AC heaters ultimate hum limit

[CJ]

once the AC gets inside the tube, how much hum will you still have?

in other words, assuming zero hum induced from the external wiring, how much hum would you still have?

 

[trobbins]

Ian,
That pcb appears to have elevated heater voltage (R23-24, C17), but only a fixed humdinger? 

If you are going to AC power the heaters then a pot is the better option, than fixed.  If you are going to use AC heaters,  then some better 0V shielding around the heater traces could be achieved. 

Are the red traces on the pcb bottom.  I the remainder of the bottom flooded with 0V ground plane?  Is the remainder of the pcb top flooded with 0V ground plane?

Have you chosen a fast diode version for BR1 (and BR2)?  Have you chosen C18 as a tuned snubber for the power supply secondary winding (ie. the winding has enough ESR to appropriately snub its leakage inductance)?

 

[ruffrecords]

once the AC gets inside the tube, how much hum will you still have?

in other words, assuming zero hum induced from the external wiring, how much hum would you still have?

However much leaks from the heaters to other electrodes of the tube I guess.

Cheers

Ian

 

[ricardo]

once the AC gets inside the tube, how much hum will you still have?

in other words, assuming zero hum induced from the external wiring, how much hum would you still have?

On EF86, a maximum of 5uV referred to the grid ... 1.5uV with centre-tapped heaters.  On lesser 'LN' valves, 6-10dB worse.

That's assuming you follow the rules from Mullard's "High Quality Audio Amplification"

[list type=decimal]
[*]A low-Loss valveholder must be used. A nylon-loaded phenolic holder, preferably skirted, is generally adequate
[*]The external leads to the heater should be run as a twisted pair to neutralise their external magnetic fields. (The heater of the EF86 is itself made in the form of a spiral.) The heater wiring should be kept as close to the chassis as possible.
[*]Earth returns should be made to a busbar. Connections can be made to the central spigot of each valveholder in the first instance. The busbar should only be connected to the chassis at one point, preferably near to the input terminal. Hum picked up from currents circulating in the chassis is then kept to a minimum. Wherever possible, an external earth connection should be made.
[*]Where a recommended layout for the component is given, there will be no danger of hum being generated by the magnetic field of the mains transformer acting directly on the electrode supports. If some other layout is tried, the possibility of this form of hum occurring should be considered very carefully.
[/list]

 

[zephyrmic]

Excellent tips! In the "dark ages" when I was building valve audio gear, in addition to the advice from ricardo, I always included a flat wire-wound 100 ohm trimmer between the two AC heater leads, with the slider to ground, and then adjusted the slider for minimum hum. Worked well for me. The other thing I found useful was to shield the EF86s - the input valves - by a grounded metal shield around the valve envelope, and that proved effective against stray hum. Hope that helps.

Kindest regards,
zephyrmic

 

[ruffrecords]

Ian,
That pcb appears to have elevated heater voltage (R23-24, C17), but only a fixed humdinger? 

If you are going to AC power the heaters then a pot is the better option, than fixed. 

Yes, elevated heaters but fixed hum dinger. I will experiment with a pot version but so far altering humdinger has made little more than 3dB difference.

If you are going to use AC heaters,  then some better 0V shielding around the heater traces could be achieved. 

Are the red traces on the pcb bottom.  I the remainder of the bottom flooded with 0V ground plane?  Is the remainder of the pcb top flooded with 0V ground plane?
[

Two red traces in centre are heaters. They are on the bottom. Bottom is flooded with 0V, top is not.

Have you chosen a fast diode version for BR1 (and BR2)?  Have you chosen C18 as a tuned snubber for the power supply secondary winding (ie. the winding has enough ESR to appropriately snub its leakage inductance)?

No and no. HT hum/interference is not a problem as operation on dc heaters demonstrates.

Cheers

Ian

 

[ruffrecords]

once the AC gets inside the tube, how much hum will you still have?

in other words, assuming zero hum induced from the external wiring, how much hum would you still have?

On EF86, a maximum of 5uV referred to the grid ... 1.5uV with centre-tapped heaters.  On lesser 'LN' valves, 6-10dB worse.

The original hum level was about 10uV. at the grid of the first valve. The twisted heater wires mod reduced the 50Hz component by about 6dB so say it is now around 5uV. 6 to 10dB worse than 1.5uV is 3 to 4.5uV so it looks like there is not much  more that can be achieved.

Cheers

Ian

 

[merlin]

I will experiment with a pot version but so far altering humdinger has made little more than 3dB difference.

