ISI Vision3 Image Problem: always changing between Black Hot / White Hot

[HB9HUR]

Hello Everyone
I'm from Switzerland so in advance my English isnt perfect.

Now to my problem I was able to purchase a defekt ISI Vision 3 Thermal Camera on ebay.
First I tried to start it with 6V because that's what the old batteries had written on them. It startet for like 2seconds and than shut down. I than open it up to have a look at its insides. I tried starting it a few times than than per coincidence I felt that a tantalum smd capacitor was a little to warm.
Which I than replaced. After that the start up was still wonky, than I found out that with 6.7V I have the most success. After I showed it a friend who knows electronics, he meant I should but a 1000mircoFarads condesator parallel to the 6.7V power input.
And than it finally started and worked surprisingly good.
But now out of now where it started to change between Black Hot / White Hot with the frequency of the chopper wheel. I retraced the analog out back to the Raytheon board on which it also looks the same so the problem isnt in the post processing side. And now im hardstuck I read about 100 posts here searching for a similar problem but its seems i'm one of the first. 

Here is a video of its current state.

Start up video:
https://youtu.be/LE_qvKfzUG0

More or less clean analog video which ran through an 2003ish Panasonic DVD player and than captured it with my Blackmagic design Intensity Shuttle. Here you can clearly see how it always changing between Black Hot / White hot.

https://youtu.be/XP9HZexhf08

Here in this Google Drive are more Pictures from the disassembly:
 https://drive.google.com/drive/folders/1dsrjLtpKz_IUvKlWptUCu2EEtvB6rm82?usp=sharing

Thanks in advance for every input you guys have.

[Bill W]

Start with the good news - it sees hot things !!

Note all BST take a lot of current to start up as the Peltier gets the sensor to temperature, as well as filling all those tantalum on the PCB.  A common fault was the tantalum not being surge rated and going pop - so a good swap out.
I'm surprised you get away with 1A 6.7V. would expect a bit more.  Are you current limiting ?

The video looks like the chopper wheel running at the wrong speed - probably too fast because it is not getting the right feedback from the position sensor.
It might be reflective (with a bright dot on chopper wheel) or transmissive (with a notch on edge of the chopper wheel)

Has the sensor been knocked ?
Has the bright dot fallen off ?
Can you scope the position sensor output ?
Have you got a bright light shining in there ?

Bill

[HB9HUR]

Hello Bill

I am not current limiting.

"Has the sensor been knocked?"
Not that I now of. I renewed the thermal paste behind the sensor but that was already after the Black Hot / White Hot problem.

"Has the bright dot fallen off?"
It's a transmissive copper wheel with two diodes.

"Have you got a bright light shining in there?"
I just tried with no light and flashlight Right at I didn't see a difference.

"Can you scope the position sensor output?"
I'll try that next.

Regards
HB9HUR

[HB9HUR]

In this first picture i measure with 5microsec and 10mV per square I think this ist something like the "sending signal" side.
 P1030256.JPG (2676 kB. 4000x3000 - viewed 42 times.)
Here I measure 4.95V DC on my Multimeter.

Here I measure with 10microsec and 20mV per square.
When I touch the copper wheel to slow it down a bit you can see it on the oscilloscope. So I guess it's the receiving side.

https://youtu.be/b4h4rr3nn_4

Here I also measure 4.95V DC on my Multimeter.

I'm an amateur so I don't really know what i'm looking at.

With my Laser speedometer I'm measuring somewhat around 2850rpm on the edge of the copper wheel.

Regards HB9HUR

[Bill W]

"Has the sensor been knocked?"
Not that I now of. I renewed the thermal paste behind the sensor but that was already after the Black Hot / White Hot problem.

I mean the position sensor for the chopper wheel - not the thermal sensor IC.

[/quote]

[Bill W]

In this first picture i measure with 5microsec and 10mV per square I think this ist something like the "sending signal" side.
(Attachment Link)
Here I measure 4.95V DC on my Multimeter.

