capacitance of T4 opto ?

I know many of you have played with the T4 photo coupler from the LA2A or LA3A, some have even built their own.

Now, has anybody measured the capacitance of the electroluminiscent foil in there? I've built a similar thing from a night light, and the capacitance seems very high (around 6nF). This means considerable loading, in both the tube and the transistor /transformer version. I know the control path has some frequency correction to even out the HF loss from this capacitance, but this is only a small signal correction. In the LA2A at least, the driving impedance being around 10kOhm, these 6nF would seriously limit the maximum voltage swing for high frequencies, even if the CV path has some compensation. (Slew rate limiting).
If the original has less capacitance, I can change the circuit to bhave the same with my higher capacitance (by making the driving stage "stronger"), but I have to know _how_much_ smaller the original capacitance is in comparison.

Anybody knows this?

(BTW, I had started a thread about a LA3A type step up transformer. By now, I think about going high voltage directly, instead. That would be using a 400V MOSFET for the CV path, complete with a little high voltage PSU (basically a LA2A CV path without tube), and the rest done at low voltage like the LA3A.)

JH.

just measured one of my LSI panels... 10nf ... I'll take it into work tomorrow to see how it reads on the sencore

[quote author="Scenaria"]just measured one of my LSI panels... 10nf ... I'll take it into work tomorrow to see how it reads on the sencore[/quote]


Thanks for sharing this!

Wow, that's even more than I expected.

You've measured an original T4 or something you built yourself ?

JH.

JH, have you seen my T4 page?
http://www.vacuumbrain.com/docs/t5.html
There use to be no cap in series with the panel, (inside the T4 can), now tthere is in the UA units, wich will influence the compression frequency response when compressing.

The 6AQ5 can drive a pretty heavy load, so don't worry about the EL cap bronging it down.
cj

[quote author="cjenrick"]JH, have you seen my T4 page?
http://www.vacuumbrain.com/docs/t5.html


Yes, great page! Seeing this gave me the kick to finally take this nightlight I had bought some years ago and actually build something now.


>There use to be no cap in series with the panel, (inside the T4 can), now >tthere is in the UA units, wich will influence the compression frequency >response when compressing.

Yes, but I didn't mean the extra capacitors, I meant the capacitance of the original electroluminscence foil. Just to compare if your 10nF and my 7nF are anywhere close to the original.


>The 6AQ5 can drive a pretty heavy load, so don't worry about the EL >cap bronging it down.

I'm not familiar with that specific tube (nor with the RC circuitry around this 3rd grid), but I figured it was some power pentode. I worried about the rather large anode resistor that has to charge the EL foil's capacitance every period, which becomes harder for higher frequencies.

Also, with 10nF EL element capacitance, and 10nF coupling capacitor from the anode to the EL element, that would mean the voltage is divided by 2, which seems a bit much.

I don't even know if there are different versions of this EL foil. Mine comes from a German Nachtlicht which runs from 230V. It had a 4k7 resistor in series with the EL element. (which I removed, of course.)

What is a typical voltage swing across the T4 EL element at "full compression"?

JH.

120 ac will light up the panel pretty good.
Scenaria is testing the original equipment from LSI, so your light should work fine since it matches up pretty close capacitance wise.

>Scenaria is testing the original equipment from LSI, so your light should >work fine since it matches up pretty close capacitance wise.

Ahh, I see. That's good.


>120 ac will light up the panel pretty good.

Sure, but my concern was that the LA2A driver circuit will no way create a voltage as high as 120V RMS.

120V RMS means 340Vpp. Even if the T3B driver would create a voltage swing of his full supply voltage, that would only be 275V (hard to read in the schemos). Even if this would be the pp driving voltage, that would mean 97V RMS. With the capacitive divider (10nf coupling, 10nF element) the EL element sees 49V RMS maximum. At 15kHz, even less because of slew rate limiting. (Is this right so far?)

I guess a little glow will already give considerable gain reduction via LDRs, so probably a voltage below 50V is just fine. But now I think it's actually this little glow, not bright lighting, that is to be expected. (?)

