• WANTED: Happy members who like to discuss audio and other topics related to our interest. Desire to learn and share knowledge of science required. There are many reviews of audio hardware and expert members to help answer your questions. Click here to have your audio equipment measured for free!

Sharp/Optonica SM4646 - Protection Puzzle - Guidance needed

Old_School_Brad

Major Contributor
Joined
Sep 10, 2024
Messages
1,231
Likes
1,479
1728742152149.png

1728742190907.png

1728742248944.png


I’ve had this old, beautiful, heavy integrated amplifier sitting on the shelf for years. Recently, I had some spare time and decided to give it a try.

The problem: After powering it on, it goes into protection mode after about 10 seconds.

When I flip the power switch: it turns red, stays that way for 5 seconds, turns green, and after 10 seconds, it goes red again (indicating protection mode).

If I turn it off and back on, it stays in protection mode (the relay doesn’t close). But if I leave it off for 1-2 minutes and then turn it on again, it comes out of protection, only to go back into protection after 10 seconds.

The issue persists even with all four 4A fuses removed from the two main transformers (which supply the bridge rectifiers).

All measurements I’ve taken were with the components disconnected from the boards, except for the voltage checks.

- Power packs (Darlington, rebranded SK ST100): Good
- Pre-driver transistors 2SC1628 and 2SA818: Good
- Diodes: Good
- Capacitors on both the protection and amplifier boards: Good
- Rectifier diodes (three sets of four for three transformers): Good
- Voltages from all three transformers: Good
- Fusible resistors: Good

I’m hoping for guidance from someone more experienced to point me in the right direction. I'm surely missing something crucial.

I’ve attached the service manual, though I’m not sure if pdf and uploads work as it should.. -Let me know and I'll upload it elsewhere.
 
Something builds up heat is the logical assumption.
Have you measure it for DC while it works?
Do you happen to have a thermal camera to observe?

(that's what newbie me would do after observing all components up close,measuring for voltages,continuity,etc )
 
Something builds up heat is the logical assumption.
Have you measure it for DC while it works?
Do you happen to have a thermal camera to observe?

(that's what newbie me would do after observing all components up close,measuring for voltages,continuity,etc )
Nothing gets hot and no components seem to have been hot.

It outputs a rising DC voltage at the terminals before going into protection mode.

I noticed I haven't described the behaviour quite exact enough; when the four 4A fuses (which supply the rectifier circuits from the two main transformers) are in place, the system never comes out of protection. However, if I remove them, it comes out of protection for about 10 seconds before going back into protection.
The DC after the rectification (the rails) is clean. Can't measure any AC.

I'm missing something here. Obviously.. :)
 
Does the issue remain if you install shorting plugs on MAIN IN? That would dramatically narrow down the list of suspects.
It outputs a rising DC voltage at the terminals before going into protection mode.
Both equally? Could be a bad / missing ground connection then. Do you reckon someone might have been in there before you or it might have been dropped at some point?

Make sure neither R421 nor 422 are open, and dried-out C411/412 could reduce the gain to the point of making the amp unstable (or if they are leaky, cause considerable DC offset). There's also C403-6 that could make the amp unstable if dried out.

I also spot a number of fusible resistors in the power amp that should be checked.

I won't be surprised if protection has a problem of its own, but make the amplifier basically work first.
 
Meter for DC mV at TP403 and TP404 located at the Zobel network(s) at the output of the Darlington amplifier ICs (IC401 & IC402). This will check for offset voltage. It should be less than maybe 30mVDC. If there is more than tens of millivolts DC and especially if there is tens of volts DC then the output amp IC(s) is maybe fried. This is the most common fault with older IC type amps so we are cutting to the chase by metering this. You may also check for power supply voltage of +/- 47VDC at pins 2 & 7 of each IC401 and IC402.
Screenshot 2024-10-17 194025 TP403-404.png
 
Thank you both @Doodski and @AnalogSteph

I'm fairly certain I've already measured and checked some of the things you're suggesting. The fusible resistors are good. However, I can't recall the details right now, and I don't have my notes with me as we're currently at our vacation home.
The unit appears not to have been dropped. It looks untouched inside. No evidence of non-stock soldering. There was just a fine layer of decades old smelly dust. No signs of liquid ingress. No signs of any component overheating or overdriven. It's in pretty mint condition and that's why I kept it with the intend of getting it up and running.

I'll check and get back to you at a later time. Thank you for your inputs.
 
Alright, I’ve decided to stop troubleshooting the amplification stages. I didn't find any faulty components, only a suspect cap. that I've changed.

The unit still enters protection mode even with the four 4A fuses for the L+R main transformer supply removed. My next focus will be on the protection circuit, though it's not as easily accessible. I plan to check Q501, Q502, Q503, Q504, and Q505, which control the relay, and monitor the DC offset.

All grounding connections are good, and the overall state is quite clean. I’d really like to bring it back to life.
 
This week, work hasn’t been too demanding, so I’ve managed to measure a lot of components out of circuit. I’ve been investigating fluctuating DC on the left channel output -cycling between 0 to 400 mV -which led me to focus on the feedback loop.

