5
\$\begingroup\$

Power supply for headphone tube amp gives me 10-12V lower voltage comparing to the schematic (my measurements are in red for amp under load). Current under load is about 70 mA.

power supply schematic

Using simple formula : 150V * 1.414 - 2V = 210V also gives approx voltage shown in the schematic.

I even tried to do simple simulation by using www.partsim.com, and it also gives about 170V for B+ and about 210V on the first filter capacitor.

simulation

I don't know what is wrong. I've triple checked all diodes, capacitors and resistors. They are all fine, and are connected correctly. I am still missing 10-12V.

Edit : transformer data enter image description here

\$\endgroup\$
13
  • \$\begingroup\$ WhAt is AC ripple on each stage? \$\endgroup\$ Apr 3, 2021 at 14:31
  • \$\begingroup\$ I've used two different multimeters, both shows same voltages. It is sinusoidal source, transformer primary is connected to the mains (240V/50Hz). Transformer is toroidal with 2 secondary windings of 150V, which I connected in parallel (antekinc.com/content/AS-05TC150.pdf). \$\endgroup\$
    – buggy08
    Apr 3, 2021 at 14:43
  • \$\begingroup\$ @TonyStewartSunnyskyguyEE75 : is it possible to measure AC ripple with multimeter ? \$\endgroup\$
    – buggy08
    Apr 3, 2021 at 14:44
  • \$\begingroup\$ Yes. But your filter is too lossy . What DC tinyurl.co.m/ykyl5drt . drop and AC ripple would you like in mV? \$\endgroup\$ Apr 3, 2021 at 14:59
  • \$\begingroup\$ Multimeter shows AC voltage 0.53V on the first capacitor, 0.03V on the second one, and 0.00V on the third one. \$\endgroup\$
    – buggy08
    Apr 3, 2021 at 15:00

4 Answers 4

1
\$\begingroup\$

The mains waveform these days is 'flat-topped' due to non-linear loads. This means that the peak voltage (which has a big influence on the diode current) is less than sqrt(2) times the RMS. Since your amplifier is working well, and a good design would be insensitive to a reduction of only 10 V in the supply voltage, I think you can stop worrying and enjoy the music.

\$\endgroup\$
1
\$\begingroup\$

The lower voltages you are seeing are to be expected. Your transformer is incorrectly specified for your application.

All the specs for that transformer are assuming it is driving a resistive load.

A bridge rectifier with large amounts of capacitance on the other end is about as far from a resistive load as one can get.

The conduction angle is much less than the full wave - it only conducts while the wave exceeds the filter capacitance voltage. The capacitor also presents a low impedance load to the transformer. Essentially, these effects conspire to force the transformer to supply the same power but compressed into just part of each AC cycle.

In other words, the peak current draw on the secondary winding is going to be a lot higher than you would otherwise expect. It doesn’t draw this current for that long, so the average current will be the what you expect, but your transformer is small enough that the high peak currents are having a large impact on the output voltage of the secondary.

When using transformers with rectifier circuits, one must always derate the current and voltage spec of their transformer to account for the nonlinear behavior this introduces.

If the peaks currents exceed the rated current and your core goes into saturation, you’ll get a lower average output voltage than you expect as a result. And unless you’re using a true rms multimeter to measure the AC output voltage of the transformer, you may measure a higher AC voltage on the secondary than is actually there.

So, long story short: get slightly bigger transformer or one with more voltage headroom.

\$\endgroup\$
0
\$\begingroup\$

I will suggest an RLC filter than will only lose a couple volts. KVL indicates too much loss in DC for 80 dB 100Hz ripple attenuation. If your amp has some PSRR in the preamp you may compute that or measure it and reduce all R’s to 2.2 Ohms for now. When you come back with specs for DC drop and AC ripple, then a better sol’n is possible like an LDO or active filter.

\$\endgroup\$
3
  • \$\begingroup\$ This power supply schematics is not my own, but is part of kit for headphone tube amp already sold in probably thousands of copies. As it is, it already sounds very good for the money. I have almost zero knowledge about electronics, so unfortunately I don't know to answer to your questions about ripple specs. All I know that I need abt. 170V DC at the output of this PS, and I currently have 159V. I've already completed this amplifier, and it sounds perfectly well, it is dead silent, even when potentiometer is turned to max (if input is not connected). \$\endgroup\$
    – buggy08
    Apr 4, 2021 at 6:03
  • 1
    \$\begingroup\$ Also there is no room in my enclosure to accommodate chokes instead of resistors. Even though I am aware this PS is not perfect, and there exists better designs, as you proposed, I am not sure I can change original power supply on the way you proposed. My only intention was to get 170V, and to try to understand why I am having 10V lower voltage then schematics shows. Yesterday I tried to put additional 270R resistor in parallel with the first one, and I've got 168V at the output. That means that if I replace both 270R with 200R resistors, I could get about 170V at the output. \$\endgroup\$
    – buggy08
    Apr 4, 2021 at 6:10
  • \$\begingroup\$ I have also measured power supply voltages without load (I took out tubes), and voltages and transformer secondary was about 155V AC in that case, and capacitors were on 209/206/203 V DC respectively, I guess it would be expected to have higher then 209V after rectifier, if AC is 155V ? At this condition (no load) there is probably no saturation at transformer, and secondary could be considered as fully sinusoidal and MM readings might be considered correct ? \$\endgroup\$
    – buggy08
    Apr 4, 2021 at 6:22
0
\$\begingroup\$

Your DC out is down compared to what you expect. The average DC volts on the first filter cap will be down because the transformer does have source impedance which will be mainly resistive. The diode current is peaky so it does not take too much resistance to drop the average DC volts. Put some resistance in series with the generator on the sim like say 47 ohm and you will lose more than 4 VDC. Now you can measure the normalized transformer source resistance and things will show much less discrepancy.

\$\endgroup\$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.