SX-3900 Powers on, but no audio output...

[markthefixer]

Let's not get ahead of ourselves, we get the power supply working, then evaluate the rest of the unit.

Sorry, but things like this take time.

And the 3900's a hard beast to cut your teeth on, mainly due to it's layout, although the amp's circuits are rather advanced.

 

[SolutionRoom]

Let's not get ahead of ourselves, we get the power supply working, then evaluate the rest of the unit.

Sorry, but things like this take time.

And the 3900's a hard beast to cut your teeth on, mainly due to it's layout, although the amp's circuits are rather advanced.

I am patient. And yes, the SX-3900 is a rather intimidating piece of electronic art. I will be eternally grateful once this thing fires up and puts out pristine sound. My work is cut out for me no doubt.

 

[markthefixer]

your transistor choices look good overall.
exceptions:
Q7 & Q8 +/- 32v series pass transistor choices:
Q7: ksc2310 0.8w (to-92 case) or ksc2682 (to-126 case)
Q8: ksa910 (to-92 case) or ksa1142 (to-126 case)
the 1845 has more gain but only .5w dissipation vs 2310 's .8w with less gain

Q9, Q10 :313 vs 313p, the 2073 will cover both fine...

"101" tiny ceramic 100pf leave it be....

i have no idea where you got 0.01 at 7-10-14 with 224k250
should be metal film poly (probably) 0.22uf at 250v. it's to keep rf noise OUT...
NOT tantalum they go to about 50v, this is non-polarized

agc-004 leave it be

you left out a lot of Cxx location information and i don't have the time to do it for you so that i can recommend voltage upgrades, which i think i already did on the 7000 and will apply 100%.

 

[SolutionRoom]

Thanks Mark. I will certainly take your advice on the Q7 and Q8. As far as the caps, is it safe to bump up the voltage rating a tad, so long as the capacitance remains the same?

I will get these all on order and go from there.

 

[markthefixer]

Thanks Mark. I will certainly take your advice on the Q7 and Q8. As far as the caps, is it safe to bump up the voltage rating a tad, so long as the capacitance remains the same?

I will get these all on order and go from there.

Yes it is safe to bump the voltage rating up. Most of my power supply cap voltage increases are to prevent an over-voltage condition if something in the power supply fails, so IF there's a fault, the damage is minimized.

the to-126's are a cooler running bet, if you look at the specs they are rated for 1.2 watts bare and 10 watts with a heatsink, with the added bonus of no exposed connections where the heatsink is attached - thus the danger of an accidental short is reduced.

 

[SolutionRoom]

the to-123's are a cooler running bet, if you look at the specs they are rated for 1.2 watts bare and 10 watts with a heatsink, with the added bonus of no exposed connections where the heatsink is attached - thus thr danger of an accidental short is reduced.

I assume you mean to-126... I will get those instead of the to-92. I like cooler and more efficient, especially on this power supply board, where there has been some real obvious heat, and this appears to be a common issue for the SX-3900 receiver.

I am gonna try and get these parts all ordered right now.

EDIT: I ordered all of the Caps and transistors from Mouser. They have a nice "project" section to save the items in. Should I replace the resistors too, just to do a complete rebuild?

 

[SolutionRoom]

I want to replace the main Power caps on the Rectifier board as well while doing this rework. The originals are Nichicon 15,000uF 71v, 40mm x 100mm.

I am looking at the Panasonic T-HA 15,000uF 80v, P/N ECET1KA153FA. They are at Digi-Key P/N P10647-ND at $13.58 ea. They are 40mm x 80mm.

Are these an excellent replacement cap, or is there something better suited for this application? The space seems pretty limited to those dimensions, as far as I can tell.

 

[markthefixer]

Resistors with cut leads, yes.

The slightly shorter caps may need a spacer to get the board up to clear the transformer down there - and WATCH OUT soldering on that board, there is a thermal fuse under the big resistor on the board, nearby soldering heat (large foil areas make soldering difficult with a small iron) pops it and it's 2 bucks for a replacement that is tricky to solder in.

 

[SolutionRoom]

Resistors with cut leads, yes.

The slightly shorter caps may need a spacer to get the board up to clear the transformer down there - and WATCH OUT soldering on that board, there is a thermal fuse under the big resistor on the board, nearby soldering heat (large foil areas make soldering difficult with a small iron) pops it and it's 2 bucks for a replacement that is tricky to solder in.

Thanks for the heads up on the thermal fuse. I take it then :scratch2being as there was no comment on the type) that means the Panasonic caps are a nice replacement. I will order them up. What is the best way to put a spacer in there to increase the length? What material?

