Heater Wire Ground
Moderators: pompeiisneaks, Colossal
Heater Wire Ground
If the power transformer has a center tap to ground for the 6.3 VAC filament power, would it still be a good idea to put the dual 100 ohm resisters from the pilot light to ground, or would this be dumb and cause problems?
Re: Heater Wire Ground
It's redundant. Use one or the other. I prefer the resistors.
-
sluckey
- Posts: 3079
- Joined: Sun Jul 22, 2007 7:48 pm
- Location: Mobile, AL
- Contact:
1 others liked this
Re: Heater Wire Ground
I like the pretty colors. But there's also a technical advantage. It's not uncommon for an output tube such as a 6L6 to have an arc between pins 3 and 2. This arc is usually a result of playing loudly and losing connection between the amp and speaker. Many times this arc will cause a short between pins 2 and 3 of the output tube (plate to filament short). The short provides a path to ground for the plate voltage. Now if your heater has a real center tap you run the risk of burning up the PT, but if you used 100Ω 1/2 watt resistors the resistors will likely just pop removing the danger to the PT.
Re: Heater Wire Ground
(1) And the heater center taps are sometimes not too accurate. A couple of resistors may well make a better center tap, with lower induced hum.
(2) And using two resistors with a pot lets you tune the residual hum - assuming it's line frequency hum, not ripple.
(3) And not tying the CT to ground means that you can protect the heater winding and the resistors with a single fuse on the heater wires. A hard CT requires two fuses, one on each end.
(2) And using two resistors with a pot lets you tune the residual hum - assuming it's line frequency hum, not ripple.
(3) And not tying the CT to ground means that you can protect the heater winding and the resistors with a single fuse on the heater wires. A hard CT requires two fuses, one on each end.
Re: Heater Wire Ground
That's twice you've said that recently. I think two fuses is redundant. The same filament current flows from top end of the filament winding through the filaments and back into the bottom end of the winding. No current flows into the center tap. It only provides a ground reference.
Re: Heater Wire Ground
Why do you think that the minimum hum is on the middle of very- too much precision- resistors ? Give us some technical reference and measurement.
Re: Heater Wire Ground
Re: Heater Wire Ground
actually power rheostat 5W is the best, but R.G. is stating that we need veery precise resistors, lets hear from him why.
Re: Heater Wire Ground
More precise resistors would be closer to the average center over a lifetime of tube swapping, no? I agree that accuracy probably doesn't matter in the grand scheme, I usually use 5%, rarely a trim pot.
Re: Heater Wire Ground
not, because every tube swapping (even 1 preamp tube) is the new situation
Re: Heater Wire Ground
It's not for the actual heater currents. It's in case the CT is grounded and one heater wire breaks or otherwise gets shorted to the chassis or signal ground. In that case, the half-heater winding tries to melt the chassis - and can melt itself in the process. Since you don't know which half will break/short, you need a fuse in both. If there is only one winding section (that is, CT is open or only resistor-connected to ground) then a single fuse on either side will prevent the heater winding from overcurrent.sluckey wrote: ↑Sat Mar 23, 2024 6:02 pmThat's twice you've said that recently. I think two fuses is redundant. The same filament current flows from top end of the filament winding through the filaments and back into the bottom end of the winding. No current flows into the center tap. It only provides a ground reference.
Granted, only one fuse does cut the probabilities down by 50%, so if the one fused side shorts, all is well; but if the other side shorts, and CT is grounded, the unfused side has no fuse. It becomes a coin flip. It becomes a question of what if, and how likely - just like all fault prevention/mitigation issues.
Re: Heater Wire Ground
I've been doing this stuff since '65 and have never witnessed such a failure. Have you? This just sounds like one of those new age theoretical exercises where you fuse everything just because it 'may' happen. In reality you can be over cautious to the point where the fuse becomes more of a liability than the circuit it was meant to protect.
Re: Heater Wire Ground
Sure, glad to. I have personally wound and measured imbalances in heater windings with CT of up to 7.5% of the total heater voltage. This is easy to beat with resistors, and especially with a pair of resistors and a middle pot with a grounded winding. I've measured less on other transformers. So, measurement down.
Next, technical reference.
I don't know if you've wound power transformers much, but I did have to at one point. The calculations for getting a low voltage heater winding with the correct number of turns for a nominal voltage, including resistive and referred-wire resistance losses under load get tricky. It's easy to end up a long set of turns-ratio calculations with a low voltage winding that needs fractional-turn more or less. So you, the winder, has to pick slightly too many turns or a fraction too little.
In a mains frequency transformer, the volts per turn is set by the incoming AC mains voltage and the number of turns needed to keep the core magnetized under the start of saturation. Once you have this number of turns for your chosen core lamination size and stack, you must then use that number of volts per turn for every other winding. Well, you don't HAVE to, but Mother Nature will make it come out that way, so you might as well try to get with Her program.
The nominal volts per turn gets bigger as cores get bigger. Small cores, you get many, many turns on the primary, and small volts-per-turn. Big cores, the volts per turn goes down as the increased volume of iron and bigger core area allow you to use fewer primary turns. So the number of turns per secondary goes down in proportion.
That's not bad for the high voltage windings, as a turn more or less doesn't make as big a proportion of voltage on the winding. But on heaters, half turns can matter.
Example: say you have picked a core which needs two turns per volt on the primary to keep the magnetizing inductance high enough and hence the magnetizing current low enough. If you're trying to make 6.3V... oops, you can pick 12, 13, or 14 turns for it. 12 turns gets you 6V, 13 gets you 6.5V, and 14 gets you 7V.
Oops#2: you want a center tap. You can't use 13 turns, as that requires 6 and 1/2 turns per half-winding. And you can't wind a half-turn by only sticking the wire through one of the core windows - that unbalances the flux in the core, with some other ugly effects. Well, you could make the CT come out the opposite side of the core than the other heater wires.
That's downright difficult to explain to the people who commissioned the transformer. Going for 14 turns lets you have seven turns per half-heater, and all the wires come out the same side of the core. Better. Only now you get 7Vac on a winding that was supposed to be 6.3V. Probably you get away with this, because the primary voltage wobbles up and down a few percent anyway, and the heavy heater current will cause winding losses to drop it a little. So seven it is.
Now you gotta wind the winding. Heaters use big wire, and it's difficult to bend the wires into place. There are two ways to wind a CT winding. One is to wind enough turns, bring out a loop for the CT after you've wound half the winding, and then wind the second half of the winding on top of or along side of the half you've already wound. It is rare that this kind of tapped winding is really balanced with the CT exactly in the middle. The wire resistances are different, their leakage inductances to the primary and their positions in the winding window are different, so they will have slightly different voltages on the ends when you get done. It may be a fraction of a volt, but it will get worse with increasing load, and heaters are high-load windings. So you in general get inaccurate center tap voltages, even if you avoid the half-turn problem.
The other way to get truly accurate center taps is to wind both halves of a winding bifilar, side by side. This way they can have exactly the same turns, fractional turns, and leakages. This is the way CT windings in output transformers get wound (well, or SHOULD get wound) in output transformers. But no transformer maker will wind a heater winding bifilar unless you pay extra for it.
So the heater voltages will be slightly imbalanced. I have personally measured 0.45V of imbalance on a nominal 6.8v (at no load) heater winding under a 4A load on the winding, That's a 7.5% imbalance. Two 5% resistors, even unmatched, will likely get closer. Any attention to matching or putting a small pot in the middle of two resistors can get this down to vanishingly small.
Did that help you understand? I can go deeper into it if you like.