I am here once again asking about Sony vfets.
I just got my hands on the very rare Sony TA-F7 (not B) which, according to the previous owner, should have been serviced and should work just fine.
To be fully sure, and before turning it on, I plan to service it completely, from a recap (if needed) to a good clean.
I found mentioned in the forum that it is advisable the precautionary replacement of all the bipolar output transistor in the bjt-vfet cascades, and my question is: is there anything else I must do before powering the amp on, to make sure the vfets will last for decades more?
Thanks
I just got my hands on the very rare Sony TA-F7 (not B) which, according to the previous owner, should have been serviced and should work just fine.
To be fully sure, and before turning it on, I plan to service it completely, from a recap (if needed) to a good clean.
I found mentioned in the forum that it is advisable the precautionary replacement of all the bipolar output transistor in the bjt-vfet cascades, and my question is: is there anything else I must do before powering the amp on, to make sure the vfets will last for decades more?
Thanks
N7 points to look out for:
1) output cascode BJTs WILL fail. (D44H11/D45H11 turned out to be excellent replacements) If they do they can take a VFET with them.
2) all fusible resistors (tend to drift up in value or randomly go open) - replace with modern ones or with comparable power rating metal film or metal oxide (NOT carbon). The point of these was to go open without flaming if a part of the amplifier is fatally damaged and prevent cascading damages. In particular be sure to replace the VFET gate resistors (on the output transistor boards) as if they go open, the original BJTs WILL die and possibly take the VFET with them. The consideration here is more about non-flamability than actual fuse behavior - carbon film resistors DO catch fire, literally, as carbon burns in air.
3) the input dual JFET drifting out of balance (this does not happen often but it's not easy to replace as there are no direct equivalents, requires some machinery and skill to construct a replacement in that case, out of matching singles).
4) The power amps have local front end power supply regulation set to 50-something V and 50V electrolytic caps on the outputs of the regulators, so the caps are working at or just over maximum rating. Replace with 63 or 100V rating caps.
5) the original main power supply filter caps have almost certainly gone below tolerance if the unit has been operating at 115/230VAC, in this case they get very close to maximum rated voltage. Out of the 22000uF usually 18000uF survive but the ESR goes way up - I've done multiple units so far and this seems to be a rule. Replace with 80V or even better 100V rated caps. The simple rule for replacement is: if the capacitance is given with +X, -Y% tolerance, if it measures below -Y% replace it. Also, if ESR becomes 2x nominal (this is the usual spec from the manufacturer - if no data is available, look up similar size and spec caps. For this sice it will be in the 0.0x ohm range!). However this does not include brand new off the shelf caps which can be marginal, because they need to re=form slightly by being exposed to a working voltage, before the capacitance and ESR settle to nominal values. Also, for some cases the simple rule has exceptions, just as any rule - this is what makes it a rule 😛 - and this is where experience and circuit insight comes in (see warning at the end).
6) Get new high quality silicone thermal transfer pads (NO cheap chinese ebay stuff!!!) for all the BJTs and VFETs, and also a little bit of thermal grease. When installing the pads, smear a very small amount of the grease all over both sides (eg. by passing the pad between your fingers having put a very small amount of grease on them - be sure to wash your fingers thoroughly afterwards as this stuff can be toxic). This very slightly increases the efficiency of the pad but it's not the main idea, rather, if for any reason the transistor needs to be taken off the heatsink, it saves the pad. Of course, before re-applying the pads, be sure to thoroughly clean all mating surfaces.
Other than that it's the usual stuff common to all amps of this vintage.
Oh, and surely you meant 'clean up' before 'recap' 🙂
A little trick, since the amp uses wire wrap - do not take te wires off the wire wrap posts, rather de-solder the whole wire wrap post strip from the board, together with wires. The wiring is quite delicate so be careful, but de-soldering the whole thing gives you boards you can comfortably take out, clean and service.
WARNING:
I am assuming you are VERY good at soldering and de-soldering. Proper restoration is not for beginners but for (very) experienced techs, and may well require very good understanding of the operation of some rather advanced circuitry, or one can cause more problems than are solved.
1) output cascode BJTs WILL fail. (D44H11/D45H11 turned out to be excellent replacements) If they do they can take a VFET with them.
