12 to 5 AC Rectification

[woodrough]

Hi,

I am currently constructing a preamp with a 5AR4 rectifier and ran into the issue of needing to convert the 12vac secondary to 5vac to run the filament of the 5ar4 to start up the whole system.
Long story short, I had to utilize the 6.3vac output of the transformer for supplying the regular heater lines. the remaining secondary for the main tube rectifier is a 12v @ 2A line. I dont have space for another (115v to 5v) transformer. I am currently looking for a 12vac to 5vac transformer (hopefully smaller than 115:5).

Another method that can be much smaller is SS rectifying the 12vac and voltage dividing the created DC. not bad...

another method is :
AC Voltage Dropper
the peak is scaled down nicely to 5v... but the RMS would be some pathetic 3v due to the incredible notching created by diode rectifying. Anyone know how to smooth out the output to raise the RMS to a realistic sinewave?

another method could be using resistors to stomp down the 12V... but it could get hot(?) Also, would this draw more Amps? My transformer can only put out 2A @ 12V. I dont mind heat from the resistors too much.. but I def dont want to over saturate the transformer.

Hopefully I am not all that screwed over. Any tips?

[Mooly]

If you can get the values right consider a wattless dropper by using the reactance of a capacitor to drop the voltage. Or a combination of C and R.

Series resistors are fine and the current drawn is the same as the filament on its correct voltage... its just a series circuit.

[Mooly]

Like this,

[Mooly]

Whats the filament current for this valve ?

[Mooly]

Google tells me it 2 amps (I don't do valves normally )),

This gives around 5 volts RMS across 2.5 ohm

[woodrough]

Wow! That simple?
It's a sovtek 5ar4 where the heater is running with 1.9A.
With your proposed capacitor set up, would it hurt the transformer by trying to fight the 7 volt loss? Also the winding is rated to 2A. Any harm?

[Mooly]

Its an old technique...

I can't really see there being an issue in practice. I'm just trying to visualise whether the current in the cap being 90 degrees out of phase with the voltage is an issue... and how it relates to the inductance of the transformer.

It just relies on the capacitive reactance of the cap (a sort of AC resistance value).

xc= 1/2*pi*f*c

I choose two caps in series because thats how a normally non polarised cap is constructed from two electrolytics but I'm thinking do we need that. There is no DC or DC bias and so I can't see why one cap wouldn't be OK.

[Mooly]

Here's with a single cap.

Practical points. Maybe use two or three parallel caps to split the current. 2 amps RMS is within a typical good quality caps spec (typically around 2800ma but check the data sheet for any caps used) but it makes sense to make it easier for the caps. Two 2200 uf give 4400uf total. Or three 1000uf caps. Adjust the resistor to suit.

Transformer dissipation wise I think its OK. It's a series circuit and the "real" power and current is dissipated in the resistances. There will be a theoretical ah but aspect which depends on the inductance of the transformer but I think its OK.

[Andrew Eckhardt]

I'd try to adjust with capacitors rather than a resistor. After going the trouble of reactive current source to being with, why have more lost in an adjustment R than in the load?. Go low with the value of the main cap bank by a few 100u and then tack on required small value in parallel to dial in if necessary. Probably should look for a couple times the ripple current capacity in the caps than the actual load, if the caps are expected to last as long as the transformer. I don't think the transformer would complain about the current phase shift any more than the limited conduction angle method, and the caps will make a lot less noise.

[Mooly]

Good thinking on the trim resistor Andrew. Must admit I had a big surprise when I saw the current requirement of the filament.

I can't really see a practical problem either with the phase shift and dissipation.

Hopefully it will give woodrough something to work with.