Partially lit/covered LDR

[matseng]

I've looked at a few datasheets but couldn't find any info about how LDR reacts when illuminated on just a part of the sensing area.

If I light up, say 25%, of the full area will I then get any significant decrease in resistance? Or are they like PV panels where even a small shadow will kill the overall efficiency markedly?

[SeanB]

Depends where you illuminate. The light must reach the CdS coating and must be a bit of the coating that bridges the contact fingers. From most CdS cells I have checked you will be safe illuminating just a 3mm diameter spot in the middle of the cell, and there you will get a lower resistance when illuminated, though it will probably never drop below the 50k mark unless you illuminate the entire cell surface. Can be 1M or higher in the dark, and if you are using a LED ( preferably a blue LED. but almost all will do aside from IR ones) in a 5mm case it will illuminate most of the sensor area. You will not get it to sunlight resistance of 100R or so, but it will get down to under 10k depending on the cell and the led drive current.

[T3sl4co1l]

Most are interdigitated, so the channel length is small and the response is uniform.

For a given intensity (I == W/m^2), conductivity should be about proportional with illuminated area.

Tim

[matseng]

Thanks, that sounds great. 

I'm thinking of doing a punched paper tape reader, but using regular thermal receipt printer paper with black and white areas on it - and using only parts that might have been available back in the 1960's.

[mikerj]

The Mullard OCP71 phototransistor was available in the late 50's

[T3sl4co1l]

Phototubes and PMTs were available since long before then, as well.  You might need some optics since these usually have larger areas than semiconductors.

Tim

[matseng]

The OCP71 is cooler than a simple LDR so I'll  order a dozen or so of them from eBay.  From what I just read many of them seem to have a problem with excessive leakage current at dark - hopefully I can get 9 good (or at least usable) parts from my batch....

[matseng]

And then I'll use BC107/177 for the schmitt-triggered comparators - they both look retro and was introduced in the late 60's and they are still available at Element14 so there's no need for getting expensive NOS from ebay....

[NoItAint]

I like this idea of building something that didn't exist in the 60's from parts that existed then.  Kind of a Steam Punk theme.  Seems like fun.

Back in the 70's I built a Wah-Wah pedal that used an LDR.  A tapered cut of metal would fully cover the LDR in the pedal up, and as you pushed down the taper would expose more light to the LDR.  I think the kit was made by Paia?  It worked, but it didn't work so great, and it was temperature sensitive.  As a bar gig filled up and the temperature would rise from amps, lights,  and people, it'd change filter range to where it wasn't giving a good wah.  I couldn't wait to save up enough for a rack and pinion type crybaby wah.  I still have the crybaby but the Paia is long gone.

Maybe add some temperature compensation to your triggers.

[T3sl4co1l]

You might want to investigate paper tape reading mechanisms -- they had been electronic (not just electric) for many years.  Doing it with print rather than hole I don't think is going to make much difference (at least if the illumination is enough).

Tim

[Richard Crowley]

Optical paper tape readers used special photo-sensitive receptors which were physically the proper size for the dimensions of the paper tape (1 inch wide, and 1/10 inch spacing in both directons).  Clever designs used the sprocket hole size (which is smaller than the data holes) as the "clock" signal to trigger the sample point and latch the byte.

If you use thermally-printed black spots instead of holes, you will have some fiddly adjustments to reliably distinguish between 0 and 1.  And you will need to do some tricky mechanical/optical design to get your photo-sensitive devices to align with the bit tracks.

[matseng]

If you use thermally-printed black spots instead of holes, you will have some fiddly adjustments to reliably distinguish between 0 and 1.  And you will need to do some tricky mechanical/optical design to get your photo-sensitive devices to align with the bit tracks.

I had a bag of 100 small (like 6 mm) LDRs and I tested one of them connected as a voltage divider using a 4k7 resistor.  Lighting it up with a 3 watt IKEA LED-desk lamp from about 20 cm and a 12 volt supply I get a healthy voltage swing between 3.6 and 5.8 volts at the junction when changing from black to white thermal paper.

Seems like the heated/black thermal paper, even if it looks more kinda dark greyish than black, blocks the light quite well.

With a swing of over two volts I think it will be not too hard to detect the two logic levels.

Last night I hacked together a small test for the reader assembly out of some masonite from the back of an old picture frame and a credit card.  It turned out quite well and the paper can be fed thru it and kept quite fixed in the non-pull direction.   With some laser cut black acrylics or delrin I'm sure it will look and work rather decent.

[matseng]

You might want to investigate paper tape reading mechanisms -- they had been electronic (not just electric) for many years.  Doing it with print rather than hole I don't think is going to make much difference (at least if the illumination is enough).

Tim

Something that will punch real holes in the paper would be fab, but making that machine would be very expensive and require a lot of machinery and skills. 

I've been thinking of the mechanics required for some different methods of making holes in the paper without real punches.  An array of 9 pcs of 1 watt lasers would work, but it would take several seconds to burn a hole.   Jabbing a servo controlled rather coarse (sewing)needle into the paper could work, but it would leave ridges on the paper that could fold back since the paper is just crunched up rather than being removed.  Heating the needles to an incandescent temperature  would burn up the paper in the jab and leave a nice black ring around it, so that might work, but the heaters (9 of them) would make it hard to reach the paper in a single row of 9.

So I think I'm going for a more easy to build and wallet-friendly approach here with just spots on thermal paper - unless something else turns up.

[BennVenn]

Plasma approach? Cheap CCFL inverter will put a healthy current across a 1mm arc. At the very least it will turn your thermal paper black at the discharge spot, at best a nice hole burnt through.

[matseng]

Interesting idea, I wonder what voltage that would be required to do the initial ionisation/arc through the paper...  Maybe the CCFL 3KV is enough if the points are almost touching the paper....

[BennVenn]

I've had the traces on the PCB arc over when overdriven without a load, I'm sure the width of thermal paper wont be a challenge. I've found that once a point has be burnt into the paper the arc will begin to travel around the circumference of the hole, slowing increasing its size in what is very close to a circle in shape.

[Richard Crowley]

An array of SMD resistors would make a nice thermal "print head" for specific hole positions.

[matseng]

I just tried to push between 300 and 500 mW thru a 0805.  It gives a nice coloring effect of the paper, unfortunately it has considerable thermal mass so it will continue to color the paper for almost 2 seconds after I turned of the power.

Maybe that can be solved by mounting them on a pcb with plenty of thermal vias to the back and a peltier mounted there.

I'll think I'll make some PCBs for this and see how it turn out.

The Peltier was discovered 100+ years ago, so it could have been practically done in the late 60'ies. Right?

And a tiny wirewound heating element in a ceramic casing sounds like it would have been achievable at the same time period. Even I could make one small enough at home with clay and nichrome wire I think.