2) all fusible resistors (tend to drift up in value or randomly go open) - replace with modern ones or with comparable power rating metal film or metal oxide (NOT carbon). The point of these was to go open without flaming if a part of the amplifier is fatally damaged and prevent cascading damages. In particular be sure to replace the VFET gate resistors (on the output transistor boards) as if they go open, the original BJTs WILL die and possibly take the VFET with them. The consideration here is more about non-flamability than actual fuse behavior - carbon film resistors DO catch fire, literally, as carbon burns in air.
3) the input dual JFET drifting out of balance (this does not happen often but it's not easy to replace as there are no direct equivalents, requires some machinery and skill to construct a replacement in that case, out of matching singles).
4) The power amps have local front end power supply regulation set to 50-something V and 50V electrolytic caps on the outputs of the regulators, so the caps are working at or just over maximum rating. Replace with 63 or 100V rating caps.
5) the original main power supply filter caps have almost certainly gone below tolerance if the unit has been operating at 115/230VAC, in this case they get very close to maximum rated voltage. Out of the 22000uF usually 18000uF survive but the ESR goes way up - I've done multiple units so far and this seems to be a rule. Replace with 80V or even better 100V rated caps. The simple rule for replacement is: if the capacitance is given with +X, -Y% tolerance, if it measures below -Y% replace it. Also, if ESR becomes 2x nominal (this is the usual spec from the manufacturer - if no data is available, look up similar size and spec caps. For this sice it will be in the 0.0x ohm range!). However this does not include brand new off the shelf caps which can be marginal, because they need to re=form slightly by being exposed to a working voltage, before the capacitance and ESR settle to nominal values. Also, for some cases the simple rule has exceptions, just as any rule - this is what makes it a rule 😛 - and this is where experience and circuit insight comes in (see warning at the end).
6) Get new high quality silicone thermal transfer pads (NO cheap chinese ebay stuff!!!) for all the BJTs and VFETs, and also a little bit of thermal grease. When installing the pads, smear a very small amount of the grease all over both sides (eg. by passing the pad between your fingers having put a very small amount of grease on them - be sure to wash your fingers thoroughly afterwards as this stuff can be toxic). This very slightly increases the efficiency of the pad but it's not the main idea, rather, if for any reason the transistor needs to be taken off the heatsink, it saves the pad. Of course, before re-applying the pads, be sure to thoroughly clean all mating surfaces.
Other than that it's the usual stuff common to all amps of this vintage.
Oh, and surely you meant 'clean up' before 'recap' 🙂
A little trick, since the amp uses wire wrap - do not take te wires off the wire wrap posts, rather de-solder the whole wire wrap post strip from the board, together with wires. The wiring is quite delicate so be careful, but de-soldering the whole thing gives you boards you can comfortably take out, clean and service.
WARNING:
I am assuming you are VERY good at soldering and de-soldering. Proper restoration is not for beginners but for (very) experienced techs, and may well require very good understanding of the operation of some rather advanced circuitry, or one can cause more problems than are solved.
Sorry… I went and got a copy of the service manual and ALL of that advice (taken together) sounds like a recipe for introducing the kind of SNAFU you're trying to preëmpt to begin with.
Its like if I have 6 pennies, and a gun.
I toss the pennies in the air 32 times.
If any time I get all heads, I shoot my dog.
Poor dog.
Sure, its ¹/₆₄ chance on any flip that I shoot the dog.
Or it is ⁶³/₆₄ chance that I don't.
But the chance you're going to shoot the dog ANY of the rolls is…
(⁶³/₆₄)³² = 0.60 … leaving a 40% chance I'd shoot the dog.
That's the point. As improbable a harm as any one of the operations above might seem to hold, taken together, they seem to approach certainty to introduce harm, error, mistake somewhere.
Just saying.
GoatGuy
Its like if I have 6 pennies, and a gun.
I toss the pennies in the air 32 times.
If any time I get all heads, I shoot my dog.
Poor dog.
Sure, its ¹/₆₄ chance on any flip that I shoot the dog.
Or it is ⁶³/₆₄ chance that I don't.
But the chance you're going to shoot the dog ANY of the rolls is…
(⁶³/₆₄)³² = 0.60 … leaving a 40% chance I'd shoot the dog.
