Bring me up to speed on step-up transformers.

[ak_47_boy]

I have been haveing a bit of trouble understanding a few things.

1.
What is the average voltage output of tube amps?

2.
There are two ways to step up an audio signal right?

Are my diagrams correct?

3. Assuming my amp outputs 40V:
I need 12k volts for my bass panel, this would require a 1:300 ratio?
I need 6k volts for mid, this would require a 1:150 ratio?
I need 2k volts for high, this would require a 1:50 ratio?

4.
Can i use neon sign transformers? 1:250 is going to be hard to get otherwise.

5.
Will i notice the sound quality difference between option 1 and 2?

6.
Does all this sound right?
12k volts for bass (10mm spaceing)
5k volts for mid (5mm spaceing)
2k volts for high (1.5mm spaceing)


thanks for so much help

[maudio]

Originally posted by ak_47_boy
I have been haveing a bit of trouble understanding a few things.

1.
What is the average voltage output of tube amps?

Before the output trannie: depends on the supply voltage and saturation voltages in the tubes. A few hundred V. Way too small to direct drive an esl.

After: P = U^2/ R. Around 28Vrms (40Vtt) for 100W in 8 ohm.

2.
There are two ways to step up an audio signal right?

Are my diagrams correct?

Yes, both will work. When you connect the windings with correct polarity

3. Assuming my amp outputs 40V:
I need 12k volts for my bass panel, this would require a 1:300 ratio?
I need 6k volts for mid, this would require a 1:150 ratio?
I need 2k volts for high, this would require a 1:50 ratio?

Yes. Assuming 1:150 means that 1V in = 150Vout measured between the outer taps of the sec (there is sometimes some confusion about this).

4.
Can i use neon sign transformers? 1:250 is going to be hard to get otherwise.

Nop. Neon transformers are designed to have very poor voltage regulation so that the current is limited. Which is done by designing the trannie with an air gap in the core, resulting in high leakage inductance. Therefore bandwidth will be very poor.
Look at it this way: even at 50 or 60Hz it's unable to sustain full current into the load.
1:250 is not a good idea anyway. Unless you plan to use this for bass only. Very hard to get decent bandwidth with such high stepup. Will not sound very good on fullrange.

5.
Will i notice the sound quality difference between option 1 and 2?

You mean between 1 and 2 transformers? Depends on more things.
The one with 2 seperate transformers has the advantage of lower stepup per transfo (good for bandwidth at the high end). This way you can build a 1:150 using 2 1:75's.
The disadvantage (sure there is one ) ) is lower primary inductance (bad for bandwidth at the low end) as both primaries are paralleled while not sharing the same magnetic circuit. To compensate you need more turns, which partly cancels out the advantage mentioned. Still the 2-transfo option is better in theory. Even better is to combine two 1:75 coils on one core.

6.
Does all this sound right?
12k volts for bass (8mm spaceing)
6k volts for mid (4mm spaceing)
2k volts for high (1mm spaceing)

Doubling the spacing requires 4 times the voltage. Do the math... Also, more spacing needs more bias.
I would say that any spacing above 3 mm (stator to membrame) is pointless as you will never be able to drive the membrame that far. The max drive voltage is limited by real-world transformers, isolation and amplifiers.
Spacings below 1 mm are impractical because of build tolerances.

thanks for so much help

You're welcome

[ak_47_boy]

Wow, thanks. I feel much smarter now
So:

12k volts for bass (4mm spaceing)
8k volts for mid (3mm spaceing)
2k volts for high (1mm spaceing)

[maudio]

It's difficult to predict how much voltage is actually needed but my guess is that you are a bit too optimistic.

Some real life examples I know of:
- Old Audiostatics (2.9 mm spacing) have a 1:150 stepup that is doubled to 1:300 for low freqs by an additionial autotransformer. Gives around 12kV tt drive voltage (100W amp).
- Quad's 63 has 1:240 stepup and 2.6 mm spacing. More or less the same numbers... The 55 uses very high stepup for bass (1:290 if I'm correct) and very small spacing (2mm?), but it was designed for a 15W amp.
- My own panels have 2 mm spacing and will just about reach full stroke with a 1:120 on my 55Vtt amp. By the way 2mm is enough to play down to 80 Hz very loud (0.3m^2 area), no need to go wild here.

Anyone have more experience to add to this list?

Looking at these numbers I'd say that you need well over 20kV drive for 4 mm spacing. You'll find it very difficult to find a transformer that will deliver that much. Even if you do find it, something is bound to break down very soon (transformer or panel).
If you want more dB's better go for more surface area. Or use a dynamic sub.

Why do you want to use different panels for high and mids?

[maudio]

Quoting myself...
Doubling the spacing requires 4 times the voltage.

I just realized that I overlooked something here. It should read: doubling the spacing requires 4 times the voltage to reach the same force on the membrame. Which makes things even worse.

So, when a 2 mm spacing panel needs 5kV for full stroke, a 4mm panel would need 20kV for 2mm stroke. For full stroke it would need 40kV.

Please correct me if I'm wrong

[maudio]

That is of course, assuming the same bias.

When you double the bias (which you can at double spacing), you're back at 20kV.

Drive force is proportional to Vdrive * Vbias / d^2

[invisible force]

Hello maudio,

one question is haunting me for a long time regarding basic ESL operation :
leaving step up transformer aside, how is acoustic output translated in a parallel resistor across the plates (otherwise there only would be blind power) ?
regards, Philipp

[maudio]

Of course there is a factor representing the actual acoustic output. But from what I remember from all theory, the actual acoustic loading is quite a complex story, resulting in pages of complex mathematical formulas . One of the problems being that it is very difficult to predict the acoustical impedance of the esl. For sure it can't be modelled by something as simple as a resistor...

But the good news (that I do remember ) is that you can forget about it, as it is completely neglectible. The 'blind power' swamps it all.

I have some literature about it somewhere, have to dig it up..

[ak_47_boy]

Originally posted by maudio
Why do you want to use different panels for high and mids? [/B]

I am going to triamp. All 3 panels will be one piece.

If i want to build a accustat 2+2 sized esl would that mean there is no need for a large stator spaceing? I am going to make a diagram of what i was planning to build.

On this site there is a small stator spaceing/voltage chart.
http://amasci.com/esloud/eslhwto.html
Is is wrong right?

[invisible force]

Hello maudio,

I am interested in the literature..