Hi Sony VFET amp experts, I am rebuilding my Sony ta 4650 VFET amp. After replacing the old capacitors , zener diode 1T43M with with a 7.5v zener diode , replacing all other diodes I am getting +-53vdc on the 2 large 47,000uf/50v caps instead of -+45vdc what might be the cause of this high voltage , your kind help is greatly appreciated . Thanks
Hi , sorry I wrote one more zero the 2 large main caps are 4700uf/50v the number written on this zener diode which look like a regular small black transistor with middle lead cut is what I wrote in my message 1T43M I think it might be a 7.5v Zener diode Sony does not mention how in service manual what voltage it is. The other two 20vdc and 98vdc going to pre amp boards are correct and can be adjusted by the new 2.2k bourn pot on the power supply board. The whole unit pre amps boards and power amp board and power supply board had been recapped . I am getting source voltage of 53 and 51 at source of the power amp board the gate voltage can be adjusted by the two 2.2k pots on the power amp board from 50 to 75 volts . Do you think it might be okay to install VFETs with 68ohm 5 w resistors in + and - b supply to source and adjust to pots so gate voltage is 66-68 range and source voltage is 53-50 voltage range ? I don’t know why the source voltage at power supply in +- 53 vdc B1 instead of of 45-48vdc range this voltage does not change even if disconnect the power amp board from power supply board .
I have never done that. I measure the V-fets if they are ok. From drain to source a few ohms, the gate to source only to test as a diode: neg on G, plus on S for the N-ch, reversed fo the P-ch. If all is ok, engage with all fingers and toes crossed.
The gate voltages must exceed the source voltages no matter what!
Without the fets installed, turn RT301 to the highest gate voltages ('blocking mode').
On the supply board, have you replaced D403 with the 'dual-diode' modification?
Have you replaced the suicide diodes (blue ellips)? There are three in the 4650. Must do.
I've also replaced C306 & C307 with non-pol caps.
The main board draws only a small amount of current from the supply without the fets.
Still curious those very high main voltages - the bias current is not that large with the fets installed.
Attachments
Hi citizen, yes all the 3 death diodes on have been replaced with twin 1N4158 didoes also the D403 on power supply board replaced with double didoes and the center of these two diodes soldered to the common ground of the power supply . I checked the secondary of the power transformer by disconnecting is from the power supply board to see what was the voltages the secondary was producing . The voltage between black wire and one of the two red wire measures a little more 39 vac . The other voltage from this same black wire to the other red wire is the same 39 voltage. Removing all the small didoes on power supply board but leaving in place the large 4 diodes which are new 1N5406 rated at 3A 400v each the voltages jump to -B1 =-53vdc and +B1=+53vdc. All diodes , all caps , resistors and transistors on power supply are including all new smaller diodes which are 1N4005. The Zener diode 7.5v .
The reason for installing each 68ohms/5w resistors in the +-53vdc line is to restrict the current through the VFET to protect them from self destructing in case there is a fault in amp.
I did what you suggested I adjusted the blue new multi turn 2.2k 1/2w bourn bias trimmers for maximum gate voltage at the VFETs . I have not yet installed the VFETs my present readings of gate voltages is +-74vdc and source voltage measures +- 53 and +—49vdc these readings are with black lead of digital fluke true rms 189 meter connected to the common ground of the power supply and the red lead connected to 4 gate and source pins. The voltage between source and gate can be adjusted from +-12vdc to +-24dc with meter black and red leads on gate and source pins . The amp produces beautiful music through the headphones even without VFETs installed but I want to hear the VFET magic through speakers . Do think it safe to install the VFETs now ?! Thanks for your kind help! It is greatly appreciated!
The reason for installing each 68ohms/5w resistors in the +-53vdc line is to restrict the current through the VFET to protect them from self destructing in case there is a fault in amp.
I did what you suggested I adjusted the blue new multi turn 2.2k 1/2w bourn bias trimmers for maximum gate voltage at the VFETs . I have not yet installed the VFETs my present readings of gate voltages is +-74vdc and source voltage measures +- 53 and +—49vdc these readings are with black lead of digital fluke true rms 189 meter connected to the common ground of the power supply and the red lead connected to 4 gate and source pins. The voltage between source and gate can be adjusted from +-12vdc to +-24dc with meter black and red leads on gate and source pins . The amp produces beautiful music through the headphones even without VFETs installed but I want to hear the VFET magic through speakers . Do think it safe to install the VFETs now ?! Thanks for your kind help! It is greatly appreciated!
Pretty much all signs on green now.
Did you do the C306/307 replacement (#5) also?
It's to early to disclose my 'Qube-Vfet replacement option to test those amps - working on it to publish as a paper or a publication.
Lots of formulas to solve, eg "Y/Yo = (-q/µ -g)/Xo = (q/µ +g)/-Xo". µ is the wanted triode amplification factor. This took me this week to disclose, but now it's logic how it works - for triodes, V-fet's and my qube circuit. There's nothing in paper libraries, not existant on the web; low pace.
Just in case your fets incinerate, I've a 4650 overmodified but with two good pairs to donate. For p&p only, though.
And the 747, queen of the skies... (I'm an old man).
