Time travel - EEV P4428 teardown

[Bill W]

First of all, here is the 'victim'.

Obtained correctly from London Underground (hence the logos) and they are very happy with the Argus2 in return.




The 'pan / chop' switch is the only real oddity here.  This goes back to earlier cameras where there was much less image processing and it was worthwhile having the bare tube image shown.  because the tube is only sensitive to changes in the scene, the user had to keep moving the camera to get a picture.  It did at least double the sensitivity.

In chop a chopper wheel goes round phase locked to the tube scans to ensure constant change of scene.  Any one 'pixel' spends on field warming up looking at the hot thing, is then read out and spends the next field cooling down and is read out again.

With an inverting video amplifier and choosing the fields to display a 'always white hot' image is obtained.

[Bill W]

Now to get inside.
There is no need to 'undress' the camera, as the body comes out with the front, leaving the rear case, cables and jacket behind.  I really hate trying to put the heat jacket and straps back on and usually get it wrong 3 times  before I get it right

So having loosened the jacket and unplugged the power cable (see above) turn the white ring and the case splits.

What you see next is a copper screen for EMC, it is soldered at the front and screwed to the rear.  The main problem is the very low signal from the tube, unlike now the EMC is keeping EM out not in

[Bill W]

This firefighters' camera was the 'reduced weight' development from the P4228 bring it down (  ) to 160mm diameter, 3kg camera on the arm.  as the batteries were attached to the firefighters harness.


A few background links from fire-tics:

main page
http://www.fire-tics.co.uk/4428.htm

brochure
http://www.fire-tics.co.uk/datasheets/P4428_brochure.pdf

datasheet
http://www.fire-tics.co.uk/datasheets/P4428_datasheet.pdf


Bill

[Bill W]

So to the internals...

It has a sheet metal chassis with two large PCB's, all through hole, and the two tubes (one sensor, one display CRT) in the centre.

This is the scan PCB which drives both tubes and also has ancillary functions like battery indicators and the iris & chopper drives on it.  It is a two sided PCB hand soldered.
The power input is the 8 way plug which is wired to the front plastics.
A 14 way connects to the iris motor, limit switches and chopper motor & sensor


Main areas of the board noted.  There's 1kV loose on this as well as a few hundred volts on those FET tags.  Fortunately all high impedance so you don't get too many uA from them.  Keeps you awake though.

There are also a lot of adjustments available, no two tubes or yokes are ever the same so all need setting up individually.



The back of the scan PCB with the CRT attached  Here it is clear that the PCB was designed using red and blue sticky tape, none of the CAD business back in 1985 !

[Bill W]

The Video PCB....



On here we have basic timing (Harris 22402 sync gen, only resolving to 1us) the video amplifiers / selection / adjustments and a large area of digital which is a 256x128 pixel x 6 bit field store.  No DAC, just a HC374 latch and 6 resistors !

The need for the field store is as follows.  As noted above, to get a constant image in 'white hot' the tube is chopped and alternately the heating (white hot) and inverted cooling (black hot) images are selected.  This is far from a perfect image as the tube has a 'background' of shading non-uniformity which is also being inverted so would give a horrible flickering.  By adding the current field to the previous (stored crudely at 6 bits) this was cancelled out.

So the tube output looking at a hot object is (white + shading), then (black + shading)
The video chain produces (white + shading), then INV (black + shading) = (white + shading) then (white + INV shading)
So the hot object is always white but the shading now alternates.
4 of the video adjustments add in line and field ramps and parabolae to the video try to cancel out the shading as best it can.

Adding two consecutive fields therefore gives a constant 2 x white and no shading.

[Bill W]

Top view, showing the CRT and magnifier lens.  This allowed both eyes to see most of the CRT screen, and the brain joins the images up.

[Bill W]

Finally the bottom view, showing the thermal tube (Pevicon) and scanning coil set



Optical leveling and focus was by moving the tube.  As can be seen, this one was only just in spec as it is set near maximum rotation.  The tube, yoke and PCB were all built in house at EEV Chelmsford. Even the sensing material (TGS crystals) was grown on site.  Metalwork from local subcontractors, the lenses varied depending on the GBP-USD exchange rate

The chopper sensor gives feedback to a phase locked loop that keeps the chopper synchronised with the tube scans, an IR beam is broken by the black painted outer edge of the chopper blade that cuts through between lens and sensor.

[Fraser]

Bill,

Thanks for taking the time to produce this thread. As a past owner of this camera model it was great to see the pictures and description of the internal parts. As you know, that pan/chop switch had me scratching my head until you explained it to me.

I may not have the P4428 any more but I do have it's descendant, the ARGUS1. I have three that will need my attention so any information pertaining to these Pyroelectric Vidicon cameras is always of interest to me.

These tube based cameras may seem very 'old school' when compared to modern microbolometers but, as you well know, uncooled thermal sensor technology was very much in its infancy when this camera was designed and advances came in small steps. Many clever people understood the principles of the venerable Vidicon monochrome television camera tube, but some clever clogs thought to use a thermally reactive target material to form a thermal scene sensitive television tube.  Genius Building on what you know and understand is often a good start and this technology proved itself in several cameras of the period.

On the Fire TICS forum you kindly documented returning one of these tube based thermal cameras to service through correct setting of the various tube bias voltages etc. Thanks to your efforts, some people who either have a somewhat tired tube based camera, or those who cannot afford the later BST/microbolometer based cameras, have a chance of owning a useful thermal camera based on this technology.

Thanks again

Fraser