LDRs as pots?

[tbryanh]

Originally Posted by Gregg

I don't have the "damping" filter because I don't have any idea how to determine those values.

The patent explains that the damping filter is to compensate for the delay in the resistor part of the LDR responding to a change in light intensity from the LED part of the LDR.

Without the damping filter, the LED will oscillate between full on and full off.

I am guessing that all you need to do is view the signal to the LED with an O'scope and use a capacitor substitution box to dial in the smallest value capacitor that stops the oscillation.

You might want to increase the value a little past that point, but probably not much.

The patent explains that you want the delay from the capacitor to equal the delay from the LDR.

Using too large of a capacitor would probably cause an oscillation too.

In summary, If the op-amp responds too quickly or too slowly, an oscillation will probably occur.

[tbryanh]

On second thought, too large of a capacitor probably will not cause an oscillation.

Probably what will happen is the response time will be slower than what it could be; i.e. slower than the response time of the LDR.

So, to summarize again.

Using an oscilloscope and capacitor substitution box, dial in the smallest value capacitor that stops the oscillation.

Then increase the value a little to insure the oscillation does not happen when temperatures change and devices age.

If you prefere a slower response time, increase the value of the capacitor until the slower response time is reached.

[tbryanh]

To clear things up a little more.

The issue of finding the smallest value capacitor refers to the LDR Variable Resistor http://subdomain.freeservers.com/sch...e_Resistor.JPG

For the LDR Pot http://subdomain.freeservers.com/sch...t_Touchup2.jpg an RC network is used. This might be a little more comlex to find correct values of resistor and capacitor.

It might be that a simple capacitor can be used for the LDR Pot instead of the RC network.

So now the question is, why was a simple capacitor used for the LDR Variable Resistor and an RC network used for the LDR Pot?

The reason the resistor is there might be to set the gain of the op amp to scale the range of the LDR Pot.

[tbryanh]

Sorry, the URLs got parsed. Here they are again.

http://subdomain.freeservers.com/schematics/LDR_Variable_Resistor.JPG

http://subdomain.freeservers.com/schematics/LDR_Pot_Touchup2.jpg

[tbryanh]

I was able to upload the LDR Variable Resistor and the LDR Pot today.

[tbryanh]

Here is a touched up version of the LDR Variable Resistor.

[tbryanh]

For the LDR Pot (Figure 3), I am thinking that maybe the feedback resistor and capacitor should be in series rather than in parallel. (See items 26 and 28 and items 27 and 29).

I am just guessing, but I was looking at some other op amp circuit for a power supply regulator, and it showed the capacitor and resistor in series for what appears to be a similar situation.

[Don Moose]

Originally Posted by tbryanh

For the LDR Pot (Figure 3), I am thinking that maybe the feedback resistor and capacitor should be in series rather than in parallel. (See items 26 and 28 and items 27 and 29).

I am just guessing, but I was looking at some other op amp circuit for a power supply regulator, and it showed the capacitor and resistor in series for what appears to be a similar situation.

Certainly seems reasonable - like I said, I'm out of my depth there.

Do you have a SPICE program you like? (SuperSpice by Kevin Aylward is a pretty nice one) Since I move to Mac, I haven't found one that really works for me.

[tbryanh]

I never used SPICE. I went to the website and checked it out. A circuit simulator.

http://bwrc.eecs.berkeley.edu/Classes/IcBook/SPICE/

[tbryanh]

Originally Posted by tbryanh

For the LDR Pot (Figure 3), I am thinking that maybe the feedback resistor and capacitor should be in series rather than in parallel. (See items 26 and 28 and items 27 and 29).

I am just guessing, but I was looking at some other op amp circuit for a power supply regulator, and it showed the capacitor and resistor in series for what appears to be a similar situation.

Now that I understand the other op amp circuit better, I think that for the LDR Pot, the feedback resistor and capacitor should be in parallel as shown. this setup should be added to the LDR Variable resistor too.