3dB is a lot when you're chasing the absolute minimum!

No and no. HT hum/interference is not a problem as operation on dc heaters demonstrates.

If the HT and heater windings are on the same transformer core then the 'cleanliness' of the HT switching also affects the AC heater waveform, which in turn will affect the hum picked up when using AC heaters. Don't dismiss a cleaner HT supply when you're chasing minimum hum. Separate transformers for HT and heater, with a mains input filter and common-mode heater filter, will all help to smooth out the heater waveform.

 

[ruffrecords]

I will experiment with a pot version but so far altering humdinger has made little more than 3dB difference.

3dB is a lot when you're chasing the absolute minimum!

No and no. HT hum/interference is not a problem as operation on dc heaters demonstrates.

If the HT and heater windings are on the same transformer core then the 'cleanliness' of the HT switching also affects the AC heater waveform, which in turn will affect the hum picked up when using AC heaters. Don't dismiss a cleaner HT supply when you're chasing minimum hum. Separate transformers for HT and heater, with a mains input filter and common-mode heater filter, will all help to smooth out the heater waveform.

I confess I was hoping for a more significant reduction but putting into the context of what is likely to be achievable it was probably too much to expect.

I am aware of the effect of HT rectification on heater ac waveform. Do you think it worthwhile altering the the non-linearity of the HT load in order to improve the ac heater waveform? (e.g. by introducing a series resistor to the reservoir capacitor)

I was not intending to dismiss clean(er) HT; just pointing out that the same circuit with dc heaters has no hum whatsoever so the cleanness of the HT is second order at least

 

[trobbins]

Ian,
The top layer could be used for 0V screen traces around sensitive nodes, and between channels.
I guess areas of interest would be between:
- V2/pin7 and V1/pin3, and phantom powering circuit.
- V1/pin3 and V2 circuitry
- V1 circuitry and 240VAC circuitry.

If you were to have an ac heater, then V2/pin6 trace underneath the heater would have least capacitance if traces crossed at right angles. 

Is Pin 9 the heater CT?  What was the aim of the resistor to ground?

Ciao, Tim

 

[kambo]

6922

 

[ruffrecords]

Ian,
The top layer could be used for 0V screen traces around sensitive nodes, and between channels.
I guess areas of interest would be between:
- V2/pin7 and V1/pin3, and phantom powering circuit.
- V1/pin3 and V2 circuitry
- V1 circuitry and 240VAC circuitry.

If you were to have an ac heater, then V2/pin6 trace underneath the heater would have least capacitance if traces crossed at right angles. 

Is Pin 9 the heater CT?  What was the aim of the resistor to ground?

Ciao, Tim

THanks for the tips. The pin 9 resistor is zero ohms. The 6922 has an internal screen connected to pin 9. The pin 9 resistor is to allow 6CG7 to be used instead for which pin 9 is NC.

Cheers

Ian

 

[ruffrecords]

I have revised the PCB layout taking into consideration all the advice given - pic attached. VR1 is the humdinger pot.

Cheers

Ian

 

[ruffrecords]

For added clarity on the new layout, here is the Gerber plot of the ground plane. As you can see it extends right inside the tube base and screens off the heater tracks.

Cheers

Ian

 

[trobbins]

Ian,  that's a nice layout. 

A few very nit picky suggestions could be :
C3 rotated, so that gnd is nearer the other grounds for grid/cathode.
Separate C16 trace to R23.
Top side ground partition between C1 of one channel, and C6 of other channel.

 

[ding]

I noticed this would make it very difficult to do 6.3v heaters.

 

[trobbins]

Also perhaps flood the top layer above the heater traces to act as ground screen.
Heater traces could be closer together?

 

[ruffrecords]

I noticed this would make it very difficult to do 6.3v heaters.

It is wired for 6.3V heaters.

Ian

 

[ding]

I noticed this would make it very difficult to do 6.3v heaters.

It is wired for 6.3V heaters.

Ian

Sorry, yes, I guess I was meaning 12v. This is what happens when you post 5 minutes after you wake up.

 

[ruffrecords]

Ian,  that's a nice layout. 

A few very nit picky suggestions could be :
C3 rotated, so that gnd is nearer the other grounds for grid/cathode.

Only problem is that would lengthen the cathode trace and it is already longer than I would like.

Separate C16 trace to R23.

OK

Top side ground partition between C1 of one channel, and C6 of other channel.

I am not sure what this will achieve.

Cheers

Ian