Transformer Conversion - 6.3 V-ac Heater to 5 V-ac

[CJ]

we want to build an amp with a mellow sounding 5U4 tube rectifier for that compressed sound due to the voltage drop/sag factor.

but we need 5 volts AC instead of 6.3.

a 5U4 draws 3 amps, so to drop 6.3 to 5 we need a resistor that will eat
6.3-5=1.3 volts AC.

1.3 volts at 3 amps, right? so 1.3/3 =0.433 ohms,
Power rating? oh boy, this could be bad, I squared R , 3^2 x .43 = 3.9 watts, so we need a 10 watt resistor to run all day without complaints, forget that, too much heat even if we do find such a weird value, parallel all day long and use up all of our pwr resistors? i don't think so, been there, done that, plus, the glaciers are melting fast enuff as it is,so we want to go green with this project,

plus, we do not know if the pwr trans will have enough backup pwr to run the 5U4's 3 amp heater in addition to the pwr and preamp tubes,

so, we have a 6.3 xfmr looking for a job, why not rewind for 5 volts?

this will provide additional dependability as we have better isolation for directly heated cathodes which will put the heater winding at the same potential as the B+, or the V-ac from the xfmr in the event of an arc over caused by a short or excessive current draw fro the pwr tubes, it is common to use two transformers in Ham Radio circuits where they have higher voltages being used,

here is the basic circuit>

 

[CJ]

as luck would have it, this Hammond 6.3 xfmr has a split bobbin for max isolation to the pwr line, that means we can strip turns without messing up the 120 V-ac primary,

there are so few turns of big wire that we can strip turns without doing a D-Lam,

 

[CJ]

we ca learn about winding our own heater transformers in the future also by figuring out pri turns, flux vs core size and all that good stuff,

plus we need a ballpark figure on how many turns to strip unless we want to do 1 at a time and re-measure all  night,  ???

looks like removing  about 9 turns will get us close,

 

[CJ]

we can customize this heater xfmr for our exact load, a 5U4 only,

just hook up a octal socket to the tube and we can wire the tube to the xfmr and plug it in,

use a pwr strip which has a built in breaker,

we took of 9T but still have too much voltage, why?

because as we remove turns, DCR of the wind drops also, which means less voltage being eat up by the wind, and thus more voltage hitting the tube heater, so we need to remove more a few more turns,

 

[CJ]

removing more turns kept dropping the DCR, so we had to remove 12T total.

now we have exactly 5 volts on the heater which is nice, how often do you get an exact 5 or 6.3  from a factory job nowadays? hard to do as your load will probably be different from the test load use by the maker,

now with a big core, you  might only have 10 turns of heater winding, so getting an exact 5 volts will probably not happen, the more heater turns, the easier to meet your target voltage, pic shows amount of copper removed,

since we did not disturb the lams, no need to bake and dip in varnish, winds are still varnished also as we were careful not to unwind too many turns,

ready to tape leads, using a chunk of tape measure to wrap tape around core, the "sneaky pete" approach they call this, also can be use to install upper main seals on a 49 Plymouth without dropping the crankshaft, but that's another forum, 

 

[analogguru]

....or you can simply dump the idea of using the 5U4 as a rectifier and use instead two silicon diodes, each with a 100 Ohm resistor in series - and you will get exactly the same result (without any heating, transformer rewinding, etc...).

 

[gyraf]

Good demonstration of technique, CJ

I'm thinking: Could this be done in a non-invasive way?

You know, cutting up patients is always hard to get through the ethics commitee's, and for those faint-of-heart?

I see in the first picture that there is a little air between coil and laminations - the bobbin is not completely full.

Could we just add a number of turns outside the existing coil, and add in series with the 6.3V - but out of phase - thereby reducing voltage...?

This would even be easier on toroid cores.

Yes, I know we'd get a bit more copper resistance (and thus voltage sag)- but the whole operation would (maybe) be less scary

Any other problems in this approach?

Jakob E.

 

[RuudNL]

Interesting that you mention this Gyraf.
I use this technique here to 'stabilize' the mains voltage. (We have some fluctuations here.)
By using a small 9 V transformer, I can add or subtract 9 V. to/from the mains voltage.
The advantage is that you only need a small transformer: 1 Amp secundary for each 200 VA!