What's New, LabGuy?
Got my first image dissector tube today. That's a special kind of TV camera pick-up tube. Type Li-608-1. It is of Russian origin and comes with a data sheet. In Russian , of course. This tube is a modern "in line" type that uses a conventional vidicon deflection coil assembly. At least that is what I hope and where I will start.
BREAKING NEWS! Before I completed writing this evening's update, I managed to purchase an IIT F4012 one inch diameter Vidissector tube from another collector. This one positively uses a standard vidicon deflection coil. That's a relief! So, after owning none, I now end up owning two image dissectors on the same day. What a world!
Dissector tubes have an unfair rap for being insensitive to light. This is not necessarily true. Since a dissector does not incorporate the property of charge storage, the output signal consists of all the energy falling on a single point of the target at that instant only. However, if we park the scanning and let one pixel "stare" for a full 1/30 of a second, it will accumulate the same energy value as any other tube. But, only for that one pixel. Weirdly, image dissector tubes can actually do this!
When a dissector is scanning at 525 lines, 60 fields per second, The "image aperture" is only dwelling on a given pixel for about 240 nanoseconds, that is not a lot of time to count photons. So, the signal is very low, even in very bright light. To make up for this, the small number of electrons released by the light are directed to a metal plate that is about 100 volts more positive than the aperture. The electrons fly toward this plate and impact it. These secondary electrons knock out more electrons on impact with the positively charged plate. These, in turn, are attracted toward another plate that is 100 volts more positive yet. This goes on in a cascade of up to eight or nine stages. This is the classic electron mulitplier, invented by Philo Farnsworth as the direct result of inventing the dissector and trying to get more gain from the tube.
The dissector is actually a real time addressable photo multiplier tube. Think of the raster scan as a grid of pixels. You can drive the deflection coil from computer DACs and place the point of interest anywhere in the scene at random. At any rate you desire. Slow for high sensitivity and fast for viewfinding. Star trackers on some satellites use dissector tubes to do just that. Unlike a CCD, a vaccum tube is not effected by radiation in space. Nor does a dissector tube have a heater like a conventional vacuum tube. This makes it highly reliable! These bizzare properties give me all kinds of wild ideas. Stay tuned for a potential future project around this tube.
The Tiny Ike iconoscope camera was resurrected today after a long spell of no activity. Aprropriately for the day, I called part six; [Tiny Ike, The Ressurection].
Added some nice one inch triniscope color screen shots tonight. Eight images showing good registration and awesome color. Visit the Tiny Triniscope Project page to check them out.
I'm sorry to tell you this. But, the Tiny Triniscope project is completed. It works so well, I can barely stand it! Great color and geometric registration beyond my expectations! You have got to see it. Right here: The completed Tiny Triniscope project page.
I made a short video and uploaded it to my Videolabguy channel on Youtube tonight. The first thirty seconds have no sound. This is intentional. It's quick and crude. You get to see Tiny Triniscope working in real time. Enjoy! [VIDEO: The Tiny Triniscope Color TV].
Tiny Triniscope: While holding the tubes in position by hand, the convergence of the images is virtually perfect. I can not believe the uniformity of the rasters between these three tubes! My first attempt to glue the green tube into a beginning reference position was a fail. The CRT ended up off center, biased hard to the left side and had sagged downward about 5mm. Needless to say, the red and blue tubes would have preferred that it was biased to the right, because their horizontal deflection is reversed. The center lines between the green and red/blue are now about 2mm off. If the images are overlapped, the outer band of the rasters are not. The centering magnets on the yoke were obviously welded in place. Have yet to break them free (without breaking them) so I can begin to have some positioning control. The microscopic dimensions of everything, and my old failing eyes, are not making the job any easier.
See more at the Tiny Triniscope project page here. Scroll to the bottom of the page to see color bars converged pretty well and a not so good shot of Dan Rowan and Dick Martin. The large area colors are incredible. This should give you an idea of the potential. All I need to do, is to figure out how to precisely align the three micro-CRTs!
It will be awesome! They said. Build it and they will watch. They said. It can never been done. They said. You can't possibly do it. They said. Oh, he'll do it. They said. Just who the hell are "they" anyway and why are they so negative?
The work on the, APR-1S-F001S, one pixel multi-standard-format video/audio/olfactory monitor is almost done. With a horizontal resolution of 1 and a vertical resolution of 1, the color version will sport 16,777,216 colors. All black. Light. One point five channel infrasound with zero to almost one Hz bandwidth. Olfacto-vision, a 3D smell process. Patent pending. No longer will you wonder if the movie you are watching actually is a stinker. Just like me.
