Hi Mooly.
Managed to find a site which has photos of the same SMPS which some other guy was working on (you can see the state of the blue caps in some shots!). From the underside of the board I can see C2023s, so I know this is a UK model (the US/Canadian version uses 2SC1986). Thankfully on the topside I can see Q602/Q604 look absolutely identical (dark grey body, very thin metal tab, angled at the edges) to the ones on my board, not to mention that circular Sony branded Q603. So this means Sony definitely used C1810's for those two positions and that these parts do not appear to have been replaced on my SMPS board. So we know that part number was an error in the service manual.
Here's the site: Picasa Web Albums - 895720 - SONY TA-F6B PLPS
You can JUST about see Q602 in the upper-right corner in this picture from the above album, and if you squint can just about make out the C18... so I'm pretty much happy that's sorted that one out:
http://picasaweb.google.com/895720/SONYTAF6BPLPS?feat=directlink#5325950347770120706
Is it perhaps best to desolder these C1810's, then run a multimeter check on them, and reuse if OK, or would you go ahead and use new MJE340s anyway?
Can also see R622 in those pictures and it's exactly the same markings as mine, so that's also a relief - I'll measure it again tomorrow once I've desoldered a leg and if OK I'll retain that.
Managed to find a site which has photos of the same SMPS which some other guy was working on (you can see the state of the blue caps in some shots!). From the underside of the board I can see C2023s, so I know this is a UK model (the US/Canadian version uses 2SC1986). Thankfully on the topside I can see Q602/Q604 look absolutely identical (dark grey body, very thin metal tab, angled at the edges) to the ones on my board, not to mention that circular Sony branded Q603. So this means Sony definitely used C1810's for those two positions and that these parts do not appear to have been replaced on my SMPS board. So we know that part number was an error in the service manual.
Here's the site: Picasa Web Albums - 895720 - SONY TA-F6B PLPS
You can JUST about see Q602 in the upper-right corner in this picture from the above album, and if you squint can just about make out the C18... so I'm pretty much happy that's sorted that one out:
http://picasaweb.google.com/895720/SONYTAF6BPLPS?feat=directlink#5325950347770120706
Is it perhaps best to desolder these C1810's, then run a multimeter check on them, and reuse if OK, or would you go ahead and use new MJE340s anyway?
Can also see R622 in those pictures and it's exactly the same markings as mine, so that's also a relief - I'll measure it again tomorrow once I've desoldered a leg and if OK I'll retain that.
Last edited:
They don't make 'em like that anymore 
It's always good to test the devices because it pieces all the evidence together. Even if in the end we fit new, we know just what got zapped and that helps determining how far back into the circuit problems may have occured.
R622 is very important and when the transistors fail it puts a short on the main rail with all the current flowing directly through that resistor. The volt drop across that resistor in normal operation determines the overload and current limiting of the switching transistors. It is a critical part value wise.
C606 the 0.47uf could be an odd one to get... you don't see that value nowadays in elecrolytics... not common anyway.
It's always good to test the devices because it pieces all the evidence together. Even if in the end we fit new, we know just what got zapped and that helps determining how far back into the circuit problems may have occured.
R622 is very important and when the transistors fail it puts a short on the main rail with all the current flowing directly through that resistor. The volt drop across that resistor in normal operation determines the overload and current limiting of the switching transistors. It is a critical part value wise.
C606 the 0.47uf could be an odd one to get... you don't see that value nowadays in elecrolytics... not common anyway.
Tomorrow I'm going to remove all the transistors we've mentioned thus far, and clean up the PCB in the process so it's as clean as possible. I'll probably have to extend a few wires when I solder in new components as there are a few lifted traces there by the looks of things.
I'll test that resistor, as well as Q603, 603 & 604. Keeping them crossed I'll be able to reuse all those, thereby keeping some of the circuit working to its original design.
I'll order 6 BUT11s for the remaining transistors (I won't bother with the MJs as the BUTs seem better spec'd).
