AC heaters ultimate hum limit

[abbey road d enfer]

Different makes of output transformer are more or less susceptible to hum fields.

Indeed.
Amongst many other factors, there is a definite correlation between leakage inductance (or coupling coefficient) and susceptibility to external fields. Leakage inductance is a consequence of a fraction of the winding-generated flux escaping the magnetic core; the reciprocal is how much external field is permitted to enter the core.
As I pointed out earlier, DCr is a factor; the larger the core, the less turns and the lower the DCr. It may not matter much when driven by a pentode's plate, though.
Obviously, magnetic core permittivity is also a factor, the higher it is, the lower the turns and the lower the leakage.
The geometric orientation of the core is also to be taken into account.

 

[kambo]

just curiosity :
if u recalculated your design and set to max gain to (lets say) 50db,
would u still have the same level of hum at max gain of 50db?

or is hum related to db gain, or related to max gain via pot?

did u ever measure hum only on one half of your mic pre at max gain?

edit : since you are using two identical amps, may be there is a way to cancel out the hum more efficiently!

 

[ruffrecords]

just curiosity :
if u recalculated your design and set to max gain to (lets say) 50db,
would u still have the same level of hum at max gain of 50db?

or is hum related to db gain, or related to max gain via pot?

Hum is definitely related to gain. If you turn the gain pot right down to zero there is still some hum from the second stage.

Cheers

Ian

 

[kambo]

Ian, any update on ur new boards!

 

[ruffrecords]

Ian, any update on ur new boards!

They arrived a while ago and I am in the process of building one right now. When it is finished I will use it to replace the one in the second prototype (the one whose measurments I posted in thsi thread). This should provide a direct measurement of any improvement.

Cheers

Ian

 

[ruffrecords]

Before testing the new Lo Hum version of the PCB I decided to run one last test on the existing design. It has been mentioned that the HT rectifier load can distort the waveform of the AC heater winding of the mains transformer. To test this I measured the noise and hum spectrum using the internal mains transformer and then with the heaters supplied from a completely separate external transformer. To my surprise there was virtually no difference. I can only conclude this is not a factor in the current design.

I then tried disconnecting the the HT and then the heater secondary while the unit was running. Again I was surprised at the results. Disconnecting the HT causes the 50Hz component to disappear but the 100Hz component remains. Leaving the HT winding connected and disconnecting the heaters causes the 100Hz component to disappear. Both are the exact opposite of what I expected. To make sure this was not some odd artefact of the transformer, I repeated the tests using the heater winding of an external transformer. The results were identical.

Anyone have an explanation for this?

Cheers

Ian

 

[ricardo]

I then tried disconnecting the the HT and then the heater secondary while the unit was running. Again I was surprised at the results. Disconnecting the HT causes the 50Hz component to disappear but the 100Hz component remains. Leaving the HT winding connected and disconnecting the heaters causes the 100Hz component to disappear. Both are the exact opposite of what I expected. To make sure this was not some odd artefact of the transformer, I repeated the tests using the heater winding of an external transformer. The results were identical.

Is this with a DC heater?

 

[ruffrecords]

I then tried disconnecting the the HT and then the heater secondary while the unit was running. Again I was surprised at the results. Disconnecting the HT causes the 50Hz component to disappear but the 100Hz component remains. Leaving the HT winding connected and disconnecting the heaters causes the 100Hz component to disappear. Both are the exact opposite of what I expected. To make sure this was not some odd artefact of the transformer, I repeated the tests using the heater winding of an external transformer. The results were identical.

Is this with a DC heater?

All with AC heater,

Cheers

Ian

 

[trobbins]

That indicates mains or secondary AC parasitic capacitance coupling to low level signal circuitry I guess.

Was there still a galvanic connection of heaters to ground/0V when the heater supply was disconnected?

 

[ruffrecords]

That indicates mains or secondary AC parasitic capacitance coupling to low level signal circuitry I guess.

That could be it. The HT AC connector is near the input tube and the HT AC is 240V rms. Does not need much stray C to couple that across. I might try an earthed screen between them.

Was there still a galvanic connection of heaters to ground/0V when the heater supply was disconnected?

Yes, via the heater elevation circuit.

Cheers

Ian

 

[ruffrecords]

I have just completed building the first Classic Solo Lo Hum PCB. I swapped it with a Classic Solo V2 in the prototype I have been using to make all the previous measurements and used the same tubes from that PCB. So this test is with the same mains and output transformer in the same enclosure with the same wiring. First I measured the noise via REW and my test rig with a 600 ohm load. REW reading was -62.1dBFS. SInce 0dBu on the REW is -6.5 dBFs, this equates to an actual signal level of -55.6dBu.

Next I measured the gain by feeding -60dBu in from my Lindos test set, engaging the 20dB pad in my test box and noting the signal level in REW. The REW reading was -17.3dBFs which corresponds to an actual level of -10.8dBu. Since the 20dB pad was in, the actual level at the preamp output was 20dB higher i.e. +9.2dBu. This means the total gain is 69.2dB. Since the noise is -55.6dBu then the EIN i -55.6 - 69.2 = -124.8dBu which is a very good result for a tube mic pre with AC heaters.

This is nearly 3dB better than the V2 PCB it replaces.

You can clearly see the 50Hz component reduce as you adjust the hum dinger pot but there is no minimum. I tried altering the hum dinger with the gain pot at zero so the only signal is from the output tube. The hum reduces by turning the pot in the opposite direction as expected because I had tracked the heaters of the output tube so they were the opposite of the input tube, the idea being that the hum in the input tube would be inverted relative to the output tube and so at least partially cancel.

100Hz seems unaffected which implies a different mechanism causes this.

Anyway, I am pleased to have gained a 3dB improvement.

Cheers

Ian

 

[ruffrecords]

Most lf the Lo Hum PCbs are now with beta testers. For their benefir I have created a new version fo the assembly coumentation. It is in the Classic Solo folder of the DIY tab of my web site. Here is a direct link:

https://drive.google.com/file/d/0B_n67A1hN3qtYVVDTGZCVE5EZmM/view?usp=sharing

Cheers

Ian

 

[CJ]

while turning the hum pot on a fender bassman the other day we noticed that if we reduce the line voltage a bit, the hum comes back, and we have to adjust the pot again, granted, this is a push pull circuit, so it might not make a difference in a SE circuit,  but you might try playing with heater voltage to see if the hum pot will dial out the hum at slightly different heater voltages,

 

[ruffrecords]

while turning the hum pot on a fender bassman the other day we noticed that if we reduce the line voltage a bit, the hum comes back, and we have to adjust the pot again, granted, this is a push pull circuit, so it might not make a difference in a SE circuit,  but you might try playing with heater voltage to see if the hum pot will dial out the hum at slightly different heater voltages,

Hmmm, there are so many factors at play in this I suppose I should not be surprised at this effect. I guess this means hum level can vary from day to day, minute to minute. Have to make sure my wife turns off the dish washer when I'm recording.

Cheers

Ian

 

[JTraverse]

Have you tried any DC offset?

 

[Matt Nolan]

Have you tried any DC offset?

Good idea. Heater elevation can reduce induced noise. See Morgan Jones.

 

[ruffrecords]

Have you tried any DC offset?

Good idea. Heater elevation can reduce induced noise. See Morgan Jones.

Heaters have always  been elevated due to using mu follower stages.

Cheers

Ian

 

[trobbins]

What elevated DC voltage are you using, and what voltages do the cathodes sit at, or swing to, compared to the elevation level?

The heater ends will swing about the elevated voltage.  Those swing limits should be at least 15-20Vdc away from a cathode swing limit, in order for the heater-cathode voltage to always be more than 15-20V different, and hence push the heater cathode resistance in to a high resistance region.  I'm guessing the upper triode's cathode will swing through the elevated voltage level?

The elevated voltage should be well decoupled at 100Hz to the 0V node of the mu-follower, so that ripple on the elevated voltage is not a contributor.  That could be checked by using a battery to power the heater, but still keep the humdinger and elevated supply connected to the heater, compared to using a separate battery for the lower and upper valves (and connecting the battery to the local cathode).

The mu-follower config also allows any capacitive coupling from heater to cathode on the upper triode to go straight in to the output signal.  Keeping the heater-cathode resistance of that triode high is also likely to be the main aim as resistive current will also go straight to the output signal - which would indicate that the elevated voltage should not be too high (ie. not too close to the upper triode cathode).

 

[ruffrecords]

What elevated DC voltage are you using, and what voltages do the cathodes sit at, or swing to, compared to the elevation level?

Heater elevation voltage is 25% of HT (typically 75V). This is ideal for SRPP stages where the top cathode is close to 50% of HT. The top cathode of a mu follower is typically 25V higher. The 6922 I use in these stages has a Vhk of 200V.

Cheers

Ian

 

[Whoops]

There was someone that got upset with the noisy Heaters

http://166.63.127.229/Forum/index.php?topic=13378.0

More extreme would be to do it in MuMetal.