• The move to the new server is done. There are some software and database maintenance updates in process. This has us passing the hat around to help out. We appreciate any donations. Seriously, even a dollar helps. The payment page may be found here - https://www.audiokarma.org/support.html

Sa-9100, DC offset question, Protection light

M

mpgasque

New Member
I can't belive what a great site this is for Pioneer eqipment! I have read every thread I could find on the SA-9100 amp and downloaded the SM. Fantastic!
I also have a PL-530, a SX535, and a TX-6500.

Here's where I got my SA-9100. In 1973 my Brother and Mother were going on a trip to visit Japan. I told my brother to be on the lookout for either the SA-8100 or the SA-9100 at shops in Tokyo. I even gave him the Ad brochure on them (still have them) so he could identify them. Inbetween tours he found some and the cheapest was in the gift shop of his hotel. Well he bought it, and although I intended for him to ship it back, he carried it on the plane with him! This thing weighs a TON, and he's dragging it half way around the world for me. What a guy, always owed him big time for that.

The SA-9100 has always been the heart of my system which includes a 8 channel home recording studio. Now, after 35 years the protection light will pop on in the middle of a song for a few seconds, then go off. It's happening more and more so I know it's time for some service. I've worked with digital electronics but not so much with analog. Thanks to everyone involved in all the threads on this topic. I've read them all and am just about ready to get started. After 35 years of great service, I'm going to the mat for this amp.

My inital plan is to measure the power supply voltages and the DC offset and bias current. Here's my questions:

1) The SM section 9 says to remove the output load, I assume that means disconnect the speakers. Yes?

2) Then, terminate the input terminals (#3 on AWH-023-B) with a 4.7k resistor. Does this mean to connect one end of the 4.7k to #3 and the other to ground?

3) Is the VR3 callout in section 9.1 a typo and should read VR2?

After, and depending on these results, I will replace the electrolytic caps on the power supply board (AWR-032-0), R12 & R24, and whtever transistors the voltage measurements dictate might be weak or bad.

I will also replace C3 through C6 and Q7 on the Protector circuit board, AWM-025-A. Any thoughts?

I won't be so long winded in the future, Thanks to everyone.:banana:
 
Register to hide this ad
1) yes

2) don't bother... Just go to zero volume with no inputs.

3) yes

Don't twist the pots, just check the offset and bias first. If you are feeling bold, first check the power supply voltages. AC voltage across a cap, or present on a dc power supply voltage rail would warrant further investigation.

Power supply caps are less likely to be a problem than the protection caps and relay driver transistor. The smaller (physically) ones dry out faster, of course heat accellerates the process on any cap. WATCH your polarities IF you do the p.s. caps....

Elsewhere I reccomended the addition of a protection diode across the relay coil, looks like you need one here. A 1n4004 has 1 amp and 400 volt ratings, and IIAC the cathode (band) goes to pin 10 and the anode to pin 9 on the protection board.

My bets are on the protection circuit driver transistor, from your description.
 
Welcome aboard. Nice gear, and a cool story. You're in good hands with Mark and many others who will probaly chime in, she'll be singing again soon. :thmbsp: I always like reading these as I learn along the way too.
 
Thanks for the response Markthefixer. I've read tons of your replies to other SA-9100 owners; always great advice. I also thought that just getting a reading on the DC offset and bias without touching the pots was the way to go at first. You must be a mind reader, because I saw a post talking about adding a diode to protect that transistor (Q7) and I thought it would be a good thing to do. Thanks so much for the quick engineering job. I'll take my readings and be back in a week or so. I love to read these threads also because I always learn so much. A favorite of mine: "Questions? Don't Do.... Ask!".
 
I took voltage readings from the pins on the power supply circuit board (AWR-032) and everything looked real good. The numbers in parens are my actual readings and the others are from the schematic.

Pin 2 -16 (-17.7)
Pin 3 -24 (-27.2)
Pin 4 -13 (-13.6)
Pin 5 -42 (-43.1)

Pin 6 -56 (-59.8)
Pin 7 +56 (+59.0)

Pin 8 +42 (+43.7)
Pin 9 +35 (+36.1)
Pin 10 +24 (+25.3)
Pin 11 +16 (+16.3)

No Problems here.

The DC offset was also pretty much on the money as the left channel was initially +11.0 mv and the right was +1.0mv. I couldn't help but tweak the left down to +3.0mv using VR1. The bias was also good at 23.0mv for the left and 23.6mv for the right channel. I overcame my desire to tweak these and left them as they were

So it seems that my intermittent protection problem is most likely on the protection board itself (AWM-025. I have ordered a replacement for Q7 (2SC1384) which was (among possible others) a NTE293 (Mouser part number 526-NTE293). Also I ordered the diode to protect Q7 (Mouser 833-1N4004-TP), and capacitors C3~C6. I also picked up all the caps for the Power supply board (AWR-032). I agree that replacing these is a very good idea for future reliability. Anyone familar with these transistors, please comment as to whether the NTE293 is a good substitute for the 2SC1384. Thanks.

