Artificial centre tap - resistance to ground

[überfuzz]

Is the resistance to ground in a artificial centre tap R = R1 || R2? I.e. R = 1/( 1/(R1) + 1/(R2))

Example if R1 = R2 = 100Ohm, R = 1/( 1/(100) + 1/(100)) = 50 Ohms?

What is a functional value of the resistors? Seems the lower I go the more current will sneak out...

[nickb]

Is the resistance to ground in a artificial centre tap R = R1 || R2? I.e. R = 1/( 1/(R1) + 1/(R2))

Example if R1 = R2 = 100Ohm, R = 1/( 1/(100) + 1/(100)) = 5 Ohms?

What is a functional value of the resistors? Seems the lower I go the more current will sneak out...

Functionally speaking, the values you choose should be as high as possible yet still get the job done. It depends on the type and number of tubes in your application.

Practically speaking, you see low values used commonly as a one-size-fits-all approach. What is a practical value? It's a balance between getting the job done and not wasting too much power.

For the 100 ohms case, the AC heater current flowing is 6.3/(R1+R2)= 6.3/200 = 31.5mA rms. This means that each resistor dissipates I^2*R watts ~=100mW - quite a manageable figure. The tube cathode-heater leakage current is small enough to produce negligible drop across 100 ohms. Once it's negligible there is no point in going smaller.

Hope that helps.

[Leo_Gnardo]

What nickb said ^^^.

In practical use I've gone as high as a 500 ohm hum balance pot to replace a roasted one in an Ampeg SVT - lots of tubes, lots of filament current, worked just fine. And I've seen as low as 47 ohm carbon comp half-watt in some 80's Fender amps. Any R in the range of 47 to 270 ohms should do. On occasion I've had to use an unmatched resistor pair to get minimum hum, say 82 and 120 ohms.

[überfuzz]

Hm.. Strange! Sliding off topic a bit: Turns out the amp I'm tinkering, that made me wonder, had both a centre tap in the transformer and an artificial tap in the filament circuit. I just took out the artificial tap. Right?

Thanks for the clarifying posts regarding artificial centre taps.

[DRH1958]

Not OT at all. As long as we are looking at this we might as well learn all we can. You are correct to eliminate the artificial tap as long as the tranny has one. The reason to put one in is to create one if there is not one already.
As long as we are doing this, I think having a CT gnd. and an artificial one might make it worse by creating a ground loop. What does anyone else think? Or will the 100 ohm resistors cancel out this effect. Sort of similar to grounding only one end of the shield in a shielded cable.

[trobbins]

It would be a ground loop if the ground points were different - although there would be negligible current flowing in that 'loop'.

A humdinger can be advantageous as a poor mans fuse in some fault situations (eg. pin 2 to 3 arc-over on output valve socket). A tuned humdinger (ie. pot) can be advantageous when trying to null out hum.

[nickb]

Not OT at all. As long as we are looking at this we might as well learn all we can. You are correct to eliminate the artificial tap as long as the tranny has one. The reason to put one in is to create one if there is not one already.
As long as we are doing this, I think having a CT gnd. and an artificial one might make it worse by creating a ground loop. What does anyone else think? Or will the 100 ohm resistors cancel out this effect. Sort of similar to grounding only one end of the shield in a shielded cable.

The current induced would be tiny compared to the 6.3VAC of hum current that is inherently present. Anyway, it's not a signal path so it's not of consequence.

[pdf64]

A humdinger can be advantageous as a poor mans fuse in some fault situations (eg. pin 2 to 3 arc-over on output valve socket)

That seems a drawback rather than a benefit?
If the heater line is pulled up to VB+, it may damage the heater-cathode insulation of all the tubes.

[trobbins]

Order of priorities - firstly save the PT.

Of course in most situations the humdinger was never designed or thought of as acting as a fuse - but after looking at the charcoal remnant, it can be somewhat satisfying to know that it serendipitously saved the PT.

If enough incentive then add a parallel 10k from heater to ground - that is common method across a cathode fuse to restrict cathode voltage if fuse fails.

[pdf64]

Order of priorities - firstly save the PT

Surely after safety, mitigation for the most expensive likely failure should be the top priority?
In this case, replacement of the heater ground reference and all tubes may well be more expensive than the PT.

If enough incentive then add a parallel 10k from heater to ground - that is common method across a cathode fuse to restrict cathode voltage if fuse fails.

If there's a plate to heater short, then the heater's ground reference will be subject to almost the full VB+, and a 10k resistor would probably have to be very beefy to handle it.

[trobbins]

Sure, there are lots of consequences to weigh up if trying to determine a priority. What I should have said was 'firstly try and save the PT and OT' - as that may sway the priority choice one way or the other a bit more.

A pin 2-3 arc may start as an arc that extinguishes, especially if initiated by OT plate voltage transients. From the examples I've seen, and a few forum comments, carbon tracking builds up on the holder base, and probably leads to continuous tracking current that has even been seen to eat away at ceramic holders.

Such an arc current is limited by the sag of the power supply, and the OT half-primary DCR, and the humdinger resistance if used (or the CT protection resistor if used in place of a fuse), so may only have a prospective level of an amp or so. If the sacrificial humdinger is used, and it is the first part to fail, then yes any parallel 'safety' resistor could see B+ across it if the remnant arc track was low resistance - so we start to fan out the scenarios that could happen, and it all gets a little messy with the what-ifs. Certainly it would be best if any added 'protection' didn't mask that a problem had actually arisen.

As a maintenance action (before any such faults occur) I would far prefer to see the opportunity for pin2-3 arcing avoided by specifically cleaning that region of the valve base, and adding some OT over-voltage protection or adding some extra creepage and clearance (as sometimes introduced by glueing or heatshrinking a sleeve over pin 3 junction).

