Sensor or Fresnal?

Cabin Fever Bob

Well-known member
Joined
May 9, 2002
Messages
835
Reaction score
0
My second cam rolled off the "production line" last night. I was messing around with testing the PIR(MS20/WM2) in my basement and noticed something strange. At ~20' I was satisfied with how the PIR was picking up movement as far as entering/exiting the "cone". For some reason at ~30-35' the PIR was only detecting me on say the one side of the "cone" but not the other. STRANGE!

I was wondering if the PIR was maybe "tilted" to one side a little or what the problem could be. I moved the PIR around to a few different positions but got the same results. Now I'm thinking that it could be related to how I mounted the fresnal lens itself thinking back to AEH's testing showing how important fresnal positioning is. I unglued the fresnal from how I had it mounted onto the Pelican last night but decided to bag it for the night.

I mounted(flat) the fresnal and sensor the exact same way in my first cam and had no problems. That's why I'm wondering if my fresnal positioning could be off a little so the sensor isn't looking through a "zone" the way it should be.?.?  I just eye-balled it and tried to center the hole in the Pelican in the center of the fresnal.

Any ideas? BTW, what's the best/safest way to clean "goop" from the fresnal without damaging it?
 

Jon5ja

Well-known member
Joined
Jun 6, 2002
Messages
722
Reaction score
0
I noticed the same thing on the newest MS20 I was hacking on the other night when I was playing with adjusting the gain.  The Fresnel lens and the other components are in the original housing, just as it came from the factory, and it's more sensitive on one side than the other.  The first MS20 didn't behave like that.

I haven't found the answer yet ... I'm going to take out the white shield part so I can see the PIR element, and then shine a flashlight toward the MS20 and rotate it as I watch to see where the focused light spots hit the PIR element;  hopefully I can tell if the Fresnel lens is focusing in the wrong spot.
 

Hill Hopper

Banned
Joined
Mar 10, 2002
Messages
1,570
Reaction score
0
Probibly because its a $7.88 complete PIR unit made with unskilled labor or a robotics table with a little slop. QC can't be that tight for that price, but they still work damn good for what we use them for.
 

Brian

Inactive
Joined
Jul 30, 2001
Messages
1,194
Reaction score
0
OK.  I ran into a similar problem way back and I am not saying this is what you are seeing but it was on mine.

Jon,

Since you left yours in the origianl package but made your mods to the sensor I feel pretty sure this is what the problem is.

The PIR is a dual element that uses a positive to negative pulse out to the gain stage of the amplifier.  The order in which the leading edge of the signal from the PIR is determined by the direction of motion.  So say your moving right to left the PIR would give a positive signal followed by a negative signal.  If you were moving left to right then it would be reversed.  It would give a negative signal followed by a positive one.  This is how the little robots follow motion in some of the neat little designs for chasing your dog and stuff.

Anyway if you bias one of these sides to a point lower than the trigger point on the comparator stage the sensitivity from a direction will decrease as you get further away from the PIR. I am guessing that with your sensitivity adjustment you have moved the trigger point of one of the comparator sides inadvertly causing one side to be less sensitive at longer range. I tested this with the Glolab circuit and found consistent results with lower one side of the comparator vs. another.  If I lowered one side it would not detect from that direction at long range.  Same if I lowered the other.  The key to getting around this is you have to adjust the sensitivity in the amplifiation stage and not in the comparator stage.  If you look at the Glolab circuit you can see the way to do it.  You might even be able to pull up the correct stage or down if you know which on is being affected.  I even looked at this on a scope and you could see the lower signal from the side that was being affected.

I am not famaliar with the MS20 but I know that the PIR design requires the exact same format.  I will email you the specs on the RE200B so you can see the signal response from the PIR.
 

Jon5ja

Well-known member
Joined
Jun 6, 2002
Messages
722
Reaction score
0
Negatory :)

I didn't change the biasing at all -- the bias is set by the source resistor, and sits at 9.1 uA on the one I was playing with.  I could tell a difference in sensitivity side-to-side even before I started adjusting the gain (which didn't change biasing).

The PIR element could be inherently unbalanced or the thermopiles uncentered, the optics could be off, or possibly the 2.5V reference that sets the idle voltage of the first op amp could be off-center compared to the references supplied to the comparator that follows the amp.

I'll see if I can figure it out this weekend.
 

ArkansasElkHunter

Well-known member
Joined
Jan 19, 2002
Messages
1,748
Reaction score
0
Wow...................You get all that Bob.  hahaha

Just kidding guys,  I always get that deer in the headlight look about this time of day.
 