The protection light actually kicked in during my testing, but only for a moment, so the problem is still with me. We'll see what happens after the new parts are installed. Thanks for everyones help.
 
#1 Rule: IF it ain't broke, don't fix it....:rolleyes:

re: swap c1451 (and it's COMPLEMENT??? where??????????:scratch2:) that's BIAS and it looks FINE...:D

Fooling with a working bias area is an invitation for a smoking ruin. I have SEEN it happen HERE BEFORE!!!!!:yikes: :tears:

Pioneer went to a lot of trouble to get the transistors and the circuits around them to play well together. The changes in the protection circuit which is a SWITCHING circuit will NOT affect audio quality, something that changing amplifier transistors cannot say. When parts can be proved bad, then they are reluctantly replaced. Operating regions for drivers and outputs are far less critical than the smaller transistors in the amp circuit, also the same goes for power supply circuits, less critical.

Those graphs in the back of the spec sheet for the transistors are IMPORTANT, and reflect MANY factors. Choose the wrong region of a transfer curve and you could pay a heavy price in an audio section, the penalty can be far less severe in a power supply circuit...

The NTE is ok, it was charlie's only realistic alternative. Generally NTE is the LAST resort.

The inductive kick from the relay's collapsing magnetic field generates a negative voltage that gets into the transistor and nibbles away at the junction, so after a while the transistor doesn't have enough "meat" to handle the load (heat usually) so it drops out, cools off, and starts over, heating up again.
The diode conducts this negative voltage, clamping it to about -0.6v at the expense of a few milliseconds extra release time.

Since Pioneer uses the exact same circuit with the same parts values in other equipment that HAS the diode, I feel that this is a prudent, tried & true, safe upgrade.
 
Well. the parts came in and I replaced C3, C4, C5, & C6 on the protection circuit board (AWM-025) along with Q7 and installed the diode between pin 9 & 10. Was not a very difficult job. However, it did not fix my problem. I'm still getting the protection circuit kicking in every ten or fifteen minutes or so. Since I had the parts, I went ahead and recapped the power supply board and tested the old caps as they were replaced on the board (AWR-032) with my old cap tester that reads up to 2000uf. The old caps were in great shape, each one 3% or so ABOVE thier rated uf. The new ones were all 5% BELOW thier rated value.

I worked as an engineer for a major capacitor manufacturer so this low cap stuff did not surprise me. They save money by always winding the caps to the low side of the rated value because it uses less foil. Then they say they can't control the manufacturing to better than +/- 20%. Pure lies. Almost any electrolytic cap you buy today will be 8% to 5% below rated cap. But back to the problem.

I didn't expect the recap to solve my problem since all the caps looked good right off the board. And it didn't. However Pin 6 now regulates at -63.5 down from -59.8, and Pin 7 increased to +62.6 from +59.0. All the other Pins moved about 3/10 of a volt away from the schematic and what I had measured before. No big problem, just not in the right direction.

I think I have an intermittent transistor somewhere that is begining to fail. I have the parts to replace Q2 & Q4 on the negitive side of the power supply board (AWR-032), and also Q1 & Q2 on the positive side. However, as was shown with recapping the power supply, you have to be carefull with replacing stuff or you might end up even worse off. I think the problem would be MUCH easier to find if it would just fail completly. Any Thoughts? Thanks.
 
This is the type of problem that drives people nuts... cuz it's very intermittent, and it won't STAY broke.

The power supply voltages are regulating, and the slightly higher RAW dc isn't a problem. You don't notice the lights flickering, do you? The ac for the lights is monitored by the protection circuit as the power on/OFF signal.

The big question is whether or not the actual DC offset of the amps is spiking, the dc offset sense is misbehaving, the bias current drive is spiking, or the current overload sensing circuit is misbehaving.

In other words, do we blame the amps or the protection circuit?

I guess that if there is no audio going through the amp, it's a lot harder to tell when it cuts out for a few seconds, so operating it at zero volume or without the speakers connected is NOT an option.

There are three diodes (D5, D6, D7) on the protection board making up a triple OR circuit, two of them (D5, D6) deal with the DC offset sensing differential amplifier circuit, the third (D7) goes to the two overcurrent sensing transistors. By selectively disconnecting the diodes, we can narrow down the source. The offending circuit WHEN MISBEHAVING will develop a much lower voltage, causing current to flow in that diode, pulling down the 20 or so volts at the base of Q5, triggering the protection and opening the relay.