Yes that 0.47uF is a pain as Farnell have no 105C versions. There are some chinese 105C ones on Ebay, but not brands I recognise. I may go for a Cerafine or perhaps a Sanyo Oscon - something that uses less than the usual amount of electrolite or none at all for safe long-term high-temp running. Hope they are low enough ESR for that particular position. Thinking about it perhaps I could use a film cap there instead like a 470nF polyester? The other caps are taken care of, but still awaiting one from Farnell for C613 which won't arrive until Feb (pain in the ***!). That's probably for the best as it ensures I won't be rushing to get it all up and running as soon as possible
Thanks ever so for all your help again today Mooly - I owe you a beer or three!!!
- John
I'll test that resistor, as well as Q603, 603 & 604. Keeping them crossed I'll be able to reuse all those, thereby keeping some of the circuit working to its original design.
I'll order 6 BUT11s for the remaining transistors (I won't bother with the MJs as the BUTs seem better spec'd).
Yes that 0.47uF is a pain as Farnell have no 105C versions. There are some chinese 105C ones on Ebay, but not brands I recognise. I may go for a Cerafine or perhaps a Sanyo Oscon - something that uses less than the usual amount of electrolite or none at all for safe long-term high-temp running. Hope they are low enough ESR for that particular position. Thinking about it perhaps I could use a film cap there instead like a 470nF polyester? The other caps are taken care of, but still awaiting one from Farnell for C613 which won't arrive until Feb (pain in the ***!). That's probably for the best as it ensures I won't be rushing to get it all up and running as soon as possible
Thanks ever so for all your help again today Mooly - I owe you a beer or three!!!
- John
Last edited:
Wattage of the pot... you can calculate it.
250vdc across the resistor chain of 100k + 1k + 2k2 gives a current of 250/ (100000+1000+2200) which is 0.00242amps. Just ohms law of I=V/R
That gives a volt drop across the pot of 0.00242*1000 which is 2.42 volts.
And watts is I*V which is 0.00242*2.42 giving 0.0058watts or around 6 milliwatts. So no problem.
I would think a film small poly cap would be OK for the 0.47uf. Sometimes the "poor" (by todays standards) charactersistics of original specced components can be an advantage in certain circuits where those characteristics are part of the design. That won't apply to this cap though but it certainly can to semiconductors.
C613, I nearly mentioned the value of that as 33uf is a value not perhaps so common these days. A 47uf would be fine here.
If a part doesn't fit such as lead spacing on that 0.47 cap for instance, then it's best if possible to drill a new hole for one of the legs so it connects to the original print.
C801 and C805 (the 1000uf 200 volt caps) are in series on 240vac mains. The 220k resistors across each should be checked. These serve only to keep the voltage across each cap the same as any difference in capacitance and leakage current would cause a huge imbalance and cause one cap to see an overvoltage. It's an old trick.
250vdc across the resistor chain of 100k + 1k + 2k2 gives a current of 250/ (100000+1000+2200) which is 0.00242amps. Just ohms law of I=V/R
That gives a volt drop across the pot of 0.00242*1000 which is 2.42 volts.
And watts is I*V which is 0.00242*2.42 giving 0.0058watts or around 6 milliwatts. So no problem.
I would think a film small poly cap would be OK for the 0.47uf. Sometimes the "poor" (by todays standards) charactersistics of original specced components can be an advantage in certain circuits where those characteristics are part of the design. That won't apply to this cap though but it certainly can to semiconductors.
C613, I nearly mentioned the value of that as 33uf is a value not perhaps so common these days. A 47uf would be fine here.
If a part doesn't fit such as lead spacing on that 0.47 cap for instance, then it's best if possible to drill a new hole for one of the legs so it connects to the original print.
C801 and C805 (the 1000uf 200 volt caps) are in series on 240vac mains. The 220k resistors across each should be checked. These serve only to keep the voltage across each cap the same as any difference in capacitance and leakage current would cause a huge imbalance and cause one cap to see an overvoltage. It's an old trick.
Haha thanks - get the popcorn ready!
It was your superb thread over at AK that got me interested in buying a TA-F6B in the first place
"A 47uf would be fine here."
Ahaa that's good news - Farnell told me my 33uF 400V cap won't ship until April due to not being in stock - ridiculous time to wait so I've cancelled that part. It's good news in that I have a Panasonic 47uF 450V in the parts box which is less than 6 months old. Only thing is there's a very slight dent in the casing which is why I never used it before - assume that dent (not severe) won't effect the way it works? I may just order a new one to be on the safe side though.