[Gregg]

Hi,

After I built that LDR pot /answer #63/ for testing purposes and did some DMM measurements to see if it's going to behave like a real pot I finally found some time to test it in a real tube amp circuit. I connected it as Gain pot in a typical SLO100/Mesa DR type of high gain preamp. The result? It works. First I didn't like the response of the pot but then realized it was a linear one. Changed it to log pot and it worked OK. It was set to be a 1M pot. At center position it reads 102k which means it's a 10% taper log pot. I'm planning to run some measurement to see how close to a log pot this LDR pot behaves but that will be just out of curiousity because I'm satisfied with it's response. After all this is a guitar amp not a Hi-Fi project.
I ran the circuit without the time correction filter in the feedback loop of the opamp without any side effects that I could notice. No delay, no oscillation or anything else that I could hear. I will test it in the clean channel as well to make sure no side efects are present.
For this test I used Perkin Elmer's VT935 series of photocells which along with other types I got from Allied. There are other better candidates for this job to be tested.
Because when ON these photocells measure 1.5-2k there's still some signal leaking out when pot is at zero position. This can be fixed by using in the lower position other type of photocells with lower ON resistance - some Vactrols for example.
However I'm having some doubts about using these LDR pots in a regular Marshall type tone stack. Although they are rated at 100V peak I'm not sure how will they handle the voltages in a real situation. Mesa used them after the first stage where the voltage swing is not high unlike the Marshall TS where I measured ~60Volts across the MID pot /when fully open/. I gues the best way to find out is to use them in a real TS.

[tbryanh]

Excellent job. I would like to duplicate what you are doing. If you didn't mind posting a schematic and parts list, that would be cool.

Thanks

[Gregg]

This is how the current schematic looks like:

[tbryanh]

Thanks for posting the circuit. Very brilliant. I’ll be spending some time to fully understand it.

I was wondering if you had a split supply (+ 15v and –15v) for the Op Amps, or are you using a single-ended (+ 15v only) supply?

I am curious what you are using to draw the diagram? It looks real good.

I modified your circuit a little bit (see attached) to show the idea of using the Perkin's Elmer VTL5C Series Vactrols. The series has Vactrols with on-resistances that go well below 100 ohms. Some of them can handle up to 200v too. http://optoelectronics.perkinelmer.c...toisolator.pdf

I hope you don't mind me modifying your diagram a little bit. Great circuit you made. Thanks again for posting it.

[Steve Conner]

My Magic 8-ball says "Don't connect LEDs in parallel, series is better"

[Gregg]

Thanks for posting the circuit. Very brilliant. I’ll be spending some time to fully understand it.

The circuit is actually not mine - what I did is just to put the pieces together.
I think the schematic for a Gain/Volume control using a VTL5Cx/2 series Vactrols should look like this:



I used +5V single power supply.

The schematic is drawn with sPlan. Check out also Sprint Layout for PCB drawing:

http://www.abacom-online.de/uk/html/produkte.html

[Don Moose]

Originally Posted by tbryanh

I modified your circuit a little bit (see attached) to show the idea of using the Perkin's Elmer VTL5C Series Vactrols. The series has Vactrols with on-resistances that go well below 100 ohms. Some of them can handle up to 200v too.

2 gotchas here -

1) as Steve mentions, diodes (including LEDs) don't parallel well at all, you need to put the LEDs in series (and use a higher supply voltage on the OpAmp) or give each LED its own series resistor.

2) The LED datasheets I've looked at allow a 3:1 range in brightness for a given current. Unless the P-E LEDs are matched, you're going to have to match your own pairs.

Otherwise, well done and best of luck!

[tbryanh]

Originally Posted by Don Moose

2 gotchas here -

1) as Steve mentions, diodes (including LEDs) don't parallel well at all, you need to put the LEDs in series (and use a higher supply voltage on the OpAmp) or give each LED its own series resistor.

