SA-7500 Protection

QSilver

Super Member
I was using my SA-7500 yesterday with the preout connected into a SPEC 4. After about an hour I started to notice that the SA-7500 was getting quite warm and this amp usually runs only slightly warm. After another 30 minutes I began to hear the protection relay clicking on and off. I turned the amp off and open the case and hooked a DMM into the speaker outputs and watched the DC voltage and powered it back up.

For the first 30 minutes it stabilised but the power transistors on the right of the amplifier were starting to get warm. After 30 minutes or so, the DC started to dance around rapidly and the relay began to click again - at one point the DC peaked at about 1V as the protection kicked in.

I turned it off and left it over night and when I come back to it now with the multimeter on, the amp has been on an hour and is staying cool and the DC is moving around 3.2-4.6mV.

I find it strange that its running in and out of protection one day and not the next. The amp was made in 1976 and is all original as far as I can see.
 
Was (is) the line voltage steady? Depending upon how and where you plugged in the SPEC4 for power it may be pulling down the line so the unit being used as your preamp may be going undervoltage. Just a WAG.

It seems odd that only one channel would suffer, however, but there is always one weaker sister of the two (or four).
 
The sa-7500 has 2sa725 transistors in the equalizer AND control amps, teh 2sa726 AND 2sc1451 transistors in the power amp.

ALL THREE are known troublemakers. The 1451 is #1, the 725 #2 and 726 # 3....

I have replaced the a725 and a726's with:
512-KSA992FBU (ln)to-92 ecb 120v .05a .5w 100mhz 150-800hfe $0.05 ea
(and the 2sc1312,2sc1313 complements with
512-KSC1845EBU (ln)to-92 ecb 120v .05a .5w 100mhz 150-800hfe $0.05 ea)

while the
2sc1451 with:
512-KSC3503estu to-126 ecb 300v .1a 1.2/7w 150mhz 40-320hfe $0.28 ea
(and it's complement where appropriate
512-KSA1381estu to-126 ecb 300v .1a 1.2/7w 150mhz 40-320hfe $0.27 ea)

Pulling the pre-out / amp-in jumpers may help isolate it to the preamp or power amp
BUT
when these transistors fail, they shoot current pulses back into the power supply which can upset the protection even with the jumpers pulled.
 
Thanks for the replies, While testing the machine its on my bench rather than hooked in with the SPEC 4 so possible this was causing some sort of upset. Again, I've had it running idle for what is now two hours and no problems DC has risen to 6.6-7.3mV and the amp has warmed very slightly but no spiking yet. Will keep it running and see what happens.

I took some voltages from the power amp board pins but everything looks stable there besides Pins 3 and 17.

Pin 3 (0V) Moves between 0 and -1.2mV
Pin 6 (35V) 37.5 dropping to 37.4 every so often
Pin 9 (-40V) -41.2 Steady
Pin 10 (-13V) -13.1 Steady
Pin 13 (-40) -41.2 Steady
Pin 17 (0V) Moves between +3.5 and +7.4mV

Idle Current for Left channel is 26.7mV
Right Channel is 28.1mV

Will keeping running it for a few more hours and see what happens. Though I'm starting to think that maybe the Spec 4 was pulling it down as they are both running from the same wall socket but not during these tests.
 
scheduled for Obsolescence

Just looked up the replacement you suggested for 2SC1451 on mouser and it says its scheduled for obsolescence. If the problem persists, I'll need to get these quick!
 
That gain code "e" only, in the c d e f gain sequence the "d" is safe for now...

original number:
512-KSC3503ESTU

use
512-KSC3503DS bulk packed
512-KSC3503DSTU in tube carrier
 
When I came back to the amp after posting last, it began working perfectly again and so left it as it wasn't high on my priorities.

However, yesterday it did the same thing again only this time it wont come back out and the right channel has 36VDC running on it. There is also a small crackling noise coming from the area around Q7 so checked solder joints and re tested but no change and still the crackling noise. The Left channel behaves normally and gets down to 1mV within about 5 seconds of powering up but right peaks at 36V and moves between that and 1V erratically.

