With transformer, amplifier will need higher voltage swing to get same output current in driver, higher voltage swing will draw higher current than lower voltage swing.
Basically my point of view is : with transformers you gain something (a more benign load on the amplifier and power supply) and loose something (you add a quite non linear part into the chain). Transformers are only interesting if you gain more than you loose. I'm not quite sure it's the case with typical solid state headphones amplifiers.
We cannot restrict the analysis to subjectively perceived dynamics.
The minimax, the Gilmore, the CKIII are all three push pull class A. If properly built, the last two are absolutely top notch amplifiers subjectively (I didn't hear the minimax).
I think there is a semantic game going one here. "Adequate" doesn't mean "average", "just good enough". It is used as the opposite of your "inadequate power supplies for driving low impedances" (post 2).
Sorry, but I don't get the question ?
As for the secondary effects, I agree. Actually, I see those effects as probably more significant than a reduced loading of the power supply.
@barleywater: Abraxalito is perfectly right about current drawn. Tubes amps wouldn't be possible otherwise.
Let's take an example, 50r headphones on a 4:1 transformer (voltage ratio). To get 1V into it without transformer, you need 1/50= 20ma. To get the same 1V into it with the transformer you need 4V before the transformer but the load is multiplied by the square of the voltage ratio. So we get : 4/800= 5ma.
In each case, the power required is the same:
- 50*0.02*0.02= 0.02
- 800*0.005*0.005=0.02
Understood, and agree with those as the issues. I don't have measurements about distortion for my trafos, so I don't know how much is lost objectively. However I do know they sound very dynamic on the headphones I used - so I guess its down to the individual as to what they prefer, great measurements or engaging sound. I'd love to present measurements which correlate with dynamics, but I'm not there yet.
What interests me is whether they're the type of class A which presents an unvarying load to the power supply. The only kind I know of (I'm a noob to this) which does this is balanced, current source loaded SE. To describe better - its a pair of current source loaded EFs, per channel and the signal output is taken between the two emitters. Such an arrangement could only work with 4 wires to the 'phones, or alternatively through transformer isolation and 3 wires.
If the amp is push-pull it seems to me that the PSU load is going to vary with signal and in that case, a trafo is still probably going to help the dynamics.
The question was asking why trafos were overkill in this case. Do you have a non-overkill solution which produces the same or similar subjective effect to the use of trafos?
Understood - are you going to test your hypothesis that these are the ways that a trafo achieves better subjective sound? What experiments might you be able to run to test this?
I'm familliar with the topology. But once again, it does seem to me like more trouble than it is worth. For an unclear benefit (the unvarying load on the PS), you need that troublesome 4wire output. In the realm of DIY, I'd rather design a PS that doesn't mind a varying load (not that complicated since headphones aren't heavy loads to start with). For speakers, it could be different.
Just in order to be complete: this kind of unvarying load can also be obtained by using a three channels topology of the beta22 kind (there is a lengthy discussion on head-fi about this, led by Amb).
The problem is the subjectivity of it, isn't it ? What is dynamics that isn't measured by thd ? I doubt it even exists.
Two huge problems here: the isolation of factors and the subjectivity of it.
One thing might be to built a good amp with two supplies, a good one and a bad one. If the transformer affects the amp with either supply in the same way, then PS loading isn't to blame. If the transformer mostly improves the amp with the bad supply, then PS loading is the dominating factor. Then you'd have to compare with and without transformers in both cases. It would take extensive dbt testing, which isn't easy with headphones.
But first I'd have to wind or buy good transformers. Honnestly I don't have the energy to do all that right now. Too many projects on the backlog right now (one involving transformers btw, no choice with tubes).
edit: would be much easier to get a grasp of it objectively: run a sim of an amp with varying loads, one with a completly perfect DC source, one with a bad supply and check for thd pattern. Sims ain't perfect but quite consistent for this kind of things.
Last edited:
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.