[Enzo]

If pin 2-3 connect, then your 480v (or whatever) is grounded by the heater winding. if you have a real center tap, your winding has about half an ohm resistance. Try to measure the 6v winding on an amp PT. It is low. So that usually blows any HV fuse right now. But in any event it is not likely to put much B+ voltage into the tubes. If the HV is not fused, it will sag to very little, after all it is grounded. SO the rectifier will be seriously stressed or destroyed. But the transformer heater winding won't probably care. What is an extra couple hundred milliamps when it is ready to carry several AMPS all the time? One might damage the HV winding, but why was it not fused anyway?

[Old Tele man]

Purpose of "artificial center-tap" is to "balance" the AC-voltages (yes, plural) from the two sides of PT-filament winding output, because the windings themselves are typically NOT perfectly "balanced" within the field-core wrappings.

A better wording might be to call it a "nulling" potentiometer...

[R.G.]

nickb
Functionally speaking, the values you choose should be as high as possible yet still get the job done. It depends on the type and number of tubes in your application.

Practically speaking, you see low values used commonly as a one-size-fits-all approach. What is a practical value? It's a balance between getting the job done and not wasting too much power.

For the 100 ohms case, the AC heater current flowing is 6.3/(R1+R2)= 6.3/200 = 31.5mA rms. This means that each resistor dissipates I^2*R watts ~=100mW - quite a manageable figure. The tube cathode-heater leakage current is small enough to produce negligible drop across 100 ohms. Once it's negligible there is no point in going smaller.

Yes.

überfuzz
Hm.. Strange! Sliding off topic a bit: Turns out the amp I'm tinkering, that made me wonder, had both a centre tap in the transformer and an artificial tap in the filament circuit. I just took out the artificial tap. Right?

Maybe. It turns out that the quality of a center tap is variable. Transformer winding to get an accurate center tap is tricky, especially on windings with few turns, like the ones making 6.3Vac. For some transformers, the no-load center tap may not be really centered. Even if it is, the LOADED center tap may not be actually centered, as the two halves may have the same number of turns, but differ in wire resistance or leakage inductance. So a resistive center tap may be more accurate.

DRH1958
You are correct to eliminate the artificial tap as long as the tranny has one. The reason to put one in is to create one if there is not one already.

See above. Artificial may give better results in some cases.

As long as we are doing this, I think having a CT gnd. and an artificial one might make it worse by creating a ground loop. What does anyone else think? Or will the 100 ohm resistors cancel out this effect. Sort of similar to grounding only one end of the shield in a shielded cable.

Only if you use them both. Only use one. If you use the CT, you may or may not need artificial. If you use artificial, disconnect the real CT.

trobbins
It would be a ground loop if the ground points were different - although there would be negligible current flowing in that 'loop'.

Maybe. Things get strange with low voltages and resistances. Large and unexpected currents can flow. However, if only one of the "CTs" are used, it fixes the issue.

A humdinger can be advantageous as a poor mans fuse in some fault situations (eg. pin 2 to 3 arc-over on output valve socket)

It's poor economy. Even a poor man ought to save up and put one fuse in each side of the heater winding. The math required to make a resistor act like a real fuse and protect things is complex enough that anyone who can do it is probably not a poor man.

A tuned humdinger (ie. pot) can be advantageous when trying to null out hum.

Yes. If your amp has a fixed hum, you can do some nulling as long as the hum you're trying to null is in/out of phase with the AC line. If it's been shifted by capacitive or inductive transmission, or being amplified by a stage with some phase shift at power line frequencies, you may not be able to null very well. And it's not very helpful for hums that are from ripple or any kind of burst-buzz.

trobbins
Order of priorities - firstly save the PT.

Could not agree more. The PT is one of the two most expensive single electronics parts in the amp. It's probably worth putting in fuses to save each winding, not relying on the AC mains fuse, which really can't do the total job.

pdf64
In this case, replacement of the heater ground reference and all tubes may well be more expensive than the PT.

Maybe. There's a lot to be said for making the protection method protect what it's supposed to, and not relying on collateral "protection". If you want to protect the PT, fuse each winding to just over it's start up surge and max-load in normal use. This usually also prevents the transformer's death throes from damaging the rest of the things connected to it.

Protection inside a tube amp needs a comprehensive approach. An hour or two spent with an experienced tech (and maybe a six-pack) will get you a wealth of information about what fails and how.

Enzo
If pin 2-3 connect, then your 480v (or whatever) is grounded by the heater winding. if you have a real center tap, your winding has about half an ohm resistance. Try to measure the 6v winding on an amp PT. It is low. So that usually blows any HV fuse right now. But in any event it is not likely to put much B+ voltage into the tubes. If the HV is not fused, it will sag to very little, after all it is grounded. SO the rectifier will be seriously stressed or destroyed. But the transformer heater winding won't probably care. What is an extra couple hundred milliamps when it is ready to carry several AMPS all the time? One might damage the HV winding, but why was it not fused anyway?

See. I told you so.

Old Tele man
Purpose of "artificial center-tap" is to "balance" the AC-voltages (yes, plural) from the two sides of PT-filament winding output, because the windings themselves are typically NOT perfectly "balanced" within the field-core wrappings.

A better wording might be to call it a "nulling" potentiometer...

Yep - another perfectly correct way to look at it.

[Enzo]

Just noticed this, you might find it of interest.

http://bmamps.com/Schematics/crate/C...oo_120_HBb.pdf

Crate Blue Voodoo 100 head, there is an artificial CT of 470 ohm resistors, but the winding also has a 120v SIDAC to ground to siphon off any high voltage that gets that far. Under normal low voltage it sits there invisible.