Archilochus

Well-known member
Joined
Mar 27, 2001
Messages
1,285
Reaction score
0
Hi Jon, CFB,
Since you guys didn't alter the references, I'm casting my vote for mis-match between the two sensor elements.  In the specs I have for high-quality Heimann samples (individually hand-tested!!) - they state the elements will have a maximum 'mis-match' of up to 10%.  I think(??) the MS-20 uses the cheaper Nippon Ceramic RE-200B(??)  The very minimal specs I've found for these parts show much looser specifications - I'm guessing that the mis-match could be up to 15 or 20%.  Combine that with slightly 'off' optics - and you could see reduced sensitivity at one extreme of the FOV of the sensor.

Archilochus
 

Cabin Fever Bob

Well-known member
Joined
May 9, 2002
Messages
835
Reaction score
0
AEH, Wa-Wa-Wa What did they say?? :help::raise-eyebrow::wow-eyes:

So fresnal alignment, sensor distance, or........?

I was messing with both the fresnal alignment and sensor distance today trying different things. The best range that I can get with it is ~35' max. My first cam with this same setup can get ~55'.?.? I just don't get it....
 

Jon5ja

Well-known member
Joined
Jun 6, 2002
Messages
722
Reaction score
0
I still haven't had a chance to decide what IS the problem with mine, but I don't think it's the Fresnel lens.  With the lights out in the room, I opened the door a crack so I had a vertical band of light coming in from the hall.  I could see where the bands of light were focusing on the PIR element, and there were LOTS of bands and they went across the PIR element from one extreme to the other.  I couldn't tell any difference in the appearance of the light pattern from one side to the other.

The PIR element on one MS20 was about 1 mm offset from the other one, but after looking at how broad the light patterns swept, I don't think that is the problem either.

The "neutral" point for the first amp/filter measured 2.501 volts, and the two references for the window comparator (of sorts) measured 2.366 and 2.605 volts.  There are 1N4148 diodes in series, so the output of the first amp/filter will have to be below or above these values, respectively, by the forward voltage drop of the diode.  Actually, this is very close to the calculated values according to the schematic, but a little surprising that it's not centered by design.

Watching the output of the first amp on the scope, as I swept a shadow over the sensor, it appeared to "slam" negative a lot harder than it did positive, although since the op amp was running off 9 V and the "neutral" point was 2.5, it may have just been clipping the negative half and not the positive.  I need a digital storage scope!

I really think Arch is right -- I think it's the PIR element.  I know the two MS20 have different-looking PIR sensors...  here's a picture of the two:

The flash was more centered over the one on the left (I had the boards laying side-by-side and cut these out of the same picture), but you can see the one on the right is more tarnished-looking, and has a wider "window", and is slightly shorter.  The one on the left was the one that is more sensitive on one side than the other;  I'm not sure whether one MS20 is more sensitive overall than the other... maybe I'll take them outside tomorrow and see.

Looking at the pictures I just took, it also looks like there may be some "crud" on the window of the one on the left... maybe I should clean the element and try again... or, maybe I should vary the crud density to balance the two sides :)
 

ArkansasElkHunter

Well-known member
Joined
Jan 19, 2002
Messages
1,748
Reaction score
0
I think you should definitly vary the crud density and maybe even alter the crud distribution.  I have a crud distribution matrix I can e-mail you if you need it.



Just kidding again.  Keep it up you guys, you know I think I am even picking up a little of this as you go.  I have always been just as happy as a hog in the mud to follow the recipie for modifying this stuff, blindly going where ever it took me.  A base knowledge of how it works would sure come in handy sometimes for trouble shooting.  
 

Jon5ja

Well-known member
Joined
Jun 6, 2002
Messages
722
Reaction score
0
Well, CFB, I did some tests today, but I don't really think it'll help.

I'll refer to the MS20's as "A" (the one on the left in the picture I posted, with the shiny PIR sensor and the smaller window) and "B" (the duller looking one on the right without any writing and with the wider window).  It was somewhat cloudy and 86 degrees F outside.

Both devices would detect me to at least 50' away at the center of the field of view.  My fence was 50' from the picnic table I had the sensors sitting on, so I didn't go farther out.

"A" would detect me at 25' at about 45 degrees off center on its right side, and it'd detect me at about 35' at 45 degrees off center on its left side.  There was a pretty broad angle at which it'd detect me at 45' away (maybe 45 degrees wide).

"B" was more balanced side-to-side, but seemed to be much narrower than "A";  I had to be about 12 or 15 feet away to trigger it at 45 degrees off center.  At 45' away, it would only detect me over maybe a 30 degree wide range.

I did clean the window on "A" with an alcohol swab before doing the test, and I think that helped.  Looked like there was some residue from where I had stuck a piece of black tape over it at some point during testing.

Anyway, FWIW.  I think ****** may have the right idea.
 
Top Bottom