TA-4650 Power Supply Problems
After recapping the boards, I started powering up the amp, first without VFETs.
+20 and +78 Volt outputs, as well the +/- 43Volt rails for the VFETS.
However the +/- 73 Volts for the BIAS were not symmetrical. Voltage shifted to one side…, e.g. it became +100 / -60.
I found also others reporting in forums, who had the same problem: unsymmetrical bias-voltage.
After long searching and testing in wrong places, I found following:
The voltage doubler in the Sony power supply is a charge pump. It is not regulated. Any non-symmetrical load on it’s +/- outputs results in the observed voltage shifts.
I suspected a fault first, but confirmed through circuit simulation in Spice: It happens also in the simulation – so it is ‘feature’ of the supply.
Two questions arose:
- What in the pre-amp board causes the not-symmetrical load ?
- Can I change the supply to a regulated one – which would be also safer for the survival of VFETs during potential next 30 years of life of this amp…
I searched some time to find good and matching replacement connectors for the on-board, partially oxidized 1970ies connectors. I found a good replacement: Hirose DF5, which is a drop-in replacement due to it’s 5mm spacing.
TA-4650 non symmetrical load
Started to investigate why there is non symmetrical load on the +/- bias voltage supply, which causes the sensitive power supply to derail. My suspect where the bias transistors Q307-Q310 (channel 1) and Q357-Q360 (channel 2).
Removed, measured all of them:
Channel 2 seems quite ok. If you look at the amps schematic: Q357 and Q358 take care of upper half, connected to +75V, while Q359 and Q360 take care of lower half, connected to -75V.
Both current and Hfe do match very nice:
Upper half: Q357+Q358 = 5,18 + 6,91 = 12,15
Lower half: Q359+Q360 = -7,77 – 4,68 = -12,45
But in channel 1 there are more significant differences in Ic and Hfe.
At the moment I think that this could be the cause of load differences on positive and negative rail of the power supply. I will replace the original transistors with new, matched replacement types…let’s see how it goes….
Tested my changed transistors today. As a result the initial symmetry of the supplies +/-75Volt is much better: It only changes much slower to unsymmetry as before…
which could mean it takes a few minutes longer now until the VFET’s get destroyed…..(if I had them in the amp….which I haven’t, yet).