[sdiy] Re: diodes/LDR

[Scott Bernardi]

1MA probably means 1 milliamp (mA).

The circuit around U2A is actually a voltage controlled current source. The
inverting input is a virtual ground, so the voltage coming off the RV3 pot and the
RV2 pot are converted to currents through R32 and R33. Say the voltage at K8 is at
-1v, you'd have -1v/1M or 1uA being pulled out of the vitual ground. Since the
opamp input draws negligible current, this current flows through R31 and produces a
voltage drop 1uA*1Mohm = 1v. The polarity is such that the voltage at the junction
of R31 + D1 + R30 is +1v.  D1 is reverse biased so you can treat it like it is not
there.
So we have +1v at the top of R30, meaning the current through R30 is 1v/100ohm =
10mA.  Most of that 10mA (all but the 1uA through R31) goes through Q1's emitter,
and the collector current is approximately the same. So basically as you vary the
voltage at K8 or K9, is is converted to a current by R30 and is pulled through the
LED via the transistor.
The configuration of the RV3 dual pot gives an approximation of a reverse log pot.
R34 and R35 were chosen experimentally to set the
top and bottom of the current range.
When you get near the low current/dark end of the LDR range, it becomes more
sensitive (i.e., small changes in current produce large changes in resistance). The
reverse log response at K8 gives greater control at the long echo end.

Charles Twist wrote:

> Hello,
> Thanks for the answer.
> I once  had the idea of a delay that would span 4 orders of magnitude from
> 0.2 msec to 2 sec, then I realise it was basically unfeasible or that I was
> just as well off to buy  a ready-made unit like the Ibanez AD202. So I
> abandoned the flanger/chorus idea and concentrated on the delays which has
> more fun potential for me. If it runs from 30 msecs to 1 sec and doesn't
> cost megabucks, then I understand that language. I like the fact that in
> this design the feedback is explicit and accessible. So I am keen to give
> this one a go. The Vactrol runs to 5K5 at 1MA (any idea what an MA is?). It
> sounds like it could go to low enough resistance values; what about the
> other end of the spectrum and long delays? I am not sure whether these
> things fade to darkness or suddenly go dark at one point.
> As you say these are cheap and are a lot easier than trying it myself.
> To change the currant flowing through it, do you change the resistance in
> the feedback of the opamp R31? Or the resistor to ground R30?
> Thanks,
> Charles
>
> At 05:16 10/01/2003 -0800, you wrote:
> >The key to the LDR to give low resistance is to run it at pretty high current.
> >My circuit with the PT2399 has an upper limit of about 35mA for the diode
> >current.  The CDS cell I used was just some I got from Electronic Goldmine,
> >however, their website is showing them as sold out. In the combination of
> >yellow LED and CDS cell I used, the minimum resistance was about 700 ohms.
> >A Vactrol (encapsulated LED and photoresistor) is a better idea. BG Micro
> >(www.bgmicro.com) has them real cheap. Part #ICSVTL5LC2 on page 8, 2 for $1.
> >Small Bear Electronics that carries the Princeton echo chips also has
> >vactrols,
> >but for much much more (like $7-$8 each).
> >You are not going to get very short delays (like for doing flangers or chorus)
> >out of the Princeton chip. Even with 0 resistance at the tuning pin, the delay
> >time is too long. You can get a pretty decent slapback echo, though.
> >I was actually more interested in some of the longer delays. This is a bit
> >more
> >tricky because the dark resistance is harder to control. For one thing,
> >sealing
> >up your LED + photoresitor is hard to do completely - ambient light gets
> >in and
> >affects the longer delay.
> >The PT2399 is only spec'd to about 300mSec, but I got it to give almost a
> >second delay with acceptable distortion.  I also picked up a PT2395 which is
> >spec'd to go to 800mSec with an external 256K DRAM chip. I'm going to try it
> >out with a Vactrol.
> >
> >
> >Charles Twist wrote:
> >
> > > Hello,
> > > My name is Charles Twist. I am in the process of putting together a few
> > > circuits for a home-made modular. I was attracted to your delay module as
> > > it is one of the few out in webland. I have no trouble understanding the
> > > various parts, but I am uncertain about the diode/LDR set-up. My main
> > > problem is knowing exactly how much light diodes put out. Also I do not
> > > understand the lux and ftc business on the LDR front. As I see it, we want
> > > the LDR resistance to vary from 1 to 100 K. Not many cover that range; most
> > > stop at 4K which is a shame as it does not allow short reverbs. I could go
> > > for a lower LDR resistance, then the range is 1-40K for 100-1ftc. Is there
> > > a diode that will function well at low light levels, ie under 1 ftc or 10
> > > lux, whatver they may be (giving stable delay times)?
> > > Any help on this matter will be greatly appreciated.
> > > Thanks,
> > > Charles