Here is what I would do:
1) hook up CHEAP, replaceable speakers (just in case)
2) open up ALL THREE diodes, and see if it still does it.

then if it does, then we have narrowed the focus considerably.

otherwise, then I would connect the two offset overvoltage diodes D5 & D6, and again monitor for failure.

If it doesn't, then I would disconnect D5 & D6 again, and connect the over-current diode D7, and monitor for failure.

More steps to follow.... after the problem's source has been narrowed.

OR

a can of freeze spray and a hairdryer can possibly stress the components to failure. sort of a shotgun approach.
With your background, I don't have to warn you about TOO hot, or too much condensation...
 
I have been hesitant to recommend surgery of that type to those that I have no idea of their abilities. I have learned caution the hard way here. That's how my avatar was awarded to me by EchoWars. The most harrowing experiences happened when folks tried to jump ahead of instructions into the unknown. One bad experience was in the bias circuit area, and you can imagine the results.

I have been researching the Fuji 2sc1451's you are suspicious of spiking things, what have you successfully used for a replacement?
(the old recommended replacements and their available replacements are all over the map)

original Fuji 2SC1451 ( npn 150v 50ma 700mw 130mhz )
I have used:
fairchild ksc1845 npn 120v 50ma 500mw 110mhz lo noise (pn # 512-KSC1845EBU mouser.com (usa) $0.05 each )

But it's worth modifying the instructions,


THANK YOU, inductor
 
inductor has raised a good point, causing me to re-think the procedure, and this makes a cleaner separation between the amplifier and the protection circuits,

I am hesitant to undo wire wraps, which never go back right, need to be resoldered and the wire almost always breaks extremely short. Resoldering the wrap always ends up looking like a hack job, plus the soldering can open up the pin to pc board connection, making a mess down below.

That said there are three wires per channel that go to the protection circuit, and eventually we will have to open them up to figure out IF the protect circuit is found to be stable, and if it is, which channel is pushing things into protect.

To open the wires up, since you HAVE had the protection circuit board up to do the caps, would be to MARK the wire going to each of these pins on the protection board with the PIN number, and then unsolder and remove the pin from the board ( I use needle nose pliers ) and INSULATE it.


markthefixer said:
<snip>

Here is what I would do:
1) hook up CHEAP, replaceable speakers (just in case)
2) open up six pins, the left channel is pins 1, 5 & 6 and the right channel is pins 3, 7 & 8 , and see if it still does it.

then if it does, then we have narrowed the focus considerably.

otherwise, then I would re-connect the Three left channel pins 1, 5 & 6 and again monitor for failure.

If it doesn't fail, then I would disconnect Three left channel pins 1, 5 & 6 again, and connect the Three right channel pins 3, 7 & 8, and monitor for failure.

More steps to follow.... after the problem's source has been narrowed.

<snip>
Click to expand...


If we do work on the amplifier board, then the fragile (and almost unobtainable) bias diode connections are in extreme jepoardy, as they go from the board to the heatsink, along with the heavier wires to the output transistors.
 
First off, thank you for your time. I know that an intermittent like this is very time consuming and can be frustrating. I do think I follow Inductors logic, but I hate to admit that disconnecting the power supply from the power amps would require a lot of study on my part before I would try that. Be assured, however, that in time I will try EVERYTHING, and I thank you for your suggestions. By the way, it has done this while I had the dmm on various components, but I could'nt tell what started it, all the readings just went heywire, and then right away the readings went back to normal. the protection pops up and then goes right away, like you just turned it on.

My first test will be to put on other speakers to see the problem happen with them. I'm sure it will, and when it does, I will focus on D7 & R9 and R10 in the manner you describe. I will take detailed notes and report back. I've got a big hammer waiting for this bug, and when I find it... Thanks again.
 
mpgasque said:
First off, thank you for your time. I know that an intermittent like this is very time consuming and can be frustrating. I do think I follow Inductors logic, but I hate to admit that disconnecting the power supply from the power amps would require a lot of study on my part before I would try that. Be assured, however, that in time I will try EVERYTHING, and I thank you for your suggestions. By the way, it has done this while I had the dmm on various components, but I could'nt tell what started it, all the readings just went heywire, and then right away the readings went back to normal. the protection pops up and then goes right away, like you just turned it on.

My first test will be to put on other speakers to see the problem happen with them. I'm sure it will, and when it does, I will focus on D7 & R9 and R10 in the manner you describe. I will take detailed notes and report back. I've got a big hammer waiting for this bug, and when I find it... Thanks again.
Click to expand...

