Pioneer SA-7800 - Help replacing transistors!

How do you know the undamaged channel is really undamaged?

re: heat sinks - your fingers will tell you. q5 and q6 most likely to be hot, but their VAS cousins q9, q10 on the other side of the circuit should have similar power dissipation, thus may need them too. The use of the larger to-126 cases MAY obviate the need for heat sinks, they were mentioned as a POSSIBILITY, not a necessity.

The original fault in the circuit was they had too much heat to dissipate for the case size, now the case size has been significantly increased, and they shouldn't get as hot.
 
markthefixer said:
How do you know the undamaged channel is really undamaged?
Click to expand...
I can't say that for sure. I just found it to be functional, but yes it is possible this channel to be "affected" too.
I do have all the replacements for both channels and I won't let the "good" channel as it is right now.
I just wanted to have a reference (and also to do the job step-by-step). Is it a bad idea?

markthefixer said:
re: heat sinks - your fingers will tell you. q5 and q6 most likely to be hot, but their VAS cousins q9, q10 on the other side of the circuit should have similar power dissipation, thus may need them too. The use of the larger to-126 cases MAY obviate the need for heat sinks, they were mentioned as a POSSIBILITY, not a necessity.
Click to expand...
Oh Yes, I forgot q9, q10 peers!
I have enough heat-sinks for TO-220, so I think it won't do any harm if I will stick some heat-sinks to q5, 6, 9 & 10.
I just have to check if TO-220 heat-sink will fit well the TO-126 case...
Thanks,
SG.
 
Ok, the other channel operates, do one at a time...

Just my usual caution:
I have had damage cross over from one NSA channel to the other using the protection circuit as a bridge. It happens when the outputs blow and the emitter resistors go a whisker faster. Huge voltages develop, and blooie, crossover.
 
Wow!
Sometimes forgetting to take pictures before an operation is a very-very good thing!

I finished the deforestation of the LEFT side of the board...
Then I was about to solder a KSC1845 and... surprise: this transistor is ECB, while the original SC1914 is BCE !!!
Holly Molly, having a BEFORE pic could have me tricked... I could have been tempted to solder the new transistor in the same way as the old one!
And then (probably) BOOM & SMOKE after powering up!

So, AKers with expertise, any other traps I should be aware of during the restoration of the amp ? :scratch2:

Oh, one more question:
Can I operate the amplifier without the FL display module ? (better ask than sorry...)
 
Last edited:
You NEVER use the original transistor layout for the replacement, you use the replacement transistor's data sheet and the symbol on the board.

I have lost count of the number of times I and others have said this.

ok, you're gonna get it:

Use the symbol silk screened onto the board.

markthefixer said:
markthefixer said:
every Mouser transistor has a data sheet available for download.... download them and look them up. then install as per the symbol printed on the pc board at that location.
never ever use the original transistor to determine how the replacement will go in.

from an earlier post:



First you need to learn about what the three transistor leads are about, and identifying which is which and where it goes according to the symbol. Then you will see that my parts list include lead arrangements - which assume the transistor lettering (and flat) are facing you and the leads are facing down, then go left to right.

Q13 2sc938 150v 0.05a 0.6w
512-KSC2383YTA to-92L ecb 160v 1a .9w 50mhz 160-320hfe $0.17


attachment.php


The big TO-220's with the metal mounting plate are hard to reverse, but the smaller ones are, but my specified replacements are usually E C B as a rule of thumb BUT we ALWAYS CHECK...

But you really should start here and follow ALL the links...

http://www.audiokarma.org/forums/showthread.php?t=331932&page=5

here's a preview, DO YOU SEE these symbols on the boards you are working on?:


80px-BJT_PNP_symbol.svg.png


80px-BJT_NPN_symbol.svg.png


e is always the arrow on one angled line (direction can be out (npn) or in (pnp) towards the base)
b is always the flat line the other two lines come out if at an angle
c is always the angled line without the arrow

letters can be smudged or illegible, while the symbols are hard to mistake and pretty language independent
pioneer doesn't use letters, if there is a symbol. sometimes in a linear hole arrangement, there is only an 'e'.... with no room for a symbol.



attachment.php



-----------------------------------------------------

now a quick lesson on installing a small transistor in a traditional triangular hole arrangement with a symbol printed in the center of the triangle

1. find the LEFTmost lead on the package and identify it's designation (emitter? base? collector?)
2. look at the symbol on the board, find the corresponding hole
3. insert that lead into that hole
4. find the RIGHTmost lead on the package and identify it's designation (emitter? base? collector?)
5. look at the symbol on the board, find the corresponding hole
6. insert that lead into that hole
7. the center lead will now need to bent forward or back to go into the leftover hole.