What could that crackling noise be?
 
Magnetostrictive effects of high current pulses (caused by the 1451), if the transistor isn't literally self destructing. The emitter resistor R21 frying under fast high current pulses? in any case, it's NOT GOOD.

Get those 2sc1451's OUT before you have an expen$ive meltdown. Like in destroying the output transistors (even though they ARE TO-3's and easily replaced)....

How much clearer can I make it? A few bucks in transistors - unfortunately more for shipping...

You can mitigate the bite by getting some ksa992 and ksc1845 transistors for other spots in the amp (the a725 and a726 transistors are almost as bad as the c1451, and the ksc1845 complements are for "replacement symmetry" in the circuits for the c1312 and c1313's) and visually inspecting for any sky blue sanyo caps which also have a proven reputation of eventual failure.

if q7 & q9 are damaged:
512-KSA940TU pnp to-220 bce 150v 1.5a 25w 4mhz 40-140hfe $0.62
512-KSC2073TU npn to-220 bce 150v 1.5a 25w 4mhz 40-140hfe $0.66
are good replacements...

512-KSA992FBU (ln)to-92 ecb 120v .05a .5w 100mhz 150-800hfe $0.05 ea
512-KSC1845EBU (ln)to-92 ecb 120v .05a .5w 100mhz 150-800hfe $0.05 ea

........534-4662 to-3 mica insulators $0.16 each.................
863-MJ21193G pnp to-3 250v 16/30A 250W 8-75hfe 4mhz $4.56
863-MJ21194G npn to-3 250v 16/30A 250W 8-75hfe 4mhz $4.23

512-KSA1381estu to-126 ecb 300v .1a 1.2/7w 150mhz 40-320hfe $0.27 ea *** just in case, here's the complement ***
512-KSC3503estu to-126 ecb 300v .1a 1.2/7w 150mhz 40-320hfe $0.28 ea
 
Last edited:
You can mitigate the bite by getting some ksa992 and ksc1845 transistors for other spots in the amp (the a725 and a726 transistors are almost as bad as the c1451, and the ksc1845 complements are for "replacement symmetry" in the circuits for the c1312 and c1313's) and visually inspecting for any sky blue sanyo caps which also have a proven reputation of eventual failure.

if q7 & q9 are damaged:
512-KSA940TU pnp to-220 bce 150v 1.5a 25w 4mhz 40-140hfe $0.62
512-KSC2073TU npn to-220 bce 150v 1.5a 25w 4mhz 40-140hfe $0.66
are good replacements...

512-KSA992FBU (ln)to-92 ecb 120v .05a .5w 100mhz 150-800hfe $0.05 ea
512-KSC1845EBU (ln)to-92 ecb 120v .05a .5w 100mhz 150-800hfe $0.05 ea

........534-4662 to-3 mica insulators $0.16 each.................
863-MJ21193G pnp to-3 250v 16/30A 250W 8-75hfe 4mhz $4.56
863-MJ21194G npn to-3 250v 16/30A 250W 8-75hfe 4mhz $4.23

512-KSA1381estu to-126 ecb 300v .1a 1.2/7w 150mhz 40-320hfe $0.27 ea *** just in case, here's the complement ***
512-KSC3503estu to-126 ecb 300v .1a 1.2/7w 150mhz 40-320hfe $0.28 ea

OK, Will take a look at this properly and order some new components next week. Just looked around for the Sanyo caps and they are dotted about all the PCBs. I tried to locate them on the schematic and the two on the power amp board appear to be C1 and C2. Problem is these are rated on the schematic as 1(mF?) @ 50V but in real life they are 47mF @ 25V. Is this correct? i.e. an equivalent?

Also, what section are these for?

