Based on my transformer, I boosted the regulator voltage by only 3 volts.
Will 2 filters after the regulator be enough to reduce hum and noise, or is it better to sacrifice the regulator and add another filter?
Heater voltage filtering.
- Thread starter vmanj
- Start date
Help Support GroupDIY Audio Forum:
Will 2 filters after the regulator be enough to reduce hum and noise, or is it better to sacrifice the regulator and add another filter?
This is very interesting, but in this version I will have to cut the tracks of the printed circuit board.
Here you need to choose what is easier to do.
RuudNL
Well-known member
Datasheets recommend a capacitor at the output of the regulator to prevent oscillation, but this capacitor can have a much lower value than 1000 µF! Usually 10 µF is sufficient to prevent oscillation.
It depends on the output voltage of the transformer if the regulator will still work for higher output voltages.
Usually a minimum voltage of 3 volts between inut and output (under load!) is recommended for good regulation, personally I prefer even en bit more. I suppose that with a 9 V. regulator you may have to reduce the value of the added filter resistors a bit, depending on the filament voltage and current of the tube.
I would never use the schematic below, because there is no regulation at all.
It depends on the output voltage of the transformer if the regulator will still work for higher output voltages.
Usually a minimum voltage of 3 volts between inut and output (under load!) is recommended for good regulation, personally I prefer even en bit more. I suppose that with a 9 V. regulator you may have to reduce the value of the added filter resistors a bit, depending on the filament voltage and current of the tube.
I would never use the schematic below, because there is no regulation at all.
Last edited:
I've used a common mode inductor placed after the regulator along with more large filter caps to get noise down an extra bit , Im still not sure if its quiet enough in this application where cathode and heater are connected .
There are lower noise regulators out there but they tend to be pricey .
https://www.diyinhk.com/shop/audio-...e-dac-power-supply-regulator-3357v-14ax2.html
https://www.ebay.ie/itm/133984803783requires an SMD resistor to set the voltage .
https://www.ebay.ie/itm/263374806928Theres another one with adjustable output voltage . Should both have less than 1uV of noise
There are lower noise regulators out there but they tend to be pricey .
https://www.diyinhk.com/shop/audio-...e-dac-power-supply-regulator-3357v-14ax2.html
https://www.ebay.ie/itm/133984803783requires an SMD resistor to set the voltage .
https://www.ebay.ie/itm/263374806928Theres another one with adjustable output voltage . Should both have less than 1uV of noise
Last edited:
Analog_Fan
Well-known member
Whould the
do the job?
| LM2576 |
do the job?
RuudNL
Well-known member
I fully understand the principle, but what is the function of the 12K resistor? (Base to ground.)
As per the article,
"The 12k resistor shown may need to be adjusted to suit your transistors and supply voltage. Reducing the value increases dissipation in the output devices and lowers output voltage. It is unlikely that any benefit will be obtained by increasing this resistor, but you may experience increased hum (hardly a benefit)."
Without it in place, the voltage drop across the capacitance multiplier would be minimal, but so would the ripple reduction. With that added resistor from base to ground, you can reduce the output voltage further, and also get the ripple amplitude lower. Couldn't explain exactly why and how though
"The 12k resistor shown may need to be adjusted to suit your transistors and supply voltage. Reducing the value increases dissipation in the output devices and lowers output voltage. It is unlikely that any benefit will be obtained by increasing this resistor, but you may experience increased hum (hardly a benefit)."
Without it in place, the voltage drop across the capacitance multiplier would be minimal, but so would the ripple reduction. With that added resistor from base to ground, you can reduce the output voltage further, and also get the ripple amplitude lower. Couldn't explain exactly why and how though
Any experiments Ive done with low level stages and directly connected heater/cathode indicated a very quiet supply was required, as Ruud rightly pointed out any noise at the cathode is subject to the full gain of the tube and any subsequent added gain The standard regulated DC tube heater supply doesnt get near quiet enough .
