In silicon, that's probably right, but CdS has a 2.8 eV bandgap; there's OTHER states than electrons and holes getting deeply involved here, or the sensitivity would be to hard UV only. If there's a variety of transitions with multiple traps inside the bandgap, it can become a much longer formula, with lots more variables.
When I get a round tuit, I plan on making a solar tracker. And was plannin g on making it somewhat like an article back in Nuts and Volts in Aug 2009. The author used Cds sensors because he said other photo sensors saturated in direct sunlight. This is really a learning exercise. So let me know i f different sensors would be an improvement.
Dan
Two small solar cells placed near each other, with a suspended shadow mask between them. Subtract the two photocurrents and use that as the error signal. Cheers
Phil Hobbs
THER
tivity
tiple
Sure. However, neither you nor I have a detailed idea of what's going on in some random piece of CdS film. There may well not be a single lifetime app licable to all, but then that's an approximation even in silicon.
The quadratic recombination rate will obtain regardless of those details, b ecause as the trap states recombine, the free carriers do too. That's enoug h to give the drool-down settling behaviour when light is removed.
The math is similar to that for fast pulses in coax: the rolloff goes as 1/ sqrt(f), leading to the characteristic drool: the first bit looks fast, but the rest takes forever.
Cheers
Phil Hobbs
s
ing on making it somewhat like an article back in Nuts and Volts in Aug 200
Photo diodes rock as light detectors.. Calibrated, high efficiency, and fast if you need it. (low light gets fussy.) If the solar tracker is for a solar panel, you could wiggle a piece of the panel (slowly) to find the maximum.... hmm bottom of a cosine, you'd want a big wiggle.
George H.
Should have been clearer--the mask is oriented parallel to the cell surfaces, covering half of each one, but spaced vertically by about one cell width. That gives you a range of +-30 degrees. Smaller spacings give wider ranges with less sensitivity.
If you go to
Cheers
Phil Hobbs
-- Dr Philip C D Hobbs Principal Consultant ElectroOptical Innovations LLC Optics, Electro-optics, Photonics, Analog Electronics 160 North State Road #203 Briarcliff Manor NY 10510 hobbs at electrooptical dot net http://electrooptical.net
And who ever it was that said the PD's saturated didn't know how to use them. You have to reverse bias 'em for maximum dynamic range.. (Something I only learned last week. :^)
I should take some PD's out into the sun sometime.
George H.
Probably he was using too large a load resistance and let them get forward biased. Sunlight is about 1 kW/m**2, so a garden variety BPW34 (7 sq mm) in direct sunlight will generate at most
0.4 A/W * (1 kW) (7 sq mm)/(1e6 sq mm) ~ 2.8 mA, which is well within its linear range when reverse biased. Smaller ones are generally better. (Of course solar cells are photodiodes as well, and some work up to 2500 suns if you cool them right.)Cheers
Phil Hobbs
-- Dr Philip C D Hobbs Principal Consultant ElectroOptical Innovations LLC Optics, Electro-optics, Photonics, Analog Electronics 160 North State Road #203 Briarcliff Manor NY 10510 hobbs at electrooptical dot net http://electrooptical.net
I am probably going to use Cds sensors first. They are nice and small, plu s I got 100 of them off AlliExpress for #2.69 including shipping.
It would not be for a solar panel. Maybe in ten years or so they may be c ost effective. This would be for a solar collector panel to warm up the ba sement. I have a wood stove in the basement, but it is kind of a pita.
I am thinking a two axis tracker would be easier than making an equatorial mount. And I may change my mind, but right now I am thinking of only check ing the tracking every 30 minute or so and when it needs to be moved, then advance the azimuth by about 20 degrees. Elevation would be less.( this as sumes the tracking is about 10 degrees off when it is checked. So it would go from 10 degrees lagging to 10 degrees leading. ) Since it is a Cosine function , this would he almost as good as continuous tracking. And would not have to worry about hunting or overshoot. Will probably use a Arduino
hey are really cheap.
Dan
The 'recorder' and source measurement issues are illustrated in the series below.
In all cases the stimulus is ~ 5uA, produced by a fixed measurement range. In order to avoid dynamic saturation of the measurement, the CDS cell is paralleled with 1M. In one case the calculated removal of this shunt resistor illustrates measurement quantization issues involved.
A ~27mS measurement interval is the best the stand-alone recorder can manage, when programmed for continuous measurement at its worst resolution(4.5dig).
160403 uses a light source with turn-on intensity issues. 160404 data shows the effect of source intensity on attack at this impedance measurement range. The lower intensity turn-on actually shows mid-switching resolution. Possibly the higher intensity turn-on might also show this in a lower-impedance measurement.Any impedance bump at turn-on is lost in other noise...an artifact visible only in an 'autoranging' display of this measurement segment alone.
