I am beginning the restoration process on the CRV59-AAE Military Iconoscope Camera (and spare Iconoscope) that I bought at the May RTF meeting.

Some time ago I ordered uo from Mouser the needed list of parts.
The only odd ones are for the power supply. First is the power transformer
for the +405 and -50 volt supplies, which is a Hammond 290EX Guitar Amp transformer. It has the requisite 50 volt (RMS) tap for the -50v. The +27 volt
filament and bias supply is a Meanwell LRF-150F-24 switching supply to which I intend to add a bit of switching crap filtration.

Last night I had again inspected the chassis, tried to remove the bad-looking iconoscope, and tried following wires and testing pots.
I could not get the iconosocpe out. It looked like the yoke was set to make that impossible without removing pots. Today I bought new batteries
for my flashlight and inspected better: yes, there is an adjustment and if moved a bit the tube comes out safely.

All the stuff I tested looks OK especially the low-ohm pots I was able to test in-circuit. Except that the bias lamp appears
missing entirely, no lamp, no wire, no socket. Huh??? The shutter was removed previously.

With the old iconoscope out I was able to test it in a tube tester for
emission and gas (transconductance too, but that's expected to be very very low). I had previously tested the new one, and it looked OK. The old one
is clearly very very gassy, and a retest of the new one under identical settings shows no gas.

There is hope!

Next up is lots and lots of photos and more circuit tracing.

Has anyone else around here other than McVoy had any experience with one
of these?

Doug McDonald

I have now replaced all the 450 volt (B+) electrolytics, C127, C142, C101, C126.
The first two are cans in the main part and were restuffed. C101 was not resutuffed
because the tabs were too stiff to twist into removal position. C126 was not resutffed out of "symmetry" with C101, and there's a nice spot for it.

The two 0.05 uF caps in the preamp area were replaced with modern ones (C113 and C114).

I have not decided whether to replace the LV (15 volt) electrolytics, C147 and C115.
C115 would be easy, C147 even more impossible than C101, to restuff, so if
they are bad they will just have their wires cut and the replacements
tacked in. These caps are usually OK.

All the tubular paper caps are now replaced.

Next up is disconnecting the 6X5 plate and catghode and replacing with
1600PIV silicon diodes.

Then I will test the +28 and -50 volt resistor chains using the 28 volt supply,
and the +400 (no filament power) using a lab supply.

It turns out that the modifications mentioned in the ham radio article,
to use AC on the filaments, were done at some point, and will have to be
undone (very easy, but its good we have a local source for ordinary resistors
and a few caps.).

The circuit changes in the +28 volt area were undone, not physically easy, as big
wires were twisted on tightly. The 6x5s (except heaters) were removed from the circuit
and replaced with 1800V PIV silicon diodes, these being specifically specified
for damper and CRT supply use.

The +28 and minus 50 circuits were checked out using the 28v supply and
appear to work OK.

Next is build the power supply.

The 405 volt supply is built and tested. It was applied to the set
and immediately showed that R210, the 1000 resistor that feeds the
gas regulator tubes, was open, These are still available.

I was able to temporarily replace it with 200 ohm resistors in series.
This showed that at least most of the B+ chain is working OK.

Next up is the -50 volt supply, the noise filter for the 28 volt supply,
and the power cable.

The power supply is complete and connected.

The electronics, without the iconoscope, was turn completely on.

All the various sections work. The vertical frequency adjust is out of range,
it wont reach as high as 40Hz.
This would not matter if I were forcing it to 60Hz, but I've not done that yet.

My soldering iron died and the old 250 watt gun is too clumsy,
so I'm out of doing changes until I can borrow one from a colleague's lab.

I got a new iron.

I got the vertical in range.

As designed it is NOT random interlace! It in fact is roughly (now)
260 x 60p. But at these extended frequencies (originally it was about 300 x 40p)
the stability is marginal.

I will continue testing before installing the iconoscope.

Doug, would you by chance have any pictires of this restoration? That would really be interesting!

I have a very complete set of "before" pictures. I'm rather tired out this evening, so its
unsafe to do more work. I'll try to take the "after" pictures and ones of the power supply,
which still lacks the sync modifier suggested bu McVoy.

Today was devoted to troubleshooting (electronic tests only ... the first
iconoscope test will be tomorrow).