DON'T THINK "WHERE IS THE FLAT FACING"
think
"outside leads" and THEIR holes...

this one's ecb....
attachment.php


and two transistor layouts, ebc and ecb with the same triangular target hole arrangement
attachment.php
Click to expand...
markthefixer said:
well, you do need to get the parts, first.... :D

and when ordering the transistors, mouser has links to the transistor's data sheets, DOWNLOAD the data sheets!!! THEY are your BEST guide as to the correct lead connections for THAT particular transistor.

Do you understand the transistor symbol? enough to identify the emitter, base and collector from it? Because most boards (Pioneer) have the symbol printed on it, with holes around it, and it is orientated so there is an emitter hole, a base hole and a collector hole. And obviously, only the transistor's emitter goes into the emitter hole, the base into the base's hole and the collector into the collector's hole. IF you want it to work... :D

oh, and by the way, the symbol usually DOESN'T have the E, C and B letters ON on it... :stupid:
there are two flavors,
pnp: arrow pointing in, 2sa,2sb, ksa, ksb part number prefixes
npn: arrow pointing out, 2sc,2sd,ksc,ksd part number prefixes

attachment.php


hint: if "rotating" it in your mind is a puzzler, orient it in the hole this way(transistor hole arrangements are usually triangular):

choose the leftmost lead, use the data sheet to identify it... for example it is the emitter
put it in the emitter hole

choose the rightmost lead, use the data sheet to identify it... for example it is the base
find the base hole, rotate the transistor line of leads around the inserted leftmost lead to point at the base hole
put the rightmost (base) lead in the base hole.

the THIRD hole is now either in front of or behind the transistor, in this example the collector is "left over"
bend the center lead forward or back to go into the hole.

easy...

notice that this example had the lead order of emitter, collector, base which is a common lead arrangement

that means that:

looking at the transistor, with the leads pointing DOWN, and the flat part (or lettering in the case of some zetex transistors) FACING YOU
going from left to right has the lead order of emitter, collector, base
Click to expand...


attachment.php
Click to expand...
 
1) Ok, now I know 'E'-letter is the key to use - when replacing transistors!

2) I didn't manage (not enough time) to fit all components of the LEFT channel.
Only 2 transistors are not yet soldered on the board: the NSA drivers... because I want to shorten their related heat-sinks.
The area is quite crowded there, and I don't wanna risk a nasty short circuit waiting to happen.

3) Question:
Can I operate the amplifier without the FL display module ? (better ask than sorry...)
TIA, Stanley.
 
groenhen said:
3) Question:
Can I operate the amplifier without the FL display module ? (better ask than sorry...)
TIA, Stanley.
Click to expand...

I can't think of any reason why not.

go ahead, if all leads are insulated.

BTW I finally found the pictures I had been looking for, the connectors you installed on the display board - the pin spacings FLOORED me. Back on page 1, post #9

What are the part numbers of these connectors and where did you get them? In that location I would consider them permissible, and if they have gold plated counterparts, there are more locations that they can be used in.

I spent the better part of a DAY looking for those pictures, and here they were, hiding under my nose....
 
markthefixer said:
What are the part numbers of these connectors and where did you get them? In that location I would consider them permissible, and if they have gold plated counterparts, there are more locations that they can be used in.
...
Click to expand...

Here it is:
http://www.tme.eu/en/details/nx5080-04sms/conectori-semnal-pas-508mm/joint-tech/a5081wv-4p/
and
http://www.tme.eu/en/details/nx5081-04pfs/conectori-semnal-pas-508mm/joint-tech/a5081h-4p/#
and, of course:
http://www.tme.eu/en/details/nx5081-tf/conectori-semnal-pas-508mm/joint-tech/a3960-tbe/#
Not golden, but good quality.
I think I may use some contact spray to avoid oxidation...

PS: Picture attached is from my beloved SA-708:
Decent connectors instead of ugly wrappings, and new e-caps too!
Cute :)
 

Attachments

  • IMG_1520.jpg
    IMG_1520.jpg
    86.3 KB · Views: 83
Last edited:
Tada,
I finally "rehabilitated" the LEFT channel:

1) I decided to leave q5 & q9 with no heat-sinks, since they were already upgraded to a bigger case;
2) However, I've put heat-sinks to those Toshiba NSA drivers;
3) I re-checked all the replacements and... found a transistor soldered in a wrong way :drool:. Corrected and moved forward;
4) I made the adjustments according to service manual, but after waiting for about 10 mins warm-up (which manual does NOT say!!!);
Well... the adjustment was a real PITA (pain in the a$$), those voltages swing quite a lot...
But, finally, I achieved relatively stable values +/-0.8mV around the values mentioned in the service manual.
5) I took some pictures (see attachment);
6) I do have a pair of test-speakers, but I forgot to take the CD-player with me, so I postpone the audio tests for tomorrow.
 

