Where is my voltage going?

[KatieandDad]

I'm currently building an Impasse Pre-Amplifier, I know that is tube based but the PSU is SS.

My question is - "Where are my volts going ?"

All the voltages below are with the PSU loaded at its designed 40mA.

The transformer is a 650VAC CT. And at the anodes of D101 and D102 I do indeed have 700VAC.

At the cathodes of D101 and D102 I only have 400VDC, I would expect 455VDC.

R103 is 470R and across it I have 18.8V (correct).

R104 is 1K and across it I have 40V (correct).

So why is the rectified 650V so low ?

R101 and R102 are small 47R resistors.

D101 and D102 are uF4007.

 

[KatieandDad]

Anyone with any ideas ?

I've checked D101, D102, R101 and R102, all are OK.

And the PSU PCB is like this:-

diyAudio

diyAudio

 

[Mooly]

Its because C101 is specified as 1uf. Use a scope to measure the voltages, not a dvm

 

[KatieandDad]

I've still only got 340V at the junction of R104 and the input to the regulator ?

There is no way that the regulator output could ever be 350V.

 

[Mooly]

C101 at 1uf means the ripple voltage is huge at that point. I haven't calculated anything, it just doesn't look right.

Just try adding a suitable electroylitic across it and see what that does.

 

[Mooly]

Lets prove it then...

V3 with C101 at 1uf comes out at 340volts (which is what you get)

V3 with C101 at 100uf comes out at 384 volts

 

[KatieandDad]

So its working then. Funny thing is that due to the tolerance of R108 (5%) my output voltage is only 325V.

If R108 was at the other end of its tolerance band, how would the regulator be able to regulate ?

 

[Mooly]

R108 sets the output voltage of the reg. Variations in tolerance just affect the exact regulator output voltage... by how much... you would have to look at the data sheet for the IC to get the relationship of R108 and R109 (which set the voltage) and then work out upper and lower output voltages for worst case tolerances).

No competent design should depend so exactly on normal tolerances that it makes the difference between it working and not.

If you use LTspice I can post the file and then you can change component values and see the effect on the voltages.

 

[KatieandDad]

Unfortunately I haven't got LTSpice.

But am I right in thinking that the regulator couldn't work if Vi(reg) was below Vo(reg).

Whether by luck or judgement mine is working because Vout is lower than designed.

I believe the Impasse was designed by SY.

 

[Mooly]

The reg needs some overhead to work properly, usually around at least 5 volts and that must include any ripple "effects" too. If Vin is less than the calculated regulator voltage then the output will just be low and follow the input voltage less a few volts. In that state it doesn't regulate. A design like this is very dependent on "absolutes". If the current drawn is a bit higher or lower than specified then the input voltage to reg varies a lot because its such a high impedance feed on the input side.

 

[Mooly]

Altering the 1k would allow adjustment of the voltages. Reducing it down in increments until you had a decent differential over the reg. Perhaps 680 ohms.

(If you wanted LTspice IV its free and installs in seconds,
Linear Technology - Design Simulation and Device Models)