The power to the iconoscope filament was inadequate. It took quite a while to
prove this, as I have no AC voltmeter that works at 16 kHz. I had to
use an analog scope or compare filament brightness between two identical tubes,
one on the camera and one on 6.3vac. Tomorrow I'll bring a digital scope that can
do true RMS AC. The problem was that the screen resistor, supposed to be 82K,
was 106K. This caused the screen voltage to drop to zero at high power, thus
limiting. Even 82K did not work, I eventually used 60K.

Then I worked more on the video amplifier chain. It was horrendously microphonic,
as if there were a broken lead connection. There was., onC154, a 1000pF mica in series
with a 0.05 uF film cap (??? WHY ... its not a high voltage area.?) It had a lead
that was broken off about 1/2 millimeter inside its case. This was not easy to find,
though tapping and poking eventually found it.

Another source of problem was found in a noisy gain pot. This is wirewound 10K.
At least for the moment is was fixed with DeOxit ... you have to take the
case off and flood it, as the noise seems to come not from the wiper on the
resistor element but on the connection of the wiper to the outside.

The shading pots are similar, but are not so bad.

The iconoscope I bought at the ETF convention is good.

The horrible white spot at the low left is present exatly the same
without the iconoscope. It is caused by the scanning rulses being to long
and ringing. Working on this is the next step.

Picture attached. It is of a window looking out
on a sunlit scene. All you see is the wooden things holding the panes of glass
and the Venetian Blind slats. At lower left is the head of a wooden turtle
(bought in Papua New Guinea!) and the corne of an RCA 9T249.

After about an hour and a half running the image faded out.

Had the iconoscope gone bad?

After leaving it off for 45 minutes and turning it back on,
changing no adjustments, I got the very best picture I have seen,
even though the sun had gone behind a cloud so it was a dimmer scene.
Maybe something charged up? Is that why a bias light from the back?
Time will tell.

Here's a better picture, made just after turning on. I have seen better.

I have to turn up the iconoscope bias as it (or something?)
warms up.

It seems to work best with the iconoscope filament runnig at
5 to 5.7 volts RMS.

Missing pictures?

Yes, missing picture. Its only somewhat better. I'll try
again tonight.

I measured the iconoscope grid current. It is OK
at turn on, then climbs to a couple of microamps when
the picture fades, in tirection that gas makes.

Thus, only a marginally good tube. Nevertheless, its a working demo.

Ad placed for better 1846 iconoscope in the "usual suspect" classifieds.

Here is the "better" screen image, and "before recapping" pictures of the chassis.

The bending at the top of the image has been fixed by adjustment of DC levels.

"after" pics of chassis in next post.

Here are the images, "after recapping" and other changes, of the chassis. Many changes are hard to see, but are there .. look at the 6X5s
for a bit of silicon.

Next is the power supply ... it needs safety covers.

Also the iconoscope. As you see, I'm lucky i'm works at all with that
opaque yellow getter rim.

Hey just wanna say it's awesome to see another one of these WWII block cameras get restored to working order and any Iconoscope camera in general, great work and look forward to seeing more on it!

I myself have built a crudely working Iconoscope camera using the same RCA 1846 tube type and driven by video/deflection circuits I made originally for a 2/3" vidicon camera, my Iconoscope camera construction log can be seen on my site starting from this page http://troysvintagevideo.0catch.com/iconoscopeproject.html .

Last week I noticed another 1846 for sale on the 'bay. It had
a good center of the getters, clerly visible, so I bought it.
It arrived yesterday and today I tried it.

It works identically to my other semi-working one. It runs for about 15
to 30 minutes and then the bias need raising and it loses the picture.
Turning off 10 minutes and the picture is back

Questions for those, especially the Aussie guy, who have gotten one to work indefinitely: is there a brown rim around the shiny getters or are they all shiny?

Have you experience of a tube acting like mine?

Do you think that maybe it could be an electrical fault since the two tubes
work or don't work identically? I have a similar appearing third tube
that does not produce any picture at all.

I have no idea about the gradual loss of picture (I have never played with one of these) except that you mentioned back bias lighting. Have you tried it? An LED flashlight, with its high blue content, might be good for a quick test.

Can you post a schematic of the first amplifier stage?

here's that part of thttp://videokarma.org/attachment.php?attachmentid=194149&stc=1&d=1489275183he schematic

Just on general principles of poking around, have you checked R114 and R126?

Just on general principles of poking around, have you checked R114 and R126?

Yes, and they are new. I've checked all bias resistors. I've studied all the
bias voltages as it heats up. They do drift, but the key grid-cathode bias
voltage does not, and all drift in sync and in small amounts. The bias
in any case need far more adjustment than the drift.

