• The move to the new server is done. There are some software and database maintenance updates in process. This has us passing the hat around to help out. We appreciate any donations. Seriously, even a dollar helps. The payment page may be found here - https://www.audiokarma.org/support.html

Sony TA-4650 help needed

A

Armandoooooh

New Member
Dear all,

I inherited a TA-4650 from my dad which had been idle for around 15 years. I used it with great enjoyment for more or less a year before noticing one of the channels was sounding more silent than the other. The difference wasn’t huge, but it was perceptible. Given the reputation of the TA-4650 as a kamikaze machine, I decided to put it aside for repair. After a few months I finally decided to start. I replaced the death diodes by 2x 1N4148 in series and replaced the 1.5kOhm resistors with 3.3kOhm as per the service bulletin. Luckily the death diodes I replaced still measured fine, so these hadn’t failed yet. I also removed the vfets to measure them. The results seem okay, except for one 2SJ18 wich measured slightly out of spec: between G and S and G and D i measured 25kOhm at 60kOhm range (spec in service bulletin states it should be >30kOhm), and when performing a diode test it was a bit lower than the rest at 0.45V. Other direction was infinite as expected, resistance between source and drain 2.1Ohm. The other VFets for comparison had an R_GS and R_GD of 33-42kOhm and 0.51-0.52V with the diode test. Is the one 2SJ18 screwed or does this seem acceptable?
Now I tried to set the DC bias (without the Vfets installed) but got -0.1mV between the test point and the +channel of the speaker A output for both channels. I had measured the DC bias on an earlier date (with Vfets installed) and here i noticed that one side I could measure and set the bias fine, but the other side I measured 0.0 mV no matter the potentiometer position. What could be the cause of the bias being zero? Does this indicate a faulty Vfet? What are the next steps I can try?
Thanks in advance!
 
Register to hide this ad
Test the trim pot. Is it one of those open type where you can see part of the resistive strip? I've had problems with that type in the past. They either went open, or wouldn't hold bias or DC offset properly. Check your work for cold solder joints, solder bridges etc. also.
 
Locate all of the trimpots. If you have any Caig FaderLube (I use the liquid version) put a little bit of the FaderLube on the end of a toothpick (or use the spray version) and apply it to each of the trimpots. Let the stuff absorb for ten minutes or so and then remove the excess with a Q-tip. You do not need to rotate the pots and risk changing any of the settings.

My ST-A6B recently got this treatment. I was amazed at how the FaderLube treatment made it work almost like a new tuner.
 
Thanks for the reply! I don’t have that exact product but I do have some contact spray meant for deoxidising electrical contacts. I can try that. I did measure the trimpots and their value seemed to check out, and slightly turning them did change their value.
 
Okay so I did some more measurements. I found a lot of the bjt’s on the boards had black legs, while others were fine. It were consistently the same transistors for the right and left channel so I suspect it depends on the type. I have read somewhere that bjt’s manufactured in the 70’s in japan sometimes have this issue. Took one of them out but it measured fine, so don’t think that’s the issue. I powered on the amp and measured different power lines. All were within a few volts of the spec, except the +75V line which was at +90V. Anybody have an idea what could be causing this?
 
Armandoooooh said:
Okay so I did some more measurements. I found a lot of the bjt’s on the boards had black legs, while others were fine. It were consistently the same transistors for the right and left channel so I suspect it depends on the type. I have read somewhere that bjt’s manufactured in the 70’s in japan sometimes have this issue. Took one of them out but it measured fine, so don’t think that’s the issue. I powered on the amp and measured different power lines. All were within a few volts of the spec, except the +75V line which was at +90V. Anybody have an idea what could be causing this?
Click to expand...
If your bias voltage reads zero volts, your protection circuit might be doing its job.
What is your DC offset on both channels?
Measure at the amp board before the protection relays.
Also, I would give little weight to any “ out of spec” vfet measurement you do with the diode test.
To test “specs” on those transistors, you need a capable curve tracer, not that small DCA pro
 
if you lost ground at the input section you would see a similar behavior, but it would be on both channels so I would rule that out, but still, makes sure those connectors are clean. The connectors Sony used on the first gen vfet amps and others, are a disaster waiting to happen.
 
so an update, as I said previously the + and - 75V are not balanced, sometimes the + rail is too high, sometimes it’s the - rail. I measured the gate voltages of the VFETS and they were varying a every time i measured them. Could go from +30-60V and -30-60V. This worries me because this would mean in some cases the gate voltages voltage was not high enough to pinch off the vfets which could have smoked them. This measurement was without vfets installed so the voltage might be different under load, but not too sure. I replaced D403 by a double diode with ground in the middle to balance the power lines, seems to have worked but only had 1 minute of testing before I had te leave. Gate voltages seemed much better and symmetrical as well now at +-59V after turning the bias resistors. More testing is needed before they go back in the circuit, I’ll do that this weekend.
 
There is a mod to do on the PS board to even out the rails using an added diode, If I remember correctly. It is in my thread from years ago . Ecowars helped me in the repair.

Let me see if I can find it.

Here it is, it wasn't in my thread but in another members thread

J

Post in thread 'Sony TA-4650 Restoration'

Connect one more diode in series with D403 and the common point between these to ground.This will make the bias supply immune to load imbalance.

This design flaw cost lot of 4650's lives.Last week I repaired a 4650,imported from japan as junk, have same problem. VFETs failed, but lucky to have some spares.
  • Like

EDIT: I see you did the added diode mod. :thumbsup:

Athanasios
 
Last edited:
+V jumping, makes me think in a ground problem, and I see it was already mentioned. Please check and fix the connection between pcb ground and the chassis , at every point you see a screw o connection from the board to chassis.
 
Okay so it's been a while, but here's an update. After verifying that all voltages were good and stable, I decided to plug the vfets back in. New ceramic thermal pads with silicone thermal paste. Installed them, turned on the power and..... it worked! Kind of. I still saw a difference in output amplitude between both channels when applying a 1kHz 1Vrms sine wave at the aux2 input. Investigating further, I saw this amplitude difference was already present at the output of the preamp. Turning the balance knob until both preamp outputs had equal amplitude resulted in a near perfect Amplifier output. This is great news, as it means there's nothing wrong with the amplifier section and precious VFETs. I checked whether changing the volume would change anything, and low and behold again the preamp output would slightly shift and have unequal amplitudes. Turning the balance until preamp amplitude was equal again resulted in perfect output from the amplifier.
So my current guess would be that the volume potentiometer is no longer properly balanced and needs a thorough cleaning or replacement. Somebody has an idea of a drop-in replacement I could use?
 
You must log in or register to reply here.
Back
Top Bottom