 

[markthefixer]

Thanks for the heads up on the thermal fuse. I take it then :scratch2being as there was no comment on the type) that means the Panasonic caps are a nice replacement. I will order them up. What is the best way to put a spacer in there to increase the length? What material?

Until you get in there and try, it is hard to say if threaded standoffs will suffice, or the support of a mechanical spacer like a shaped flat piece of wood is needed, as the caps are the main mode of support for the board. And there is a transformer in that cubic volume that may prevent the lowering of the board closer to the chassis side....

Replacing caps usually turns into a mechanical can of worms.

 

[SolutionRoom]

Until you get in there and try, it is hard to say if threaded standoffs will suffice, or the support of a mechanical spacer like a shaped flat piece of wood is needed, as the caps are the main mode of support for the board. And there is a transformer in that cubic volume that may prevent the lowering of the board closer to the chassis side....

Replacing caps usually turns into a mechanical can of worms.

I think I may be able to space out the cap mounts to make up the 2cm difference with not too much problem. That way it will keep the Rectifier board right where it normally is and also provide the proper support. I will have to take a closer look at how the cap supporting/mounting clamps are attached to the frame.

 

[SolutionRoom]

Had a tad bit of a delay... my wife had a baby on Wednesday, kinda tied me up. But I'm back at it.

The Power Supply Board is completely rebuilt, with the exception of a few of the diodes, all caps and transistors are in. Moving on and upon a detailed look at the Power Amplifier Board (AWH-097), I found 2 burned resistors (crispy). I want to replace all of the transistors on this board as well, as a shotgun approach, along with the caps. I have every single part number for the transistors on this board, except for one...

Q105, Q106 2SC2291. I cannot seem to locate a suitable replacement for this part...

Is the 2SC1583 suitable??? Looks like it is only 50v as compared to 100v for the 2SC2291... But the data sheets are in Japanese, adding to my confusion.

 

[Jon_Logan]

Had a tad bit of a delay... my wife had a baby on Wednesday, kinda tied me up. But I'm back at it.

The Power Supply Board is completely rebuilt, with the exception of a few of the diodes, all caps and transistors are in. Moving on and upon a detailed look at the Power Amplifier Board (AWH-097), I found 2 burned resistors (crispy). I want to replace all of the transistors on this board as well, as a shotgun approach, along with the caps. I have every single part number for the transistors on this board, except for one...

Q105, Q105 2SC2291. I cannot seem to locate a suitable replacement for this part...

Is the 2SC1583 suitable??? Looks like it is only 50v as compared to 100v for the 2SC2291... But the data sheets are in Japanese, adding to my confusion.

....2SC1583 Vceo is 50V, 2SC2291 Vceo is 100V. Is this for a differential amplifier? What is the collector voltage? I'd try to find something closer.

 

[SolutionRoom]

It's for the Power Amplifier Board on the SX-3900. Not sure of the specifics, sorry.

 

[markthefixer]

sx3900 amp parts awh097

Had a tad bit of a delay... my wife had a baby on Wednesday, kinda tied me up. But I'm back at it.

The Power Supply Board is completely rebuilt, with the exception of a few of the diodes, all caps and transistors are in. Moving on and upon a detailed look at the Power Amplifier Board (AWH-097), I found 2 burned resistors (crispy). I want to replace all of the transistors on this board as well, as a shotgun approach, along with the caps. I have every single part number for the transistors on this board, except for one...

Q105, Q105 2SC2291. I cannot seem to locate a suitable replacement for this part...

Is the 2SC1583 suitable??? Looks like it is only 50v as compared to 100v for the 2SC2291... But the data sheets are in Japanese, adding to my confusion.

This is one of those innocent sounding posts that is going to eat up my entire evening......... sigh......

Also I have the sneaking suspicion that I have already posted much of this somewhere else in the Pioneer forum.

I have ALREADY done the shotgun thing with the 3900's amps. They can be finicky.

FIRST: Q101, Q102, Q103, Q104, Q105 & Q106 are NOT replaced willy nilly. They have to be PROVEN BAD.
Q101 and 102 are offset compensation transistors for the output offset compensation setup.
Q103 and Q104 are the differential stage transistors, are FIVE leads each, and are TWO transistors in a single case, and they share EMITTER leads.
IF you do INDIVIDUAL eplacements, you need to find the highest gain matched pair.
Q105 and Q106 make up the current mirror portion of the differential stage, they have FIVE leads each, are TWO transistors in a single case, and they share BASE leads.

I HAVE done what you propose - shotgunning the ENTIRE amp, with those exceptions above. Don't forget to do ALL the diodes too, including the zeners. When I used ON-Semi MJE15032G and MJE15033G transistors in the Q134 - Q136 positions ( the four heatsunk to-220 drivers across one edge of the board) I had severe bias control problems, I could NOT dial the bias down to zero, and I had NEW power output transistors in place as well, which when I had FIRST used KSA940 & KSC2073 transistors the bias adjusted properly. I used the mje's as an "upgrade" so that there could NEVER EVER be a situation where the drivers were loaded down. Needless to say the MJE's came out and the KSA/C's went back in.
The outputs were On-semi:
863-MJL1302AG pnp TO-3P/TO-264 bce 260V 15/30A 200W 45-150 30MHZ $3.50 ea
863-MJL3281AG npn TO-3P/TO-264 bce 260V 15/30A 200W 45-150 30MHZ $3.50 ea
These are the smaller brothers to the ones EW used in the sx-1980 restoration



These are the six transistor types I used across the amp:

512-KSA992FBU (ln)to-92 ecb 120v .05a .5w 100mhz 150-800hfe $0.05 ea Q107, Q108, Q111, Q112, Q117, Q118, , Q103, Q104
512-KSC1845EBU (ln)to-92 ecb 120v .05a .5w 100mhz 150-800hfe $0.05 ea Q113, Q114, Q119, Q120 , Q101, Q102, Q105, Q106


************************************
start edit 9/12/2010 replacements
512-KSA910YSHTA to-92 ecb 150v .05a .8w 100mhz 40-240hfe $0.07 ea Q115, Q116, Q121, Q122 9/12/2010 obsoleted
512-KSC2310YBU to-92 ecb 150v .05a .8w 100mhz 40-240hfe $0.11 ea Q109, Q110, Q123, Q124 9/12/2010 large to-92 discontinued
these (ksa910/ksc2310) ran hot...

edit 9/12/2010 possible ksa910/ksc2310 replacements
512-KSA1013YBU to-92L ecb 160v 1a .9w 50mhz 160-320hfe $0.17 (R=60-120, O=100/200, Y=160/320)
512-KSC2383YBU to-92L ecb 160v 1a .9w 50mhz 160-320hfe $0.17
---or---
512-KSA1220AYS to-126 ecb 120/a160 1.2a 20w 155mhz 35-320hfe $0.39 (R=60-120, O=100/200, Y=160/320)
512-KSC2690AYS to-126 ecb 120/a160 1.2a 20w 155mhz 35-320hfe $0.40
---or---
512-KSA1381es x to-126 ecb 300v .1a 1.2/7w 150mhz 40-320hfe $0.27 ea
512-KSC3503estu to-126 ecb 300v .1a 1.2/7w 150mhz 40-320hfe $0.28 ea


***********************
original transistors:
2sa985/a pnp to-220 bce 120v/150v 1.5a 25w 180mhz 35-320hfe orig 180mhz nsa driver
2sc2275/a npn to-220 bce 120v/150v 1.5a 25w 200mhz 35-320hfe orig 200mhz nsa driver

these:
***************!!!!!!!!!!!!!!!!!!!!
757-2SA1930(Q,M) pnp to-220 bce 180v 2a 20w 200mhz 50-320hfe 200mhz nsa driver
757-2SC5171(Q,M) npn to-220 bce 180v 2a 20w 200mhz 50-320hfe 200mhz nsa driver
***************!!!!!!!!!!!!!!!!!!!!
old and slow:
512-KSA940TU_Q pnp to-220 bce 150v 1.5a 25w 4mhz 40-140hfe $0.54 Q135, Q136
512-KSC2073TU npn to-220 bce 150v 1.5a 25w 4mhz 40-140hfe $0.54 Q133, Q134

***************************
9/12/2010 end possible replacements

misc diodes:
512-1N5234B 6.2v zener D101 - D106
512-1N4148 D113 - D116
512-UF4004 ultra fast rectifier 1amp 400v $0.13 ea D111, D112, D117, D118
512-1N4454 D129, D130, D132, D133
512-1N5245B 15v zener D131

and two transistor types I used in the protection circuit:

512-KSA1013YBU to-92 ecb 160v 1a .9w 50mhz 160-320hfe $0.17 ??Q139 (if 2sa1100)??
512-KSC2383YBU to-92 ecb 160v 1a .9w 50mhz 160-320hfe $0.17 Q125, Q126, Q137, Q138, ??Q139(if 2sc2575)??,
512-KSC2690AYS to-126 ecb 120/a160 1.2a 20w 155mhz 35-320hfe $0.40 Q140

edit, Here's the caps:
two 4.7uf 25v amplifier input caps 647-UPW1H4R7MDD 0.14 ea
one 100uf 25v protection cap 647-UPW1H101MPD 0.36 ea
one 470uf 16v protection cap 647-UPW1E471MPD 0.68 ea
two 0.22uf 50v protection caps- changed to poly metallized film 598-DSF050J224 0.50 ea


I have decided to NOT include the original transistor specifications in this post, nor detail (argue) the reasons for my choices of either which transistor replaced what, but even why there MAY be two different replacement numbers used for the same original number - it has to do with what is doing what, and where in the amp. (hint - a few get WARM, and I'm glad I did those substitutions)

The results speak for themselves, I have about the same distortion numbers that original versions of working amps produce.