I won’t go through every component I’ve checked, as none of them were outright defective. However, I believe I’ve identified the likely culprits in the power regulation.

The differential pair transistors Q601 and Q602, shown in the attached schematic. Q601 has a measured gain of 59, while Q602’s gain is 108, creating a significant mismatch.

Second, capacitor C624 appears suspicious with an ESR of 3.4 ohms.

To access the board, I had to disassemble the entire front of the amplifier, of course.

Comments appreciated.

Circuit_6.jpg
 
Last edited:
I’ve been investigating fluctuating DC on the left channel output -cycling between 0 to 400 mV -which led me to focus on the feedback loop.
So are the supply voltages at the input pair stable and where they should be? There's fusibles all over the place. Also note that the magnitude of "-43 V" would actually be expected to be smaller than that of "-42 V" as drawn - the magic capacitance multiplier that has higher voltage coming out than going in has yet to be invented. ;)

Fluctuating DC offset on the left channel could indicate a leaky C411 or Q401 (wouldn't be the first 2SA798 to go bad, there's a whole cottage industry around them, see e.g. AK).
The differential pair transistors Q601 and Q602, shown in the attached schematic. Q601 has a measured gain of 59, while Q602’s gain is 108, creating a significant mismatch.
These are not a differential pair at all, they are complementary npn and pnp types serving as regulator pass transistors. Higher gain in the pnp at low (test) currents is quite normal, and the value for the npn seems within spec. You would be astounded at the differences to be seen in nominally complementary types.
Second, capacitor C624 appears suspicious with an ESR of 3.4 ohms.
What do others of the same type like C623 (C405, C429, ...) measure like in comparison? It's a 1µ/100V, so a few ohms of ESR are by no means unexpected.
 
So are the supply voltages at the input pair stable and where they should be? There's fusibles all over the place. Also note that the magnitude of "-43 V" would actually be expected to be smaller than that of "-42 V" as drawn - the magic capacitance multiplier that has higher voltage coming out than going in has yet to be invented. ;)

Fluctuating DC offset on the left channel could indicate a leaky C411 or Q401 (wouldn't be the first 2SA798 to go bad, there's a whole cottage industry around them, see e.g. AK).

These are not a differential pair at all, they are complementary npn and pnp types serving as regulator pass transistors. Higher gain in the pnp at low (test) currents is quite normal, and the value for the npn seems within spec. You would be astounded at the differences to be seen in nominally complementary types.

What do others of the same type like C623 (C405, C429, ...) measure like in comparison? It's a 1µ/100V, so a few ohms of ESR are by no means unexpected.
Thank you for commenting.

From my notes:
I did have a stable 43 V.
C411 measured 43 uF
Q401 I did measure a Vb=29V, Vc=68V, Ve=28.3 V, Vbe=0.7V
C623 measures 1.2 uF, 0.8 ESR
 
C411 measured 43 uF
Which on a nominal 33 µF cap tends to indicate excessive leakage current, no major surprise there (normally it never sees any DC to speak of so its dielectric layer never gets replenished). Needs to be reformed at the very least, and it may be too late for that.
Q401 I did measure a Vb=29V, Vc=68V, Ve=28.3 V, Vbe=0.7V
That's bonkers. What did you use as a 0 V reference? Make sure that input ground (at R401, R405, C407, C411) shows continuity to "official", chassis / power supply ground, and that R405 does not measure >>56k in-circuit (power off).

Also check that R421 (47k) is not open. What voltage to ground are you getting on either side of it?
C623 measures 1.2 uF, 0.8 ESR
Well yeah, then C624 may have seen better days. Small caps like that are prone to drying out over the decades if the seal isn't 100% perfect. It doesn't seem critical yet but should eventually be replaced (you may only be able to get 2.2µ/100V or 4.7µ/100V, either would be fine).
 
Which on a nominal 33 µF cap tends to indicate excessive leakage current, no major surprise there (normally it never sees any DC to speak of so its dielectric layer never gets replenished). Needs to be reformed at the very least, and it may be too late for that.
Is it? -Didn't know that higher measured capacitance is indicative of leakage. In any case, I did change both C415 and C411 as one of the very first things because I then had a broken tester and couldn't be sure.
C415 and C411 has been changed early on.
That's bonkers. What did you use as a 0 V reference? Make sure that input ground (at R401, R405, C407, C411) shows continuity to "official", chassis / power supply ground, and that R405 does not measure >>56k in-circuit (power off).
Ref. is chassis. Or rather should be. I may have made a mistake in my notes as rail voltages are much lower. I'll backtrack when I have changed C624 and Q601/602 and assembled the boards again enough to power it on.

All ground points has been measured fine.
R405 is OK. (Out of circuit)

Also check that R421 (47k) is not open. What voltage to ground are you getting on either side of it?
R421 is OK. (Out of circuit)
Well yeah, then C624 may have seen better days. Small caps like that are prone to drying out over the decades if the seal isn't 100% perfect. It doesn't seem critical yet but should eventually be replaced (you may only be able to get 2.2µ/100V or 4.7µ/100V, either would be fine).
I managed to find 1uF/100V.
 
Back
Top Bottom