512-KSA992FBU (ln)to-92 ecb 120v .05a .5w 100mhz 150-800hfe $0.05 ea
512-KSC1845EBU (ln)to-92 ecb 120v .05a .5w 100mhz 150-800hfe $0.05 ea

This amp is all original from 1976 so its due a service, the EQ circuit needs attention as the sound changed slightly (more mid range) when the tone switch is turned on with all controls defeated. Are there any caps/transistors that need replacing which would effect this?
 
OK, I've put all of that into an order, Didn't bother getting the output transistors, I'm fairly sure they're ok. I got the 512-KSC3503estu but it's compliment is obsolete. Do I need to get the gain code D instead to get a compliment?
 
OK, I've put all of that into an order, Didn't bother getting the output transistors, I'm fairly sure they're ok. I got the 512-KSC3503estu but it's compliment is obsolete. Do I need to get the gain code D instead to get a compliment?

the ksc complement is ksa, and whatever best gain code you can get for each.

Also, what section are these for?

512-KSA992FBU (ln)to-92 ecb 120v .05a .5w 100mhz 150-800hfe $0.05 ea
512-KSC1845EBU (ln)to-92 ecb 120v .05a .5w 100mhz 150-800hfe $0.05 ea

these:

I have replaced the a725 and a726's with:
512-KSA992FBU (ln)to-92 ecb 120v .05a .5w 100mhz 150-800hfe $0.05 ea
(and the 2sc1312,2sc1313 complements with
512-KSC1845EBU (ln)to-92 ecb 120v .05a .5w 100mhz 150-800hfe $0.05 ea)


This amp is all original from 1976 so its due a service, the EQ circuit needs attention as the sound changed slightly (more mid range) when the tone switch is turned on with all controls defeated. Are there any caps/transistors that need replacing which would effect this?

yes, and not an easy or quick answer...
replacement transistors to look up, as well as differentiate between ukl low leakage and upw general replacement caps.
 
Last edited:
Sorry its been a while, Its been a busy, difficult time recently,

To get back to this amp, I've been looking through the schematics to refresh my memory and I'm seeing the transistors a725 and a726 on a few different PCBs, the 1313's in the EQ board and 1312's in the Control board. Am I just going to replace ALL the transistors with these codes throughout the Amp?

Cheers,
Qsilver
 
Still getting protection

I finally got around to doing this. I changed all the a725's ,a726's, 1313's, 1312's and the 1451's.

I also changed almost all the capacitors and in some cases used new film capacitors in the smaller value areas. I powered on and let it warm up, setting the DC ofset, then the idle current to 30mV and all seemed good.

I listened to it for around 30 minutes and its started going in and out of protection again. I checked idle current and DC ofset and both are as i set them.

The only thing I can find is that the area around R46 on the control amp assembly is quite warm...
 
was relay coil driver q11 2sc1384 updated? ksc2383?

C16 and C17 0.22uf swapped with film caps?

The "in and out of protection" : was there the turn on delay, in other words: did it take 4-7 seconds to come back from protection, or was the cycling faster?
If cycling FAST, look to q11. If the delay is present, we dig deeper.

you can test and see if the protection cycling (when there is time delay) is suppressed by desoldering one end of D17 and lifting it.
if that does it, then q5, q6 2sc869 can be swapped by ksc1845's.


That's the problem with old lists, knowledge gained isn't included.
 
Yes, I repalced C16 and C17 with polyester film caps.

There is a delay in the protection coming off on power up. The instances of it going into protection only seemed to occur when I turned the volume up higher and let it play for a while. The first time, the protection clicked on then off immediately. Second time it stayed on a little longer and third time it stayed on longer again.

I powered it off and left it for a few minutes while I got my DMM and checked measurements and since then it hasn't done it again, but I didn't turn it up very loud.

No I didn't replace any transistors in the protection circuit - just on the equalizer, power amp and control amp boards.

This might be a stupid question but the replacements for the 1451's - they'll need heatsinks won't they?
 
