Emulating mirror drive

[kavermei]

Hi guys,

well I'm finally finishing my full-range ESLs which have been an ongoing project for over 5 years now

At the moment they work fine, and I'm still recovering from the original shock of hearing how transparent and detailed they really are. At this moment, I don't have anything in the primary circuit. I use a line-level baxandall shelving circuit to augment the bass and this arrangement takes out my amp during bass-heavy passages

I'm now at the stage where I am experimenting with the primary circuit, in an attempt to make the speakers easier to drive. One idea I will be trying soon is an emulated mirror drive. My transformers have dual primary windings and I have the following idea to eliminate the autotransformer that is used in the Audiostatic mirror drives: see this quick & dirty figure...



Have any of you tried something similar before? I can't see why it would not work... Can you think of possible shortcomings?

Thanks
Kenneth

[bolserst]

Originally Posted by kavermei

I'm now at the stage where I am experimenting with the primary circuit, in an attempt to make the speakers easier to drive. One idea I will be trying soon is an emulated mirror drive. My transformers have dual primary windings and I have the following idea to eliminate the autotransformer that is used in the Audiostatic mirror drives: see this quick & dirty figure...



Have any of you tried something similar before? I can't see why it would not work... Can you think of possible shortcomings?

Thanks
Kenneth

Yes, this should work. You will probably need an additional resistor in series with the C1 to damp out the high-Q resonance between the transformer leakage inductance and panel capacitance at the top end of the audio band. Alternatively, I guess you could put the resistance in the HV circuit wires going to the panel since R1 will already protect your amplifier in the event of core saturation.

Speaking of core saturation, one possible shortcoming is limited dynamic range in the LF as the core may saturate. What is the core size and the number of turns on the primaries? The Audiostatic mirror-drive transformer avoids this problem by using a relatively large number of turns on the auto-transformer.

Another possible issue is exceeding the breakdown voltage of the air between the stators with the additional 2x step up used in LF range. What is the step-up ratio? what stator-diaphragm spacing are you using? 4kV/mm is a good rule of thumb for when air starts breaking down and your music starts sounding fuzzy. I ran into this problem when trying a similar EQ technique. I have found that adding active diaphragm area is the best way to equalize ESLs if you want to keep high-SPL capability. If you are running your panel well below the air breakdown voltage, then this shouldn't be a problem.

[bolserst]

I recently ran across a US patent application for a circuit to compensate for the front to back dipole cancellation of ESLs that does not use an additional transformer like that Audiostatic mirror-drive patent.

Very little detail is given concerning the values of the different circuit components. The description of how the circuit works sounds plausible. But, I have done a bit of modeling and have yet to come up with a set of component values that produces the compensation the patent claims. I'm beginning to get a bit skeptical...

Has anybody else seen this circuit before?

[kavermei]

Hi bolserst, all,

I appreciate your thoughts. First off let me say I have made pretty big panels, about 0.5m^2 active area, wire stators, 3mm D/S spacing. So by and large, I needn't worry about overvoltage, I'm confident my amp's protection circuit will trigger first, heh.

The transformer consists simply of (2) 230V:12V+12V toroidal power supply trannies. The 230V windings are put in series so I can basically choose between 1:19 or 1:38 (ignoring losses). I know, it's low, but I plan on making a high-rail voltage amp later on

I must say that since I added electrical segmentation, the tonal balance has improved a lot! I now omit the line-level shelf circuit.

During the last couple of hours, I've been experimenting with different primary circuits. When I try the above mirror drive idea, something sounds wrong and distorted in the bass -- like that resistor is doing something funny to the sound... More experimentation is definitely needed, it seems. For the time being, the panels sound best with the primaries in parallel (1:38 ratio). Just a bit trebly.

Concerning that patent: never seen the circuit before, but it looks like a dud to me... Statement [0010] sounds strange, there is no mention of component values (not even as an example), and the MOV stuff seems thrown in out of the blue, probably to give the patent some extra weight!

[bolserst]

Originally Posted by kavermei

Hi bolserst, all,

I appreciate your thoughts. First off let me say I have made pretty big panels, about 0.5m^2 active area, wire stators, 3mm D/S spacing. So by and large, I needn't worry about overvoltage, I'm confident my amp's protection circuit will trigger first, heh.

The transformer consists simply of (2) 230V:12V+12V toroidal power supply trannies. The 230V windings are put in series so I can basically choose between 1:19 or 1:38 (ignoring losses). I know, it's low, but I plan on making a high-rail voltage amp later on

I must say that since I added electrical segmentation, the tonal balance has improved a lot! I now omit the line-level shelf circuit.

During the last couple of hours, I've been experimenting with different primary circuits. When I try the above mirror drive idea, something sounds wrong and distorted in the bass -- like that resistor is doing something funny to the sound... More experimentation is definitely needed, it seems. For the time being, the panels sound best with the primaries in parallel (1:38 ratio). Just a bit trebly.

