Heater voltage filtering.

[vmanj]

I welcome everyone!

I made the bias of my microphone (based on the Cad Trion 8000) according to this scheme.
Everything is fine, but there is a hum and noise.

When I turn off the power supply of the microphone, the noise and hum disappear (until the filter capacitors in the power supply are discharged - this is an ordinary Chinese power supply). Its circuit and voltages are in the image.
I understand that the heating voltage is poorly filtered.

How can you filter-smooth the heating voltage well so that remove the hum and noise, maybe use lm317 instead of 7806?

 

[Khron]

Is it a ripple-on-the-heater-rail issue, or some internal ground loop issue? Have you measured the ripple on the heater voltage? Or are you just assuming?

 

[kingkorg]

That is known issue with that topology and that psu. Heater needs to be extremelly well filtered. I doubt you can eliminate the hum without changing psu design. There was ioaudio kit that came with better psu schematic, but i can't find it right now.

 

[vmanj]

Unfortunately, I have nothing to measure the voltage ripple of the heater.

In an auto-bias circuit, this power supply operates without hum or noise.

In the circuit with a fixed bias to the grid (in the u67 version), I redid the voltage of this power supply - I replaced the regulator with 7906, the polarity of the diodes and capacitors.
Everything worked just as well with no hum or noise.

But according to the bias scheme as in u47 (in the image above), there was significant noise and hum, which disappears when the power supply is turned off.
Therefore, I thought that it could be noise and hum from the 7806 regulator, which does not have additional filtering, as in U67 - R9 (60Mohm) and C7 (0.01Uf).

 

[vmanj]

That is known issue with that topology and that psu. Heater needs to be extremelly well filtered. I doubt you can eliminate the hum without changing psu design. There was ioaudio kit that came with better psu schematic, but i can't find it right now.

I would try to remake my power supply (I hope my transformer will allow this)..
I would be very grateful if you give a link to this scheme.

 

[kingkorg]

I can't find it. Ioaudio links are dead in the support thread. Offcourse the noise is comming from the regulators. You have basically no filtering there. The impedance is low, so i guess you would need caps in farad range to smooth it out. You can always try to add some resistance, but you need to experiment.

 

[RuudNL]

I suppose the only way to solve this problem, is raise the voltage of the filament output and try some R-C-R-C filtering. (In every filter stage you will loose some voltage.)
Personally I would skip the idea of fixed bias altogether and use cathode bias...

 

[moamps]

Unfortunately, I have nothing to measure the voltage ripple of the heater.

Does that mean you don't have a DMM? If you have one, set it to measure AC voltage, 200mV, and measure the voltage at the heater power supply output. If the input voltage on 7806 is 13V, then the output voltage should be well filtered at the output. And did you increase the heater voltage for the voltage loss across the cathode resistor?

 

[RuudNL]

I once tried a 78xx regulator in a power supply for a U67, but the ripple was still too high for a 'clean' output.
Every millivolt of noise/ripple is amplified by the tube!
What at the end worked was a LM317 regulator with an output of 10,2 V. and an R-C-R-C filter, where R= 10 ohm and C= 1000 µF. (EF86 takes 200 mA for the filament.)

 

[abbey road d enfer]

How can you filter-smooth the heating voltage well so that remove the hum and noise, maybe use lm317 instead of 7806?

First I would suggest you increase significantly C7, from 1000uF to 4700 or 10 000 uF.
I don't believe replacing the 7806 with a 317 would improve things significantly.
Now I see that the grounding arrangement of the PSU is rather dubious, allowing rectifier current surges to pollute the regulated voltages.
Try lifting the ground connection from the negative of C7.
Both HV and heater voltage should be grounded at the very output.
See suggested mod attached; of course valid only if the actual wiring is in accordance with the schemo.

 

[vmanj]

I suppose the only way to solve this problem, is raise the voltage of the filament output and try some R-C-R-C filtering. (In every filter stage you will loose some voltage.)
Personally I would skip the idea of fixed bias altogether and use cathode bias...

If I can't filter out the heater voltage, I'll have to go back to auto-cathode bias. I wouldn't like it.
I tried different bias, but this one is the best..

 

[vmanj]

And did you increase the heater voltage for the voltage loss across the cathode resistor?

I have not increased it yet, but if I manage to remove the noise and rumble, then maybe I will increase it.

 

[vmanj]

I once tried a 78xx regulator in a power supply for a U67, but the ripple was still too high for a 'clean' output.
Every millivolt of noise/ripple is amplified by the tube!
What at the end worked was a LM317 regulator with an output of 10,2 V. and an R-C-R-C filter, where R= 10 ohm and C= 1000 µF. (EF86 takes 200 mA for the filament.)

According to the U67 scheme, it will not be possible to use a stabilizer 78 .., only reverse polarity 79 ... There should be a minus instead of a plus on the heating voltage.
My 7906 in the u67 scheme worked perfectly, without noticeable noise and hum.

 

[RuudNL]

I don't believe replacing the 7806 with a 317 would improve things significantly.

Well, it doesn't! But the only reason for selecting a LM317 was to make the voltage variable.
Because at the start it is not clear how much voltage you will have to loose in the R-C filters for optimal performance.

 

[RuudNL]

According to the Yu67 scheme, it will not be possible to use a stabilizer 78 .., only reverse polarity 79 ... There should be a minus instead of a plus on the heating voltage.

As long as you use a separate transformer winding, it doesn't matter if you use a positive or negative voltage regulator. In this case the whole stabilisation circuit is 'floating', and you can ground what you want or need.
(By grounding the output of a 78xx regulator, the center pin will become the negative 'output' voltage.)

 

[vmanj]

First I would suggest you increase significantly C7, from 1000uF to 4700 or 10 000 uF.
I don't believe replacing the 7806 with a 317 would improve things significantly.
Now I see that the grounding arrangement of the PSU is rather dubious, allowing rectifier current surges to pollute the regulated voltages.
Try lifting the ground connection from the negative of C7.
Both HV and heater voltage should be grounded at the very output.
See suggested mod attached; of course valid only if the actual wiring is in accordance with the schemo.

I checked, the diagram turned out to be drawn not quite correctly.
In fact, it looks like this.
I'll try to replace c7 with 10000mf, I don't know if this will help to remove the noise of the stabilizer 78 ....

 

[RuudNL]

I don't know if this will help to remove the noise of the stabilizer 78 ....

It won't! The only thing that will help, is add extra filtering after the regulator.
(But in that case you will have to raise the voltage of the regulator, because you loose voltage in the extra filtering stage(s) )

 

[vmanj]

As long as you use a separate transformer winding, it doesn't matter if you use a positive or negative voltage regulator. In this case the whole stabilisation circuit is 'floating', and you can ground what you want or need.
(By grounding the output of a 78xx regulator, the center pin will become the negative 'output' voltage.)

Interesting, I did not know this, I had to redo the entire circuit for 7906 (in the image).
If I understand you correctly, then you can make a negative heating voltage at 7806?

 

[vmanj]

It won't! The only thing that will help, is add extra filtering after the regulator.
(But in that case you will have to raise the voltage of the regulator, because you loose voltage in the extra filtering stage(s) )

That is, it can look like this (conditionally)?

 

[RuudNL]

C8 isn't that important, because the output impedance of the regulator is very low. (Near zero ohms.)
The capacitors after the first and second added resistor will make a big(ger) difference.

Here is an example how how can use a 'positive' regulator for a 'negative' output voltage.