single pot lfo unvibe LDR/LED

[numpty]

I can't obtain the VT5C3/2 here  but I have some LDRs (200R to 100K) and green LED's in my junk box. I think that might work as a substitiute with the stock wah pot to give some degree of speed control. I have found 2 LDR's which match up for resistance.I might use a pair of the LED/LDR diy opto couplers using heatshrink sleeving which I have seen done somewhere else My wah pot is 92K and the voltage rail is 16v approx and I intend to use a 2n5088 or 2n3904 I don't know what the LED spec is but I think they're bog standard.My problem is that I don't know the value of Re as I don't know what the vbe should be.It may seem elementary to some but I would appreciate any help anybody can offer 

[R.G.]

Bipolar transistors always have approximately the same Vbe, which is determined by the material they're made of. For silicon, this is invariably in the range of 0.5 to 0.7V for normal currents. So pick Vbe as 0.6V and you're really close.

Re is picked based on what currents you want your LED to have. That's almost certain to be less than 20ma, because the T1 - 3/4 package gets too hot for much over that. So LEDs in the standard 5mm diameter T1-3/4 package almost universally have a max current of 20ma or less.

The current through the LED is very nearly the same as the current through Re, being different only by the very small base current. So by controlling the current in Re, we control the LED current. The current in Re is equal to the voltage across it divided by Re. And the voltage across it is the voltage on the base pin minus Vbe. That gets us back to the potentiometer.The voltage on the wiper of the potentiometer varies with pot rotation between one silicon diode drop above ground to one silicon diode drop below the power supply, given only that the base current is much, much smaller than the current through the pot. We'll come back to that.

With a 16V power supply, the max base voltage is 15.4V (approximately). So the maximum emitter voltage is about 14.8V. We want the maximum emitter current to be (for a first shot) 20ma, so we now have it - Re= (Vpower - Vsilicon-Vbe)/Iled = (16-1.2)/0.02 = 740 ohms. A 750 ohm resistor is the nearest standard value, and easy to find.

That is the *maximum* current, so that's the *minimum* resistor. You may not need the light intensity from maximum current to get your LDRs to the resistance range you need. So you may need a *bigger* Re to get a *smaller* maximum current on the LED.

Did that help?

[numpty]

Forgot to add this!
http://www.geofex.com/circuits/ldrlfo.htm
I think a standard Led is about 3volt 0.25ma and the 2n3904 is about 0.8v if 15v is the V+ so l'll go for 468R  

[numpty]

Oh no my sums are wrong I just posted that now, without seeing your post.In your calculation where did 1.2v come from?

[numpty]

I might use a 1k trimmer pot and adjust to  the correct voltage, measure then swap it out for a fixed resistor.

[R.G.]

Oh no my sums are wrong I just posted that now, without seeing your post.In your calculation where did 1.2v come from?

1. 2V is the sum of one silicon Vbe in the transistor (0.6V) and one silicon diode at the top of the pot (another 0.6V).

I would make this (15V-1.2V)/0.02 = 690 ohms. a 680 would work. Notice however, that you are unlikely to need a resistor this low.

The standard univibe circuit (and the Neovibe, which is the same in this regard) uses a 4.7K resistor in series with the pots, so there is really little need to get all the way to very small resistances. In addition, there is a 220K in parallel with the pots, and so there's no need to get to very high resistances. Most LDRs go to quite high resistances and a low resistance of 100R to a few K. You'll probably find that the usable range on the LED current to get the LDRs to go from slow to fast on the LFO speed is a small current, under 1ma, up to a few milliamps. I found this in my prototype. It depends on the specific LED and LDR.

Also, the mechanical arrangement of how much of the LED light falls on the LDRs is a big issue, as is making sure that *equal* amounts of light fall on both LDRs to keep them matched if they were matched in light versus resistance to start with.

If I were you, I'd start with 1K and go up in resistance. 4.7K might well be a decent starting point, too. Try a 5K pot for that resistor to figure out the best Re.

[cpm]

You may not need the light intensity from maximum current to get your LDRs to the resistance range you need.

very true

i see LDR's resistance is not linear at all referenced to "human-perceived" light. A very faint lighting may produce a 99% variation or resistance (eg from 1M to 10k)

then if LED's intensity also is not linealy related to its current, and current pushed by the transistor is not linear to base current wich would be what we're using to control whatever device the LDR is attached to, the you have a pretty mess of non-linearities to calibrate 

[R.G.]

i see LDR's resistance is not linear at all referenced to "human-perceived" light. A very faint lighting may produce a 99% variation or resistance (eg from 1M to 10k)

then if LED's intensity also is not linealy related to its current, and current pushed by the transistor is not linear to base current wich would be what we're using to control whatever device the LDR is attached to, the you have a pretty mess of non-linearities to calibrate 

I worried about that a lot when I first thought this up. I could not figure out whether the reverse log nature of the real pot would aid or oppose the current coefficient of the transistor and the light-to-resistance curve of the LDR, so I just hooked one up and tried it.

It's not too bad. It is neither exactly a linear nor exactly a linear-feel (that is, accurate reverse log pot characteristic) in operation. But the change in speed per pot rotation is pretty good with this scheme, at least with a VT5C3/2 opto as I tested it. It's very usable. I believe the quirks are due to the competing nonlinearities, as you note.

The worst thing is that this circuit exposes the oddity of LDRs that they tend to drift up in resistance at very dark settings. So if you have it set to very slow LFO speeds, it will slowly, over a minute or so, get even slower. This is the expression of the dark adaptation quality of the LDRs. LDRs vary in how much of this they have. The worst part of this is that the LFO may stop oscillating if the dark adaptation is too bad and the LDRs drift to a value that lets the LFO stop. This can be corrected by replacing Q11 with a low-Vgs JFET like the J201 or others, and changing the 2.2M bias resistor up to maybe 4.7M. The LFO will still run at very high resistance settings then. In practice, the parallel 220K resistors on the pots in the LFO help prevent this too. But it can happen sometimes.

[cpm]

so LDRs would have a "discharge" curve sort of like a capacitor

cool

[R.G.]

so LDRs would have a "discharge" curve sort of like a capacitor

It's more like they "relax" to a higher resistance over time for any light change. This is most pronounced in the darker region of their response. And the exact amount of "light history" depends on the composition of the LDR. The wide range ones and fast ones tend to have the most light history.

Of course...