Thanks so much for the green light on VFETs 👍 will install them tommorow will be running one channel only as the other side has a missing 2SJ18 it went kaput from previous attempt to fix the 4560 was not familiar then what I was doing lol but not have educated myself quite a bit but by no mean an expert . That is very kind of you to donate the pair I am simply thrilled and thank you a million for donating it to me !👍😊 would love to have them that will be perfect I can then listen the ta 4650 in stereo please send me the cost of p&p and your PayPal info in pm and I will PayPal the payment . Regarding the 747 yes love the plane! The best Boeing ever made! Except Lufthansa no one is flying it anymore , thanks again! 😊👍
Finally installed VFETs wow! Great music ! There is one problem I cannot adjust the bias set my fluke 189 true rms digital on auto MVdc setting the bias reads just 0.5vdc and does not change when you rotate the bias trimmer rt 301 . So I left it at max voltage setting +-77vdc gate voltage and +-53vdc source voltage . The sound is crystal clear the VFETs hardly get hot , but still wondering why the bias is not adjusting , your kind help as always is greatly appreciated!
I installed left side VFETs on the right side I could not adjust the dc bias on the right either, yes that would be awesome
As the dutch bulletin is hardly readable for non-dutch, There is readable info though.
It is password-protected and an image, so ocr does not help. I'll try to make a transcription & post.
It is password-protected and an image, so ocr does not help. I'll try to make a transcription & post.
Yes I wish we could get it translated in English 🙂 after I got the voltages correct +- 77-78 on the gates and +-53 at max voltage settings of the RT301 and RT 351 and installed the VFETs and then connected the fluke 189 meter which is quality move capable of reading both dc and ac voltages at the same time, + of the meter to +of speaker +A of the left terminal the - of the meter to the bias adjust point on left to part of the front of the power board near the RT 301 vishay (yes these are made by vishay not bourn) 2k 1/2 w trimmer no signal vol turned all way down but speaker A selected on the front panel, I get 0.5mvdc reading on the meter, adjusting the trimmer does not change the mvdc reading but when when you connect a source and turn a up the volume with music playing you can see the reading change to any number of different mvdc setting like 35vdc,12 according to the signal intensity . The music is crystal clear no distortion and last night I ran it 3hrs the VFETs never became warmer than 80degF the temp being very consistent! I wondering just leave it like that any concentrate listen to the music or you think it will sound even better when I am able to adjust to Sony recommended 45mvdc at idle no signal instead of 0.5vdc at idle. I can’t figure out why?! The bias voltage is not adjustable to 45mvdc when rotating the rt 301 bias adjust trimmer at idle zero volume !
Currently halfway on page 3/3. I'll post the original text first and then the adjusted (as good as I can) english version after dragging it through G-translate. It's even for me already very archaic dutch, although I'm raised in that time!
Also the punctation will be adjusted because we tend to pronounce long sentences without breathing meanwhile so nowhere comma's. (qed)
Very curious...
Sorry there were lot of typo in the other message thanks to apple word correction lol🙂. It should have +-77 to 78vdc on the gates and +- 53vdc on source when bias adjust trimmers are moved to full end of their travel which increase to voltage to max also nearly zero bias ! With VFET installed the meter shows 0.5mvdc idle current with zero signal when you try to adjust idle current to Sony recommended 45mvdc connecting the meter as described in the service manual and trying to adjust rt 301 to increase the bias current to 45mvdc nothing happens! This is happening on both left and right sides. When you connect a source to listen you can see the bias current value jumping to all different values like 35mvdc . 25vdc, 12mvdc or any value in between according to the change in volume level and signal frequency .
We'll address that nasty bias issue later...
The format of the tables will be changed on diyaudio, so instead the plain textfiles attached.
Sony_TA-4650_5650_TAN-5550_ServBull_NL.txt
Sony_TA-4650_5650_TAN-5550_ServBull_EN.txt
(typed with notepad, no auto correction or spell checks, minor adjustments after G-translate)
The format of the tables will be changed on diyaudio, so instead the plain textfiles attached.
Sony_TA-4650_5650_TAN-5550_ServBull_NL.txt
Sony_TA-4650_5650_TAN-5550_ServBull_EN.txt
(typed with notepad, no auto correction or spell checks, minor adjustments after G-translate)
Just found time to read the translation. You are genius ! And an asset to this community !No one before has bothered to translate it English as it might have been very difficult for them! Not for you!👍😊
the four 1.5 k resistors on power board have already been changed to new 3.3k values before so this is not the issue with bias adjust problem
How many turns do you think I need to turn rt301 bias adjust 2.2k trimmer to the right clockwise direction for it present max turned to left position , before I can expect the meter to show the bias voltage changing from present 0.6vdc. I counted the turn on the vishay 2.2k 1/2w trimmer it takes about 24 turns from one end to the other
The bias range is (roughly calculated) between 1.0~1.3Vdc. But this bias does not set the current through the V-fets as in regular AB-circuits. The driver circuits (Q307~310 / Q357~360) relate the bias to the gate supply rails in a rather unusual way. Hence it is rather difficult to 'predict' the proper setting. Without the V-fets installed, one should see changing the gate voltages while turning the bias-pot.
Do you have an early US version or a later US-E-AEP version of the 4650? I have not found when the changes were made with serial numbers.
Changes are with C303/353, C306/307, C356/357 and C420. And the +78V vs +100V supply rail for the pre-amp section.
Attached some enhanced drawings for the 4650.
The board layout has corrected blue colours, the grounds are made green.
Do you have an early US version or a later US-E-AEP version of the 4650? I have not found when the changes were made with serial numbers.
Changes are with C303/353, C306/307, C356/357 and C420. And the +78V vs +100V supply rail for the pre-amp section.
Attached some enhanced drawings for the 4650.
The board layout has corrected blue colours, the grounds are made green.