I tried checking drift of filament voltage while running by comparing
the brightness and color of the tube in the camera to that of my dead
1846 running on 60 Hz and they are rock stable. I measured
the heater voltage both by comparing the brightness to the other tube
both on 60 Hz and the other one on 60 HZ and the camera one on
15750Hz, and also by measuring the camera tube voltage with a digital
scope across the filament with the HV disabled, then calculating the
RMS voltage from the scope trace (its sort of a square wave). I can't
get the voltage up to 6.3 but is very close and does not drift enough
to cause the problem.

Of course, both tubes could be equally bad by chance. But its
odd that of three tubes that have look-alike getters one does not
work at all and two work sort-of, and the same.

I've tried swapping 6AC7s around to no effect.

Oh yes, a bit off topic ... the iconoscope I just bought came with "extras".
One was a mounting block for it to display it, into which I will put my bad one.
Second was a copy of Pem Farnsworth's biography of Philo, autographed by
her in 2004 (she died at age 98!).

I'm running out of wild guesses. Slow onset and slow recovery seems like a thermal problem, but if the bias voltages aren't changing, what could it be?

What happens if you run it until the video disappears and then ground the collector ring with a clip lead/screwdriver? Could this be a bad ground connection charging up the second anode?

I've tried grounding the collector ring with a screwdriver. I've also
tried putting + or - 9 volts on the unused pins which attach to
the Cesium dispensers, and they don't do much if there is no
hum on them.

I've tried using a magnet to move the electron beam over to the
side where it hits beyond the edge of the silver dot area.
This generates a very strong signal which lasts a lot longer.
Its currently cooling down so I'll try that again.


It does seem thermal, but why cooling down much faster than
warming up?

Edit: a new try shows that the actual picture lasts much longer,
at any setting of the bias, at the far left of the screen.

...
Edit: a new try shows that the actual picture lasts much longer,
at any setting of the bias, at the far left of the screen.
...


Well, this just gets more and more puzzling.

Do you have anyone who can review the wiring with a fresh eye in case something got missed when you recapped?

I discovered by accident that even if it runs 5 minutes beyond the
picture disappearing, and I then turn off all power for 5 seconds and then
back on again, as it re-warms up for a couple of seconds an excellent
picture appears, just a bit small, lasting just an instant as it gets
full size.

This would seem to be a bias drift problem.

I spoke to Schechter last night on the phone. Amazingly, he only knows
of himself, McVoy, Aussie Bloke, the Japanese genius with the perfect
transistorized picture, and me, with 1846 cameras working at all.

Do you know of more?

If my wild guesses are just annoying, tell me and and I'll shut up...

Is there more than one connection possible to the collector ring (like through the cesium depositors)? If so, can you try measuring voltage and/or grounding with a screwdriver?

Can you pull the socket partly off and measure directly on the pins?

Another thought is to float a VOM across the cathode and G1 so you can see any differential drift on a sensitive scale. (Just be careful to have it on an insulated surface and don't try to touch it while in operation!)

Another thought is to float a VOM across the cathode and G1 so you can see any differential drift on a sensitive scale. (Just be careful to have it on an insulated surface and don't try to touch it while in operation!)

I've tried all your suggestions. But for this last one ...there is a problem.
That is, there are big blanking pulses on both cathode and grid. This means that the voltages measured on a voltmeter are not necessarily representative of
what they are during the active scan time.

I will retry this evening with the two blanking pulses disabled. And I will
look at the two pulses with a scope and see if they drift.

I've had my CRV59 camera working for at least 15 years now and it seems to be about the same today as it was when rebuilt that long ago. I agree about it possibly needing the rear mosaic lamp. I've never had the picture fade out on me as you describe as the bias lamp(s) on this camera do work. (It's been a quite a while since I've opened the camera up so I've forgotten whether or not there's just one or two inside!)

I revamped the scans circuits to standard def sweep rates so I could use a regular broadcast monitor for viewing. No problems there. The camera power supply is solid state.

Ralph, is yours a different camera from the ones I mentioned a couple of posts ago?
How long does yours take to stabilize from turn-on?

My lamp does work.

But old_tv_nut's suggestion did the trick, with my new angle.

Removing the blanking pulses generated a "correct" grid bias of -20 to -22 volts.
The picture faded when I adjusted the control, turned right on, to -26 volts.
When the picture faded with the control all the way up the
picture faded at .... -26 volts! I had not realized how much the blanking pulses
caused a measurement problem.

