LA2 Time Constant?

I never really pondered this RC thing until I redrew the schematic.
Is this a time constant thing, aka Fairchild 670 etc?
If so, I'm thinkin rotary sw if you catch my drift.

Looks more like a shelving network to me.

Hello,

the high-frequency attenuation in the side-chain is required in EL panel-driven opto-compressors, because the brightness of the electroluminescent panel (known also as "the electroluminescent capacitor"), and hence the gain reduction, depends not only on the voltage, but also on the frequency. The higher frequency, the brighter panel shines. So to avoid over-sensitivity to high frequencies we have to attenuate them.

Hmm, maybe. But if that's the case, why not just use a smaller cap without the resistor?

Maybe it's something to do with keeping two units from running too far apart in stereo mode.

If thats the case, we can rip out two more parts. Easy to test the effects of that rc on a live unit. Just one snip.

A single cap would attenuate indefinitely with increasing frequency. This probably is a decent match for flattening the EL characteristic as suggested by Lukpio.

But you, arch-snipper and all-around deconstructionist, are welcome to snip away at will :green:

OK.

> Is this a time constant thing?

The time-constant you are thinking of has to be in the "DC" section, after the signal rectifier.

The LA2 has no rectifier (or rather: the "rectifier" effect is in the lamp, which makes the same light for electricity in either direction).

Snip the cap off the 6V6 cathode so the 6V6 runs as a cathode-follower, connect that point through a big film-cap to the LineIn of a flat amp and speaker. Listen. The threshold knob adjusts the level into the 6V6. The 47K and cap add about 6dB of bass-boost (actually mid-high shelf-down). The 1Meg pot (not shown in your plan) and 0.001uFd add some bass-boost too, user-twistable. The 0.03uFd in the first cathode boost above about 10KHz.

> high-frequency attenuation in the side-chain is required in EL panel-driven opto-compressors, because .... The higher frequency, the brighter panel shines.

True, but the interstage network is pretty flat, mostly bass-bump.

Actually, I think the higher=brighter is for constant-voltage testing.

The EL panel is pretty-near a pure capacitor with negligible losses (those losses may be the real work done), right? What is the capacitance of the EL panel? Seems to me it is tenths of a uFd. The output of the 6V6 is >100K||22K= 20K. Any plausible value for EL capacitance leads to a slant-line voltage and flat current over most of the audio band. If we assume that light is proportioal to current, then the output stage is nearly flat above 100Hz, the interstage kinks the response in the bass to meet the R-C curve of the 6V6+20K+EL circuit, +/- the time-response of the photo-resistor (which has internal time-constants in audio bass).

Great! Now I'm gettin it.
Time constant, doh, somebody spiked my corn flakes.

OK, 9 caps involved in driving that EL panel if you count the panel and the cap in the T4 that UA sells.
Sheesh. Someone was doing some tweaking.

That 47k/0.01 network does slope down from about 10k, which means less bass hitting the panel, which means less bass getting grabbed off the input signal.

Here is what the Spice Girls did to a simulation that did not include the tubes or their cathode resistors. The big dipper is with the R37 pot set for 1 Meg.
The blue line is with the pot off.
So pot set to 1 Meg = Bass Dip To EL Panel = More Bass at the output
Green Curve

Pot set to Zero Ohms = Blue Curve = Less Bass Dip = More Bass at the output:



So what the heck is that Red curve? It's caused by this trimmer cap on a later schematic. Set to 380 pf, it pulls down the graph a bit arounfd the 10K mark.(see above graph)



I thought I remebered something about a futile attempt to try and get the imaging to work better on stereo LA2 compressors by tweaking this cap, but I looked all through my old docs, and couldn't find anything. The trimmer does not do anything when the R37 pot is set to zero ohms.

FYI:
Here is a graph showing the pot turnned to both extremes, with and without the 47k/0.01 network.

PRR made some really good points about EL panels in the LA-2. But there are a few more quirky things about ELs in compressors that might be pertinent to this discussion of the frequency response shaping business in the sidechain of the LA-2.


Namely, a peculiar characteristic of an electro-luminescent panel is that it also changes colour with input frequency: it starts off emitting a yellowish colour at 30Hz, and the color gradually changes progressively as the drive frequency goes up through green, blue, and on to purple by 20kHz. I don't know if it is putting out UV as well with high frequency drive or not - I'm not that knowledgeable about it - I've just played around and noted the color change. But it sure is purple. And quite bright, too, but I was using constant voltage drive, not constant current, which might make quite a difference as PRR point out.

So what is my point about this color business? Simply that many photocell materials have different sensitivity curves to different wavelengths of light (just to make the transfer function a bit more complex - ugh).

I don't know if anybody with a real LA-2 has ever actually mapped the composite frequency to resistance change characteristics or not, using the whole sidechain, electronic frequency response shaping and all. I'm sure it would be interesting. I'd like to see it if someone ever has gone to the trouble.

And because LA-2 compressor sidechains aren't normally driven with sine-wave or pink noise inputs, I expect the light spectrum the EL puts out is rather complex spectrally when driven with real audio signals, which in turn I expect would probably affect the photocell response with it's differential response to various light spectra, time constants to various photo-stimulation frequencies, and so on.

FWIW