One piece of info also missing from this and other thread tests is a part number, which some modelers tend to be pretty sticky about.
The one type used here is an anonymous bog-standard part used in 120V security light switches.
Aiming tests at a specific part might make some sense, allow for later comparison, and reduce wasted time.
RL
I have a HP (doubtless not authentic HP, but an acquisition of theirs) plotter which has exactly one power transistor (2N3055) per channel, driving its servo bidirectionally.
A pair of neon lights and TO-5 canned CdS photocells convert the front end into a chopper amp. The power transistor resides inside a FWB, and the motor is powered by AC.
Oh, for the days when transformers with a million windings were still cheaper than just adding another power transistor...
Tim
-- Seven Transistor Labs, LLC Electrical Engineering Consultation and Contract Design Website: http://seventransistorlabs.com
Yes, old lab.
I suppose I could rig up much the same again, but it's a lot of bother (and I'd guess Jim isn't paying for lab hours on this curiosity ;-) ).
Scan thru the R(t) data -- notice the heavy quantization noise. At least, as I recall, those humps come from that, and the smoothing function. Without a smoothing function, obviously(?), tau goes undefined because dR/dt = 0 in those segments.
Tim
-- Seven Transistor Labs, LLC Electrical Engineering Consultation and Contract Design Website: http://seventransistorlabs.com
I may not understand the problem, but, for a solar tracker, I would try to use LEDs in the sensing mode for the sensors. I would put a cylinder over each one, the length of which I might adjust for accuracy. I am thinking of the T1-1/4 ones.
The old chopper-based plotters were impressive.
Prior to the CdS cells, mechanical choppers did amazing things. These were metal tubes with a vibrating reed and a bunch of contacts, sort of like an auto radio vibrator. A good tube-based chopper opamp had microvolt long-term offset and occupied a good fraction of a cubic foot.
Choppers could convert DC to AC, with the AC amplified with tubes to drive a 2-phase electric motor. The motor would move a pen or an indicator or something, and a pot, with the pot generating feedback into the front end.
-- John Larkin Highland Technology, Inc lunatic fringe electronics
Is there some spec for the saturation current, (max current) I was looking at the few spec sheets I have here and I don't see any photo currents in the mA range.
George H.
Yup. My plotter has a pair of long linear wirewound potentiometers. (But like I said, DC motor, in this case.)
Now that you mention it, it's almost tempting to convert it to toob... ;-)
Tim
-- Seven Transistor Labs, LLC Electrical Engineering Consultation and Contract Design Website: http://seventransistorlabs.com
That can work OK for a one-off, sure. The angular patterns of LEDs are very poorly controlled, though, because the lens is very steep and the die centration is erratic. The shadow mask approach gives you a very well-controlled and predictable angular signal, and is no more complicated.
Cheers
Phil Hobbs
-- Dr Philip C D Hobbs Principal Consultant ElectroOptical Innovations LLC Optics, Electro-optics, Photonics, Analog Electronics 160 North State Road #203 Briarcliff Manor NY 10510 hobbs at electrooptical dot net http://electrooptical.net
Not that I know of.
Cheers
Phil Hobbs
-- Dr Philip C D Hobbs Principal Consultant ElectroOptical Innovations LLC Optics, Electro-optics, Photonics, Analog Electronics 160 North State Road #203 Briarcliff Manor NY 10510 hobbs at electrooptical dot net http://electrooptical.net
I may not understand your use of "shadow" here. If a (light-blocking) cylinder is fitted over a T1-1/4 LED is there not a shadow if the cylinder is not aimed at the source? My thought is that the light that enters the LED will be channeled by the cylinder. A shorter cylinder allows more off-axis light while a longer cylinder does not.
Not as simple as that. You have to worry about light bouncing off the sides , for one thing. (grazing incidence reflections are surprisingly bright, ev en off black surfaces). For another thing, the spatial patterns of the LEDs themselves are poorly controlled, so there are unpredictable gain variatio ns with angular position.
For a third thing, in order to get a central null, you have to aim the LEDs and the tubes just right. Otherwise you'll be off-centre, or at worst (if the patterns don't overlap) you'll get a multi-peaked response with multipl e stable servo points.
None of that is very difficult to fix for a one-off, but the shadow mask me thod is even easier and suffers none of those problems. (Check out the Amaz on reference I posted.)
Cheers
Phil Hobbs
ElectronDepot website is not affiliated with any of the manufacturers or service providers discussed here. All logos and trade names are the property of their respective owners.