Tiny Triniscope Update: Completed the color matrix circuit and it works very well. Tested the color capability of this video monitor and it is excellent. Will be moving into the convergence or image registration phase next. Will start by optically aligning all three CRTs on the prism. This is necessary before I start any electronic alignment of width, height or linearity. Hoping to make good progress over the coming weekend. I would be overjoyed to have good registration by Monday morning. But, since it is mechanical work, it may take longer. Time will tell.
Tiny Tiniscope project update. Have been working on the color matrix circuit board. Wired the first three input op amps and checked them on the o'scope. Discovered that the DVD player video was DC coupled. But, all of the black reference levels were not at zero volts. So, I add input AC coupling to the video amps. Then installed a DG211 quad cmos switch and LM1881N sync separator chip. The 1881 conveniently provides a clamp pulse called BG (burst gate). When BG is active low, during the horizontal sync back porch, the switch closes, shorting one end of the input AC coupling capacitor to ground. It is released for all of the rest of active video. This is a brute force method of setting that particular part of the video signal at precisely zero volts. The capacitor is a tiny 10uF ceramic type that will hold the clamp voltage for far longer than the occurance of the next clamp pulse. It worked perfectly first time.
Once the DC levels of the Y, U & V signals were properly set at zero volts DC reference, I then wired the red and blue recovery networks. This takes the Y signal, and the U (B-Y), amplified by 2.032 times and adds them together. The result is the original blue signal.
Did the exact same thing for the red side of the circuit. This time boosting the V (R-Y) signal by 1.14 times. This also came up working. A slight adjustment and it was also spot on, producing a perfectly serviceable red signal with sync. This is the signal that now drives the little CRT monitor of the approrpriate color.
Tomorrow evening, I will finish the matrix circuits, getting the green output working and adjusted. Then we can move on to the mounting and fine mechanical adjustment of the small CRTs.
Progress to report on the newest electronics project, the Tiny Triniscope. It has reached the point where I was able to power and feed video to all three viewfinder monitors. See more at the Tiny Triniscope project page here.
Goldmark 1 update. I have obtained a new ball bearing motor and built the third color wheel. This time, I balanced the wheel using a knife edge technique suggested by Cliff Benham. I placed the wheel on a temporary axle and placed it on narrow paralled rails. It naturally rolled to the point where the heaviest side is straight down. Add weight to the opposite side until the wheel does not roll no matter what position it is placed in. It was actually somewhat easy, if time consuming. The wheel could rock for up to 15 minutes before coming to a rest.
Once the wheel was balanced, I removed the temp axle and mounted it to the new ball bearing motor. It starts, runs and stops as smooth as silk now! Sweet! I would like to publicly thank Cliff Benham on his guidance with this technique. The results are excellent.
Added a new project page today. I got bored yesterday and decided to convert my NBTV Baird 32 line Televisor kit to color operation. What do you know? It worked!
I have completely rewritten the Goldmark 1 project pages. There is a lot more work to go. But, the pages now reflect the completed Goldmark 1 color TV and all the superfluous mucking about has been stripped out. I will add in more details as time permits. For now, enjoy the rebuild.
I have updated the Goldmark 1 project front page. Showing the progress to date. The color servo gets better with each new iteration. The wheel balance issue is going to require a slight rebuild of the wheel and motor mount. I am gong to place the wheel on an axle suspended in ball bearings like I planned to originally. The motor will be on a separate mount. This also allows me to lower the wheel slightl as the new monitor is not as tall as the original Akai monitor. It should be a lot easier to balance the wheel in the new configureation. I plan to attache it to the motor with a flexible coupler to eliminate coupling any vibration back to the motor. Stay tuned!
I took some time this afternoon to dismantle the Goldmark 1 and paint all of the support structure flat black. It came out very well. Also completed bolting down the video monitor so that the vibration doesn't make it walk away. Mounted the power supplies on end to not consume a lot of space. The last thing is to dirll the holes and attach the servo circuit board to the base board. Will need to hit the hardware store for some inch and quarter length number four machine screws. As you can see in the third photo, the Goldmark 1 makes a darn good picture after I tweaked up everything.