I will investigate the divider chain drift further, with all pulses sent to it
disabled. If I can't figure out what's the problem and fix it,
replacing a couple of resistors with Zeners should do the trick. That's unless
the blanking pulses cause problems by drifting.

Off to dinner.

I checked on the camera and I can't find the original ID tag that was riveted on, but I did find what I think is RCA Camden's inspection tag; see pix attached. Interesting that they didn't call it a camera, rather a "Conversion Unit!"

The schematic I followed in my work is exactly the same as the excerpt you showed earlier, so we're both talking about the same model of camera.

This camera is definitely not one of those you listed earlier.

One more thing: I added a square fan to the top via a replacement slide (instead of the tube access slide.) I did this after noticing how much heat there was inside the case when the original slide was in place. Considering how many tubes there are in such a small container, it might be helpful for longevity to do the same with yours even if it doesn't reduce the "blackouts" you've mentioned.

The problem is discovered and apparently fixed. I already had bought the
necessary parts.

The key, and it was absolutely necessary, was to disconnect C134 and C107,
the caps that send the vertical and horizontal blanking pulses to the grid-cathode circuit.
This done, all the voltages measured with my 200MOhm
probe made sense.

C109 and especially C108, the oil HV filter caps, were leaking badly: C108
was at best (cold) 25 MOhm, worse when warm. I replaced them with
0.47 uF 1600V film caps. The voltage on the iconoscope went too high
so I added a 1 MOhm 1 watt resistor in series with R125 (330K). The new
caps are larger than the originals, so I mounted them on a terminal strip
under the chassis. (They are polypropylene high reliability ones, and are very large.)

Its now making better images, the problem being mainly horizontal streaking
after bright or dark areas. Its quite stable.

I will now check C116, C117, and C118, the other oil caps. Only C116
is critical: if leaky it would kill V104. Well, no, I looked at the schematic
and it would just raise the cathode current. But I checked and its OK.

And I'm now the owner of two good iconoscopes.

I tried this gizmo the first time since I moved a year ago and it worked fine the first try.

Does anybody have a military or RCA manual for it? I need it for my collection.

Can we get any updated pictures of it operating?

Can we get any updated pictures of it operating?

certainly
The format of this camera is currently 262 @ 60p

Note that it is looking through a window screen.

https://www.videokarma.org/attachment.php?attachmentid=207296&stc=1&d=1748712652

https://www.videokarma.org/attachment.php?attachmentid=207297&stc=1&d=1748712652

I put the camera outdoors on my patio and get a better picture, attached at higher resolution.

This is the very best it can do. I should add that while its a clear sky and sunny, the sky is white due to
high altitude smoke from soviet canuckistan.


Doug

https://www.videokarma.org/attachment.php?attachmentid=207298&stc=1&d=1748716523

thanks

I still need a manual for this gizmo. Web searches are futile.

While running through a test of all my old tubized gear I
tested this thing. It was not working as well as last year
(i.e previous post). I never was able to get it working well with adjustments, even
in the brief periods of sunshine. It did work.

I spent four days trying to diagnose the problem with no success.
It made pictures but the shading was bad, and the horizontal
scan was too wide, i.e. the actual picture area was too narrow on the monitor.

I'm was never able to find what had gone wrong. All voltages and waveforms looked OK per the schematic.
I never got a reply to my query in the post a year ago for a real military manual. So I have no reference for most voltages.

Then the cathode resistor in thge 6L6 horizontal output stage started
smoking. I quickly measured the voltage across it and it was as normal, 17
volts AC ... there is no known correct DC value, but it was rather high.
A NOS 6L6 changed nothing. The resistor and its bypass measured as per the schematic. A new resistor didn't get hot
and nothing changed ... so I have no idea the cause, and must attribute it to a loose flake of
solder shorting something, that fell out while doing the checks.

The HV was only -700 volts, rather low (iconosocpe rating is 1250V), and
the waveform of the horizontal drive at the yoke was wrong at the first part of the scan. I fixed this somewhat by
tweeking the resistor-capacitor series pair across the yoke ( 470 ohm, .02uF) ( it was a small tweek)
and removing the 1 Meg resistor I had installed in series with the
HV rectifier diode. I tried adjusting the width control, but the best setting was at the low inductance end. So I
gave up and added a 2.56 mH "RF choke" across the width coil. This fixed those problems and got 800 volts HV-,
a passable value. It still took a LOT of tweeking to get the values of the two signal
few-pF ceramic trimmer caps set right.

