-share of lifted traces. Also a good fist-full of burnt resistors. I told Greg this wins the award for the most farked-up unit that still runs. :yes: One positive note: still had all the original output devices (2SB600/2SD555, same as SX-1250)
One thing that concerned me was one of the biasing diodes (also called 'varistors' or 'thermal diodes' in some manuals) had a wire busted and somehow the previous tech managed to get the wire reconnected with a glob of solder. Not very confidence inspiring. You see these small 'horseshoe-shaped' devices on a lot of vintage gear mounted on the heatsinks with a single screw, and thin, fragile wires leading back to the driver board. Too much monkeying around with these and the wires break off flush with the diode itself leaving no way to repair it, and these are getting very very hard to find. Not only that, but on many amps a busted lead on one of these will instantly smoke the output devices and probably a lot more. The Spec-2 has a STV-4H diode in parallel with a STV-3H diode and a 470-ohm pot, so the loss of one diode might not blow the amp, but I see no reason to roll the dice. So here's a tip for you budding repair dudes: if you find you must manhandle a driver board/heatsink assembly with a thermal diode on the heatsink, loosten the screw and carefully move the diode someplace safe so you won't be putting any stress on it. Your output devices will thank you.
Anyway, I got lucky, and a guy on eBay had a couple of STV-4H's pulled from a Sansui something-or-other. Bought them, and sat down to order the rest of the parts I needed, mostly from Digikey. Also, we're replacing all the original 15,000µf 100V caps with 22,000µf 100V Nippon Chemi-Con caps
so I got them on the way.One thing that I thought odd is the use of a crappy 4.7µf tantalum capacitor at the input of the amp. Even the Spec-4 uses a film cap here. I wanted to use a film cap, but once I got them in I could see that I was never going to fit a 4.7µf metal polypropylene cap in that space...the input connector is right there and they interfere with each other. Sooo...second best will have to do: stuck in a 4.7µf BlackGate 'N' cap.
I redressed the lifted traces the best I could, replaced all the electrolytic caps, and methodically went through and found all the burnt resistors. Another tip on old Pioneer's: don't neglect to check those resistors covered with the fabric weave...one tends to replace the ones that are obviously burnt, but the covered ones usually require desoldering to see if they have overheated. If they look burnt in any way, replace them even if they measure OK.
Lastly, all the oddball transistors get replaced.The Spec-2 is a little weird in that is uses a cascoded input differential feeding a voltage amp/biasing stage which is also a differential, but consisting of large TO-220 transistors. Pioneer feeds these transistors about 15mA of current, and with the huge supply voltage (93V with the switch in the 8-ohm position), these transistors end up dissipating a fair amout of heat, so they used transistors that are usually reserved for driver stages. These had some off-brand Chinese transistors stuck in there, so I put in some very fast Panasonic 2SA1535A's and 2SC3944A's (200MHz parts). The drivers were also some oddball mix, so I put in On-Semi MJE15032/33's (60MHz). All these transistors are much faster than the original Pioneer parts. As a matter of fact, the original Pioneer drivers and VAS transistors are 2SD610A's and 2SB630A's, which, although are very rugged devices, are only about 5MHz parts and in this case considerably slower than even the TO-3 output devices. Care must be taken when swapping out to much faster devices like this as you can introduce instability and the amp could even oscillate...the output needs to be checked with a 20KHz square wave when done to see if there is anything that needs attention.
A problem with a unit that has been operated on as many times as this one is fragile wires. Pioneer (and about everybody else in this era) used solid-core wires wirewrapped to posts, and these will only take so much flexing before they break. I fought this the whole time I had this amp apart, and had to re-do a good dozen broken connections (I usually remove the original wirewrap, solder and heatshrink a short extention to the busted wire, and wrap 2 or 3 times around the post and solder it). Sucks, but what can you do?
The 470 ohm cheapo bias pot was replaced with a Bourns 25-turn trimpot, and the current limit pots were replaced with sealed single-turn pots, as were the meter adjust pots (you can see them in the pics).
The original Pioneer output posts were trashed, so I ordered some 5-way gold posts from PE and popped 'em in. Easy on the Spec-2.
Picked up the heatsinks from the platers, and put the drivers and outputs back on 'em. Installed the new 22,000µf caps (a lot easier typed than done), and then put in the driver/heatsink assemblies in. Fired it up the first time with the 100W lamp in series, and made sure that bias was at a minimum. Here's where I discovered that one of the driver boards was not providing any idling current for the outputs. Crap...I must have missed something.
Seems that the biasing is low to the tune of about a diode drop. Methinks a shorted base-emitter junction, so if that's right, it shouldn't pass a signal. I feed it a sine wave, and it looks fine at the output. Very wierd. Long story short, the last guy to fix it replaced one of the STV-4H thermal diodes with an STV-3H diode. There's where my missing diode drop went. Hard to see WTF, since they look identical, and they aren't marked in a fashion that tells you what they are. Couple more hours down the drain... Luckily I bought two from the eBay guy, but I was hoping to have a spare since they are so rare.
OK, so I get it back together again, and now it biases fine. I set the bias, check the offset (-22mV left, -46mV right, and non-adjustable on the Spec-2
), and connect the dummy load set for 4 ohms to adjust the current limit. I get it set for both channels, reconfigure for 8 ohms, and connect the function gen for a square wave to check stability. Bottom line, 20KHz wave look really really good. :thmbsp: No problems, no instability.So I calibrate the meters at 250W, and check distortion. Measures 0.0077% at 250W, which is getting close to my measurement limit. Then I push the amp to clipping, and I am really suprised at what I read: 340WPC into 8 ohms!! Not too shabby... :thmbsp: No, I can't do 4 ohms with this much power without investing in another $150-worth of dummy loads. Let's just say that if it was around 400W I wouldn't be suprised.
The last thing to do on the amp I'll be working on next week...Retro wants to add an offset servo to each channel since the Spec-2 has no adjustment. I blew it when I ordered the parts and forgot some stuff I needed to do this, so I'll have to place another order and get the rest of what I need on the way. I'll take some more pics of the servo install and post 'em in this thread.
Last edited:
. I had Mike here in town revamp my SPEC-2 last year. I wished I had taken the bag of parts he replaced. It was alot of components I can tell you. Mike siad the front-end needed alot of attention requiring the replacement of some hard to find very small transitors among other things. He explained to me about what needed to be replaced and checked in a SPEC-2. If I was not so paranoid about shipping damage and would have had the time I would have contacted you to work on my amp. It was just a matter of convienence that Mike is here close by and did a fantastic job curing my 2's heating and going into protection problems. 
