SONY TA-4650 first power on

spaghetti

Active Member
Hi there, I've just finished refurbishing a Sony TA-4650 VFET amp. First thing I did was remove the VFETs, they all tested fine on the DMM, so I went on replacing the suicide diodes, making the PSU voltage modification in order to center the B+/B- values, and replacing four 1.5K ohm resistors with 3.3K as per service bulletin.

Right now the VFETs are still outside of the amp, I am a little nervous about putting them back and I would like some reassurance that everything looks fine.

These are the voltages at the VFETs sockets:

Q312 2SK60
G -60V
S -44V​
Q311 2SJ18
G 60V
S 44V​

Q361 2SJ18
G 58V
S 44V​
Q362 2SK60
G -58V
S -44V​

All drains are 0V. The G-S voltage difference is always about 14V, I think it can be modified by setting RT301 and RT351, and I will have to adjust those trimmers later in order to set the bias correctly.

Can anyone confirm me it's ok to put the VFETs back into the amp and power it on?

Thanks
 
Last edited:
Don't have the SM for this amp, neither ever had a SONY VFET amp. But i remember the information not to increase voltage with a variac for a first start. The VFETs would like to see directly their supposed voltages. Maybe this helps.
 
True... do not use a variac with vfets.

Think you should be reading xx.x digits on your dmm no two. I'm used to seeing perhaps a one decimal point difference or perhaps one volt but not two volts. Might try another closer ground source to the circuit and find a pcb screw mount instead of chassis. Also check said chassis screws all over. That can make a difference..
 
Last edited:
True... do not use a variac with vfets.

Think you should be reading xx.x digits on your dmm no two. I'm used to seeing perhaps a one decimal point difference or perhaps one volt but not two volts. Might try another closer ground source to the circuit and find a pcb screw mount instead of chassis. Also check said chassis screws all over. That can make a difference..

Yes my DMM reads 3 digits, one decimal, but I rounded the voltage values to integer. I will repeat measurements again if they need to be exact.

Of course I am not going to use any DLBT/variac once the VFETs are in place. Actually it might be better to put the DLBT in a drawer, just to be on the safe side ;)
 
Not yet.

Since I got no explicit "go ahead" signal in this thread yet, I am reviewing all the threads on AK regarding the TA-4650 in order to extract as much info as possible that may be relevant to my case.

I found this sentence:

I measured the S and G of the vfet terminals according to joshvi ("Make sure that the voltage between gate and source of all vfets are between 8-18 volts.") and they were ~+/-16v and holding.

which seems to confirm that the 14V~16V deltas in my amp should be good numbers.

Another sentence I've found suggests that I should increase the delta to the max value before inserting the VFETs, and reduce it later when adjusting the bias:

Before putting the VFET back, you have to vary bias preset and measure the Vgs. It will vary between 10v to 22v approx. ( K60 Vgs is -ve and J18 Vgs is +ve. Keep the preset at highest Vgs ( so no bias current will flow through VFETs when switched on)

I am away from home for a few days, so I won't be able to work on the amp until next wednesday. But I have plenty of time to scrape more info from the old threads :)
 
Thanks for posting. 16v holding is max number and probably good as said steady.

Got to aim and fire sooner or later. Allow a good low idle time to stable the fets. no inputs etc. to monitor bias.
 
.... I am reviewing all the threads on AK regarding the TA-4650 in order to extract as much info as possible that may be relevant to my case.

....

Perfect. Glad to see someone else doing this :)

You get one chance to get this right. Don't rush yourself...

It would be great if EW or Oilmaster or one of the others well versed in VFETs would chime in. But I am confident you will get it on your own.
 
Since Sony itself in the Service Bulletin 62 advises to "set the dc bias adjustment resistor on the power amp board for maximum reading", I've just set the trim pots for max readings, which happens to be between 20~21V. Others reported higher voltages, about 24V.

Now, what happens is that at power on the S-G voltage fluctuates between 12-30V, then after a few seconds it stabilizes to the above values (~21V).
Same happens when I turn off the amp. Is that normal? Since the voltages are greater than 8V I suppose this won't damage the VFETs once they're back.
 
Wish I could help.

I believe that EchoWars hasn't been around much recently. I know that Oilmaster knows his way around VFET amps...I PM'd him a link to this thread. I have no idea what he is up to currently, or whether he will be checking messages anytime soon.
 
@roger2 Thanks for the support, I appreciate it :) Today I checked all the voltages in the power amp board again, and they're within specs. I think it's about time to put the transistors back and turn it on, there seems to be no reason for not doing it. Maybe during the weekend I will make up my mind... I'll keep you updated of course
 
@roger2 Thanks for the support, I appreciate it :) Today I checked all the voltages in the power amp board again, and they're within specs. I think it's about time to put the transistors back and turn it on, there seems to be no reason for not doing it. Maybe during the weekend I will make up my mind... I'll keep you updated of course

Hello Spaghetti,
May be a little to late but i would strongly advice to power up first time after each repair with a resistor R68 5W in each supply line 45Vdc. This is sufficient for bias setting and measurements with low volume and you cannot blow up precious Vfets!
 