Another problem is that the sync pulses are a larger fraction of the
period than NTSC specifies (its actually 262@60p), because it was designed for 40 Hz, not 60. The overall shape of the
vertical scan was a bit wrong. Adding .22 uF to the horizontal output cathode fixed that and got 900V HV-.

Once done, the picture was a teensy bit better than last year
(not enough better for a new picture here), and it runs on almost all monitors.

I'm still looking for the military repair manual (though, of course, in intended use repairs were, ah, the opposite of what was desired).

One astounding tidbit. Looking at the severe noise (the S/N of a sunlit scene
is like 2:1) I noticed something and measured the frequency response. That's the frequency response looking at the width and shape of a full dark-to-light sharp edge or light-to-dark or stripe test patterns. The 3dB down point.

Its 8 MHz. That's eight megahertz. Eight, 8, e i g h t. And it shows on the high end
Ikegami monitor I'm using. Why would they make such a thing when it was supposed
to feed a transmitter that surely not that wide, nor displays that wide either.

It does demonstrate how sharp iconoscopes can be, such as the talking one in Japan :-)

Given it's a WWII guidance camera and that the military had a lot of leeway on how it used RF during the war it's possible the engineers foresaw the possibility of having a channel wider than 6MHz and higher scan rates for increased resolution for some special application and gave the video circuits enough bandwidth that one could swap the transmitter and tweak the scan to achieve that easily.

One astounding tidbit. Looking at the severe noise (the S/N of a sunlit scene
is like 2:1) I noticed something and measured the frequency response. That's the frequency response looking at the width and shape of a full dark-to-light sharp edge or light-to-dark or stripe test patterns. The 3dB down point.

Its 8 MHz. That's eight megahertz. Eight, 8, e i g h t. And it shows on the high end
Ikegami monitor I'm using. Why would they make such a thing when it was supposed
to feed a transmitter that surely not that wide, nor displays that wide either.

It does demonstrate how sharp iconoscopes can be, such as the talking one in Japan :-)

Is the horizontal scan rate 15750? If it's slower, the bandwidth will be less by the same ratio even though the horizontal resolution is high.

Is the horizontal scan rate 15750? If it's slower, the bandwidth will be less by the same ratio even though the horizontal resolution is high."

Yes, its ~15750 Hz and ~60Hz, locked together, not locked to line.
Put another way, its 262+-1 @ 60p, the latter not locked to power line (since
all the power lines are DC).

But that's immaterial for my main measurement, which actually measured
the rise and fall times of sharp edges in an image. The ringing frequency of the noise agrees.

Given it's a WWII guidance camera and that the military had a lot of leeway on how it used RF during the war it's possible the engineers foresaw the possibility of having a channel wider than 6MHz and higher scan rates for increased resolution for some special application and gave the video circuits enough bandwidth that one could swap the transmitter and tweak the scan to achieve that easily.

The lower the horizontal scan rate the higher, not lower , the resolution in pixels. In any case, I found a statement of 4.5 Mhz bandwidth for that system, and double sideband AM, black at 100% up modulation, but it didn't say whether the 4.5 Mhz was the baseband or the DB channel width. Somebody with a transmitter .... or the manuals ... should know. But nobody has manuals.

I've found that trying to run at NTSC rates rather than 14kHz and 40Hz
(progressive) was what was causing big problems with the shading corrections.

Adjusting some cap values and the width coil value, plus adding a small vertical linear correction directly to the output stage fixed that. It now makes a fairly good picture,
still with sone shading problems even if direct sunlight scenes.

It was not drifted component values.

I've found that trying to run at NTSC rates rather than 14kHz and 40Hz
(progressive) was what was causing big problems with the shading corrections.

Adjusting some cap values and the width coil value, plus adding a small vertical linear correction directly to the output stage fixed that. It now makes a fairly good picture,
still with sone shading problems even if direct sunlight scenes.

It was not drifted component values.

Nice!

Pictures???

The quality in photos has not changed much. Its more visible direct, because you can see the high frequency noise better.
Photos need more than one field exposure, which averages it out.

The quality in photos has not changed much. Its more visible direct, because you can see the high frequency noise better.
Photos need more than one field exposure, which averages it out.

Got it. Was wondering if a shading improvement was visible.

Got it. Was wondering if a shading improvement was visible.

Not for that scene in broad saylight.
I'm still working on the white stripe near the left edge. This is caused by
ringing in the horizontal deflection pulse used for shading. I was never able to get completely rid of it at 15750 Hz.