2) The LED datasheets I've looked at allow a 3:1 range in brightness for a given current. Unless the P-E LEDs are matched, you're going to have to match your own pairs.

Otherwise, well done and best of luck!

For 1) Yeah, I didn't think about that much. I was just getting a concept out quickly. Thanks for the finer points.

For 2) Yeah, I was figuring you would have to match your own pairs. I was not thinking there would be a 3 to 1 range though. Ouch! Hopefully the Vactrols are tighter than this.

[tbryanh]

Originally Posted by Gregg

The circuit is actually not mine - what I did is just to put the pieces together.

Yeah, well thats still really cool anyway, and thanks again for a great effort.

[tbryanh]

Originally Posted by Don Moose

2 gotchas here -
2) The LED datasheets I've looked at allow a 3:1 range in brightness for a given current. Unless the P-E LEDs are matched, you're going to have to match your own pairs.

Maybe if the diodes are kept in parallel each with their own series resistor as mentioned, the values of the series resistors could be tweaked to try and make the diodes match.

There is also the issue of the light dependent resistors matching too.

I am thinking that the matching may not need to be real precise. If they are close, it is probably good enough.

If you are Mesa Boogie and you want all of your Triaxis's to be interchangable with each other, then precise matching might be needed. But for the typical one-of-a-kind amps we make, precise matching is probably not be needed.

[tbryanh]

Also, the resistor pairs that will replace P1 can be tweaked too.

All this tweaking is not good for mass production, but we are not doing mass production.

[Tiago]

Hi everybody,
At first I was sceptic but now I'm really curious about this...
Has anyone made oscilloscope/spectrum spectrum analyzer measurements?

[Gregg]

I am thinking that the matching may not need to be real precise. If they are close, it is probably good enough.

If you are Mesa Boogie and you want all of your Triaxis's to be interchangable with each other, then precise matching might be needed. But for the typical one-of-a-kind amps we make, precise matching is probably not be needed.

Well, I currently need them for DIY purposes but if I can find them in bulk at reasonable price maybe people will start thinking about limited production as well.

Somebody mentioned earlier in the thread about a manufacturer offering such 4 pin LDRs in 1000+ quantities. I think this quantity is big even for a group buy but maybe they can be custom ordered with some chinese manufacturer in smaller /sample/ quantities or even the same quantitiy at reasonable prices?

[tbryanh]

Here is another little mod to Gregg's circuit.

[tbryanh]

2) The LED datasheets I've looked at allow a 3:1 range in brightness for a given current. Unless the P-E LEDs are matched, you're going to have to match your own pairs.

I am thinking the 3:1 range is probably a result of comparing LEDs between different manufactures. Each manufactuer has its own process. This is probably what causes such a wide range.

The chances are that the Vacrols from Perkin Elmer will be fairly matched to each other, especially those from the same lot (wafer) They are made with the same process.

[Don Moose]

Originally Posted by tbryanh

I am thinking the 3:1 range is probably a result of comparing LEDs between different manufactures.

Sadly, no. I found this on individual datasheets for LEDs by the likes of HP, Sharp and at least one other. To put it another way, the HP parts I looked at had a 3:1 range for min/max at given current(s). So did the Sharp LEDs, and so did at least one other mfr.

Now a 3:1 range for min/max doesn't sound so bad if you're buying enough quantity to get them binned (matched to some subrange of that range). If you're buying hobbyist/boutique quantities ...

Where I've been spending some time is trying to visualize a clean, repeatable way to EVENLY illuminate two photoresistors from one light source, all in a compact package. Then I only have to worry about matching the photoresistors.

On the other hand, I spent all that energy worrying about an oscillating control amp, and that turned out to be nothing.

Anyway - hope this helps!

[tbryanh]

Where I've been spending some time is trying to visualize a clean, repeatable way to EVENLY illuminate two photoresistors from one light source, all in a compact package.