As RuudNL mentioned in post #23 the regulator needs to be bypassed to prevent oscillation. C7 and C8 are large value capacitors, intended to filter out power line frequencies, i.e. 50 or 60 Hz and the first few harmonics. Since the oscillation of the 7806 regulator can be a very high frequency, very small capacitors are also needed in addition to the large filtering capacitors C7 and C8.
To prevent oscillation, a very small capacitor (0.1uf) should be placed as close to the regulator as possible, from the output pin to the ground pin. Similarly, another small capacitor (0.33uf) should be placed as close to the regulator as possible, from the input pin to the ground pin.
This bypass technique is shown on page 4 of the L7806CV datasheet from STMicroelectronics.
https://www.mouser.com/datasheet/2/389/cd00000444-1795274.pdf
In addition, I would recommend bypassing all eight diodes (D1-4 and D7-10) in the bridge rectifier circuits with 0.01uf capacitors, one capacitor in parallel with each of the diodes, placed as close to the diodes as possible. This will help reduce the noise floor because the switching transients from these diodes create some very high frequency hash, which sometimes sounds like power line noise, but other times can sound like tape hiss or buzzing. This simple fix made a big improvement in my unregulated U47 PSU.
To prevent oscillation, a very small capacitor (0.1uf) should be placed as close to the regulator as possible, from the output pin to the ground pin. Similarly, another small capacitor (0.33uf) should be placed as close to the regulator as possible, from the input pin to the ground pin.
This bypass technique is shown on page 4 of the L7806CV datasheet from STMicroelectronics.
https://www.mouser.com/datasheet/2/389/cd00000444-1795274.pdf
In addition, I would recommend bypassing all eight diodes (D1-4 and D7-10) in the bridge rectifier circuits with 0.01uf capacitors, one capacitor in parallel with each of the diodes, placed as close to the diodes as possible. This will help reduce the noise floor because the switching transients from these diodes create some very high frequency hash, which sometimes sounds like power line noise, but other times can sound like tape hiss or buzzing. This simple fix made a big improvement in my unregulated U47 PSU.
Have you tried removing the ground wire on the South West of the schemo?
RuudNL
Well-known member
IMHO it is nothing more than an R-C(-R-C) filter, followed by a darlington.Without it in place, the voltage drop across the capacitance multiplier would be minimal, but so would the ripple reduction. With that added resistor from base to ground, you can reduce the output voltage further, and also get the ripple amplitude lower. Couldn't explain exactly why and how though
I suppose that more current through the filter would mean less ripple rejection...
(In Fig 2 in the same article, it seems there is no need for a base-to-ground resistor.)
moamps
Well-known member
If the ripple voltage at the input of the multiplier is high, without this resistor the pass transistor leaves the active mode and the ripple rejection is not good enough. By adding a base resistor, the output voltage decreases, the pass transistor runs all the time in active mode so the ripple at the output of the circuit is smaller. The load of the last base capacitor with this resistor is negligible.
Rempember than in order to operate with consistent Hfe, a transistor needs enough Vce. In particular, at the negative ripple peaks, this voltage is quite low.
dmp
Well-known member
Like the other posts said, you need more filtering. A psu for cathode bias mics will not be enough for fixed bias.
Look at this schematic for the EF-47:
https://cdn.imagearchive.com/groupdiy/u/39511/58d1402a0714e.pdfThis starts with 20VAC and uses a choke to reduce the voltage drop through the filtering network. Passive design with the voltage set by a pot. No regulator.
Fixed bias is worth building but it is more difficult.
Look at this schematic for the EF-47:
https://cdn.imagearchive.com/groupdiy/u/39511/58d1402a0714e.pdfThis starts with 20VAC and uses a choke to reduce the voltage drop through the filtering network. Passive design with the voltage set by a pot. No regulator.
Fixed bias is worth building but it is more difficult.
Like this ?
Attachments
Looking what fixed bias.
I did a fixed offset as in the u67 circuit, everything worked great with the 7906 regulator ..
But with the bias option as in u47, I still can’t, I’ll probably have to remove the regulator, replace the transformer with a higher voltage for heating, and do multi-stage (5-6) filtering .. something like in the circuit you gave the link to .
The drawing shows that the South-West is disconnected, as I suggested.
Yes.