PAY VERY CLOSE ATTENTION TO THE LEAD ARRANGEMENTS. The emitters are marked on the board.

You have a choice for Q139, depending upon it's type and the variant of the protection circuit on this unit. If it's the 2SA1100, it is the newer circuit.

I also did NOT have positive results using a zetex ztx1056a in the Q140 position, it immediately pulled the relay in on powerup. It could have been a leaky D131, but I just returned the original 2sc1384, after which it operated normally.

this post is provisionally finished. see post 41 for when....

 

[SolutionRoom]

I found the other transistor and diode cross-reference Mouser P/N's from a previous post for a SX-D7000, in which you posted the information Mark.

I verified that the Q numbers from the SX-D7000 crossed over to the actual parts in the SX-3900 schematic.

It was only the Q105 and Q106 spots that were not referenced, and I wanted to do them all so as to not chase my tail.

Am I creating a monster here, or should I just replaced the known bad resistors and diodes? Or do I have to go the "out of circuit" route to test all of the transistors? I feel comfortable to test resistors and diodes, but when it comes to transistors, not so confident.

I can post the P/N's for replacements if it will help. They are all Fairchild, except for the Zetex Q140, 522-ZTX1056A. I have the KSA940 & KSC2073 listed as the ones to use.

It was at this thread

http://www.audiokarma.org/forums/showthread.php?t=161866&highlight=sx-d7000

13th post down.

 

[SolutionRoom]

The Q numbers that you don't yet have listed, I have the following:

Q101, Q102 512-KSC1845UBU
Q103, Q104 512-KSA992FBU
Q105, Q106 ?
Q115, Q116 512-KSA910YBU
Q125, Q126, Q137, Q138, Q139 512-KSC2383YBU
Q140 522-ZTX1056A

 

[SolutionRoom]

You'll notice that the protection circuit transistors are not detailed yet.There are THREE transistors in a row, either 3 2sc2575's OR 2 2sc2575's and s 2sa1100. Which do you have.

Q138 and Q139 are 2SC2575. Q137 has no visible markings on it. Not sure if they wore off, or ??? According to the schematic they are all 2SC2575, but I am aware (from reading another AK post) that newer models have a reversed polarity for this circuit.

Can I tell by the markings for the "E" on the PCB?

The Board is marked "AWH 0970 1AB"

 

[markthefixer]

Q138 and Q139 are 2SC2575. Q137 has no visible markings on it. Not sure if they wore off, or ??? According to the schematic they are all 2SC2575, but I am aware (from reading another AK post) that newer models have a reversed polarity for this circuit.

Can I tell by the markings for the "E" on the PCB?

The Board is marked "AWH 0970 1AB"

Sadly, the AWH-xxx board markings don't help. At least there are no factory modifications? Extra wires, components etc...?

Q139 is the one that would be the 2sa1100.

The ztx1056 didn't work for me, don't use it unless you want to experiment. see the above in-progress post.

You are turning this into a TWO night post. I can't keep up. Stop jumping ahead...

Q115 and Q116 are part of the UNFINISHED POST............

Transistor testing is needed out of circuit in one of this complexity - for good readings -

 

[SolutionRoom]

Sadly, the AWH-xxx board markings don't help. At least there are no factory modifications? Extra wires, components etc...?

Q139 is the one that would be the 2sa1100.

The ztx1056 didn't work for me, don't use it unless you want to experiment. see the above in-progress post.

You are turning this into a TWO night post. I can't keep up. Stop jumping ahead...

Q115 and Q116 are part of the UNFINISHED POST............

Transistor testing is needed out of circuit in one of this complexity - for good readings -

There is no sign of factory mods. No odd jumpers or piggybacked components. And no indication of prior repairs on this board.

Q139 is definitely a 2SC2575, so I suppose that rules out the 2SA1100.

I will not use the ztx1056, and await your recommendations for the replacement.

Sorry 'bout jumping ahead. I just was in the flow of going through it all and wanted to provide the newest info I had. Patience... got it.