EEVblog Electronics Community Forum
Products => Thermal Imaging => Topic started by: Ben321 on May 11, 2021, 10:41:54 pm
-
Today I just received the Electrophysics Micronviewer 7290A that I bought on eBay last week. The seller didn't show any screenshots of a TV with the camera's video output onscreen, but when I messaged him, he said the camera does output video. I got it today, and tried it by connectiong it to the composite video input on my TV, but no video output on my TV screen. I connected up its video output to my Picoscoope USB oscilloscope and tried to see if it at least output something, maybe a waveform with the wrong frequency (maybe it had 50hz instead of 60hz video output, so wouldn't work on a US TV I thought), or maybe I could see some increase in electronic noise when the power was on, or maybe even a single transient voltage spike at the instant the power was turned on. But NOTHING. The video output port always had the EXACT SAME 0-volts output (with a slight amount of noise of a few mV) on my Picoscope, even when the scope was set at its highest sensitivity setting (50mV per division).
I even checked, and double checked, all of the power and video cable connections, to make sure nothing was loose. I made sure the power switch on the Micronviewer was turned on. I checked everything that could result in the camera not being powered, and nothing was wrong with any of my equipment setup. So the output of this thing is DEAD!
There's a number of things that could be wrong with it, from a loose solder connection somewhere on the board, to the power supply for the camera having no voltage output (dead power supply), to possibly even a crack in the glass of the SWIR vidicon tube in the camera (which could actually have been caused by poor handling of the package by the UPS crew, and not necessarily the eBay seller's fault).
Any ideas on what is most likely wrong with it, and what things I should be checking to see if I can fix it?
By the way, one thing that could help, is if somebody could provide the pinout for the power connector. It uses a non-standard power connector with 4 pins arranged in a square and an outer screw-on ring to hold the connector in place when it's connected to the camera. That way I could verify that the power supply is working at least, by testing it with a simple multimeter.
-
Not output can be a lot of things. If you power it off a bench supply (or otherwise measure current consumption), mine starts at 750-800mA at 12V and then settles into the mid/high 600mA range when in operation. I've had one with an issue where it didn't not turning on far enough to see an image where after a few seconds of normal power consumption, it drops to below 600mA.
That said, using a USB dongle that can see either PAL or NTSC and reconfigure automatically, it still puts out a signal, it just never resolves to an image. It will show some gray noise but will correctly output frame timing info required to get a video lock even when the tube isn't reading anything real, the output formatting electronics will do their thing fairly quickly after power on.
It's worth checking the obvious stuff though: change the gain setting knob, sometimes with some settings just a black screen is output, and listening to the slight whine of the switching electronics that drive the tube voltages - should be audible shortly after power on.
As for the power connector, if it's the 4 pin arrangement, it should be just power and ground, using a nominal 12V and connecting directly to a 7810 linear regulator on the inside for a 10V operating voltage for the boards. There is another - I think seven pin - input connector that is for the versions with an integrated battery (which is NiCd and guaranteed to be dead), but the connections should be obvious if you take the lid off and look at the wiring on the inside face. Maybe some of the documentation can show which pins are used.... or maybe even one of the threads? I converted mine to a barrel jack and never looked back.
-
It's worth checking the obvious stuff though: change the gain setting knob, sometimes with some settings just a black screen is output, and listening to the slight whine of the switching electronics that drive the tube voltages - should be audible shortly after power on.
That's one of the things I'm NOT hearing. I don't hear the high voltage high frequency tube power supply circuit running. So I suspect that either the internal HV supply is dead, or the external 12V power supply is dead, or there's a loose solder point somewhere in the power circuits in the camera somewhere between the power-input jack and the HV power supply.
As for the power connector, if it's the 4 pin arrangement, it should be just power and ground, using a nominal 12V and connecting directly to a 7810 linear regulator on the inside for a 10V operating voltage for the boards. the connections should be obvious if you take the lid off and look at the wiring on the inside face.
What are the minimum number of screws that need to be removed to open it? There's 4 screws on top, 4 screws on the back, and 8 screws on the bottom. I wish I could get a service manual for it or something, that would show exactly the right way to open it. I want to be able to put it back together again.
It would be great if Fraser could show up in this thread. He usually has info on exotic IR equipment, info that I never thought anybody outside of the companies that make the equipment even had access to. Such info would be great in this thread, as it might help me fix my Micronviewer.
-
I think he's got one, but I've got one on the bench beside me so I can give you at least some info ;) I think the only manuals and things that have turned up have been linked in the main 7290A thread.
You only need to remove the four screws on top and the filter holder if your unit has one. If your unit has the side handle (unlikely if you don't have the internal battery) then one of the screws for the handle goes into a standoff on the board and also needs to be removed.
-
Quick observations.....
This will start as a standard ‘dead unit’ investigation. That is to say, Visual inspection for physical damage or visible component failures (including fuses!) followed by simple checks on power and control. If the basics are OK, you can then begin the fault tracing process on the PCB’s.
1. Always check for power at the connector, preferably inside the camera, to ensure that power is reaching the power input regulator board.
2. Check power output of regulator board mounted in the rear of the camera
3. Monitor current draw at a convenient point in the power supply rail.
4. Check that wiring from video connector to the PCB is intact.
5. If power is present, check for heater glow from the rear of the Vidicon tube.
If you have power at the output of the regulator board then the problem can be more complex in nature and further investigation of the PCB’s will be required.
A broken Vidicon tube will not cause a ‘no signal’ condition at the video output as other circuits produce synchronisation pulses even if the Vidicon signal is absent
WARNING : If you are not familiar with Vidicon cameras, be careful where you poke your fingers as high voltages are used for the tube and they will ‘bite’ if touched. Painful but not deadly.
Fraser
-
In case you are wondering, SOFRADIR refuse to release the schematics or service information for the 7290A camera. It is basically a standard ‘late design’ of Vidicon CCTV camera though, so not hard to reverse engineer if you have the time. Start by downloading the data sheets for any chips found inside and that gives you test points to check ;)
Fraser
-
As a side note, you may wish to start searching for a suitable lens for your camera, if it was not supplied with one.
The lens needs to designed for the old 1” Vidicon standard to achieve full coverage of the target. 2/3” and 1/2” lenses can cause vignetting. I bought a Fujinon 25mm F1.4 1” Vidicon lens with built in manual IRIS. It was not cheap ! These are excellent lenses but not able to extend much further into SWIR than 1.6 um as they are standard optical glass. They do not have an IR coating as found on some modern lenses and this is important. You want plain old fashioned uncoated optical glass. The good quality older CCTV lenses have been in demand in recent years as photographers use them in 4/3” photography. Prices are sadly higher as a result.
Fraser
-
Quick observations.....
This will start as a standard ‘dead unit’ investigation. That is to say, Visual inspection for physical damage or visible component failures (including fuses!) followed by simple checks on power and control. If the basics are OK, you can then begin the fault tracing process on the PCB’s.
1. Always check for power at the connector, preferably inside the camera, to ensure that power is reaching the power input regulator board.
2. Check power output of regulator board mounted in the rear of the camera
3. Monitor current draw at a convenient point in the power supply rail.
4. Check that wiring from video connector to the PCB is intact.
5. If power is present, check for heater glow from the rear of the Vidicon tube.
If you have power at the output of the regulator board then the problem can be more complex in nature and further investigation of the PCB’s will be required.
