Virtual Center Tap Resistor Value

[mbratch]

Engineering is about using proven solutions to get the required outcomes, minimising risk and development cost - experimentation is for scientists and should be the last resort. Therefore as a value of 100 ohms for the balancing resistors is know to work, there's no benefit in risking the use of 150.
Whereas you don't have those engineering constraints, and can experiment to your heart's content, so give your 150 ohm resistors a go in this application.

I thought engineering was about designing to pre-calculated outcomes, but mixed with some experimentation to refine empirical results at times when those theoretical results are difficult to come by.

I was hoping some other enterprising soul had already perhaps performed said experiment and let me know the expected results in advance.

[mbratch]

I don't see any reason in the world why 150 wouldn't work. But if you're worried, why not use two 150s in parallel for each resistor, that's 75 ohms.

Thanks Steve. I reckon I'll give it a go. The reason I asked in the first place was to find out if, from anyone else's experience or theoretical knowledge, such an attempt would be a waste of time. Sounds like it's worth a shot.

[Chuck H]

Well, on the macro view what's come out of this at least indicates that your 150 ohm resistors should be just fine in this application. Just stick 'em in. No need for matching or concearn about circuit stress or extra noise. Now...

Since the OP's Q seems to have been answered I'm not really hijacking, am I?

Whenever I read about the addition of an elevated DC bias for the filament CT it's usually derived from the preamp side of the B+ rail. I assume this is because ripple is lower there. But since such circuits are usually bypassed by another cap on the grounded end of the divider can't it be taken fron the main filter as well. I ask because I have been doing it. Since I include a bleeder circuit in my designs for safety reasons I use a 220k with a 22k dropper bypassed by a 100/100 cap off the main filter node. I figured that since the bleeder circuit is already in place why not use it for this rather than build a second circuit. Besides, From a grounding perspective isn't it better to send any ripple to the PS ground than the preamp ground? Should I be doing this differently?

P.S. Happy Holidays everyone.

[mbratch]

Since the OP's Q seems to have been answered I'm not really hijacking, am I?

As the OP, I offer my permission, if that means anything lol

[Steve Conner]

Since I include a bleeder circuit in my designs for safety reasons I use a 220k with a 22k dropper bypassed by a 100/100 cap off the main filter node. I figured that since the bleeder circuit is already in place why not use it for this rather than build a second circuit. Besides, From a grounding perspective isn't it better to send any ripple to the PS ground than the preamp ground? Should I be doing this differently?

I don't see anything wrong with that method. The ripple current associated with the 100/100 cap is negligible because it has to come through the 220k resistor. So say there was 22 volts of ripple, there'd be 0.1mA of ripple current going to ground through that cap.

The filament rail isn't THAT sensitive to hum either, especially once you've plonked a positive DC bias onto it. So it doesn't (IMO) matter if the positive bias isn't perfectly filtered, there'll still be less hum than with no bias at all.

The amount of ripple introduced by a wrong ground connection of the filament bias bypass cap is almost certainly negligible.

[mbratch]

Just to close the loop on the original topic, the 150 ohm resistors worked perfectly.

[mooreamps]

Just to close the loop on the original topic, the 150 ohm resistors worked perfectly.

Good to hear. Perhaps one of the "few" threads I've seen on here where it was not "pulled off topic" by others.... :|

-g

[redelephant]

With so many answers this is probably redundant, but...My guess on why the resistors need to be pretty low in value is normally u want a reference network, which such a resistance is, to be relative low in resistance compared to the rest of the circuit it is the reference for, that being the heaters. But lower than 100ohms and u start pulling power into waste, so it is probably a compromise value. My next guess is any value up to several hundred ohms will work fine.

Agreed, centertapping, and better yet having an adjustable reference via a potentiometer, provides better hum attenuation in high gain circuits than simply connecting one side to GND. So when center tapping you only have some 5volts going negative, while connecting one end to GND u have 10volts, and on sensitive circuits that makes a difference.
After my own experience, and checking tube books, I kind of irritate easily when I see heaters referenced to GND and not a positive voltage. But I run short on ammo and having tried to explain so many times I give up...