Attachments

  • IMG_1645.jpg
    IMG_1645.jpg
    126.4 KB · Views: 109
Last edited:
rcs16 said:
Well done!!
Click to expand...

Thanks!

BTW: I forgot to say I am very impressed how few e-caps SA7800 has! :thmbsp: (for example, SA-708 has exactly 49 e-caps spread all over its boards)
Now, checking the picture in my previous post, you can see the LEFT channel has ZERO e-caps (idem for RIGHT channel).
Actually, that power amp board has ONLY 2 e-caps that belong to protection circuit...
Hmmm, my mistake: I bought from mouser UKL instead of UPW:
C25 = 330uF / 6V -> Mouser: 647-UKL1A331KPDANA
C26 = 100uF / 50V -> Mouser: 647-UKL1H101KPD
Could that be a problem?
 
Last edited:
Yup I see only a few, on the tone, C51/52 (22uf/16V/NP) what did you use? Could parallel them with a stacked polyester film 0.47-2.2uF, what ever fits, Sansui does this all the time.
C39-42 in the tone feedback, also NP types.
There you go 2sk129A jfets in the tone ckt. Too bad they did not use them in the phono too. Sansui used them all the time as they must have felt they sounded better than bjts in that position.
UKL is a better performing ecap than a UPW, not too sweat.
Nice little amp, hope it serves you well.
 
rcs16 said:
Yup I see only a few, on the tone, C51/52 (22uf/16V/NP) what did you use? Could parallel them with a stacked polyester film 0.47-2.2uF, what ever fits, Sansui does this all the time.
C39-42 in the tone feedback, also NP types.
...
Click to expand...

I intend to replace all NPs you mentioned above with UES (Muse?) from Nichicon.
To be honest, I'm kinda purist, so 99% of the time I use the TONE DEFEAT switch.

Just curious: Why does Sansui parallel those NPs with stacked film (0.4-2.2uF) ?!
 
Last edited:
Sansui usually used two "brown written on the schematic" ecaps (Elna/Sanyo?) back to back to make a bi-polar ecap and then put a film cap in parallel, with those two series ecaps, why, well their thinking was to use the ecaps to pass the low freq information and use the film to pass the hi-freq information. I assume because the films have less loss and distortion thus higher sound quality. I am not sure if Sansui used stacked film or wound poly. I guess I can check a few out to measure dissipation factor, another day perhaps.
Pioneer/Marantz used a lot of those sky blue tantalums, in the signal path, where as, I have never seen Sansui use any at all.
 
Last edited:
Whooopeee !!!
IT WORKS !!!
:banana::banana::banana:
Audio opening belonged to Mr. Billy Idol - Eyes Without A Face! :music: Nice!

Bass are not as expected, probably because I used some modest test-speakers... (SONY SS-H-177)
The only annoying thing: after relay engages, I hear a little fart on the RIGHT channel (the one that I didn't touch!).
Not sure, maybe it was there even before I made changes to the LEFT channel, I didn't pay too much attention at that time...
Anyway, I am happy - great achievement so far! :tresbon:
Next step: same changes on the other channel + protection circuit :scratch2:
 
Last edited:
1) Didn't manage to replace all components on RIGHT channel... Maybe tomorrow.

2) Power-On lamp is dead! (it used to light intermittently before). RIP!
I have a nice warm white LED, so I'm thinking about LED conversion...
I only have to find a place to hide a small bridge rectifier, also not to interfere with audio paths... :scratch2:
 
Danger... the 8v 50mA lamp is run off of 49v ac with R85 and R86 (860 ohms, 6 watts) taming it. Your intended bridge (which won't work as you expect) will give about 63v peak voltage (the lamp rides it (the extra current) out, the led won't). Your resistors will have to dissipate about 2 watts at 20mA, and thus there are better places to tap that voltage than adding a bridge rectifier somewhere...

Tap it off of the 46v unregulated main capacitors (they won't miss the extra 20mA), while tapping the raw 63v or regulated 48v will increase the dissipated heat even more. Tapping off of DC will eliminate the flicker potential that I infer you were trying to avoid by using a diode bridge.

Quickest and easiest with the least modifications is to change R86 to 3200 ohms 2 to 3 watts and substitute a 1n4004 for r85. That gives 20mA for the led. It may be too bright. 6400 ohms will give 10mA and cut heat dissipation in half. As current goes down, so does the power dissipation, while the resistance required skyrockets (we like that).
But there still could be flicker. Lifting the R85 inner board lead end will disconnect from the ac feed and allow a flying wire from the resistor lead to tack onto the foil at the base of pins 29 or 30 for that raw +46v .
 