There was also some progress toward wheel balance. I have traced the wobble to the color wheel mounting puck. When the set screw are tightend, the mounting flange is not centered on the motor shaft. I put a single layer of scotch tape on the opposite side of the motor shaft, from the set screw faces, and reinstalled the wheel. Run out is now almost zero and the vibration is noticably less. I think it is down to truly balancing the color wheel now. Stay tuned for that.
Earlier that same day....
In the OLDVTRs discussion forum, the subject of color fringing on moving objects came up. The scan converter I use suppresses the effect very well, making it dificult to demonstrate. But, while recording video, the fast movement of the camera produced the effect. You can clearly see the separation of the various scans. But, normally this is not viewable.
Goldmark 1 update. The video monitor achieved first light this week and there are now two YouTube videos showing it in operation.
There are few issues left to resolve. Balancing the color wheel is now number one priority and I will devote some time to this, this coming weekend. Then the monitor must be addressed. The CRT is out of focus and the scan raster is being modulated by the vibration of the near by very large DC motor. This blurs the image even more. At 80% the resolution of NTSC, we have no margins to lose here.
Goldmark 1 works! We made our first color pictures today at 4:45pm PST. There is still a lot to fix. But the basic operation was finally demonstrated. First, the new video monitor is scanning well enough to produce a fine undersized picture. All of my new circuits worked virtually first time. Initially had the video signal inverted. That took only five minutes to fix. Adjusted the pulse width of horizontal drive and got the monitor to operate at its normal voltages. The scans were short in both axis as I expected. Horizontal needs an inductor changed to widen it. Vertical requires a capacitor change. I don't anticipate any problems there. The picture is currently scanning about two inches diagonal and nothing is pulling abnormal current or running hotter than normal.
Then I put it all together for the first time. The video monitor, servo board and color wheel assembly. Turned it all on and proceeded toward success from there. It is making pictures. But, that does not mean it is finished. The servo is very slow to reach speed and it is sloppy when it is locked. It drifts in and out of phase constantly. I will balance the speed vs phase servo summing ratios to resolve this.
The darned thing was hard to photograph clearly because the wheel still needs to be balanced. The whole rig was trying to walk off of the table! Still, this is an excellent milestone. I will update the project article accordingly in a few days.
I fixed the horizontal pulse timing and now have the entire picture on the screen. It was folded over previously. Now I can set the drive pulse width to where the flyback and yoke are working perfectly and also have full image centering control now.
Goldmark 1 project update. Completed the mounting of video monitor number two. It was made by Sony for Tektronix in the mid 1980s. The Akai CRT is on the side lines for now. We will be testing this 3.5" CRT monitor pulled out of an obsolete logic analyzer. It requires plus and minus five volts. So, I added two new voltage regulators and heat sink to the servo board just for this purpose. We will need to add an inverting video amplifier with some unknown amount of gain. I have three spare video op amps on board for just such an emergency. Next step is to measure and make two interconnect cables for between the video monitor and the main servo board. Then add the approriate connectors and new sync and video circuits to the servo board.
I feel some trepidation about getting this new CRT monitor to work at 29.16KHz. It was running at near 15.75KHz in the analyzer. I have the full service manual and monitor schematics. So that helps. But, who knows? It just might work.
Goldmark 1 project update. I have obtained another CRT video monitor to try in this TV. It is an instrument monitor rmoved from a logic analyzer. I fear it will also be a dud in the end because it turns out to be a Sony tummy TV. But the ciruits are stripped to the bare minimum. This might mean that bending the horizontal scan rate won't be as difficult as with the Akai monitor. Fingers crossed there.
I started the day by adding the new plus and minus five volt regulators to the color wheel servo board. The new monitor use this for its primary power supply. Simple enough. Mounted a large solid heat sink for the regulators and had it all wired before lunch time. Have been slowly assembling the monitor support frame. I have created the base platform to fit where the Akai monitor previsouly did. Mounted the CRT bezel to some C channel extrusion and bonded that to the front of the base. Tomorrow, a trip to the harware store for three inch spacers to hold up the circuit board and to completely mount the CRT. Then it will be safe to power up. Please stand by.[Sneak peek 1] [Sneak peek 2]
Happy New Year!
Added an update status report to the bottom of [Goldmark 1, part 7] page. It details the latest progress in designing a horizontal magnetic scan circuit for the three inch field sequential color TV project.
To see years 2007 to 2013 What's New pages, [CLICK HERE]. Scroll to the bottom of each year to find a link to the previous year.
Created: April 6, 2012, Last updated: April 02, 2014