Chuck H, ripple on the elevated heater potential is no concern. As long as the ripple never goes below the cathodes in potential, nothing leaks into the circuit. The reason preamp voltage often is used to tap a divider on to is probably both b/c the voltage is already a little lower i value, and b/c it's a good idea to have some bleeder circuit at the end of the power supply string. But nothing wrong in tapping voltage off any other node in the supply.

[R.G.]

The idea of using a potentiometer for the heater reference is based on hum being either impossible or uneconomical to stop with good build practice. A pot using the wiper for referencing the heaters can be deliberately unbalanced to one side or the other to provide an actual hum cancellation if the hum is power line frequency. It works well as long as it's the hum inside the amp that's being cancelled, and as long as the hum is not twice-power-line ripple voltage, which it can't cancel at all.

Raising the heaters above ground works because the hot wires have some amount of electron emission, and if the heater is made positive, the electrons are pulled back into the heater before they get loose into the rest of the tube.

[Alex R]

...on the one-side-to-ground thing, there's a little old single-ended amp we see in the UK sometimes labelled 'Leo', an old catalogue practice amp from the 60s/70s, with 1 x EL84 and a simple preamp. They hum a lot, what with being single-ended, and also have one side of the filaments grounded. A virtual centre tap does reduce the hum quite noticeably on those little things. You have to hack the PCB a bit to lift the filaments from ground, not too hard to do.

[redelephant]

Allrighty then! I didn't notice much difference the time I used two resistors to make a virtual gnd compared to just connecting one side. But I really noticed a difference when biasing the heaters to +10V and up.
I guess what works works

[Enzo]

And remember, all these methods fight POTENTIAL hum. Not every amp is going to hum if the heaters are not elevated. it is a way of preventing it in the first place, or correcting it if pre-existing. But if your layout is successful, it becomes redundant.

And as Steve points out, that balance control is to null out hum, not necessarily just to balance the 6.3v. It may work best off center because it is nulling heater hum against some other source of hum by actually allowing some hum into the circuit.

And likewise, in some amps grounding one side of teh 6v will indeed cure hum, while in other amps it won't.


Oh, and thanks for correcting me, Gary, I guess .3969 is nowhere near a half. I'll let you slide on the extra significant digit.

[mooreamps]

Oh, and thanks for correcting me, Gary, I guess .3969 is nowhere near a half. I'll let you slide on the extra significant digit.

No worries.. It's probably pretty close enough to 1/2 watt regardless... :|

-g

[mbratch]

I have a 5E3 clone and a Matchless Spitfire clone. My experiences with them so far is as follows:

My Spitfire has 100R + 100R virtual center tap values on the heater wiring. I found that adding a hum balance (humdinger) helped me minimize hum. I tried all sorts of other techniques described in terms of grouping grounds, ground methods, etc. In the end, just about everything sounded the same. But when I rebalanced the virtual center tap values, I was able to remove more hum. The balance ended up being asymmetrical. It has a wee bit of hum, not much. Acceptable level. I did try elevating the DC of the filaments by connecting the virtual center tap to the power tube cathode RC rather than ground. It had absolutely no impact on the sound. So I went back to ground.

I thought I'd try the same thing on the 5E3 built some time ago. It doesn't hum much, but at very high gain you can here some. It currently has 100R + 100R virtual center tap resistors. After playing with values a bit, I discovered that hum was minimized when I fully grounded one side.

Interesting. I'm not sure if this is a design dependency, or something to do with my particular build.

[Tom Phillips]

In my experience, the exact tubes installed can affect which method (or which setting of the humdinger pot) works best. This assumes there are no other major problems. For best flexibility, I therefore like having the pot for final adjustment as tubes age and are changed out.