Last edited:
markthefixer said:
Quickest and easiest with the least modifications is to change R86 to 3200 ohms 2 to 3 watts and substitute a 1n4004 for r85. That gives 20mA for the led. It may be too bright. 6400 ohms will give 10mA and cut heat dissipation in half. As current goes down, so does the power dissipation, while the resistance required skyrockets (we like that).
But there still could be flicker...
Click to expand...

Oh, ooops - I forgot to mention: I modified R85, R86.
Now they are R85 = R86 = 3.3 kohm (3W each).
(Replacing one resistor with a diode, yes - it gives an annoying flicker... Maybe things are ok in US where power grid is 60Hz?)
Now I'm trying to find a place to hide the diode-bridge...
1) solder the bridge ~/~ directly to pins 37 and 60 and then solder the LED wires to bridge +/-. Maybe reinforce it (on the foil side) using a glue gun...
or
2) wrap the bridge in therm-shrinking tube (hide it somewhere) and use some wires to connect both ways: ~/~ to pins 37, 60 and +/- to the LED.

Dunno...:scratch2:

PS: I have a looong experience with white LEDs. They work better in AC than DC.
The only way to run them in DC with NO problems is to use a constant current power supply.
 
Last edited:
groenhen said:
Oh, ooops - I forgot to mention: I modified R85, R86.
Now they are R85 = R86 = 3.3 kohm (3W each).
(Replacing one resistor with a diode, yes - it gives an annoying flicker... Maybe things are ok in US where power grid is 60Hz?)
Now I'm trying to find a place to hide the diode-bridge...
1) solder the bridge ~/~ directly to pins 37 and 60 and then solder the LED wires to bridge +/-. Maybe reinforce it (on the foil side) using a glue gun...
or
2) wrap the bridge in therm-shrinking tube (hide it somewhere) and use some wires to connect both ways: ~/~ to pins 37, 60 and +/- to the LED.

Dunno...:scratch2:
Click to expand...




markthefixer said:
Danger... the 8v 50mA lamp is run off of 49v ac with R85 and R86 (860 ohms, 6 watts) taming it. Your intended bridge (which won't work as you expect) will give about 63v peak voltage (the lamp rides it (the extra current) out, the led won't). Your resistors will have to dissipate about 2 watts at 20mA, and thus there are better places to tap that voltage than adding a bridge rectifier somewhere...

Tap it off of the 46v unregulated main capacitors (they won't miss the extra 20mA), while tapping the raw 63v or regulated 48v will increase the dissipated heat even more. Tapping off of DC will eliminate the flicker potential that I infer you were trying to avoid by using a diode bridge.

Quickest and easiest with the least modifications is to change R86 to 3200 ohms 2 to 3 watts and substitute a 1n4004 for r85. That gives 20mA for the led. It may be too bright. 6400 ohms will give 10mA and cut heat dissipation in half. As current goes down, so does the power dissipation, while the resistance required skyrockets (we like that).
But there still could be flicker. Lifting the R85 inner board lead end will disconnect from the ac feed and allow a flying wire from the resistor lead to tack onto the foil at the base of pins 29 or 30 for that raw +46v .
Click to expand...

NO BRIDGE!!

There's center tapping and grounds to consider on the foil pattern where you have the resistors!!.

Pay attention, you are almost there...

You already have the LED running, by your flicker comment, so:

Just lift the named (inner lead - NOT the lead closest to the edge of the board) R85 lead, connect a wire and fly that wire over to and tack it onto pin 29 or pin 30.

Light should be fine with no flicker - the difference in voltage shouldn't result in much of a brightness difference.
 
Last edited:
I ran my LED off the 3v AC feed for the fluroscan display. Just another way of doing things.

Lee.
 
You must log in or register to reply here.

Similar threads

bass3d
Pioneer Sa 7800 power transistors burned out
Replies
0
Views
199
bass3d
bass3d
VintageMad89
SA-7800 NSA driver transistors
Replies
11
Views
743
VintageMad89
VintageMad89
P
Pioneer SA-7800 Output Transistor Question
Replies
2
Views
375
Pubcha
P
70smidfiguy
Pioneer SA-7800 replaced output transistors ....
Replies
19
Views
3K
70smidfiguy
70smidfiguy
mazinkaiser
Pioneer SA-708 replacement output transistor help!!
Replies
8
Views
1K
mazinkaiser
mazinkaiser
Back
Top Bottom