LDR vs. Transcondution Question

[Pablo1234]

What are the pros/cons between an LDR and a transcendence amp?
I have a small stone phaser that I love but I hate the noise, is this from the Amps or the all pass topologie? Would LDR's be quieter/noisier?

I currently have designed a programmable LFO/Envelope Follower/Expression system and have plans for Comp, Wha, Phase, and Chorus(x2) but with all the work I have put into it I would really like to have the quietest, easiest FX to adapt them too.

The Comp has more programing involved with Threshold, Attack, Level and Curve shape so this is where I am starting. I know LDR's have an inherent Attack time of greater than 50 ms but for guitar I think that is quick enough. I may be wrong though so any insight would be helpfull.

[petemoore]

What are the pros/cons between an LDR and a transcendence amp?
  OTA? [operational transductance amplifier?]
  LDR needs a light shield...and a bit more room on or over the board, they both have their applications and layouts/circuit topologies, sometimes a ''peripheral parts'' [components needed to make the chip operate] count matters, or the sound of one fits the position on a pedalboard or an instruments recording track.
I have a small stone phaser that I love but I hate the noise, is this from the Amps or the all pass topologie? Would LDR's be quieter/noisier?
  As a general rule, audio chip design puts low noise on the 'front burner' of characteristics. The 'goodness' of the power supply, the lack of a ground loop...the 'peripherals'...what makes the chip operational may contribute to noise more than the chip.
  Still, there are numerous ''lowest noise phaser''..s:
  http://www.bing.com/search?q=lowest+noise+phaser&form=AARTDF&pc=MAAR&src=IE-SearchBox
I currently have designed a programmable LFO/Envelope Follower/Expression system and have plans for Comp, Wha, Phase, and Chorus(x2) but with all the work I have put into it I would really like to have the quietest, easiest FX to adapt them too.
  The Comp has more programing involved with Threshold, Attack, Level and Curve shape so this is where I am starting. I know LDR's have an inherent Attack time of greater than 50 ms but for guitar I think that is quick enough. I may be wrong though so any insight would be helpfull.

[petemoore]

The Comp has more programing involved with Threshold, Attack, Level and Curve shape so this is where I am starting. I know LDR's have an inherent Attack time of greater than 50 ms but for guitar I think that is quick enough. I may be wrong though so any insight would be helpfull.
  The LDR's will be as fast as they are/no more. Some are quicker than others.
  The phaser sweep circuits [all varieties] are outside of the audio path, though the power consumption can be 'lumpy' enough to get into the audio path through the supply, good layout/filtering/etc. can eliminate it from being percieved as noise, click or thump in the output.
   SS can be somewhat hissy, though noise in audio comes from 'mysterious and numerous' sources...until the mystery is solved and the process of elimination found and eliminated all of the noise sources, be they multi-faceted, dependant on other equipment or even 'intermittently-temperamental'.
  No ground loops, ripple-free DC supplies, and excellent signal routing/layouts is a great place for low noise circuits to try to pop up.

[Rob Strand]

It comes down to specifics but it's probably easier (and cheaper) to get low noise using an LDR/opto-coupler.   

The older transconductance devices can be noisy.  You really have to tune the operating levels to maximize dynamic range.   Another trick is to use pre-emphasis and de-emphasis - that is what the dynacomp uses.

A given design can give the impression of high noise if the gain at low input signals is too high.  You can either limit the maximum gain or not compress at tiny inputs.  Studio style compressors have threshold setting.  Minimal compression occurs for signals below the threshold.

> I know LDR's have an inherent Attack time of greater than 50 ms but for guitar I think that is quick enough.

That's not entirely true the response of LDRs (both attack and release) depends on the specific model of LDR/opto-coupler.  You can get fast attack models, the release is usually controlled by the filter cap in the compressor's rectifier.  The use of a feed-forward or feedback compressor can also affect the time constants. 

[Pablo1234]

I'm gona start with the LDR I think, I have an interesting idea for the feedback loop to maintain gain while changing ratio.

I have 2 ADC inputs for the envelope, one pre-Comp and one post-comp, this will give me the ability to mix or average them between feed-forward and feedback styles. I just found 2 OTA's in my garage so I will give them a shot also, their from an old DoD FX-17 wha pedal I always hated. I just think the OTA's will require more programing in the end though so they may just get put off to the side till I am complete with this system.
The attack and release will be in software control, I am doing an active full wave rectifier for 70 - 300 hz, 300 - 800 hz and 800 - 3k each and mixing them to rms then summing them into an adc pre and post comp then outputting a current controlled signal to the comp device itself.
The rectifiers and micro work really well but I have spent so much time on it I really want to get the compression right first time around.

Thanks for the advice.

[PRR]

What you want is the ratio of hiss to maximum undistorted output.

In a compressor, both limits change with compression action. It is a mind-puzzle.

Maximum input to a OTA is 20mV-100mV depending on details. That's quite small. So even though the hiss level is actually not high, the low maximum input means you usually need lots of gain

Photoresistors will take about a volt of signal really clean, and many volts not too dirty. The hiss level is higher, and highest at no-compression, but you can run large levels and drown it out.

Matchbook figuring says the S/N isn't very different, OTA or opto, but the opto will need larger signal level to be optimum. Indeed when you throw in R-FET and tube limiters, none of them have outstanding S/N ratios unless you go to heroic (expensive) designs (eight dual-triodes just for VCA, plus a 10-Watt sidechain).

[Pablo1234]

CA3080 says rail to rail
Supply Voltage (Between V+ and V- Terminal) . . . . . . . . . . . . . 36V
Differential Input Voltage . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5V
Input Voltage . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . V+ to VInput

Did you mean the VCA(JFET) has low input voltage? I personaly hate VCA designes, thier too much a pain imo. I use j111 for switching and I will use them occasionally for an amp but that's about it.

[PRR]

> CA3080 says rail to rail

That's damage rating.

'3080 is a simple LM3700. Same input stage. Bare-naked BJTs obey natural law. Look at the distortion and noise plots:




> I personaly hate VCA designes, thier...

"VCA" is the _general_ term for Voltage Controlled Amplifier/Attenuator.

If I train my dog (ha!) to move a volume slider according to a voltage she feels, that's a VCA. While the dog VCA may be impractical, there is a class of motorized potentiometers which work the same (with less shedding).

> Did you mean the VCA(JFET) has low input voltage?

The "linear resistance" zone of JFETs is Vp-Vgs. In an attenuator, the least-attenuation condition has Vgs nearly equal to Vp. The "linear zone" becomes very small, a few mV.

National did not re-publish that paper on the WWW. The old FET book has other data but my copy isn't handy.

Nevertheless, very-fine results are possible with a JFET: 1176LN.

Silonex sells LDRs but has a nearly-fair summary of the usual audio VCAs:

http://www.silonex.com/audiohm/levelcontrol.html  

[Pablo1234]

I meant VCR

[Pablo1234]

what datasheet do you have, my datasheet is lacking distortion vs dif input V