- John
Ahaa that's good news - Farnell told me my 33uF 400V cap won't ship until April due to not being in stock - ridiculous time to wait so I've cancelled that part. It's good news in that I have a Panasonic 47uF 450V in the parts box which is less than 6 months old. Only thing is there's a very slight dent in the casing which is why I never used it before - assume that dent (not severe) won't effect the way it works? I may just order a new one to be on the safe side though.
- John
Last edited:
"I would think a film small poly cap would be OK for the 0.47uf"
I'm wondering why they used an electrolytic here anyways as I'm sure small film caps of the same value were available back in '78. And in the high temp. environment of an SMPS I would have thought a film cap more reliable too...
Odds bodkins!
I'm wondering why they used an electrolytic here anyways as I'm sure small film caps of the same value were available back in '78. And in the high temp. environment of an SMPS I would have thought a film cap more reliable too...
Odds bodkins!
Hi John hope your feeling a bit better.
The Wima cap should be OK. Used to see a lot of small electros in consumer gear years ago, sometimes down to 0.1uf, all down to size and cost I guess.
Seen and worked with loads of dented caps over the years, never a problem.
C604 is another important one to change, 22uf 400 or 450 volt.
Just better mention (I'm sure you know this, others may not) that these PSU's are not isolated from the mains on the part we are working on. It's always at half mains potential when plugged in so never connect any grounded test equipment or even touch any of the components as they are all live irrespective of the polarity of the live and neutral.
The Wima cap should be OK. Used to see a lot of small electros in consumer gear years ago, sometimes down to 0.1uf, all down to size and cost I guess.
Seen and worked with loads of dented caps over the years, never a problem.
C604 is another important one to change, 22uf 400 or 450 volt.
Just better mention (I'm sure you know this, others may not) that these PSU's are not isolated from the mains on the part we are working on. It's always at half mains potential when plugged in so never connect any grounded test equipment or even touch any of the components as they are all live irrespective of the polarity of the live and neutral.
Also, when testing, remember to that the ground connection is made by the box being screwed to the chassis and if you want to troubleshoot while it's out of the chassis, make sure to use an isolation transformer (better safe than sorry) and also to make that ground connect with an aligator clip from the main PLPS heatsink to the main amp chassis.
Thanks Mooly Been suffering from an auto-immune illness for the past 3 years. 'tis a sod always seems to strike just when I want to get started on something - feels like severe jetlag, and then some! Grrrrr! Hopefully tomorrow or Tuesday I'll be able to think straight & get stuck in again. All parts ordered now anyways.
thefragger: Thanks for the reminder - yes it's in the manual so I'm going to get a couple of croc clips this week and make up a cable... I'll probably still switch it on wearing rubber gloves and a long stick hahaha (My tongue is firmly stuck in cheek folks. No I usually just wear a frogmans suit....! ). I need to get an isolation transformer for future projects like this one - unfortunately the finances are low at present.
I have a rather large favour to ask you, and I COMPLETELY understand if you decline - no problem at all. But if you ever have the lid off your TA-F6B again I could really do with knowing the dimensions of that switching transistor clamp on the underside of the PLPS. I've taken a guess at the dimensions by looking at various pictures and applying that to the PLPS in front of me, but it's always nice to have the precise dimensions. They again it doesn't look critical by any means. AFAICS I guess even a rectangular one would do if need be as it's just to press the transistors against their heatsinks embedded into the PCB side. From your pictures over at AK I guess it's pressed steel, though I think 1mm thick aluminum should work just as well (crosses fingers!).
Cheers.
- John
Been suffering from an auto-immune illness for the past 3 years. 'tis a sod always seems to strike just when I want to get started on something - feels like severe jetlag, and then some! Grrrrr! Hopefully tomorrow or Tuesday I'll be able to think straight & get stuck in again. All parts ordered now anyways.thefragger: Thanks for the reminder - yes it's in the manual so I'm going to get a couple of croc clips this week and make up a cable... I'll probably still switch it on wearing rubber gloves and a long stick hahaha (My tongue is firmly stuck in cheek folks. No I usually just wear a frogmans suit....!