Concerning that patent: never seen the circuit before, but it looks like a dud to me... Statement [0010] sounds strange, there is no mention of component values (not even as an example), and the MOV stuff seems thrown in out of the blue, probably to give the patent some extra weight!

As you say, with 3mm spacing and only 1:38 step up ratio, you are no where near the voltage gradient at which air starts to break down.

You didn't mention what the core size was, and number of primary turns. If you don't know, the wattage of the transformers would give you some indication of how hard and/or low you can drive them before you start getting distortion.

What value of series resistor were you using when you noticed the bass sounding distorted? Distortion in the transformer as saturation is approached manifests itself as the current(spikes) not following the input voltage. As the series resistance gets larger, this distorted current produces a larger and larger distortion voltage across the resistor which is passed on to the secondary and the panel. Perhaps this is what you are hearing.

I know Calvin is familiar with the limitations of using toroidal power transformers for driving ESLs. Perhaps he might be able to help you determine if your transformers are capable of being used full range.

[kavermei]

Hi,

Originally Posted by bolserst

As you say, with 3mm spacing and only 1:38 step up ratio, you are no where near the voltage gradient at which air starts to break down.

Absolutely! I need a bigger amp capable of higher voltages

You didn't mention what the core size was, and number of primary turns. If you don't know, the wattage of the transformers would give you some indication of how hard and/or low you can drive them before you start getting distortion.

I couldn't find 'em with dual 115V+115V windings, so I use two toroids of each 230V:12V+12V, 200VA. So total core size, so to speak, is 400VA. I don't know the number of turns on them, but, for power toroids it's typically very low.

Actually, I've been experimenting all weekend, and got some nice progress. Instead of the mirror drive, I now have inserted a line-level equalizer to flatten the response. Just running the primaries in parallel... It's unbelievable how low they go!! (Though not at earth-shaking sound pressures.) I think it helps that my panels are 24" wide

Also I was shocked to find that my EL84 push-pull amp (15W maybe?) drives these panels to pretty good loudness levels! This is really unbelievable, with a big full range panel with 3mm D/S, go figure!

Okay, really loud bass is a problem, then the ESL trannies saturate. I don't want to roll off the bass too much, however, because at reasonable levels, it goes so satisfyingly low! I'll just need to use discipline when turning the volume knob, or maybe shell out big $$$ for a better tranny.

What value of series resistor were you using when you noticed the bass sounding distorted? Distortion in the transformer as saturation is approached manifests itself as the current(spikes) not following the input voltage. As the series resistance gets larger, this distorted current produces a larger and larger distortion voltage across the resistor which is passed on to the secondary and the panel. Perhaps this is what you are hearing.

It was 3.3 ohms, just sth. I had lying around. I'll try to do some 'scope measurements next weekend! Thanks!

I know Calvin is familiar with the limitations of using toroidal power transformers for driving ESLs. Perhaps he might be able to help you determine if your transformers are capable of being used full range.

All info is much welcomed, but as above, I don't know the # of turns on my transformers, just that they're 2x200VA...

Kenneth

[bolserst]

Originally Posted by kavermei

It was 3.3 ohms, just sth. I had lying around. I'll try to do some 'scope measurements next weekend! Thanks!


All info is much welcomed, but as above, I don't know the # of turns on my transformers, just that they're 2x200VA...

Kenneth

I would think a 400VA transformer would have a core large enough for full range use. However, the primaries may not have enough turns. Most full range ESL transformers have at least 50-60 turn primaries to keep the primary inductance up and keep the core from saturating. An easy way to determine the number of turns in you primaries is to wind 10 turns around the core. Feed 1.0 VAC 1kHz signal in to the 10 turn winding and then measure the the voltage output on the primaries. The number of turns can then be calculated as 10 x Voutput.

If you can hook your scope up across the series input resistor it is quite interesting to watch the sine wave distort as core saturation is approached.

[bear]

Anyone point me to this "mirror drive"??

The circuit presented looks a lot like what Audiostatic was touting early on in their literature.

My friend who owns Audiostatics and has opened up the interfaces has found nothing other than a straight single transformer driving the panels... the roll off for the bass panels is achieved by series resistors. Nothing exotic.

I'd consider using the patented Acoustat method myself...

Power transformers tend to have rotten HF performance... be forewarned.

_-_-bear

[kavermei]

Hi bear,

look up US Patent 4,461,931. Basically, they use an autotransformer and RC circuit to step up the bass a factor of two, with regards to the treble, before sending it to the step-up transformer.

Not all Audiostatic models have it, but I don't know which do and which don't -- others here will know that no doubt.

I will look up what Acoustat did, thanks!

I know power transformers are far from the ideal solution, it was mainly a budget call. Instead of getting more expensive ones later, I'm contemplating going direct drive instead (I'm not kidding! Seeing as how little voltage they need to give sound pressure levels which are acceptable to me...)

[bear]

If that is the Audiostatic patent, I have seen it... which is what i referred to in my post, indirectly.

the patent is likely under Jim Strickland's name... a better solution imo, and of course no one else could do it after he patented it.

_-_-bear