The specific thing I need to know is:
The first time, the protection clicked on then off immediately. Second time it stayed on a little longer and third time it stayed on longer again.

what you call on is where there is NO sound, while it's a different time delay I am seeking:

NOT the time delay starting when the speakers start delivering sound until the sound cuts out,

but rather the time delay starting when you LOOSE the sound, until the sound comes BACK on AGAIN.

IF the protect circuit is doing it's job, WHEN it cuts the sound off, it should take 4 to 7 seconds minimum to come back. Faster is the relay driver transistor, 4-7 seconds means a fault further back into the circuitry.
And the protect circuit thinks it is doing it's job.

Quite frankly, four ksc1845's are called for, as well as a ksc2383 and a ksc2690, these will get it fixed in one fell swoop. UNLESS the amp itself is misbehaving.

awm078 board parts for protection circuit section

Q5, q6 2SC869-C = over current detectors
ksc1845

Q7, Q8 2SC945-Q = DC offset voltage detectors
ksc1845

Q9 2SA733-Q = logic inverter - RARELY needs to be changed
ksa992

Q10 2SC945-Q = timing capacitor discharge transistor, this needs to be more robust than the original
512-KSC2383YTA to-92L ecb 160v 1a .9w 50mhz 160-320hfe $0.17


Q11 2SC1384-R = relay driver transistor, this needs to be more robust than the original
512-KSC2690AYS to-126 ecb 120/a160 1.2a 20w 155mhz 35-320hfe $0.40

C16, C17 0.22uf
598-DSF050J224 CDE Metallized Polypropylene Film Capacitors 0.22uF 50V 5% $0.14

C18, C19 330uf 6v 647-UPW1A331MPD 330uf 10v $0.20 ea

C21 2.2uf 25v 647-UPW1H2R2MDD 2.2uf 50v $0.12

C20 100uf 16v 647-UPW1H101MPD 100uf 50v $0.22

don't forget to add the 1n4004 flyback diode across the coil, with the banded end
connected to the side of the relay coil going to pin 53,
there is no convenient pin to name for the switched side of the coil - the collector of q11
which gets the anode, UNbanded end of the diode.

512-1n4004 conventional rectifier 1a 400v $0.06 ea
 
Last edited:
I've been running the amp everyday since I last posted for around an hour each time at a good volume level and it hasnt gone into protection once. The only thing that has changed is that im now playing my PD-91 through the amp instead of my pc. I guess there is nothing to be done until it goes into protection again?
 
The amp started going into protection again yesterday and the protection circuit does take at least 4 seconds to come back out. So the fault is further back? Should I check the idle current and DC offset or would is it going to be better to connect a probe somewhere?
 
Subtractive or suppressive troubleshooting works best for intermittent protect triggers.

If it goes away, you're getting close. Better than trying to catch it in the act.


If the dc offsets are good, temporarily open D15, D16 to eliminate the dc offset protect to see if q7,q8 are triggering falsely
and
for the over current sense
D17 to see if q5 or q6 are triggering falsely or leaky.. as long as c16 and c17 are not the originals.

of course you have to be careful with part of the protect circuit not working... like no good speakers.. etc...

or just replace the 4 (q5, q6, q7, q8) preemptively. All four can be ksc1845's.
 
Last edited:
I pulled D15 and 16 and powered on the amp. After around 2 to 3 hours it started to trigger the protection circuit again. Meanwhile I was monitoring the DC offset and Idle current. Offset didn't reach anything higher than 6mV and the idle current is 30mV on both channels - I adjusted the offset to 0mV.

Meanwhile the protection circuit randomly clicks on and off and at the moment its less than 4 seconds. So I guess the ksc1845's are in order? The capacitors have all already been changed so is there anything else I might need?

EDIT: Since then I disconnected D17 too and have left it on all day. Amp is still stable and the protection circuit hasn't triggered either so likely that its triggering too early or falsely?

Looking into ordering the protection transistors, the 512-KSC2383YTA is on back order and out of stock. I checked around for an equivalent that is in stock - is the 512-KSC2383OTA a good substitute?
 
Last edited:
Back
Top Bottom