). I need to get an isolation transformer for future projects like this one - unfortunately the finances are low at present.I have a rather large favour to ask you, and I COMPLETELY understand if you decline - no problem at all. But if you ever have the lid off your TA-F6B again I could really do with knowing the dimensions of that switching transistor clamp on the underside of the PLPS. I've taken a guess at the dimensions by looking at various pictures and applying that to the PLPS in front of me, but it's always nice to have the precise dimensions. They again it doesn't look critical by any means. AFAICS I guess even a rectangular one would do if need be as it's just to press the transistors against their heatsinks embedded into the PCB side. From your pictures over at AK I guess it's pressed steel, though I think 1mm thick aluminum should work just as well (crosses fingers!).
Cheers.
- John
Last edited:
The heatsink arrangements puzzle me a little... just because I can't see the whole unit.
Is it a common heatsink for all four with a "rubber/neoprene"type insulating pad isolating each device. From the photo it looks like the transistors sit on a plate of some kind inset in the pcb. Is that then isolated ? Picture of the top of pcb might be useful. The missing plate almost certainly is steel and as such allows it to be tightened to the point of "bending" such that it applies firm pressure to the transistors. There may well have been a pad of some sort here too.
I know I mentioned BUT11AF's at the start. These are 100% plastic and as such could be secured via a screw direct into the heatsink with no isolation problems if it came to that.
Is it a common heatsink for all four with a "rubber/neoprene"type insulating pad isolating each device. From the photo it looks like the transistors sit on a plate of some kind inset in the pcb. Is that then isolated ? Picture of the top of pcb might be useful. The missing plate almost certainly is steel and as such allows it to be tightened to the point of "bending" such that it applies firm pressure to the transistors. There may well have been a pad of some sort here too.
I know I mentioned BUT11AF's at the start. These are 100% plastic and as such could be secured via a screw direct into the heatsink with no isolation problems if it came to that.
You have it sussed - there's a large silpad on the top side of the PCB, which is underneath a cast alloy mini chassis, upon which the transformer and - I think - a choke are mounted.
Underneath, there are four holes, which have the silpad exposed. Into each of these holes is placed a mini-heatsink. The switching transistors are pushed gently into contact with the heatsinks/silpad/chassis via that F-shaped clamp. I would have thought a 1 or 2mm thick clamp fashioned from aluminium would be just as effective. There's nothing on the underside of the clamp either - it makes contact with the bodies of the transistors directly.
- John
P.S. I ordered the BUT11As which I believe aren't insulated, so the mounting arrangements will be the same as the stock items. Parts arriving tomorrow
Underneath, there are four holes, which have the silpad exposed. Into each of these holes is placed a mini-heatsink. The switching transistors are pushed gently into contact with the heatsinks/silpad/chassis via that F-shaped clamp. I would have thought a 1 or 2mm thick clamp fashioned from aluminium would be just as effective. There's nothing on the underside of the clamp either - it makes contact with the bodies of the transistors directly.
- John
P.S. I ordered the BUT11As which I believe aren't insulated, so the mounting arrangements will be the same as the stock items. Parts arriving tomorrow
As long as the transistors are in good contact with the heatsink aluminium should be OK and it doesn't need much pressure over the face of the transistors so as long as the aluminium holds it's form over time. Those transistors should run fairly cool anyway, that's the beauty of the SMPS. They are either fully on or fully off so power losses in the device are small... in theory It's the other two that will get hot as more current is drawn so they do need to be in good thermal contact with the heatsink.
For initial testing I would rig up a 60 or 100 watt bulb... solder wires to the terminals, then remove the mains fuse (f801) and solder the bulb across the holder.
It's the other two that will get hot as more current is drawn so they do need to be in good thermal contact with the heatsink.For initial testing I would rig up a 60 or 100 watt bulb... solder wires to the terminals, then remove the mains fuse (f801) and solder the bulb across the holder.
Right - so you mean solder wires directly to the terminals on a light bulb? And one is replacing the F801 fuse with the bulb, and then turning on as normal. And leaving F802 in place?
I guess first step after that is to then set the voltage on the SMPS to 98-99V as per page 9 of the service manual. Assume this is with no source or speakers connected - the manual doesn't appear to indicate this.
I guess first step after that is to then set the voltage on the SMPS to 98-99V as per page 9 of the service manual. Assume this is with no source or speakers connected - the manual doesn't appear to indicate this.