Solid State Guitar Amp Forum | DIY Guitar Amplifiers

Solid State Amplifiers => Amplifier Discussion => Topic started by: Bakeacake08 on June 29, 2014, 03:01:00 PM

Title: Transformer Selection
Post by: Bakeacake08 on June 29, 2014, 03:01:00 PM
I've been designing a tube/SS hybrid amp. I based the power supply on running at 50 volts max as per my device limitations. At 50 volts of rectified DC, I was going to need a transformer with about a 35 volt RMS secondary (50V pk/1.414=35V RMS). So I designed in a 36 volt transformer, but I didn't really understand how the 50 volts was going to get there; I thought that it would be running at the RMS value and somehow I had to just know that the peak value was going to be higher and I had to keep it in mind when choosing components. Then I built a simple string of LEDs for my coat closet because our doorbell transformer sticks out of the wall in there and was begging to be more useful. After running some experiments to figure out voltages with a diode rectifier wired in, I realized that the peak voltage BECOMES the DC output when rectified, and the AC voltage goes away (save for a little bit of ripple voltage). So after I put a rectifier and the proper capacitor in, the system will be running off the PEAK voltage of the secondary, not the RMS voltage.

I don't know how common that misunderstanding is, but I went through it so I thought I'd explain how I figured it in case that helped someone else starting out figure out how to design a power supply. My biggest advice (which others repeat constantly) is to actually build the circuits so you can see how the theory/math actually works and if it doesn't, you can start to figure out why not.

Anyway, that was the helpful part of my post; now on to my actual questions.  :)

I decided for now that I'm just going to stick to the 35V design instead of re-figure out all the other components. I actually found a 25.2V transformer at Radio Shack that would be cheaper than buying one online with shipping and everything. My question is about VA rating: through my reading I have learned that the DC amps drawn by the circuit will actually draw more AC amps from the secondary (by a factor of about 1.6, as I understand it). So if I expect my circuit to draw 500 mA, I should actually make sure my transformer is rated for 500mA x 1.6 = 800 mA. Am I understanding this correctly?

My second question is more theoretical, but the transformer in question is center tapped so each side of the CT is at 12.6V. That got me to thinking: Is there any reason one couldn't wire up the center tap as some sort of reference voltage? I was thinking maybe there's a way to make a bridge diode between the two outside leads and one between the "ground" lead and the center tap or something like that. My gut tells me that there's some obvious reason why it wouldn't work, but it also seems like it *should* so I thought I'd ask people who actually know what they're doing.
Title: Re: Transformer Selection
Post by: J M Fahey on June 29, 2014, 07:37:20 PM
Congratulations, you got it pretty right and on your own.  :dbtu:
Tes, a typical PSU "makes" DC out of AC and the caps charge to peak AC voltage.
That drops somewhat afterwards under load, but, hey, Real World is where we actually live, so we´d better get used to that (and many more things :) )
As of transformer current raring yes, it´s good practice to have, say, 50% xtra capacity so as not to force anything close to limits.
As of the center tapped transformer, it´s a common way to get symmetrical power rails (voltages) and many amps work that way.
So if you ground the center tap and feed 2 caps in series, with their center "tap"  also grounded, you´ll have roughly +/-16V supplies.

Just look at this 12W Marshall amp, which makes  + and - 19V rails from a 14+14VAC transformer.
(http://www.amparchives.com/album/Marshall/Schematics%20&%20Layouts/Marshall%20Amp%20Schematics/Miscellaneous%20Solid%20State%20Heads,%20Combos%20and%20Mini%20Stack/3005%20Micro%20Stack/slides/3005.jpg)
Title: Re: Transformer Selection
Post by: Bakeacake08 on June 30, 2014, 12:39:17 AM
That makes sense about making the +/- rails. With this setup, would it also work to power a load between the + and - rails (skipping the ground rail), so the load sees 38 volts?
Title: Re: Transformer Selection
Post by: J M Fahey on June 30, 2014, 01:23:36 AM
Yes, if your transformer has a  center tap but you don´t use it you can charge a, say, single 50V cap to 38V or whatever the peak voltage is.
The winding must be "floating", no part of it must be grounded.
Not even the center tap.
In that case, you ground the negative leg of the filter cap and the positive leg is +38V .
What heathkit does in the TA16 .
Title: Re: Transformer Selection
Post by: Roly on July 01, 2014, 11:22:09 AM
Very well nutted out.   :dbtu:


Search: rectifier waveforms ->  (lmgtfy (http://lmgtfy.com/?q=rectifier+waveforms+image))


(http://www.ami.ac.uk/courses/ami4966_emclvd/u02/images/Fig2_8.gif)

Red: pulsating DC from a bi-phase or bridge rectifier.

Blue: rectified "DC" voltage showing sag due to connected load between recharge bursts.

Black: recharge burst current around loop (may be many amps in a big amp).

Can you guess where "rectifier buzz" originates?
Title: Re: Transformer Selection
Post by: Bakeacake08 on July 01, 2014, 08:50:14 PM
So what would happen if you made it so it was +/- 19 volts with the center tap and also put a load from +19 to -19?
Title: Re: Transformer Selection
Post by: J M Fahey on July 03, 2014, 11:20:15 AM
Exactly the same.
The +19 to -19 connected load does not know or care that there is a center tap somewhere.
Title: Re: Transformer Selection
Post by: Roly on July 03, 2014, 11:43:31 AM
Quote from: Bakeacake08 on July 01, 2014, 08:50:14 PM
So what would happen if you made it so it was +/- 19 volts with the center tap and also put a load from +19 to -19?

That's pretty much what happens with op-amps in preamps.  They actually run from the +/-Vsup rails and the ground reference for correct circuit operation is actually pretty incidental to the power flow.  This is why op-amps should be locally bypassed with a cap direct from Vsup+ to Vsup-, and not via ground.