This post has been in flux and is now closer to what you need to do. I am almost done researching the new procedure I recommend, NOT opening up components, just disconnecting 6 wires on the protection board.
I had hoped to finish BEFORE you got online....

edit: it IS safe, as long as the wires are insulated from EVERYTHING.... and easier and quicker than some things I thought of.

Don't sweat my stick-to-it-iveness or patience this recent thread on an sa-9800 took 336 posts from april 21 through june 25 to resolve successfully (that is a personal record) and I have NO regrets, in fact I am both satisfied and happy to get that sa-9800 back singing.
 
Speaker Test
After hooking up another pair of good speakers and wires, the problem continued, so the Amp is not going into protection mode because of a speaker issue ( Dynaco A-25, old, but still sound good).
 
mpgasque said:
Speaker Test
After hooking up another pair of good speakers and wires, the problem continued, so the Amp is not going into protection mode because of a speaker issue ( Dynaco A-25, old, but still sound good).
Click to expand...

GOOD point... Don't overlook the obvious....
 
Well, I can't tell you HOW it works, but the service manual says this:

"The protection circuit comprises 7 transistors. Of these, Q1 and Q2 server to detect excessive current in the power amp and short-circuiting of the output load (speaker leads etc.). Each transistor servers one channel.
Q3 and Q4, constituting a differential amp, detect DC potential at the power amp output; this circuit is common for both channels. Transistors Q5 and Q7 are the relay driving circuit."

I thought that these old Dynaco speakers (30 years old) might have some problem with the voice coil shorting out or something. Speaker shorting is what the protection circuit should protect the power amp from. Other newer speakers had the exact same problem so I think the Dynacos are OK.
 
Well, I have seen (on AK and not too long ago) a problem where a tangle of wires to the speakers was actually shorting intermittently, and the sound was was in all likelyhood providing the miniscule movement necessary to make intermittent contact...

But since it did it when you were reading the offset voltages (no sound, no speakers??) it seemed to be internal to me.

Ok, due to the intermittent nature of the problem and inductor's (you are a shop, right?) experience jumping ahead doesn't seem too irresponsible in this case.

The transistors inductor refers to are q6 and q7 on the power amplifier boards (two channels, one on each side of the chassis ) and are the round metal ones. They appear to be unreliable over time. In this amp I have to respectfully disagree with inductor's choice of replacements, a big TO-126 10 watt 140 mA transistor, as the 2sc1451 is being used in the bias circuit (I calculate Ic @ 6.7 milliamps, with Vce=40 we get 0.25w ) as opposed to a pre-driver or driver application (hundreds of milliamps) where I would heartily recommend his choice of the 2sc2911. (I CAN be pursuaded otherwise, the power dissapation is a bit high for a .5 watt package, but the low noise and higher gain tip the balance, plus the SiZe considerations)


Instead get:

fairchild ksc1845 npn ecb 120v 50ma 500mw 110mhz hfe400-800 lo noise pn # 512-KSC1845EBU mouser.com $0.05 each
(the 2sc1451 has an unusual ebc layout, so lead rearrangement is necessary)
(but if we change our minds, bdent.com has the 2sc2911's at $0.41 each with a $15 minimum order
or the fairchild ksc2682 npn ecb 180v 100mA 8W 200mhz hfe160-320 pn # 512-KSC2682YSTU mouser.com @ $0.31 each)
or get both... they're cheap....


when ordering from mouser.com, specify First class mail shipping and you can get away with spending less than 4 bucks... you will need at least four, two for each channel.

The heat sink has the amplifier board mounted to it, and after the heat sink is freed, the assembly should have enough play to free the amplifier board and hinge things open for access. BE VERY CAREFUL OF THE LITTLE HORSESHOE SHAPED THING (bias diode) WITH THE **FRAGILE** LEADS on the heatsink, THEY BREAK... and are almost irreplacable. I unscrew it from the heatsink, and make one careful bend, tucking it away in a safe area, then start working on things.

The bias diode and these transistors are part of the bias circuit, which is the most dangerous (to the amp) area to work in. If something goes wrong, the outputs are converted to a smoking ruin because the outputs are BOTH turned on HARD, conducting massive, destructive amounts of current.

When the work is done, the bias will have to be re-set too.

When you do this, we will discuss procedures and lead arrangements while the parts are in transit, instead of me trying to cram everything in one post.
 
Well, since they have worked in a 9100, better safe than sorry.

Go with the ksc2682's.
 
You must log in or register to reply here.
Back
Top Bottom