Here I measure with 10microsec and 20mV per square.
When I touch the copper wheel to slow it down a bit you can see it on the oscilloscope. So I guess it's the receiving side.

https://youtu.be/b4h4rr3nn_4

Here I also measure 4.95V DC on my Multimeter.

I'm an amateur so I don't really know what i'm looking at.

With my Laser speedometer I'm measuring somewhat around 2850rpm on the edge of the copper wheel.

Regards HB9HUR

Wheel must be 30Hz - 1800rpm -  it goes around at the same speed and time as the sensor readout scanning down the sensor
It is phase locked by the position sensor.

The sensor out signal should be >4V mostly and <1V at the detection point, once per chopper revolution (FRAME = 30Hz) as either the reflective dot or the cut in the wheel go past the position sensor.

Can be the other way up i think - it is phase locked to an edge.  The control behaves a bit like a 4046 / 74C932 IC but I think buried in the giant ASIC.

I have not got a BST core and running just now - but can scope out next week sometime.

Bill

[Bill W]

"Has the bright dot fallen off?"
It's a transmissive copper wheel with two diodes.

Your image 2811 adjusted attached with the position sensor arrowed.
Thanks for the teardown photos - recommended viewing for others into the various approaches to Raytheon BST end products

Position sensor normally blocked by the edge of the wheel and should see the cut in the edge of the chopper wheel ..... BUT ONLY if it is still in the right position.

They have not used the Raytheon support part - but then neither did we at Argus:
http://www.fire-tics.co.uk/project3/images/4.jpg
That is not the default location though - I moved round there for mechanical reasons.  Argus2 was directly below.

[HB9HUR]

Thanks for the help Bill
Theres a chance that the optosensor is broken, I unscrewed it form the chopper wheel and obscured the sensor and looked what's happening on the oscilloscope and it kinda did nothing.
Is it a common part? Can I buy it new?

I also updated the Google drive with some more disassembly pictures from the "Raytheon sensor side"
https://drive.google.com/drive/folders/1dsrjLtpKz_IUvKlWptUCu2EEtvB6rm82?usp=sharing

[Bill W]

Might be as simple as the 0V side missing or the LED+ / 5V leads shorted to output.  Does not look like they are sleeved

However although the Argus ones are mechanically a bit different I have probably got a bin full of them.  Argus cameras were also wheel edge detection - await info.


Bill

[Bill W]

Found the chopper sensor information for Argus3 / OEM kits

It is HOA1875-2
Datasheet:
https://eu.mouser.com/datasheet/2/187/HWSC_S_A0000006417_1-3073038.pdf
Now obsolete, although some close equivalents offered.
As suspected I have some spares, so could send you one for GBP6 (post included) - PM or use fire-tics contact address and I will advise my payPal

Normal operation in a camera below, video in blue, chopper sensor output yellow, 5V/div
This is from an Argus3, with the sensor at '9 o clock' position. 
There is a shift circuit to get the pulse where the BST core 'expects' it.

[HB9HUR]

Hey, Thank you very much for the help Bill

It was the optosensors fault. The camera is now working like a charm.

https://youtu.be/7pF3PYQI8Yc



Best Regards
HB9HUR

[dennho]

I'm happy to see another ISI Vision 3 user here. I know that touching the edge of a spinning chopper wheel will indeed cause that Black/white fluctuation. It's a pretty cool effect, but I don't want to be too mean to these cameras lol.

I have video output problems on mine with "noise static" on the alternate video output that's used for the wireless video transmission video source.

Even if the analog signal is directly connected to a screen, I get this "interference" noise, I think I have a dirty ground? I know there is an internal copper shield inside the camera case, with a ground wire soldered to it. That wire disconnected off it's solder joint, could that be the cause of this noise?

Since the original post has been resolved,  I hope you guys don't mind me asking now.