Elevated vs cathode elevated: results

[ampnewbie]

I built a spitfire clone that came out great. I was playing around with heater voltages, and below is link to a sound sample. The first second or so is the voltage elevated via a divider on the B+ (around 65v). The last second or so is the CT connected to the power tube cathode (13v). The first method is definitely quieter. This is recorded with guitar plugged in, volume maxed, with my iPhone a few inches from the speaker.

https://drive.google.com/file/d/1wQ79f4K5AF1yccWJ1LPn8POHSkEG-yk4/view?usp=drivesdk

[Backwoods Joe]

Thx for posting this. Now... Was wondering about the 65 volts from the divider. How did you determine the optimum elevated voltage?

[tubenit]

Do you have a schematic showing what you tried, please?

With respect, Tubenit

[sluckey]

Thx for posting this. Now... Was wondering about the 65 volts from the divider. How did you determine the optimum elevated voltage?

Does this example explain?

[2deaf]

Well now, if you raised the heater reference up 100V, that would put the heaters right at the positive max for all the other 12AX7 stages.  If you bought into TDSL's quick reference, you would put the heaters at 90V so that the CF is 10V under the negative maximum and the other stages are 10V under the positive max.

For reasons that I don't recall at the moment, I always figure the heater reference should be no more than 100V over the cathode reference and no more than 200V under the cathode reference.  To be safe, let's stay within 2/3 of the maximums.  This will put the high end of the elevation range at +67V because of the non-CF stages.  The CF's cathode is sitting at 180V and our 2/3 maximum is 133V, so we need to raise the heater reference to at least +47V.  Under these criteria, our elevation range is +47V to +67V.     

[sluckey]

Clips from three RCA Receiving Tube Manuals...

[Backwoods Joe]

Thanks Sluckey & 2deaf. Wonder why the old classic circuits (5f6a etc.) didn't raise heater voltages?

[d95err]

Thanks Sluckey & 2deaf. Wonder why the old classic circuits (5f6a etc.) didn't raise heater voltages?

Good point!

A huge percent of all guitar amps made since the sixties, including most Marshalls and derivatives, use cathode followers without elevation. If running a high cathode voltage way over the datasheet limit was really an issue, I think we would have found out by now...

Elevating the heaters is good for hum reduction though.

[ampnewbie]

Sorry for big delay in response here. For the divider, I ran the heater center tap from the junction of I think a 220k connected to the positive side of B+ and a 100k to ground. I also have a 22uf across the 100k. For the cathode version, it’s a stock matchless spitfire with the CT connected to cathodes.

[trobbins]

How much hum/noise comes through the heater-cathode can depend a lot on the tube itself, so tube rolling is always a good idea when trying to summarise if your amp is quieter or not with a given setup.

Aging tests have shown that the heater cathode interface can change/fail over time.  One degradation mechanism was identified back in the 1950's that led to a 'short' when a voltage was applied that exceeded the datasheet voltage limits, but it does take quite some time of sustained over-voltage.  Testing for that 'short' failure basically set the datasheet limits, but the test used did not allow anything more than a pass/fail outcome.  There is another degradation mechanism that lowers the insulation resistance over time - that is seen in many applications where datasheet limits are not exceeded, but insulation resistance falls over time.

So it is likely that manufacturers who ran their equipment at borderline heater-cathode voltages did not see noticeable warranty returns, or repair statistics that pointed directly at a particular design issue - especially given the nature of how repairs were done back then.  Whether that still holds for modern manufacture or not ........

[Soulfetish]

25µF is a low value bypass capacitor in the output stage.
At large input signals and clipping, there will be pretty significant bias excursion at low frequencies, and the cathode voltage will fluctuate.
Since this excursion will only happen during a portion of the duty cycle, it may cause the heater noise to modulate at a high frequency and couple into the audio as a nasty buzzing sound. 25µF might be enough. But, I'm curious if there might be a noticeable drop in noise with a bypass cap value of,.. say 220µF vs 25µF (both at idle and while pushing the amp). 

[ampnewbie]

I had a 22 laying around, and valve wizard says 10 and up should be ok. No issues yet!:

http://www.valvewizard.co.uk/heater.html

[ac427v]

I have used elevated heaters with both of those methods because it allows me to use vintage tubes without adding hum. Both circuits seemed to reduce hum but I have not done an A/B test like you did to see if one was more effective than the other. Thanks! I choose the "divider on B+" circuit when the amp has high voltage on a cathode such as the Tweed Bassman V2 example. I see the Spitfire circuit does not have that issue. Based on your experiment I may use the divider on all future builds. I'm happy with a 10-22 uf smoothing cap as Valve Wizard says--with a 150 volt rating.

[Soulfetish]

I had a 22 laying around, and valve wizard says 10 and up should be ok. No issues yet!:

http://www.valvewizard.co.uk/heater.html

right. But he's talking about the cap value for smoothing out the power supply ripple attenuated by a divider from the B+ in that case to provide a steady voltage reference. That is not the same as what is happening at the cathode voltage, where there can be a significant difference in voltage between when the amp is at idle, vs driven into clipping.

but even so, if a bypass cap value of 22µF in this spot sounds fine, it is fine.