Converting old CRT tube signals into a digital image

[the_anomaly]

I have a very old JEOL scanning electron microscope.  I'm looking for a way to get digital images out of it.  Attached is the schematic for the photo unit where you can put film and do a long exposure for a crisp image. 

The photo unit receives horizontal and vertical coil scan signal (saw tooth waveform I believe) and the brightness and contrast signal. 

Does anyone know of a plug-and-play style converter unit that could convert these primitive signals into a digital format like VGA/HDMI/Composite video?  Or even an IC chip to build a circuit from?

Thanks!

 

[johnnyzoo]

Interesting project.

So... the signals come from the detector (?) of the electron microscope itself, into this photo unit? Did I understand it correctly?

Do you know if the original manufacturer sold any accessories that would produce a standard video signal?

If not, I would imagine that finding a plug and play solution is unlikely.

Conversion to standard analog video such as PAL or NTSC composite might be very complicated:
Basically, this would mean converting CRT drive signals into a standard video signal. CRT televisions and monitors do the opposite.
Something like this happens in an analog video camera, but using existing camera electronics for this would probably be difficult or impossible. I'm not familiar with this technology so I can't comment much.
Even if some camera electronics could be modified for this purpose, the analog conversion from whatever format this device uses (number of lines, scan rates, etc) into a standard format would be challenging.

So perhaps it would be easier to digitize the signals right away, and do the conversion in software. If you can get high speed DAQ equipment, you can record the signals using your computer and write software that turns them into digital images. I remember seeing some people using this approach to record the digital output of old CGA video cards. Analog capture has some additional challenges.

Modern high performance oscilloscopes might also have enough memory to capture a complete frame at once, so if have access to one, you may not need any additional capture hardware.

If you want a standalone converter, the next step might be using ADCs and a processor of some sort to do this without a computer, but it might be easier to try it on a computer first... unless you are already familiar with high speed AD conversion.

The signaling format must be sorted out. Voltage levels and ranges, number of lines, rate of frames per second and so on. I would assume that one of the deflection signals is sawtooth and the other is "stairs" but I'm not sure. I'm not very familiar with CRT technology so I don't know how the brightness/contrast signals work. Brightness = intensity level, contrast = ratio of high/low intensity levels?

Once you know how to safely digitize the signal, reverse engineering gets a bit easier. Be careful - the CRT uses high voltages and besides injuring yourself, you can fry sensitive electronics if you accidentally plug them into wrong place. Also be careful with correct reference potentials when you connect oscilloscope or data acquistion equipment.

edit: Rewrote my post, the original was written in a hurry.

[Da_Stier]

A first look at the signals you have would be very helpful I guess.
By looking at the schematics, I see H and V as well as brightness.
In theory this is all that you need to get a monochrome VGA signal.
However the H and V timings need to be just right.
(as well as rectangular pulses, but that shouldn't be to hard to do with a few comparators)

I guess one possibility would be to use something like a RetroTINK converter.
https://www.retrotink.com/
These things are made to support all kinds of obscure old video formats.
There are other ones as well, made to adapt all kind of strange sync frequencies to VGA.

If you are able to get an off the shelf solution, something like this would be my best bet.


Greetings,
Michael

[klugesmith]

>>Do you know if the original manufacturer sold any accessories that would produce a standard video signal?

The SEM has its own raster line count and H and V frequencies.
Probably variable under user control, within ranges that make sense for the microscope physics.
Why would they match, or be able to match, NTSC or PAL video standard?
Maybe to give option for SEM to work with standard off the shelf monitor, which doesn't seem to be an obviously valuable feature.

Are there off-the-shelf capture systems that can accept analog computer-monitor video signals,
which have a wide range of raster resolutions and frequencies?

[Twospoons]

I think you are going to have to build your own digitizer, which is pretty trivial these days.
As has been said, the signals are unlikely to conform to any broadcast video standard.

[Hydron]

How fast is the scan speed? If it's on the order of seconds-per-image for high-res long-exposure photo capture then you may even find that the relevant signals sit in the audio range, which could massively simplify the process of getting them into a computer.

[the_anomaly]

Do you know if the original manufacturer sold any accessories that would produce a standard video signal?

I don't.  Its a JEOL JSM-840 which I believe is circa ~1986?  Reading a paper dated 1996 states that it is not economical to retrofit analog microscopes to become digital ones and the 840 is definitely analog.  Now today that might be different but I have not found anything from JEOL.  I've done more research since my first post and aftermarket people talk of using 'frame grabbers'.  So that might be the path I want to head down.  I was thinking convert the raw H/V and brightness 

A first look at the signals you have would be very helpful I guess.

I owe you guys some o-scope screen shots, I'm waiting on longer BNC cables to arrive so I can capture some images. 

I found this video where the author does exactly what I hope to do:
 

/>
Except in his case, he found a composite video signal and as mentioned he was able to use comparators and an ADC to record and send data to a PC.  I am still looking through the schematic book to see if I can find something similar. 

So... the signals come from the detector (?) of the electron microscope itself, into this photo unit? Did I understand it correctly?

Yes, the signal comes from the secondary electron detector which is a photomultiplier tube with a aluminum/phosphor filter that converts reflected electrons to light. 

Apologies for the scan quality.  I am waiting on a wand scanner which I hope will produce better than my phone.

 [ You are not allowed to view attachments ]

Here is a top level view of the connection diagram.  The SEI is sending the raw photomultiplier signal (after an internal pre-amp) to the image control unit.

 [ You are not allowed to view attachments ]

Looking at a more detailed connection diagram, you see the image control unit connects to the CRT display unit through NA8/JC2 and NA7/JC7.  I picked those two for first interest because they are BNC connections and when looking at the CRT display circuit:

 [ You are not allowed to view attachments ]

(top left) ...they are marked as image1 and image2.  I'll post what image 1 signal looks like when I get proper BNC cables.  It is a waveform that I did not see any timing in. 

How fast is the scan speed?

16.7mS / 63.5uS = 263 but since its interlaced x 2 for 526 lines? 

Long story short, I cannot get any reasonable looking images out of it even at very low magnifications.  I was thinking if I could use the longer scan times, this would clean up the image quite a lot.  But recently I saw this image of a 'mostly the same' microscope as mine at an auction site:



The image displayed on the monitor is far clearer than anything I have been able to obtain.  My image is completely filled with shot noise.  Past 500X mag. I can only see noise.  So, I'm starting to think there is something wrong with the image generation circuit.   

Now the microscope has an oscilloscope like function where it can show the horizontal sweep waveform.  It should look like this:



But mine it filled with fuzz.  I can see a waveform inside the fuzz but no amount of control/alignment of beam seems to have any impact.  So perhaps my digital quest is too early.  I'll be able to look at more signals from the SEI detector and so on when my cables order arrives. 

I appreciate everyone's time looking at this and giving suggestions.