I can't find the article, but I remember somebody showing on the internet how to connect an LED to a single photoresistor, and I think this would work for connecting an LED to two photoresistors.

They used a glass shaft that was specially made to conduct light. I think it is available through Edmund Scientific, and it is probably also available at many other places.

The glass shaft was the right diameter so that it fit into the face or the round photoresistor. This provided good alignment.

The LED was connected to the other end of the glass shaft by drilling a hole in the center of the end and sticking the LED in it.

It apparently worked real good.

The setup could possibly be modified to make LDR Pots.

Instead of drilling a hole for the LED at one of the ends of the glass shaft, the hole can be drilled on the side of the shaft in the middle of the length. Then a photoresistor would be connected to each end of the shaft.

Hopefully there would be enough light radiating out of the sides of the LED to illuminate the photoresistors, and hopefully the light would radiate evenly on both sides.

I am not actually sure if the shaft was glass or not. It could have been some type of plastic too, but it was a shaft specially made to conduct light.

There also might be something in the field of Fiber Optics that could be used instead of the glass shaft, but chances are the glass shaft would work.

Ideally there would be some way of dividing the light from the end of the LED into two equal beams and then direct the beams down to each end of the shaft.
I guess you could use some sort of tiny mirrors to do this, but it sounds like the machining and placement of the mirrors into the tiny hole for the LED would have to be ultra precise. This does not sound within our reach, but the simple approach just described might work well.

[tbryanh]

Now a 3:1 range for min/max doesn't sound so bad if you're buying enough quantity to get them binned (matched to some subrange of that range). If you're buying hobbyist/boutique quantities ...

The bottom right of each page that shows the Typical Perfomance Curves for Output Resistance vs. Input Current for each of the vactrols in the VTL5C Series of Vactrols says“At 1.0 mA and below, units may have substantially higher resistance than shown in the typical curves.”

This does indicate that characteristics vary widely between vactrols with the same part number, but the sentence that follows says “Consult factory if closely controlled characteristics are required at low input currents.”

This indicates that the factory will sort through the vactrols and match them for you. The question now is: What is the minimum quantity you must buy to get matched vactrols and what is the price per matched vactrol?

It sounds like one of us needs to contact the factory. I will volunteer to do that.

http://optoelectronics.perkinelmer.c...toisolator.pdf

[tbryanh]

It might be best to ask the factory for matched pairs of vactrols instead of matched batches of vactrols. Its probably easier to sort out matched pairs than to sort out a handfull of vactrols that are matched to each other.

[tbryanh]

I just sent the following email to Perkins Elmer with the subject heading "Need Matched Vactrols"

I noticed the datasheets for your VTL5C Series of Vactrols says to
contact you if I need closely controlled characteristics at low input
currents.

I am a member of a community that builds boutique vacuum tube guitar
amplifiers, and we are interested in using your vactrols to make
programmable potentiometers.

We need matched pairs of vactrols. The characteristics can vary
amoung pairs, but the vactols in each pair must match.

What is the minimum quantity we must purchase in order for the
vactrols to be matched, and what is the price per matched pair?

Thanks

[Enzo]

Just curious, how many of these things do you estimate there will be a demand for? Who would be the customer for them?

[tbryanh]

Originally Posted by Enzo

Just curious, how many of these things do you estimate there will be a demand for? Who would be the customer for them?

I'm going to order 500 for you. What is your credit card number?

[Gregg]

This is what a google search returned for the original part number:

http://www.eleson.net.cn/eleson_Product_1357357.html

Chinese anyone?

[Enzo]

tbr, I wasn't being flippant, though maybe I often am. I seriously wondered how many of these you think there will be a market for, and who would find them interesting commercially. I realize the few individuals here in the discussion will find them of interest, but as a product in the marketplace, who might be the intended target?

[tbryanh]

The whole interest in this thread is probably inspired by Mesa Boogie's Triaxis Preamplifier.

Just check out the Triaxis for an idea what LDR Pots can do.

Thanks