A broken Vidicon tube will not cause a ‘no signal’ condition at the video output as other circuits produce synchronisation pulses even if the Vidicon signal is absent
WARNING : If you are not familiar with Vidicon cameras, be careful where you poke your fingers as high voltages are used for the tube and they will ‘bite’ if touched. Painful but not deadly.
Fraser
I was just going to start my testing, but found the 9V battery in my multimeter is dead, and I don't have any spare 9V batteries available. I need to buy some more.
I was able to get some pictures inside though. I numbered and named the pictures. I hope that maybe some defect on the boards (maybe some solder point that had come undone or something) that was to small to be obvious to me, may be easily seen by somebody else, and that these pictures will help them to help me pinpoint the issue with my camera.
-
I have no info on this camera but some experience in video.
Are you terminating the video out line with 75 ohm resistive?
I have seen video output circuits that rely on a DC path to ground via the 75 ohm load at the 'far end of the coax'. Connecting a capacitively-coupled or high impedance load results in zero video output. I have no idea if this is the case with the Micronviewer but I have seen other perfectly good kit scrapped because the technician had tested it with a non DC-coupled monitor and decided, erroneously, that the device was dead.
-
Ben321,
You can use your Picoscope to measure DC voltages so you can do the described tests without the multimeter.
Fraser
-
I have no info on this camera but some experience in video.
Are you terminating the video out line with 75 ohm resistive?
I have seen video output circuits that rely on a DC path to ground via the 75 ohm load at the 'far end of the coax'. Connecting a capacitively-coupled or high impedance load results in zero video output. I have no idea if this is the case with the Micronviewer but I have seen other perfectly good kit scrapped because the technician had tested it with a non DC-coupled monitor and decided, erroneously, that the device was dead.
I didn't even hear the high pitch tone of the flyback transformer, for the vidicon tube when I was testing it. Furthermore, my first test was one where I actually hooked it up to the composite video input of my TV, not my Picoscope with a high impedance impedance. The specification for composite video inputs them to have a 75ohm DC resistance to ground. So if that was the only issue with my camera when testing with a picoscope, then it still would have worked with the TV, and if it worked with the TV, I never would have needed to test it with anything else, and this thread asking for help wouldn't exist. And the camera wasn't having an issue no video content either, as in that case the sync pulses would still be present, and my TV would display a black screen without any error messages. However with this camera, my TV displayed a "signal not found" type of message, meaning that there was literally no signal coming out of the camera's video-out port.
-
If it's this dead and you paid for a working one, I personally wouldn't hold onto it unless it was heavily discounted, but if you want to check voltages, you can do that with a scope and a 10x probe, as mentioned, making sure you're within your probe's operating specs.
As said before, input should be 12VDC across the two connected terminals of the jack, they should go straight into a 7810 regulator (looks like it's on the board attached to the back panel on your unit, but I've seen it strapped to the bottom chassis as well) where you should get 10V out. On the smaller of the two side boards there is a TO-220 with a piece of metal heatsink which goes across the tube, and this is the primary 5V regulator, so it should be 10V in, 5V out, and system ground.
For a more in depth check, looking at the larger of the two side boards, you can check the circled and numbered pins near the potentiometer marked "Target". Each circled pin with a number is the corresponding voltage test point, and all of these are derived from the main switching transistor (a TO-220 and an inductor in the lower left of the larger board when looking at it installed on the camera), so if there are voltages in the right ballpark present, then the tube is being driven by something, at least. Don't probe the 300V or 450V marked ones if your probes are not rated for it, and probably don't probe any of them (maybe with the exception of the target voltage itself) if you're using a 1x probe. The only units I've worked with do have a slight whine from the switching or from the Hsync generation, but the units I've worked with also show a blank output with signs of doing something rather than no signal when powered on, so if you have 10V and 5V on the board, your problem probably lies within video signal generation, not driving the tube and not reading from the tube.
Again, unless you have an arrangement with the seller it seems odd to be poking around in there, but none of what you describe is normal behavior for a working or nearly working unit, in my experience.
-
Ok got a new 9V battery for my multimeter. I measured the voltage across the pins in the power supply and 0V. So either the power supply is dead, or else whoever wired the power supply's output connector soldered the output wires to the wrong pins on the output connector. The connector has 4 pins, but only 2 of them are used. So my next step is to cut output cable from the power supply, and directly measure the voltage without the end connector.
-
Ok so I cut the power supply output cable and directly measured the voltage coming out of the cable. It is 12V just as it's supposed to be. As for the power output connector, I used the resistance setting on my multimeter and found that there's infinite resistance from all 4 pins to both wires in the cable! So that means NO PINS on the power supply's output connector are actually wired into the power supply! I have no idea how that happened, but at least this seems to be the problem, meaning the camera itself is likely still good. I could spend hours trying to manually solder the power connector myself (though I don't even know how to open the power connector to expose its solder points, and even if I could the pins are so close together that my hands wouldn't be steady enough, as I tend to have a tremor in my hands when trying to do precision work, which ironically I DON'T have when NOT trying to do precision work). It seems that my best bet may be to simply contact whoever makes (or used to make) these cameras, and ask if they can send me a new power supply for it.
-
In case the information is useful to someone. I attach pictures of the dismantled connector and a link to the thread that identifies it and where to purchase one. The connector is from the Hirose (HRS) SR30 range.
https://www.hirose.com/product/series/SR30# (https://www.hirose.com/product/series/SR30#)
https://www.eevblog.com/forum/chat/mystery-industrial-camera-connector-help-please/ (https://www.eevblog.com/forum/chat/mystery-industrial-camera-connector-help-please/)
When soldering such a connector, it is relatively easy to make up a ‘jig’ to hold the connector and wires in place whilst applying the soldering iron and solder to the pins. It can be achieved with nothing more than some pieces of cardboard and tape ;)
Fraser
-
A little hint for anyone who suffers from hand shake when micro soldering..... I use a small wheat pillow out of a microwaveable hot pad as a stable supporting surface on which to rest my hand whilst soldering. Any sort of relatively dense small pillow will work but I had some microwave heat pillows laying around. The wheat pillow also works well for supporting a PCB at an angle for soldering as it is heat resistant and mouldable to shape :-+ These heat pads come in many sizes and can be expensive. Mine came out of a soft toy ‘Hottie’ Cat that is used where a hot water bottle might normally be used. There are all manner of shapes for placing around the neck and on limbs to treat pain. Do not overpay for such an item. My cat was around £7 and a neck warmer is about the same on Amazon. As an added bonus you can microwave it in winter to keep your hand warm or use it on aches and pains when not soldering :-+
https://www.amazon.co.uk/Microwave-Cushion-Reliever-Fragranced-Lavender/dp/B00NXZ7UV6/ref=asc_df_B00NXZ7UV6/?tag=googshopuk-21&linkCode=df0&hvadid=207995677378&hvpos=&hvnetw=g&hvrand=14394440764132575171&hvpone=&hvptwo=&hvqmt=&hvdev=t&hvdvcmdl=&hvlocint=&hvlocphy=9046142&hvtargid=pla-807623317674&psc=1 (https://www.amazon.co.uk/Microwave-Cushion-Reliever-Fragranced-Lavender/dp/B00NXZ7UV6/ref=asc_df_B00NXZ7UV6/?tag=googshopuk-21&linkCode=df0&hvadid=207995677378&hvpos=&hvnetw=g&hvrand=14394440764132575171&hvpone=&hvptwo=&hvqmt=&hvdev=t&hvdvcmdl=&hvlocint=&hvlocphy=9046142&hvtargid=pla-807623317674&psc=1)
Fraser
-
I could spend hours trying to manually solder the power connector myself (though I don't even know how to open the power connector to expose its solder points,
Just go onto the red and black wires inside with clips from a bench supply - then you'll see if it fires up and is worth the effort.
Bill
-
Ok, so the power supply I ordered FINALLY came. I bought it from Cascade Laser (the same company that used to distribute these Micronviewer cameras). They charged me the RIPOFF PRICE of about $200 for the power supply! While I'm upset they decided to price gouge me, they did say that they had to contact their manufacturer to get one made (apparently these aren't a stock item, now that the camera itself is seems to be obsolete), so they may have considered it worth more as a custom item or whatever.
So, I got the power supply and plugged it in, and got some pics with my USB video capture dongle. I've attached these pictures to this post. They include pictures with no lens and AGC, and also with a lens with all 4 gain settings (AGC, MGC low, MGC medium, and MGC high), and with the lens capped on the first 3 gain settings (AGC, MGC low, and MGC medium). The no-lens picture (entire sensor illuminated with room light), showed an interesting rectangular feature that's at an angle. I'm not sure what it is. Is it normal or some defect?
The various gain pictures, show that not only is the gain (amplification) of the signal altered, but so is the offset (blackpoint). By the time you have the gain up to high, even WITH the lens cap in place (I didn't bother to test it with the lens cap off, as the entire picture would be washed out and useless), the entire picture appears white. Likewise, when the manual gain setting is medium, it appears a flat gray color across the image with the lenscap on, and with the lens cap off, the entire picture's blackpoint is this gray color (nothing appears any dimmer than that gray color). I suspect that this is NOT how the gain should be working. AGC however seems to work properly (adjusting only the gain, while keeping the proper blackpoint).
With the lens cap on, and AGC also on, I noticed that the last image the camera saw, appears to be "burned in" to the tube sensor temporarily, even if that image wasn't from a bright light source (just normal room illumination). It takes a minute or 2, for the image to finally return to normal (which with the lens-cap on is a flat gray, as the AGC finds the max and min levels are approximately the same, and so adjusts the picture to be mid-gray as the image brightness level).
Also, no matter how I adjust the lens, I can't get it to be very sharp. I suspect this may be an electron beam focusing issue in the image sensor tube. I may need to somehow calibrate the electron beam focusing.
Also, in case you are wondering, the object this camera is pointing at is my bed. Above my bed is a solar system poster, and there also happens to be a cardboard box on the floor (I should have moved it before capturing these pictures). Also, I was only able to capture anything at all, because this camera can see both visible light and the NIR+SWIR parts of the spectrum. My room lights are incandescent, so emit almost 100% visible light. I'm still going to need to get an NIR or SWIR longpass filter for this camera. That will cost about $80 though, so I think I will wait a bit, as I've already spent $500 on the camera, and $200 on the replacement power supply.
-
One more thing I noticed, is that the image appears to be offset (vignetting only on one side of the image). I don't think the lens holder is out of place, so I suspect this is an electron beam aiming issue (horizontal deflection coil calibration issue).
-
Also, as can be seen in the attached picture, a bright light source, such as the IR LED in an TV remote control, even if not pointed directly at the camera, produces a dark artifact off to the right of the light source. Any idea why?
-
Also it has some less than ideal response times for darker objects. I've attached 2 pictures. In the first picture, I'm moving my hand VERY slowly in front of the camera. My hand is darker than the background. Because my hand is moving at all, instead of staying absolutely still for like 10 to 20 seconds, my hand looks like a ghost image. Meanwhile, when I move my IR TV remote control in front of the camera, while holding down a button to activate its IR LED, every single bright pulse is captured completely, and leaves an after image that fades slowly.
Is these the kind of response times I should expect with this camera? Or is this a defect?
-
One more issue I found. If I touch the metal part of the camera lens (but no part of the metal case), I see electronic interference in the image. I don't know why. I can only assume that the high speed signals (such as image generation electronics, and beam sweeping electronics) are somehow close enough to the lens, that the metal parts of the camera lens are capacitively coupled to these circuits, and that when I touch the lens, by body is actually altering the electronic properties of the circuit by changing a capacitance in the circuit (like a capacitive touch screen). Alternatively, my body may be acting as an antenna for other enterfering signals from other things (like switching power supplies in other electronic equipment), and while that's blocked by the metal case of the camera normally, when I touch the lens, which is appreciatively coupled into internal circuits, I'm basically acting as a receiving antenna, and the camera's electronics are being impacted by the external signals being coupled into them.
Is this something that can be compensated for? Is this a defect in my unit, or a common problem with Micronviewer units?
-
So, as you can see, I have lots of issues here, and I don't if all of them are defects, or if some are just the way the camera works. I really need help identifying the defects, based on the pics and comments I've posted. And where the defects are, I'd like to know which ones could be fixed by myself, versus which would require sending my camera in to be repaired professionally. For the ones I can fix, I'd like some tutorials on how to fix those issues.
Maybe you @Fraser could help me. You think you know a lot about these tube-based cameras. Right?
-
Some of these are discussed in the other thread, but they sound mostly normal.
For the rotation, it's either the focus pot, the pot with a similar function on the yolk board, or it's the rotation of the coils around the tube (there are set screws in the frame that can be loosened to change this).
For the dark artifact, I'm guessing it's that too high an exposure is causing an artifact because of saturating the detector in the region. If you keep it moving, it probably won't still do it, and using a narrower iris setting on your lens or an off angle view of the LED will probably get rid of it.
Finally the interference - do you have the case on? There's a reason there are several chassis to board grounding wires that are in place and the short of it is that the frontend is pretty sensitive to EMF and when you touch the lens or even get your hand near it, the EMF picked up on your hand is enough to cause visible noise. The frontend FET is located on the side board, closest to the lens, which is on the right side of the camera when the lens is pointing away and you're looking down on the top, so you'll probably see the most noise getting close to that. My unit shows the same when the cover is off, and none with it on, so if yours is showing it with the cover on, it's probably good to check the various ground tie points to the chassis, the shielding on the tube, and the boards to make sure they are all in place.
-
Also, as can be seen in the attached picture, a bright light source, such as the IR LED in an TV remote control, even if not pointed directly at the camera, produces a dark artifact off to the right of the light source. Any idea why?
Input amplifier frequency response being shown up by the HUGE signal going in. A similar effect can sometimes be seen on old Vidicon TV images where spot lamps are visible.
It looks reasonably OK to me, as the only adjustment was usually a capacitor value (based on Pevicon TIC cameras).
Bill
-
My room lights are incandescent, so emit almost 100% visible light.
No, will be 5% visible and a lot of NIR through to LWIR 'heat'.
That is why a 10W LED is as bright as a 100W incandescent - 5W of heat vs 95W of heat
-
Many of the things you're experiencing are just down to the fact that it's a vidicon-based camera. In particular, the interference when your hand is near the front of the tube is perfectly normal. It's mainly down to the (necessarily) extremely high input impedance of the first video amplifier.
May I suggest you buy an old vidicon based security camera such as a Pye Lynx (or even the camera that the MicronViewer is based on) and experiment with that? Apart from the different frequency response a visible light vidicon camera will behave much the same as the NIR one, except that you won't care if you damage it.
-
or even the camera that the MicronViewer is based on
Wait. So there's also a visible-light version of the Micronviewer? What company made it. And what's its model number?
-
My room lights are incandescent, so emit almost 100% visible light.
No, will be 5% visible and a lot of NIR through to LWIR 'heat'.
That is why a 10W LED is as bright as a 100W incandescent - 5W of heat vs 95W of heat
I made a big OOPS there. I meant to say since my room has LED lights INSTEAD OF incandecent lights. That's why I was able to verify that the specs that say this SWIR camera can also see visible light, are certainly correct. The lights in my ceiling emit nearly 100% visible light, because they are LEDs.
-
or even the camera that the MicronViewer is based on
Wait. So there's also a visible-light version of the Micronviewer? What company made it. And what's its model number?
I believe I read on EEVblog that the MicronViewer was based closely on a CCTV camera but I can't remember which. A quick search should find the answer.
It doesn't surprise me that the camera can see in the visible spectrum:
The 7290A uses a PbS coated Vidicon tube and has a bandwidth of 0.4um to 1.9um.
0.4um is violet, verging on UV. The tricky thing with the MicronViewer is keeping visible light out - an 850nm or 950nm screw-in photographic-type filter from eBay is a good place to start. Expect to pay <US$20, but do be careful to get the right physical size to screw into the front of the lens (ie not the tube end). 850nm is usually cheaper than 950 - I don't know why - but I do think you will get better imaging 'value' from a 950nm filter.
I have done a lot of playing with 700, 750 and 850nm filters on 'full-spectrum' DSLRs and the world does look different with the different filters. Have a look here (https://www.flickr.com/photos/ultrapurple/albums/72157617230652040) for a few sample images taken with these filters. There is no benefit in paying top dollar for a top brand filter: the resolution of the Micronviewer is far lower than even the shoddiest filter's performance.
Why 950nm? A 'full-spectrum' digital camera with a silicon sensor will respond down to almost 1.1um so by putting onto the MicronViewer a low pass filter as close as you can get to the silicon cutoff, the less overlap you will get. 'Full-spectrum' DSLRs are fairly easy to obtain, or you can just (destructively) modify a used camera. I use a converted Nikon D600 as my main full-spectrum camera, but I also have a couple of Fuji IS Pro (https://en.wikipedia.org/wiki/FinePix_IS_Pro) bodies.
-
I believe I read on EEVblog that the MicronViewer was based closely on a CCTV camera but I can't remember which. A quick search should find the answer.
Yes, at least the 7290 (non A version) I got is based on the ITC510 CCTV camera: https://www.eevblog.com/forum/thermal-imaging/swir-electrophysics-micronviewer-7290a-user-manual/msg3265658/#msg3265658 (https://www.eevblog.com/forum/thermal-imaging/swir-electrophysics-micronviewer-7290a-user-manual/msg3265658/#msg3265658)
As to seeing in the visible spectrum, these camera's are essentially a vidicon based camera, where the vidicon has an improved sensitivity (see attached datasheet) towards the IR, but they operate as a "normal" vidicon in the visible. That is why it can be very useful to install a long pass filter on the camera lens.
-
I believe I read on EEVblog that the MicronViewer was based closely on a CCTV camera but I can't remember which. A quick search should find the answer.
Yes, at least the 7290 (non A version) I got is based on the ITC510 CCTV camera: https://www.eevblog.com/forum/thermal-imaging/swir-electrophysics-micronviewer-7290a-user-manual/msg3265658/#msg3265658 (https://www.eevblog.com/forum/thermal-imaging/swir-electrophysics-micronviewer-7290a-user-manual/msg3265658/#msg3265658)
As to seeing in the visible spectrum, these camera's are essentially a vidicon based camera, where the vidicon has an improved sensitivity (see attached datasheet) towards the IR, but they operate as a "normal" vidicon in the visible. That is why it can be very useful to install a long pass filter on the camera lens.
Wait, there's a non-A version? I know there's a standard version of the Micronviewer (with model number which ends with an A), and an extended wavelength model that can see beyond 2000nm (which ends with an A-EX if I remember correctly, with EX meaning extended). I never knew a model ever existed without the A at all.
-
On another note, this camera DOES appear to require a 50ohm impedance on the TV (or video-capture card) to work correctly. In order to try to further debug my Micronviewer (read some of the various problems I've encountered, even after getting a working power supply, on the first page of this thread), I decided to connect the camera up to my Picoscope. Even though I bought the cheapest Picoscope, which has a theoretical maximum sample rate of 1 million smp/sec, I found it has an actual maximum sample rate of 6.25 million smp/sec. This lowers the horizontal resolution I'm able to capture at (about 395 total pixels per line including all h-sync and blanking, with only about 325 image pixels per line). However, this is still plenty of pixels to get a good idea of what an image should look like. After using my Picoscope as a high speed digitizer to capture raw samples to a WAV file (using my own custom Picoscope recorder software that pushed the Picoscope to its maximum possible samplerate, which the official software doesn't do), I used Goldwave audio editor to center the signal values on the blanking level, and then used some of my own custom software to read the sync pulses to get all the lines to line up (end result looks like a film strip with many pictures), and then loaded that into Gimp's raw image viewer and cropped the result to be the size of a single field (I didn't do any interlacing to get a frame, just cropped it to a single field).
The result, while not as good as a dedicated video capture card, still should give a good idea of what you would see on a TV (and has for many other video sources I've used this technique with). But it didn't. I have captured many pics using a USB video capture dongle (that's how I got the pics from the first page), and what I saw from this Picoscope capture shows no similarity in the image content. The sync and blanking areas are correct, but the image content area is messed up real bad. I suspect this is because my Picoscope is a high-impedence input (I think it's 1MOhm) while I know that video signal sources expect to be connected to a low input impedance device (usually about 50 Ohms), and some (like this camera it seems) will not even generate a usable image UNLESS they are connected to a low input impedance device.
I've attached the picture captured using my above described method to this post.
Do you know if there's a 50 Ohm BNC or RCA adapter I could get, that is designed to go between an oscilloscope and a video signal source? It would be like a normal coaxial adaptor or extender, except instead of having a completely isolated pin connector and outer shield connector, it would have a 50 Ohm resistor between the center pin and the outer shielding connector, to give it the 50 Ohm DC resistance that video signal sources are usually plugged into (allowing all video signal sources to work with high-impedance input oscilloscopes).
-
Video signals are typically 75 ohms, and this one should be no exception. I'd put a T connector on the scope input and a 75 ohm termination one one side, then the camera input into the other. If you need to capture the video output as well, the video receiver should take care of the termination, so you'd just want your scope in a high impedance input mode. If your picoscope is really only getting that low of a samplerate, its input bandwidth is probably too low for the signal and subsequently the input capacitance too high to capture the signal properly - a video signal could take up to 6MHz of bandwidth (don't think this one goes that high), so to make sure your measurement instrument isn't effecting signal quality too much, I'd probably look for a minimum 50MHz bandwidth frontend with a 1M ohm or more rated input. Overkill from what's required, but good enough that you shouldn't have any significant video detriment from the frontend of your scope.
-
Video signals are typically 75 ohms, and this one should be no exception. I'd put a T connector on the scope input and a 75 ohm termination one one side, then the camera input into the other. If you need to capture the video output as well, the video receiver should take care of the termination, so you'd just want your scope in a high impedance input mode. If your picoscope is really only getting that low of a samplerate, its input bandwidth is probably too low for the signal and subsequently the input capacitance too high to capture the signal properly - a video signal could take up to 6MHz of bandwidth (don't think this one goes that high), so to make sure your measurement instrument isn't effecting signal quality too much, I'd probably look for a minimum 50MHz bandwidth frontend with a 1M ohm or more rated input. Overkill from what's required, but good enough that you shouldn't have any significant video detriment from the frontend of your scope.
Thanks. I just ordered those now on Digikey. It should take 4 days for shipping it said, so when I get it I'll let you know the results.
-
I have resisted commenting on this thread as there are other forum members with far more practical knowledge of this camera than me. I will, however, make few comments now….
The first sample picture you showed us displays a rotated rectangle. That is image burn and, at fist look, appears to be a rectangular mask created geometric burn with the imaging Vidicon tube having been rotated after the burn occurred. I then thought more about this and it could just be that he camera spent its life observing a rectangular test table and the tables image has burnt into the cameras target. The correct Vidicon tube orientation may be checked as detailed in the data sheets. Vidicon tubes are designed to be orientated in a certain way within the deflection yoke.
The images produced by your camera appear reasonably good with only the unusual vignetting on the left hand side causing concern. Please check the camera with the filter cartridge removed from its slot ro ensyre that he filter cartridge is not causing the problem though incorrect positioning in its carrier. That vignetting has the appearance of such a circular filter holder misalignment.
The camera is supposed to be 75 Ohm output impedance and too high an impedance just causes the amplitude of the signal to increase beyond the specification for a standard composite video signal. The effect of this is to really mess up the video signal image content when recorded by a video capture system.
The Micron viewer was originally released as the 7290 model and it was basically a modified commercial Ikegami CCTV camera. As the Vidicon tube of a CCTV camera and SWIR camera have similar needs this was an obvious path for Electrophysics to take and saved development time.
The 7290 appears to have existed with at least two different ‘Donor’ CCTV camera platforms as might be expected as models became obsolete. The 7290A appears to be a completely new custom camera design, likely built under contract for Electrophysics when the original ‘Donor’ CCTV cameras became obsolete and no new models were released due to the introduction of CCD digital imaging technology. The 7290A model uses more modern components in its design and it was the same hardware up until the camera was discontinued. The 7290A could be fitted with the standard or extended range SWIR Vidicon tube as you know and there was also a handheld portable version that was battery powered and fitted with an electronic viewfinder….. much like very early monochrome home video camera systems.
-
I have resisted commenting on this thread as there are other forum members with far more practical knowledge of this camera than me. I will, however, make few comments now….
The first sample picture you showed us displays a rotated rectangle. That is image burn and, at fist look, appears to be a rectangular mask created geometric burn with the imaging Vidicon tube having been rotated after the burn occurred. I then thought more about this and it could just be that he camera spent its life observing a rectangular test table and the tables image has burnt into the cameras target. The correct Vidicon tube orientation may be checked as detailed in the data sheets. Vidicon tubes are designed to be orientated in a certain way within the deflection yoke.
The images produced by your camera appear reasonably good with only the unusual vignetting on the left hand side causing concern. Please check the camera with the filter cartridge removed from its slot ro ensyre that he filter cartridge is not causing the problem though incorrect positioning in its carrier. That vignetting has the appearance of such a circular filter holder misalignment.
The camera is supposed to be 75 Ohm output impedance and too high an impedance just causes the amplitude of the signal to increase beyond the specification for a standard composite video signal. The effect of this is to really mess up the video signal image content when recorded by a video capture system.
The Micron viewer was originally released as the 7290 model and it was basically a modified commercial Ikegami CCTV camera. As the Vidicon tube of a CCTV camera and SWIR camera have similar needs this was an obvious path for Electrophysics to take and saved development time.
The 7290 appears to have existed with at least two different ‘Donor’ CCTV camera platforms as might be expected as models became obsolete. The 7290A appears to be a completely new custom camera design, likely built under contract for Electrophysics when the original ‘Donor’ CCTV cameras became obsolete and no new models were released due to the introduction of CCD digital imaging technology. The 7290A model uses more modern components in its design and it was the same hardware up until the camera was discontinued. The 7290A could be fitted with the standard or extended range SWIR Vidicon tube as you know and there was also a handheld portable version that was battery powered and fitted with an electronic viewfinder….. much like very early monochrome home video camera systems.
The amplitude is not actually exceeding the specs of my scope. It maybe in the camera itself overdriving some kind of limiter circuit, but on the scope end it wasn't overdriving anything.
As for the images captured with a video capture dongle, I'm not so sure that the vignetting is caused by the misalignment of the filter holder, because when I remove the lens the filter hole appears exactly centered. My hypothesis is that it's an H-Offset adjustment issue with the deflecting coil.
Another issue I notice is at the left edge of the image where it fades out has a serrated edge to it, not just following the contour of the darkest parts of the image.
Also there appears to be a signal gain/offset issue in the image. where even the darkest parts of the image (even with the lens cap in place) aren't actually black. Instead they always are above black as a dark gray. The only part that is truly black is outside of the actual image (just to the left of where the vignetting is occurring on the left side of the image).
-
So just an update now. I have done some repair work on the camera itself.
See this picture uploaded previously for reference to what the image was like before repairs.
https://www.eevblog.com/forum/thermal-imaging/whats-wrong-with-my-electrophysics-micronviewer/?action=dlattach;attach=1226891;image (https://www.eevblog.com/forum/thermal-imaging/whats-wrong-with-my-electrophysics-micronviewer/?action=dlattach;attach=1226891;image)
Notice that the actual image is offset to the right of the frame.
Notice that edge of the vignetting isn't circular, but instead suddenly drops from gray to black, where the darkness of the image is too low, which depends on not only vignetting but also image content, making an irregularly shaped vignetting area.
Notice how the edge of the left edge of the image, where it drops into vignetting, has a serrated look to it (kinda like the teeth on a comb) instead of simply being a normal edge.
As for the repairs, my first one was to correct the horizontal offset of the electron beam. My hunch was that the reason the picture was shifted too far to the right, was more than just a lens miss alignment. The filter tray wasn't pushing the lens or anything, although the lens itself (if it had maybe been dropped by the previous owner, as I bought the lens with another CCTV camera on eBay) could have its internal optics slightly out of alignment. However, I figured that even that optical missalignment wouldn't account for such a large image offset as I was observing here. So I went with my hunch that the reason the image appeared too far to the right was that the electron beam was scanning to far to the left, and so I adjusted the H-center potentiometer to move the image to the right. My hunch was correct! I was able to bring the right part of the image back into the frame, something that would not have been possible if the lens was simply out of alignment, and the lens was projecting the image onto the tube's surface outside of the light sensitive area. I didn't completely recenter it though (just pretty close to center, enough to see the full right side of the actual image), because this lens is used and does have trouble focusing as good as it should (though it does focus pretty good), and so some of the optics in the lens may be slightly out of alignment (thus somewhat contributing to the horizontal offset issue), and I don't want to accidentally overcorrect for horizontal offset using the electron beam scanning controls, because if I get a new lens that's better quality, I don't want to need to re-correct the horizontal offset again.
My next repair was to fix the strange brightness issues that caused the brightness of a given pixel to jump from gray suddenly to black when the amount of light exposing that pixel was too low. For this I adjusted the black pedestal potentiometer (which I suspected I needed to do. Black pedestal refers to the 7.5 IRE offset for black above blanking level (0 IRE) in the NTSC specification. I had a lot of trouble with getting it set, because for this to be set, I needed to keep the image black, but even a capped lens wouldn't have worked for this, because with the case off (needed to adjust the potentiometer) the tube was still exposed to light. I eventually gave up on this effort, put the black pedestal potentiometer back to where it was to start with, and decided to close the case, just being happy that I'd recentered the image.
After putting the case back on and testing it, I noticed if I bumped the case by accident (for example while handholding the camera and pointing it around at different things, but accidentally bump it into something nearby) the image would flicker briefly with horizontal lines through it, and sometimes the overall image brightness would change (sometimes for just a second, but other times it would stay at a new brightness level). And it didn't even need to be a hard tap. Any kind of momentary mechanical force, a "bump" (but not a constant pressure being applied to the case), even if only a very LIGHT TAP, was able to cause this. This seemed to indicate that SOMETHING is loose SOMEWHERE (hopefully not a bad solder point, because I would never try to do a solder repair on something this rare and valuable, because it could be broken even worse if I made any mistakes). This issue of what must be a loose connection somewhere, has actually been an issues since I first powered on the camera to test it after I bought the replacement power supply for it.
I thought to myself then, I wonder what happens if I INTENTIONALLY tap it, will that flickering get better or worse? So I tried tapping the case several more times, and the coolest thing happened after I stopped tapping it after the last tap. I noticed that ALL of the remaining issues were gone! No more sudden dropoff from gray to black. Vignette was now perfectly circular. And the left edge of the image was no longer a serrated edge like the teeth on a comb. Just like in a movie, where somebody gives a few whacks to a broken piece of equipment, and it suddenly starts working, this ACTUALLY WORKED IN REAL LIFE for this camera!
Of course, if there's a loose connection still, this may have temporarily made the connection better, but I wonder how loose the connection is, and if it's still loose, how strong of an accidental tap will it take to bring back the same problems I fixed by intentionally tapping the case. Whatever the case, I immediately captured a picture from it just after getting it fully working, and I've attached that picture to this post.
-
Of course, if there's a loose connection still, this may have temporarily made the connection better, but I wonder how loose the connection is, and if it's still loose, how strong of an accidental tap will it take to bring back the same problems I fixed by intentionally tapping the case. Whatever the case, I immediately captured a picture from it just after getting it fully working, and I've attached that picture to this post.
Good to see it looking a lot better.
The problem could be a bad pot. The images you posted show rather prehistoric open frame carbon track pots. They can get very 'noisy' and touch sensitive over time.
Bill
-
Not wanting to cause unwarranted concern but I have seen similar ‘image offset’ behaviour in cathode Ray tubes that have been dropped in transit. The cause of the offset was found to be a misaligned electron gun that was no longer in correct axial alignment with the tube body. The electron gun is often mounted on a combination of stiff wire and glass supports that form a cage on which the various parts of the electron gun and associated grids are mounted. If a cathode Ray or Vidicon tube are exposed to a drop event that does not result in implosion, the whole ‘cage’ that supports the electron gun can shift downwards towards the point of impact due to kinetic energy.
You may be able to correct fir this gun offset with the positioning controls that act upon the deflection coils and it would appear that you have been able to do so in this case.
Regarding intermittent connections in equipment….. there is an age old technique fir finding such…… use a small plastic handled screwdriver …. Hold it by the metal tip and use the insulated handle as a miniature hammer to gently tap the the PCB’s in a methodical manner working across their surface and monitoring the effect on a display. Some microphony is normal. The vidicon tube may be tested in a similar manner by tapping its deflection you’re gently. Vidicon tubes are microphonic so you will see some weird effects on the image and this is normal. This technique will often identify an area of a PCB that is sensitive to vibration and may contain a dry joint or failed component. It is also worth removing the Vidicon tube from the deflection yoke and visually inspecting the electron gun end of the tube and the target contact that is mounted at the front of the deflection yoke assembly.
-
Last comment…. As you likely already know, your current lens is designed for a small sensor so you are seeing vignetting. You need a lens that is designed for a 1” Vidicon tube but you may get away with one designed for a 3/4” tube as well. I found the 25mm 1” Vidicon tube compatible lens to be the most easily obtained. Anything smaller, like 12mm or 8mm is often designed for a 3/4” or smaller tube.
-
For information… pictures of the internal electron gun arrangement in camera tubes…..
They are quite robust in terms of alignment but they are not designed to withstand ‘drop event’ impacts.
-
On another note, this camera DOES appear to require a 50ohm impedance on the TV (or video-capture card) to work correctly. In order to try to further debug my Micronviewer (read some of the various problems I've encountered, even after getting a working power supply, on the first page of this thread), I decided to connect the camera up to my Picoscope. Even though I bought the cheapest Picoscope, which has a theoretical maximum sample rate of 1 million smp/sec, I found it has an actual maximum sample rate of 6.25 million smp/sec. This lowers the horizontal resolution I'm able to capture at (about 395 total pixels per line including all h-sync and blanking, with only about 325 image pixels per line). However, this is still plenty of pixels to get a good idea of what an image should look like. After using my Picoscope as a high speed digitizer to capture raw samples to a WAV file (using my own custom Picoscope recorder software that pushed the Picoscope to its maximum possible samplerate, which the official software doesn't do), I used Goldwave audio editor to center the signal values on the blanking level, and then used some of my own custom software to read the sync pulses to get all the lines to line up (end result looks like a film strip with many pictures), and then loaded that into Gimp's raw image viewer and cropped the result to be the size of a single field (I didn't do any interlacing to get a frame, just cropped it to a single field).
The result, while not as good as a dedicated video capture card, still should give a good idea of what you would see on a TV (and has for many other video sources I've used this technique with). But it didn't. I have captured many pics using a USB video capture dongle (that's how I got the pics from the first page), and what I saw from this Picoscope capture shows no similarity in the image content. The sync and blanking areas are correct, but the image content area is messed up real bad. I suspect this is because my Picoscope is a high-impedence input (I think it's 1MOhm) while I know that video signal sources expect to be connected to a low input impedance device (usually about 50 Ohms), and some (like this camera it seems) will not even generate a usable image UNLESS they are connected to a low input impedance device.
I've attached the picture captured using my above described method to this post.
Do you know if there's a 50 Ohm BNC or RCA adapter I could get, that is designed to go between an oscilloscope and a video signal source? It would be like a normal coaxial adaptor or extender, except instead of having a completely isolated pin connector and outer shield connector, it would have a 50 Ohm resistor between the center pin and the outer shielding connector, to give it the 50 Ohm DC resistance that video signal sources are usually plugged into (allowing all video signal sources to work with high-impedance input oscilloscopes).
Regarding the above quoted post, and its attached picture, I have now gotten a BNC T adapter and a 75ohm BNC terminator. When using the same technique to get an image from the raw signal from the Picoscope I still get some kind of waveform added to the image (in the actual picture portion of the video signal), but less strong now, and also at a higher frequency than before. I've attached an image of the captured frame (or more accurately the first field of the frame). As with before, the width of the image is limited to only about 300 pixels wide, due to the lower than ideal sample rate of 6.25 MSPS. This also could be adding in some unfortunate lower frequency noise if there are strong, instead of weak, signals near the upper part of the video signal's bandwidth, because it would be at a frequency greater than the nyquist limit of the scope, and thus the frequency would appear to be lower than it actually is. The only real way around this issue, is to get a more expensive Picoscope with a wider bandwidth. The specs on my Picoscope (which is the cheapest one) is that it officially has a maximum samplerate of 1 MSPS, but I contacted someone there and they told me it could be pushed up to 6.25MSMP but would likely drop some buffers because of the scope's small internal memory (a problem I haven't had on my computer, which appears to be because my computer is fast enough to get the data before the scope's hardware buffer overflows). However, even at 6.25 MSPS, the nyquist limit is less than the highest possible frequency from a video signal following NTSC specifications.
By the way, I noticed in the spectrogram for this signal that there's a strong wideband signal appears in the middle of the spectrum, so it is at about 1.5 times the nyquist frequency (3 quarters of the sample rate) of my Picoscope.
-
Then I remembered something important. My Picoscope has a 6.25MSPS limit for streaming mode (capturing a long signal up to several seconds). But it has a higher limit for triggered block mode (only sending a buffer of data when there's a trigger, instead of as fast as possible). This mode will miss things between buffers during the trigger reset period, but it can run much faster. And Picoscope official software itself can use this mode quite well and display the graph. This mode runs much faster easily capturing signal at 500ns per division (with 10s of samples per division, so the sample rate is very high). In this mode I can see that the video signal's image data is NOT a continuously varrying voltage level, but instead it seems a series of PULSES. are these being sent at the rate of the camera's pixel clock? Not sure, but the video image signal is definitely chopped up. Here's a screenshot of the signal in the official Picoscope software.
Any idea what might be causing this? Is this a defect in the camera?
-
Given the pieces of textual information and attached images that I've provided in the last 2 posts, is there anybody here with enough technical knowledge of vidicon tube camera techonlogy, that they would be able to diagnose the latest problem issue I've noticed with my camera? It is an issue with the video signal output from the camera that only becomes obvious when actually measuring the signal with an oscilloscope (even with a 75ohm resistor in parallel with the scope's input). My USB video capture dongle has NO PROBLEM with this. What I'm noticing is that in the image portion of the signal (not blanking or syncing portions), there appears to be a series of pulses at a frequency that I'm guessing must be near the pixel clock frequency. These pulses also I've observed have TWICE the voltage that they should for a given optical input to the camera. For example, in a region that's pure white, which should have a signal voltage of 0.7 volts, I'm observing on my oscilloscope a voltage of about 1.4 volts! And, yes the scope does have its 75ohm resistor in place on a T adapter (white voltage pulses would have been about 2.8 volts without that adapter).
I think that my camera must be expecting some kind of capacitance across the input of the video input device it's connected to, and that capacitance would average out the pulses, creating a constant signal with half the amplitude of the pulses. Do most video input devices have a smoothing capacitor across their input like that? I knew about the 75 ohm parallel resistance, but was unaware of any parallel capacitance that would be required for monitoring the output of a video camera with an oscilloscope.
Can somebody here tell me what is going on? Is this just a defective camera? Is there actually supposed to be a signal smoothing capacitor across the output of the camera, inside the camera, but the capacitor is broken?
-
Come on people. SOMEBODY HERE must know what's going on with the signal here. I even posted a screenshot showing its strange waveform in my Picoscope software. Yet nobody seems to be able, or willing, to help here. And that's despite the fact that I was SURE that with all the people here with extensive technical knowledge, that SOMEBODY would be able to help me. Yet NOBODY seems to be able to. Or maybe its just my long paragraphs are making you not want to read them? Well guess what. Those long paragraphs are long, because they contain IMPORTANT INFORMATION regarding my latest issue. Lots and lots of important information, that you will need to read to fully understand my latest issue. And you will need to understand my exact issue if you are going to help me. So please don't just run away because you see a long paragraph. I NEED YOU GUYS' HELP to fix this latest issue.
-
Before anyone can help you they need reliable, trustworthy, data to analyse. We currently do not have enough accurate data to comment.
Some food for thought…..
1. Does the camera display a normal image on a monitor without any evidence of herring bone pattern or noise bars ? If so, why do you think you are not seeing the unusual signal that you think you have from your oscilloscope readings ? Are those oscilloscope readings trustworthy ?
2. Attach your oscilloscope to the video output of another camera, video recorder or other trustworthy source of a known good video signal. Capture the signal and analyse it by comparing to the many examples of a baseband video signal that are available on the internet (use Google to find them) Only when we are certain that your Oscilloscope system is giving you readings that may be trusted will we be able to comment on them.
3. People have busy lives and we cannot spend time analysing other people’s electronic issues if we are not confident that the supplied data is accurate. We need to see a complete picture. In this case I would want to see a complete oscilloscope image of a line scan and field scan waveform in order to assess the signal. I would also expect the oscilloscope used for such to have adequate bandwidth and sampling rate for a 6MHz complex signal..
I would not spend time analysing a partial signal of unknown validity as that could just waste my time and does not help you if incorrect assumptions are made by me about what I am seeing.
I strongly suggest that you spend some time on eBay and find a decent 20MHz Analogue oscilloscope as these are quite common and inexpensive. An analogue oscilloscope rarely lies to you and is a great general purpose test instrument. A DSO introduces the challenge of both true bandwidth and sampling rate. If the Sampling rate is inadequate you basically get an inaccurate display of a signal. You need the decent sampling rate for single shot complex signals and not the over sampling spec for signals that do not change. For video work I would want a DSO with at least 60 Ms/s true, single shot, real time sampling rate.
Please analyse your cameras image and if there is no evidence of unusual ‘noise’ then suspect your oscilloscope captures as potentially invalid.
The change in video signal amplitude is interesting however. A video signal needs to stay within the video standard if a monitor is to correctly display it. An incorrect video signal level would certainly be worth further investigation.
-
Before anyone can help you they need reliable, trustworthy, data to analyse. We currently do not have enough accurate data to comment.
Some food for thought…..
1. Does the camera display a normal image on a monitor without any evidence of herring bone pattern or noise bars ? If so, why do you think you are not seeing the unusual signal that you think you have from your oscilloscope readings ? Are those oscilloscope readings trustworthy ?
Mostly normal image on a TV or video capture card, but I have seen some vertical bar noise on the camera at times (usually just to the right of a very bright area in the image), and very prominently over the entire image with the case off (which is necessary to make potentiometer adjustments). Usually it's not a problem for normal viewing (so long as the case remains on), but that still doesn't mean that the signal is in spec for an NTSC video signal source. It just means that so far the display I've tested it on and the dongle I've tested it on have not had a problem with the out-of-spec signal, because it was "close enough" that the dongle could compensate. By the way, there's one semi-professional dongle I tried to use, but it never detected the signal. It definitely is built to exact NTSC specs, and expects any sources connected to it to also use proper NTSC specs. It rejects any signals out of spec. And that was a $200 USB video capture device from the professional video equipment company Black Magic Design. The funny thing is the dongle that did work actually was a cheap dongle I bought for like $5 at a yard sale, and comes from some company called SIIG that I'd never heard of before. I just Googled that cheap dongle, by the way, and found the dongle in question costs $50 from SIIG's own website, but that's still only one quarter the cost of the more expensive one.
2. Attach your oscilloscope to the video output of another camera, video recorder or other trustworthy source of a known good video signal. Capture the signal and analyse it by comparing to the many examples of a baseband video signal that are available on the internet (use Google to find them) Only when we are certain that your Oscilloscope system is giving you readings that may be trusted will we be able to comment on them.
I have never before had an issue with another NTSC video signal source on my PicoScope. The high frequencies are always weaker than the low frequencies (due to the camera having a proper lowpass filter that blocks frequencies above about 4.5MHz and attenuates signals smoothly up to that point (the higher the frequency the lower the amplitude, evidence of a simple RC filter, as every video source is supposed to have on its output to band-limit the signal), so even if the signal bandwidth exceeds the nyquist frequency this would only add a small amount of low amplitude medium frequency noise to the rest of the signal. This Micronviewer is either designed well out of spec for the NTSC signal bandwidth, or has drifted out of spec overtime. I hope to put it back in spec.
3. People have busy lives and we cannot spend time analysing other people’s electronic issues if we are not confident that the supplied data is accurate. We need to see a complete picture. In this case I would want to see a complete oscilloscope image of a line scan and field scan waveform in order to assess the signal. I would also expect the oscilloscope used for such to have adequate bandwidth and sampling rate for a 6MHz complex signal..
Good point. I'll probably end up buying a better Picoscope at some point, but the one that would best capture the full bandwidth of an NTSC signal costs about $1200. I'm not quite ready to spend that much money yet, but probably will sometime later. My current Picoscope is the cheapest one, and cost about $120.
I would not spend time analysing a partial signal of unknown validity as that could just waste my time and does not help you if incorrect assumptions are made by me about what I am seeing.
Fair enough.
I strongly suggest that you spend some time on eBay and find a decent 20MHz Analogue oscilloscope as these are quite common and inexpensive. An analogue oscilloscope rarely lies to you and is a great general purpose test instrument. A DSO introduces the challenge of both true bandwidth and sampling rate. If the Sampling rate is inadequate you basically get an inaccurate display of a signal. You need the decent sampling rate for single shot complex signals and not the over sampling spec for signals that do not change. For video work I would want a DSO with at least 60 Ms/s true, single shot, real time sampling rate.
Unfortunately I have a space issue. I don't have enough space for any more large devices in my room, like an analog CRT oscilloscope. And I can't exactly put my equipment outside of my room. My room is my bedroom, and I am living in an apartment with my parents, so the only place for me to store my personal equipment is in my own bedroom.
-
I have never before had an issue with another NTSC video signal source on my PicoScope. The high frequencies are always weaker than the low frequencies (due to the camera having a proper lowpass filter that blocks frequencies above about 4.5MHz and attenuates signals smoothly up to that point (the higher the frequency the lower the amplitude, evidence of a simple RC filter, as every video source is supposed to have on its output to band-limit the signal), so even if the signal bandwidth exceeds the nyquist frequency this would only add a small amount of low amplitude medium frequency noise to the rest of the signal. This Micronviewer is either designed well out of spec for the NTSC signal bandwidth, or has drifted out of spec overtime. I hope to put it back in spec.
The image looks like a video instability, maybe caused by having the laptop and picoscope connected. The tube signal is only a few nA into meg-ohms.
As for not being detected by a 'strict' video dongle, I would be more suspicious of the sync timing than the vertical amplitude. The vidicon era thermals were all mono (RS-170 not NTSC) and might use a 'close enough' sync generator at relatively low clock rate.
What is the sync generator and crystal ?
CD22402 known not to be great.
-
The image looks like a video instability, maybe caused by having the laptop and picoscope connected. The tube signal is only a few nA into meg-ohms.
My Picoscope has a BNC T-adapter with 75ohm terminator resistor connected to it now, so the camera no longer sees the Picoscope as a high impedance load, and while that has improved it, it hasn't fixed it. Do most vidicon era cameras expect to have a smoothing capacitor in parallel with the load, either at the camera end or load end of the coaxial cable?
As for not being detected by a 'strict' video dongle, I would be more suspicious of the sync timing than the vertical amplitude. The vidicon era thermals were all mono (RS-170 not NTSC) and might use a 'close enough' sync generator at relatively low clock rate.
What is the sync generator and crystal ?
CD22402 known not to be great.
I have no idea what the sync signal source is for an Electrophysics Micronviewer. There's a couple ways the level could be an issue. If the level is too high, the capture dongle might intentionally reject the signal due to it thinking it's not a valid video signal. Also since using pulses of too high of a level (though usually outside of the picture area, unlike from my camera where they are inside the picture area) is a technique used by Macrovision for copyprotection, and since modern consumer video capture equipment is required by copyright law to respect copyprotection signals, it could be mistakenly seeing the out-of-spec image signal levels as being copyprotection signals, and thus refusing to capture the signal.
As for the difference between RS-170 and NTSC, the fact is that NTSC was actually developed to be compatible with RS-170 (using 59.94Hz field rate instead of 60Hz field rate) so that NTSC color video would still be viewable on older monochrome TV sets. So even a strict NTSC video capture device should not be so strict as to block RS-170 monochrome signals. If it's that strict, that seems like a defect in the capture device, or possibly that the camera itself is outputting a signal that as well out of spec for either NTSC or RS-170.
-
Do most vidicon era cameras expect to have a smoothing capacitor in parallel with the load, either at the camera end or load end of the coaxial cable?
No. A capacitor that could smooth out that noise would also kill the syncs.
I have no idea what the sync signal source is for an Electrophysics Micronviewer. There's a couple ways the level could be an issue. If the level is too high, the capture dongle might intentionally reject the signal due to it thinking it's not a valid video signal. Also since using pulses of too high of a level (though usually outside of the picture area, unlike from my camera where they are inside the picture area) is a technique used by Macrovision for copyprotection, and since modern consumer video capture equipment is required by copyright law to respect copyprotection signals, it could be mistakenly seeing the out-of-spec image signal levels as being copyprotection signals, and thus refusing to capture the signal.
As I said, level is unlikely to be an issue. NTSC chroma can hit 1.4V after all. Junk where Macrovison or similar are coded might be possible:
https://forum.videohelp.com/threads/170667-What-Macrovision-looks-like
As for the difference between RS-170 and NTSC, the fact is that NTSC was actually developed to be compatible with RS-170 (using 59.94Hz field rate instead of 60Hz field rate) so that NTSC color video would still be viewable on older monochrome TV sets. So even a strict NTSC video capture device should not be so strict as to block RS-170 monochrome signals. If it's that strict, that seems like a defect in the capture device, or possibly that the camera itself is outputting a signal that as well out of spec for either NTSC or RS-170.
The CD22402 is WAY out of spec for RS-170 or NTSC. 4.0us syncs for a start. However RS-170 kit was never that fussy (and maybe also had a CD22402 inside).
-
The other big issue is the image is chopped into pulses. This was noticible on my Picoscope 2000 series scope, but just a week ago I got my Picoscope 5000 series scope, and it has a much higher maximum sample rate. Attached to this post is a screenshot PNG from the Picoscope software. The h-blanking and sync to the left of the active image line looks fine, but look at the image signal. It's chopped into pulses. The peek of each pulse is where the correct brightness is, but between these pulses it goes back to zero level.
I captured this with my Picoscope running at 250 million smp/sec. This sample rate gives me a bandwidth that is high enough to contain all the harmonics from the image chopped into pulses artifact I was describing. Bitdepth is set to 12bits per sample (needed to be this low to get the sample rate as high as I did, but still 16 times the vertical resolution of the Picoscope 2000 at 8bits). Voltage range is set to +/- 2 volts, and capture duration is 100ms. In the attached screenshot, I've zoomed in so you can see the detail I was talking about.
I also have a copy of the PSDATA file which has the full waveform, but it's 30MB in size, and that's larger than I can attach here. If you would like a copy of the PSDATA file (a proprietary Picoscope format which contains the full waveform, I believe using GZIP compression, as well as metadata like sample rate) so you can further examine the waveform, in an effort to help me debug my Micronviewer, please let me know where would like me to upload it so you can download it easily.