Hello Spaghetti,
May be a little to late but i would strongly advice to power up first time after each repair with a resistor R68 5W in each supply line 45Vdc. This is sufficient for bias setting and measurements with low volume and you cannot blow up precious Vfets!

Hi, no I've not powered it up yet, I set the amp aside temporarily to complete other easier projects. What do you mean with R68? Do you mean 68ohm resistor? Should it be placed between B+ and S of Q311 and Q361 output transistors?
 
Now, what happens is that at power on the S-G voltage fluctuates between 12-30V, then after a few seconds it stabilizes to the above values (~21V).
Is this happening on all supplies to all VFETs?

Do a thorough search on this site - I recall one experienced AK'er did a rebuild of this amp or the 5650 and he had fluctuating voltages UNTIL he reinserted the VFETS. Unfortunately you wont know if this is your issue until you risk reinserting the VFETs, which I would be cagey of.
 
Reply #12, 13, 14: May be a little to late but i would strongly advice to power up first time after each repair with a resistor R68 5W in each supply line 45Vdc. This is sufficient for bias setting and measurements with low volume and you cannot blow up precious Vfets!
Hi,
a) Indeed always, after every mod, temporary protect Vfets with 2 current limiting resistors aprox 43 á 68 Ohm 5W in both + and - 43VDC supply lines to the amplifier board in order to effectively limit max current through Vfets. Without temporary resistor and with a fault the current could be as high as ~90A (43V / 0,47R) . Vfets with very small S/D resistance can only have 5A max each according spec! Try-out without protective resistors is like Russian Roulette i discovered to my regret.
The by some repairman advised 12V halogen lamps instead of resistors is not advisable, since resistance of cold lamp is very low and in the approx 200 msecs time for heating up the lamp wolfram filament, the thin silver lead wires inside the Vfet to the multi-fet dye can be molten away. Mostly this is the failure with a defective Vfet.
PS: When difficult to obtain, several small resistors in parallel can be used to make a 5W resistor.
b) Fluctuating voltages: In several TA5650 i noticed normal stabilizing fluctuation after switch-on and switch-off during approx. 6sec with Vfets in place. Protective circuit with speaker relay will protect speakers and Vfets for heavy currents during stabilizing. I dont know fluctuations without Vfets, not tried! When fluctuations occur with shorted signal input AFTER stabilizing time, something is wrong. Main errors are the known unreliable faulty double diodes and faulty signal input circuits.
Succes!
 
Last edited:
Is this happening on all supplies to all VFETs?

Do a thorough search on this site - I recall one experienced AK'er did a rebuild of this amp or the 5650 and he had fluctuating voltages UNTIL he reinserted the VFETS. Unfortunately you wont know if this is your issue until you risk reinserting the VFETs, which I would be cagey of.

I don't know about experienced, but that might be my rebuild you are thinking of. TA-5650 but similar beast. OP, please tread with caution. Just because my unit was only stable with the VFETs in, doesn't mean it is a common thing.
 
Today it was VFET day...not a good day unfortunately. I found one of the 2SK60 that I stored away months ago to be defective: it reads about 70 ohm GS and vice versa. So, at the moment I could not complete the amp even if everything else was OK.... which doesn't seem to be the case.

I added 2x33 ohm (~ 66 total) protection resistors on each VFET rail (+45, -45V), as suggested by @RealWorld , inserted a transistor pair in the left channel, powered it on, no relay click. The resistor would get very hot fast.
Same happened when I moved the transistors to the right channel. Luckily the transistors are still good, no damaged occurred.
 
I refurbished one a couple of years ago thanks to help from members of this board. I'm not certain where you are right now with yours but from my own refurb thread, these are the readings I got on mine and was told they were good, were safe to reinstall the VFETs and power on.


-B1: -45.7V
-B2: -91.4
+B2: 91.2
+B1: 45.2
B3: 95.5
B4: 20V

60 Source: -58.8V
60 Gate: -45V
18 Source: 45V
18 Gate: 56.1V

18 Source: 44.8V
18 Gate: 61V
60 Source: -44.8V
60 Gate: -64.4V

Everyone seems to see greater voltage than 75V on B2 until the VFETs are installed, correct? If not, I'll list a few more readings, and mark with asterisks those which differ from what the schematic indicates:

Q45 E: -44V
*Q402 E 95.6V
Q401 E: 8.15V
Q410 E: 19.91V
*R405 96.8V, 99.5V (each side)
R402 132.6V, 99.5V (each side)
R319 60.4V, 77.3V (each side)
R369 55.6V, 77.2V (each side)
R330 64V
R380 58.1V
R413 78.7V, 81.9V (each side)
 
ref:"powered it on, no relay click. The resistor would get very hot fast"
Very hot temporary protection resistors is more or less normal due to current flow. Measure voltage over resistors and you know the current and disipated power. Should be approx. the bias current 50 ma.(50?) If considerately more, than vfets are conducting (open) . Measure carefully gate voltage of both Vfets Should be high enough to close Vfet (+ and- 60V?). Turn the bias pot for adjustment. Attention: When one Vfet is conducting, the other will also conduct in order to compensate! Note-1: Bias voltage in normal operation 75(?) V is always higher without Vfet-consumption and goes to max. 90 -100V.
 
Back
Top Bottom