Pioneer SA-9800 restoration and upgrade
- Thread starter leesonic
- Start date
larryderouin said:
Hey Larry. I'm the one who should be sorry, I probably over-reacted. It is impossible to convey tone of voice, inflection, or whatever those things are called through just plain text. Please accept my apology, maybe it was the 50v that put me in shock?
Re-reading what I wrote, I wrote it down wrong, I should have put 0mv plus/minus 60mV.
larryderouin said:
I think I might just go ahead and replace all the outputs, seeing as they aren't THAT expensive, certainly not compared to the main supply caps. I looked up the On-Semi MJL3281 and MJL1302 that Brent (SolutionRoom) used for his 9800, they stock them at Mouser, but not the insulating washers it seems. Reading the data sheet, it isn't obvious whether the tabs are insulated from either B C or E, but I bet they aren't.
Lee.
SolutionRoom
...***...
Well... I would test the existing output transistors out of circuit as the next step. Then and only then would I replace them. The On-Semi replacements, while a great option, require pretty extensive modification to fit/connect. IF you do go with the On-Semis, you can use the existing silica pads under the stock transistors, but you'll need to trim them with scissors. Also, the collector is the portion that will ground out on the heat sink if there is no electrical isolation. If the stock outputs test good, I'd go with Larry's course of action.
hogdaddy
Super Member
Thanks! I should have read threads here and saw them posted but anyway this is a great thread and I like this amp.
Did some more work on the 9800 yesterday, sadly I only get a few hours a week where a soldering iron won't be grabbed by tiny fingers...
Here is a picture of my output transistors on the "bad" channel, remember, this was the one putting out 50v of offset. I removed the wire wraps, and also desoldered the wire links that connect the two bases and collectors together.
Using a meter with a diode tester, and going on the "transistor is like two diodes connected back-to-back (or front-to-front)" like they taught us in Physics at school, I tested these four outputs. They all tested fine, which is good, although I did just buy eight new On-Semi outputs from Mouser.
So I figure something on the amp board is bad. I thought about shotgunning it, replacing all the components. I already have all the resistors, electrolytic capacitors, plus the transistors specified in SolutionRoom's thread. What I need is some suitable equivalents to what's left.
Here's what I can up with after extensive searching of the threads :
Transistors
Q2 2SC1775A 512-KSC1845EBU (already have some of these)
Q4 2SA979 unobtainium?
Q6 2SC2291 unobtainium?
Q8 2SA912 512-KSA1220AYS TO-126 (ordered these already)
Q10 2SA726S 512-KSA992FBU (ordered these already)
Q12 2SC1885 512-KSC2690AYS TO-126 (ordered these already)
Q14 2SC1400 KSC1845U
Q16 2SA750 512-KSA992FBU
Q18 2SA750 512-KSA992FBU
Q20 2SC1400 KSC1845U
Q22 2SA904A 512-KSA1013YBU
Q24 2SC1890A 512-KSC1845UBU
Q26 2SA733A 512-KSA1013YBU
Q28 2SC945A 512-KSD1616AGBU
Q30 2SA733A 512-KSA1013YBU
Q32 2SC945A 512-KSD1616AGBU
Q34 2SC1913 512-KSC2073TU
Q36 2SA913 512-KSA940TU_Q
Most, if not all of these came from looking at threads by MarkTheFixer. Man, I hope that guy is OK, he really is helpful with matters like this.
Diodes (no idea about these)
D2,4,8 - 1S1555
D6 - MZ-061
D10 STV2H
D12, 14, 16, 18 - 1S1555
D20, 22, 24, 26, 28, 30 - 1S2471
As an aside, I may have unwillingly stumbled across a replacement for the mica washers on the factory outputs. I ordered Mouser part number 951-SP900S-009-00105 to go with the On-Semi MJL3281 (863-MJL3281AG) and MJL1302 (863-MJL1302AG) outputs. When the box arrived, I found I couldn't use them with the new outputs, but offering them up to the exisiting outputs, it looks like they will fit (see picture below).
The picture shows the amp board removed, and with all the wires taped off, I powered the thing up. Sure enough, after 5 seconds, the protection relay clicked in.
Lee.
Here is a picture of my output transistors on the "bad" channel, remember, this was the one putting out 50v of offset. I removed the wire wraps, and also desoldered the wire links that connect the two bases and collectors together.
Using a meter with a diode tester, and going on the "transistor is like two diodes connected back-to-back (or front-to-front)" like they taught us in Physics at school, I tested these four outputs. They all tested fine, which is good, although I did just buy eight new On-Semi outputs from Mouser.
So I figure something on the amp board is bad. I thought about shotgunning it, replacing all the components. I already have all the resistors, electrolytic capacitors, plus the transistors specified in SolutionRoom's thread. What I need is some suitable equivalents to what's left.
Here's what I can up with after extensive searching of the threads :
Transistors
Q2 2SC1775A 512-KSC1845EBU (already have some of these)
Q4 2SA979 unobtainium?
Q6 2SC2291 unobtainium?
Q8 2SA912 512-KSA1220AYS TO-126 (ordered these already)
Q10 2SA726S 512-KSA992FBU (ordered these already)
Q12 2SC1885 512-KSC2690AYS TO-126 (ordered these already)
Q14 2SC1400 KSC1845U
Q16 2SA750 512-KSA992FBU
Q18 2SA750 512-KSA992FBU
Q20 2SC1400 KSC1845U
Q22 2SA904A 512-KSA1013YBU
Q24 2SC1890A 512-KSC1845UBU
Q26 2SA733A 512-KSA1013YBU
Q28 2SC945A 512-KSD1616AGBU
Q30 2SA733A 512-KSA1013YBU
Q32 2SC945A 512-KSD1616AGBU
Q34 2SC1913 512-KSC2073TU
Q36 2SA913 512-KSA940TU_Q
Most, if not all of these came from looking at threads by MarkTheFixer. Man, I hope that guy is OK, he really is helpful with matters like this.
Diodes (no idea about these)
D2,4,8 - 1S1555
D6 - MZ-061
D10 STV2H
D12, 14, 16, 18 - 1S1555
D20, 22, 24, 26, 28, 30 - 1S2471
As an aside, I may have unwillingly stumbled across a replacement for the mica washers on the factory outputs. I ordered Mouser part number 951-SP900S-009-00105 to go with the On-Semi MJL3281 (863-MJL3281AG) and MJL1302 (863-MJL1302AG) outputs. When the box arrived, I found I couldn't use them with the new outputs, but offering them up to the exisiting outputs, it looks like they will fit (see picture below).
The picture shows the amp board removed, and with all the wires taped off, I powered the thing up. Sure enough, after 5 seconds, the protection relay clicked in.
Lee.
Last edited:
Only thing you did not do was to measure the voltage on the two wires you removed from the bases of the transistors (after all, that kinda completes the t-shooting circle, does it not? 
).
The 2SA979 and 2SC2291 are tough to find. Littlediode carry them, but you'll pay a premium. The rest of your choices look OK, except the drivers, which are hella slow. I prefer the Toshiba 2SA1837 and 2SC4793 in this case (Digikey).
I'll look at the diodes and post back, but except for the zener and the bias diodes (D9,10), looks like most are vanilla 1N4148 signal types.
).The 2SA979 and 2SC2291 are tough to find. Littlediode carry them, but you'll pay a premium. The rest of your choices look OK, except the drivers, which are hella slow. I prefer the Toshiba 2SA1837 and 2SC4793 in this case (Digikey).
I'll look at the diodes and post back, but except for the zener and the bias diodes (D9,10), looks like most are vanilla 1N4148 signal types.
Only thing you did not do was to measure the voltage on the two wires you removed from the bases of the transistors (after all, that kinda completes the t-shooting circle, does it not?
Yeah, I could have done that. That would have shown whether one of the outputs was being turned on by something further upstream, yes?
The drivers are the 2SC1913 and the 2SA913 correct? Just wanted to make sure we're talking about the same thing. I figured I'd try replacing all the transistors except the 2SA979 and 2SC2291, do you think that is wise? If they are bad, perhaps they might take out the new stuff further downstream.
Thank you so much for your help.
Lee.
I'd say it's just about a certainty.
Yes.
It appears that there is a transistor that is shorted, and that has the output stuck at the + rail. You'll likely confirm this by checking the voltage on the removed wires.
The amp is a fair bit more complex than most, with its 'non-switching' topology, so a wholesale swap of transistors makes for a pretty busy day. I don't think I'd personally go that route (I'd be looking for the failed component, and checking the resistor values and looking carefully for bad solder connections), but I'm not the tech on the spot. That'll have to be your call.
SolutionRoom
...***...
So how is the 9800 going Lee? Any time to work on it further? Wondering if you went with a shotgun approach or trying to narrow it down.
I was kinda waiting to see what recommendations anyone would like to make on the diodes before I placed an order for them. I did send Mark The Fixer a link to this thread in my "get well soon" message, so far he hasn't replied.
However, I also have to put the wheels back on my Mini Cooper, just in case we move house. It's looking likely, as it's only been on the market two weeks, we've already had six showings, and two offers. The Mini is just a shell on a wooden frame, no hubs to mount the wheels to, no suspension to mount the hubs to, and no subframes to mount the suspension to...
I might see what I can do tomorrow.
Lee.
However, I also have to put the wheels back on my Mini Cooper, just in case we move house. It's looking likely, as it's only been on the market two weeks, we've already had six showings, and two offers. The Mini is just a shell on a wooden frame, no hubs to mount the wheels to, no suspension to mount the hubs to, and no subframes to mount the suspension to...
I might see what I can do tomorrow.
Lee.
The 1S1555 and the 1S2471 can be replaced with a 1N4148. The MZ-061 is a 6.1V 1/2W zener diode. The STV-2H is rather rare and I know of no real replacement (possibly two UF1004 diodes in series, but because the original is screw mounted to the heatsink, replacement is problematic).
Last edited:
Caminokid
Pioneer and Realistic
Hey Lee, I am having the same issue is getting just one knob for a SA608. Of course when I went and got this one....the selector knob was gone. I dont know why there is thousands of knobs out there for every other Pioneer made. The SA Series is the hardest to get knobs for.
Here are some before and after pics of the problematical amp board. I pretty much shotgunned it, there's only a few of the original components left on, these being the five legged transistors, the ceramic and mylar caps, and the zener diode (couldn't find a 6.1v one).
You will notice that I've upgraded some of the resistors. Places where there were 1/4w resistors up on stand-offs, I replaced these with 2w ones. All the resistors are now 1% metal film, except for the odd few that I couldn't get in the correct values, these are 5%. Not obvious from the pictures, but I left the transistor leads long to space them off the board a bit, hopefully to get some air circulating around them.
Hopefully I'll get to install it sometime over the weekend, or this time next week...
You will notice that I've upgraded some of the resistors. Places where there were 1/4w resistors up on stand-offs, I replaced these with 2w ones. All the resistors are now 1% metal film, except for the odd few that I couldn't get in the correct values, these are 5%. Not obvious from the pictures, but I left the transistor leads long to space them off the board a bit, hopefully to get some air circulating around them.
Hopefully I'll get to install it sometime over the weekend, or this time next week...
Last edited:
SolutionRoom
...***...
Sure looks pretty. Nice work.
IMPORTANT:
if there was a triangular symbol with an exclamation point in it next to the part both in the schematic AND in the parts list,
there are no substitutions,
1/4 W is meant to blow quickly (a disconnect) upon a fault, for safety...
as in FIRE...
the spacers on the 1/4 W are to preserve the board when they pop...
NOT to get their measly 1/4W of heat off the board.
there are also flameproof ( 1/4PMF) resistors... but those are mainly a varnish coating....
sorry....:sigh:
Last edited:
SolutionRoom
...***...
I just said it looks pretty....
it IS A VERY NICE JOB. NO criticism.... i even apologized for bringing it up
But I would be remiss if I hadn't said something immediately.
Last edited:
SolutionRoom
...***...
:ntwrthy:
It's ALIVE!
After a lot of time sitting in the basement while other projects came and went (just search for threads started by leesonic to see what I mean), I finally spent some time getting this amp board back in. Fingers crossed, I powered it up, and was rewarded with a relay click about five seconds after. I followed the adjustments procedure in the manual, and managed to get the DC offset hovering around +- 1.5mV (-11.5mV for the original non-repaired left channel), and the idle current at 71.mV (57.5mV for the left). I have a few observations though.
1) The adjustment procedure in the manual says to adjust the DC balance first using VR2 and measuring across the terminal 11 and ground. Then adjusting idle current across terminals 18 and 20 by first adjusting VR4 and then VR6. Is there a better was of doing it? By this, I mean is there some way of measuring what VR4 is adjusting, then VR6? I gave it time to warm up between adjustments, but it just seems odd to adjust two trimmers in different parts of the circuit by measuring at the same place.
2) For some reason, the right channel heatsink seemed to be warmer than the left. I realized that the left channel isn't adjusted right, but it seemed like the heatsink seemed too warm. The amp was under no-load, exactly as specified in the manual.
3) Still remaining with the heat issue, are there amps meant to run hot or something? The power supply board has two transistors on heatsinks which I replaced (see page 1), but the replacements seem to run hot. There are also four resistors in close proximity (R5, R21, R22, R23) that run hot as well. Looks like R23 and R22 are just used for the pilot lamp, if it's going to cut down on the heat build-up, I'd sooner go with an LED pilot lamp.
Lee.
After a lot of time sitting in the basement while other projects came and went (just search for threads started by leesonic to see what I mean), I finally spent some time getting this amp board back in. Fingers crossed, I powered it up, and was rewarded with a relay click about five seconds after. I followed the adjustments procedure in the manual, and managed to get the DC offset hovering around +- 1.5mV (-11.5mV for the original non-repaired left channel), and the idle current at 71.mV (57.5mV for the left). I have a few observations though.
1) The adjustment procedure in the manual says to adjust the DC balance first using VR2 and measuring across the terminal 11 and ground. Then adjusting idle current across terminals 18 and 20 by first adjusting VR4 and then VR6. Is there a better was of doing it? By this, I mean is there some way of measuring what VR4 is adjusting, then VR6? I gave it time to warm up between adjustments, but it just seems odd to adjust two trimmers in different parts of the circuit by measuring at the same place.
2) For some reason, the right channel heatsink seemed to be warmer than the left. I realized that the left channel isn't adjusted right, but it seemed like the heatsink seemed too warm. The amp was under no-load, exactly as specified in the manual.
3) Still remaining with the heat issue, are there amps meant to run hot or something? The power supply board has two transistors on heatsinks which I replaced (see page 1), but the replacements seem to run hot. There are also four resistors in close proximity (R5, R21, R22, R23) that run hot as well. Looks like R23 and R22 are just used for the pilot lamp, if it's going to cut down on the heat build-up, I'd sooner go with an LED pilot lamp.
Lee.
for #1, yes it seems odd, but is necessary. to delve deeper I would have to go into exquisite detail about the NSA circuits, but what really counts is the results, at the output transistors...
in the 3900/d7000, having the heat sinks at 90 degrees to their normal orientation causes them to run a lot hotter. But they normally do run warmer. 70mA is a lot of idle current for 100 (+50, -50) volts, 7 watts.
3. yes, the thermal bias diodes run at 7 mA, with supply voltages in the 50v range we are talking about some heat.
pilot lamps run at 50ma and are calculated as such. when you only use 8v of it at 50ma all those extra watts have to go somewhere. the led's aren't much better when run at their rated current of 10 to 20mA, that's why I use super high brightness leds and run them at 2mA. much less wasted wattage.
and a linear supply dissipates the difference in wattage between what the voltage the raw dc power supply is generating and the regulated voltage the circuit needs (times the current drawn). Pioneer deliberately added resistors like R1 220 ohms 5 watts, r2 150 ohms 5 watts (and others in different models) to stick some of the waste heat into the resistor, rather than dumping it all into the power supply transistor.
in the 3900/d7000, having the heat sinks at 90 degrees to their normal orientation causes them to run a lot hotter. But they normally do run warmer. 70mA is a lot of idle current for 100 (+50, -50) volts, 7 watts.
3. yes, the thermal bias diodes run at 7 mA, with supply voltages in the 50v range we are talking about some heat.
pilot lamps run at 50ma and are calculated as such. when you only use 8v of it at 50ma all those extra watts have to go somewhere. the led's aren't much better when run at their rated current of 10 to 20mA, that's why I use super high brightness leds and run them at 2mA. much less wasted wattage.
and a linear supply dissipates the difference in wattage between what the voltage the raw dc power supply is generating and the regulated voltage the circuit needs (times the current drawn). Pioneer deliberately added resistors like R1 220 ohms 5 watts, r2 150 ohms 5 watts (and others in different models) to stick some of the waste heat into the resistor, rather than dumping it all into the power supply transistor.