Is this completely wrong, or am I going into the right direction?

JH.

The obvious answer is that one of the 200 LA2A owners on this messageboard should fire it up and take some measurements :wink:

I guess the capacitance even varies when you cut anything off the foil. Did you already cut it, Jürgen?

Chris

Also, EL pannel is dynamic load...
:wink:

> In the LA2A at least, the driving impedance being around 10kOhm, these 6nF would seriously limit the maximum voltage swing for high frequencies,

But the light output of the EL panel increases with frequency.

So it makes sense for the drive voltage to decrease with frequency.

> I'm not familiar with {6AQ5} (nor with the RC circuitry around this 3rd grid)

6AQ5 is a small 6V6, used in many guitar amplifiers. As a loudspeaker amplifier it can make around 5 watts. In this resistance-coupled application it can make over 100V Peak. The distortion will be high, but distortion hardly matters here.

The 6AQ5 "3rd grid" is internally tied to the Cathode. I suspect you are looking at the 2nd grid, which is connected to a voltage divider. This just sets a reasonable voltage on the screen grid, to set bias and gain and make the tube work as a Pentode instead of a Triode. If you try an FET in there, you don't need the "screen grid" connection: FETs work like pentodes.

The LA2 seems to give current-drive to the EL panel. The LA3 seems to give voltage-drive to the panel. They could have ended up with the same effect by changing the low-level EQ networks in the side-chain.

[quote author="chriss"]I guess the capacitance even varies when you cut anything off the foil. Did you already cut it, Jürgen?

Chris[/quote]

Yes. 7.5nF before cut, 6nF after.

JH.

[quote author="ijr"]Also, EL pannel is dynamic load...
:wink:[/quote]

... but what exactly does this mean?
I figure the capacitive nature ist just a first order approximation.
As soon as it starts to glow, the impedance must get a real part
(in addition to the capacitance's imaginary part), I suppose.
But I have no idea to which extent. Thinking of it, I don't know
much about the panel's nature at all. Is there some good reading
about this topic?

JH.

> But the light output of the EL panel increases with frequency.

Very interesting! I wasn't aware of this. Seems I have do do some basic reading about EL elements in general.

> So it makes sense for the drive voltage to decrease with frequency.

Then yes of course this would make sense.

>The 6AQ5 "3rd grid" is internally tied to the Cathode. I suspect you are >looking at the 2nd grid, which is connected to a voltage divider. This just >sets a reasonable voltage on the screen grid, to set bias and gain and >make the tube work as a Pentode instead of a Triode. If you try an FET >in there, you don't need the "screen grid" connection: FETs work like >pentodes.

Thanks for the info. I got the basic idea. I was wondering about the details, especially about the RC network and its values on that grid. Does it bend the frequency response in any way, and if so, how?
(With tubes, I'm being let down by my #1 engineering tool, PSpice. I found a model for this tube on the web, but it's so terribly wrong that even the dc voltages are 100% off ...)

>The LA2 seems to give current-drive to the EL panel. The LA3 seems to >give voltage-drive to the panel. They could have ended up with the >same effect by changing the low-level EQ networks in the side-chain.

I see. The LA3's equalisation is quite straight forward: It's possible to set the HF boost just to give a flat small signal voltage-output response even for a truely capacitive load. I had (probably falsely) concluded that this was mainly the adjustment goal. But the more I read in this forum, the more I'm convinced it's a bit more complicated. I guess I will just provide a similar drive circuit as the LA3 (an IRF820 looks good for an output stage), and the low level CV processing will be with opamps, in order to make easily-calculated frequency response adjustments.

Some *good* Spice models would really be helpful for this.

JH.

In my LA-3A I measured the following voltages across the EL of the T4B at full peak reduction:

70vac at 20Hz
46vac at 1000Hz
122vac at 20kHz

With the HF contour set full the gain reduction peaks at 8K, then tapers off from 8khz to 16khz.