IR Thermopile vs. Pyroelectric

[vector_vortex]

Hi all.
I am thinking about building a UAV autopilot using the method of IR horizon sensing. However I am finding it very hard to source the solid state thermopiles that most people use. I was thinking about using a pyroelectric sensor like those used in PIR security systems. I believe that those systems react to a change in temperature of the detected object, IE produce a signal when the temperature changes, which wouldn't be such a big problem to decode. However are these type of sensors accurate enough to detect the change in temperature of the earth vs. sky.
Otherwise does anyone have a good, cheap source of the solid state thermopile sensors that people seem to be using.
Also thinking further ahead, these systems seem to be dependant on a near flat horizon or else they become tricked, has anyone experimented with scanning arrangement for detecting the horizon. I have been researching IR detection and tracking systems and found that early tracking systems eg. early AIM 9's used a lead sulphide detector which is a pyroelectric. Are the response times of the thermopile cells adequate for this kind of thing.
Has anyone used a pyroelectric cell in autopilot construction?.

thanks

Joel

[LukeZ]

Otherwise does anyone have a good, cheap source of the solid state thermopile sensors that people seem to be using.Can't answer all your questions, but you might be interested in checking out the IR sensors (http://scalerobotics.com/store/catalog/) that a local RCGroups fellow sells.


Luke

[radiohound]

I was thinking about using a pyroelectric sensor like those used in PIR security systems. I believe that those systems react to a change in temperature of the detected object, IE produce a signal when the temperature changes, which wouldn't be such a big problem to decode.

Hi Joel,

Ok, I am wrong. I though pyroelectrics were the same as thermopiles, but the following I found on the internet:

"Thermopile detectors differ from Pyroelectric detectors in several important ways.

Thermopile detector output is proportional to incident radiation while the pyroelectric detectors output is proportional to rate of change of incident radiation. In other words, the thermopile detector is DC coupled while the pyroelectric detector is AC coupled.

Thermopile detectors have low impedance while pyroelectric detectors have very high impedance requiring an internal impedance converting buffer to make them useable. Low impedance is an advantage in that associated circuitry is less susceptible to disturbance from extraneous radiation and electrical noise. "

So it sounds like you would have to come up with a different circuit than most people are using to do it.

A couple of features of the MLX90247

it has a 100 degree field of view (pretty important to be close to this).

and sensitivity range from 7.5um to 13.5um

Here's the MLX90247 's data sheet http://www.melexis.com/prodfiles/0004763_MLX90247_v2.pdf

[vector_vortex]

From my research so far:
thermopile, uses the same theory as the thermocouple junction, that of two different metals in contact with each other. just in a solid state form factor and duplicated many times via silicon processes. I think that the MLX90247 is one of these devices, and by far the most commonly used in autopilots. I think this produces a DC voltage that can then be measured.

pyroelectric.
uses a crystal, similar to a peizoelectric, (maybe the same?) and produces an output due to changes in incident IR. I really don't know much about these pyroelectric devices, their output etc.
any help would be appreciated.

thanks

Joel

[vector_vortex]

I can source these pyroelectric sensors
http://www.farnell.com/datasheets/71928.pdf
the two element one.
from the internal schematic they appear to be a pyroelectric crystal connected to the gate of a FET.
However there is no indication of the output signal.
it appears that there are two cells,
maybe this could be orientated vertically with a lens to focus the sky on one and the ground on the other. and use the temperature differential between the two to identify the horizon.
I really don't know enough about how these pyroelectric devices work to know if this will work.

thanks for the responses, any more help would be greatly appreciated

thanks

Joel

[vector_vortex]

I found this interesting article on the web
http://www.capgo.com/Resources/Temperature/NonContact/NonContact.html
and when coupled with this
http://www.andrew.cmu.edu/user/luisn/Blank%20Page%202.htm
gives some interesting ideas on how to use these PIR sensors.
Unfortunately for simple sensing, this means that a mechanical motor will need to be used for the signal chopper but in the more complicated sky temperature "mapper" which would have to have a scanning mirror anyway this could be good, also from my reading. pyroelectrics have a faster response time than thermopile sensors.