That's the point. As improbable a harm as any one of the operations above might seem to hold, taken together, they seem to approach certainty to introduce harm, error, mistake somewhere.
Just saying.
GoatGuy
Don't know exactly what your point is. If there is something wrong, you fix it, not completely mess things up, if you know what you are doing. If you don't, you have no business doing it, surely.
The points are given approximately from the most problematic to the least problematic, and it assumes that the owner wants to keep the amp working longer than just pure chance given it's advanced age. I've had a fair number of these on the bench and in fact own(ed) a few. In one case, the output BJTs (8 in case of the TA-F7) failed in 7 out of 8 cases when tested at 10% power on a curve tracer. This is a known problem. So, you might just power up the amp and it may fail in a year or in 10 minutes. Would you want to take the risk? There is only a couple of hundred (and falling) of these on the planet.
And the other answer:
This is why it's a really BAD idea for 'anyone' to just go poking around in these amps. I'm not even going to tell the number of amps in general I had to re-re-build due to people who thought they were competent enough, as in 'I know how to solder and identify components', after they have made 30 solder joints and in some cases were not sure if something was a resistor or a cap. let alone what type. This is NOT like putting together a kit or a new amp, because in the latter case the components are new and solder joints are not there already, so no heat damage, no cracked joints, no corroded switch or relay contacts, no known components that fail invisibly and are likely on their last legs, no mechanical and electrical components that are for all intents and purposes irreplaceable and need to be handled with extreme care and experience, not to mention cases of years of dirt and smoking tar deposits and remains of acidic flux on the PCB, glue that becomes corrosive and eats away components (literally), etc. And if things go wrong, it can take only one component or joint in the wrong place for hundreds of $ to go up in smoke.
For instance, poking around the output stage of one of these might cost you 3 pairs of VFETs, at least $300 damage IF you can find them, and if you do, it's a big unknown as to what you actually got vs what it says on the case. Just about the only way to be certain is using a curve tracer, so do you have one?
Many people attempting rebuilds do not know or understand even the basic stuff and overestimate their competence (see: Danning-Krueger effect, hope I got the spelling right). Like, if you change output and driver transistors, the amp should be re-biassed, starting from the lowest possible setting, meaning you have to set that and check the bias circuit BEFORE you power up the amp with new semiconductors. I've had people use the procedure for starting up completely different circuits with VFETs, which do not apply with this one (in fact will very probably instantly damage the output stage) just because it say's VFET on the front, and the e is unusual - no understanding how it works AND no understanding the service manual, which BTW is a pretty good one.
And then, if there is no damage but the thing does not work, there is a fair amount of knowledge on circuits (and sometimes more than that as these amps do contain oddities and rarities and semiconductors that do not behave the usual way).
AND, there is 0.1% probability of a local repair shop fixing it one you made a mess of it.
Examples:
Just recently I got a TA5650 VFET amp to fix, having been diagnosed by 'experts' with total output stage failure because all output transistors were shorted when measured out of circuit. Which is exactly the way perfectly functioning VFETs measure. The previous shop diligently pulled them all out and threw them into trash. 400$ minimum damage. The guy got a cosmetically challenged parts unit. Guess what, that one also had 'all transistors shorted'. So, that's 800$+ damage. For an amp that sells for less. And one that is almost notorious on the internet, so even a simple google search could have provided clues.
An Italian 'Laboratory' was given a TA-E88 preamp to re-build. Result: trashed ALPS stepped attenuator made ONLY for this amplifier, which is irreplaceable, and likely better than the best attenuators available today, easily a couple hundred $ to replace with something even close to the original. They put in a 20k ALPS blue (with motor... not connected) to hide the deed. Reason: the original has a built-in ball bearing. If you don't know that, the balls will fall out on disassembly. Their size is not standard. Lose one, and you can get away with it. Lose 2, and the wiper will scrape the gold contacts off the ceramic substrate in about 10 turns and it's dead. Guess what happened.
My point: These were top of the line components, as is the TA-N7/F7 and they are not conventional, which means additional traps lurk even for the experienced tech, let alone someone with no experience. And, I am sorry, but experience has shown there are likely problems in there that cannot be diagnosed by a simple multimeter, and they can be very serious. It's just the way it is. Would you rather have me lie?
Last edited: