AC heaters ultimate hum limit

[kambo]

close

 

[kambo]

distances

 

[ruffrecords]

Now I am beginning to recover from my recent operation I have spent a couple of hours testing the local/remote mains transformer scenario. This is using the PCB available on the emporium and configured for 6.3VAC heaters. The spectrum of the box as built is attached. You can see the 50Hz competent peaks to nearly -60dBu.

Ian

 

[ruffrecords]

Then I disconnected the transformer secondaries and wired in another transformer located just over 90cm away from the preamp. The spectrum obtained is attached. As you can see, it is a couple of dB better from which I conclude that the hum we are seeing in not principally from the transformer but more probably due to PCB layout. Next step is to build one of the new Lo Hum PCBs that arrived the other day and compare.

Cheers

Ian

 

[Whoops]

Now I am beginning to recover from my recent operation

Ian

Ian, I wish all the best for your recovery!

 

[Deepdark]

just curiosity
attached ...

It is an artefact.  The noise spectrum varies continuously with little peaks like that all over the place. I used four averages to smooth thinks out a little but deoending on when you take the snap shot you get little peaks like that.

Here is the same amplifier this time powered by AC heaters. No mistaking the peaks on this one!!!

Cheers

ian

Edit: Note that with the standard A or IEC filters for noise measurements, this 50Hz peak drops by about 20dB.

This graph is with AC heater, gain at max, with an 150-200ohm source plugged into the input, or input shorted? Those values are in dBu, I guess, so in dBfs, the big 50hz peak at -60dBu is about -70dBfs to -74dBfs, or so?

 

[Deepdark]

And lastly for comparison, the same amplifier with the gain turned right down so what you see now is just the noise and the hum from the output stage. Largest 50Hz peak is now -75dBu.

Cheers

Ian

And this one is with gain close, so your listening to the internal output noise, right? What I understand is that the hum get higher when you push the gain up, right? -75dBu means about -85dBfs, or so, right?

 

[ruffrecords]

And this one is with gain close, so your listening to the internal output noise, right? What I understand is that the hum get higher when you push the gain up, right? -75dBu means about -85dBfs, or so, right?

I think the input was shorted but even with 150 ohms it is very much the same. The figures are in dBu. What that means in dBFS depends on your converters.

The hum definitely gets higher as the gain gets higher whic means some hum must be getting  into the input tube.

Cheers

Ian

 

[kambo]

can we say
50hz sine_wave at  -60 dbu / 0.7 mv (RMS),
would be close your main hum noise ? (minus the harmonics of it)

 

[Deepdark]

And this one is with gain close, so your listening to the internal output noise, right? What I understand is that the hum get higher when you push the gain up, right? -75dBu means about -85dBfs, or so, right?

I think the input was shorted but even with 150 ohms it is very much the same. The figures are in dBu. What that means in dBFS depends on your converters.

The hum definitely gets higher as the gain gets higher whic means some hum must be getting  into the input tube.

Cheers

Ian

Gain at max, if it was on my system, would looks like -75dBfs and gain close would be about 90dBfs or so. In. Normal mixing situation, you would not hear any noise, even with gain at max, compared to the audio signal. The hum at the input could be the heater traces EMI polluting the nearest tube pins, like the grid?

I made an test with my LA2A and rewired the heaters by connecting V1 last, instead of the beginning of the chain. It help quiet a bit but when gain is to max, then hum get higher. Before I change the heater routing, i hadn't this rising hum, but the EMI was worsed at V1, so general hum performance. Maybe the heaters now are passing near an sensible part and get hit with full EMI.

 

[ruffrecords]

can we say
50hz sine_wave at  -60 dbu / 0.7 mv (RMS),
would be close your main hum noise ? (minus the harmonics of it)

Yes with 70dB of gain.

Cheers

Ian

 

[ruffrecords]

And this one is with gain close, so your listening to the internal output noise, right? What I understand is that the hum get higher when you push the gain up, right? -75dBu means about -85dBfs, or so, right?

I think the input was shorted but even with 150 ohms it is very much the same. The figures are in dBu. What that means in dBFS depends on your converters.

The hum definitely gets higher as the gain gets higher whic means some hum must be getting  into the input tube.

Cheers

Ian

Gain at max, if it was on my system, would looks like -75dBfs and gain close would be about 90dBfs or so. In. Normal mixing situation, you would not hear any noise, even with gain at max, compared to the audio signal. The hum at the input could be the heater traces EMI polluting the nearest tube pins, like the grid?

I made an test with my LA2A and rewired the heaters by connecting V1 last, instead of the beginning of the chain. It help quiet a bit but when gain is to max, then hum get higher. Before I change the heater routing, i hadn't this rising hum, but the EMI was worsed at V1, so general hum performance. Maybe the heaters now are passing near an sensible part and get hit with full EMI.

Hum can do strange things over two stages of gain. It is possible for the phase of the hum induced in one stage to be the opposite of the phase of hum in the other stage. This means that at some middle gain the hum just about cancels out but at higher and lower gains it gets bigger. Sometimes it is worth trying swapping the heater leads of each stage in turn to see which gives the lowest hum..

Cheers

Ian

 

[ruffrecords]

I think it is also important to remember that hum can get in from almost anywhere. I does not all have to come from a local transformer/power supply. Here is a spectrum of a Pultec EQ with built in gain make up that I recently finished building. This is the initial test in the same test  rack and using the same dc power supply that was used to test the original solo design as shown in the second plot of the first post of this thread.

Note how there is a definite 50Hz component at -95dBFS and a couple of harmonics at 250Hz and 350Hz. Rememeber this is with a dc supply to the heaters.

Cheers

Ian

 

[ruffrecords]

Now, although the unit in the above post was completely enclosed in an aluminium cassette, I had not made an electrical connection from the cassette (chassis) to 0V. So I did this and repeated the test which you can see in the attached spectrum. As you can see the 50Hz component has dropped by 10dB and the harmonics have changed.

Cheers

Ian

 

[ricardo]

Now, although the unit in the above post was completely enclosed in an aluminium cassette, I had not made an electrical connection from the cassette (chassis) to 0V.

All my LN stuff is in boxes hand carved from solid Unobtainium by virgins.  I never have trouble with hum because of this.

More trouble and slightly more expensive but worth it  8)

 

[Deepdark]

Now, although the unit in the above post was completely enclosed in an aluminium cassette, I had not made an electrical connection from the cassette (chassis) to 0V. So I did this and repeated the test which you can see in the attached spectrum. As you can see the 50Hz component has dropped by 10dB and the harmonics have changed.

Cheers

Ian

This is impressive result. DC heater looks to work well here

 

[ruffrecords]

This is impressive result. DC heater looks to work well here

it is good, but not as good as it could be. Attached is a noise plot of the original EZTube version of the Classic PCB at minimum gain (its output stage is a mu-follower just like the gain make up in the EQ) in the same test rack using the same external dc power supply. Here you can see there is a tiny 50Hz component at -120dBFs (about -113dBu) and no sign of any other harmonics. The only difference in the case of the EQ is that there are some unscreened Carnhill inductors in there which may be picking up a bit of hum from outside. Just shows you cannot be too carefull when it comes to hum.

Cheers

Ian

 

[Deepdark]

This is impressive result. DC heater looks to work well here

it is good, but not as good as it could be. Attached is a noise plot of the original EZTube version of the Classic PCB at minimum gain (its output stage is a mu-follower just like the gain make up in the EQ) in the same test rack using the same external dc power supply. Here you can see there is a tiny 50Hz component at -120dBFs (about -113dBu) and no sign of any other harmonics. The only difference in the case of the EQ is that there are some unscreened Carnhill inductors in there which may be picking up a bit of hum from outside. Just shows you cannot be too carefull when it comes to hum.

Cheers

Ian

You're right. The best would be external power supply haha. I didn't think about the inductors picking noise but you're right. Unscreened output yransformer must picked a bit, too. I know cinemag offer their inductors in a schielded can. I remember building some pultec and using them. I guess their is some circuit design more prone to pick up noise than other, ?

 

[abbey road d enfer]

Unscreened output yransformer must picked a bit, too. I know cinemag offer their inductors in a schielded can. I remember building some pultec and using them. I guess their is some circuit design more prone to pick up noise than other, ?

The OT is generally less critical than IT or EQ inductors (or wiring layout for that matter), but nevertheless driving the OT with a low impedance circuit is very important.
It's easy to confirm; just listen to the output of a transformer and notice the difference when the primary is loaded or not.
Ther's a good case for negative impedance drive, but it's quite difficult to implement with tubes.
The ideal being using a FB winding to produce negative drive impedance.
Using ca. 20dB of NFB around a transformer is equivalent to a good shield. It has the advantage of also improving frequency response and THD.

 

[ruffrecords]

@abbey,

Different makes of output transformer are more or less susceptible to hum fields. Holger Classen and I have both found that the  Canrnhill VTB2291 is more susceptible to hum pick up than the Edcor equivalent (XSM2.4K/600). Holger now uses a mu metal screened 2K4:600 output transformer transformer by Sowter. More expensive but no hum problems and it is also PCB mounting.

Cheers

Ian