Text and graphics on CRT

[c4757p]

I've done a fair job of getting this going as it is, but I was just hoping for someone with experience with this to round off a few rough corners for me. Usually I'd figure it all out by trial and error, but that seems a bit dangerous here.

I have a small (5") monochrome CRT (electromagnetic deflection, not from an oscilloscope). I'd like to use it in a project for displaying text and low-resolution graphics. Here is what I figured out/got working:

• Filament takes about 70-80 mA at 12V (this was from a 12VDC portable television and I am fairly certain that the filament runs right off the 12V input).
• Running the flyback transformer (I took that too) with a 50%, 31kHz square wave gets a decently bright spot on the screen; current draw is up to 300mA. The MOSFET and transformer don't run very hot (I've had it going for half an hour now with no heatsink and all is good).
• One blue wire coming off the back appears to affect the brightness. If I touch the end lightly with my finger the brightness varies unpredictably. If I pull it to ground through a resistor the brightness is fixed. Sometimes if I leave it floating, the beam goes away completely. This wire will accumulate a large voltage on it if I leave it floating for long - it bit me! - but appears to be at a high impedance because I can short it directly to ground with no issues. I'm guessing it's something equivalent to a grid in a triode, since it controls the current flow and isn't directly connected to anything?
• 17mA applied to the deflection coils sends the beam off about a centimeter from the middle.
• The beam isn't very well focused. I'd guess the remaining wire should control that, but I can't get it to do anything.

So, my questions are:

• How is this "grid" wire (not sure what it's really called) supposed to be controlled? How much voltage will it accumulate if left floating, and what is the best way to protect the control circuit from that?
• How the hell do you get the focus control to do anything? I've seen oscilloscopes that use a pretty high voltage for it, but this little TV had no other source of high voltage that I am aware of to feed it. Sadly I no longer have the circuit board to inspect, but I'm pretty sure the only other inductors or transformers on it were little RF types. The flyback doesn't seem to have any other windings, though it has a bunch more apparently unconnected pins.
• Is the current draw considered acceptable, or should I optimize the flyback driver? I pretty much just threw a waveform at it and got it to do something.
• What sort of circuit is usually used to drive the deflection coils? Seems to me that with that much current to get just a small deflection, it would have to dissipate quite a bit of heat. There's the additional problem that they have to be driven positive and negative, but that's something I can figure out.
• I have no high voltage probe. I'm looking to buy one, but until then - if I measure the inductances and leakage of the transformer in an inductance bridge and simulate the circuit in LTSpice, how accurate will its prediction of the output voltage be? Just remembered it has a built-in diode and I can't get a meaningful measurement...
• The beam is decently bright, but how much brightness can I expect to lose with it sweeping over the entire screen?

[TerraHertz]

There are many sources of circuits for old CRT TVs. Why not obtain some even if they are not for your specific unit, and study them?

Also, if you live in any kind of western urban setting, free CRT TVs are easily found. People are throwing out all kinds even large, relatively 'modern' perfectly working ones, as they buy nice LCD TVs. So you could probably find more to play with if you wanted.

As for the beam intensity control - they don't use the grid to control beam current. The grids are typically set at fixed voltages that interact with the focusing rings to give correct beam geometry. And they're not really 'grids' - they're discs with a little hole in the middle. Get some old CRTs, crack the little pump-out nipple under the rear plug molding to de-vacuum the tube safely, then smash the neck open to examine the electron gun. They're pretty!
Beware - the screen phosphors, coating on the cathode, and the substance in the 'getter' ring are all really poisonous. Don't inhale the dust, don't get in cuts, wash hands after, etc.

Beam current is controlled by varying the voltage on the cathode. The more negative, the stronger the beam.
If you look at the small PCB typically mounted on the tube rear socket, it will have either one HV transistor (monochrome tubes) or three (RGB.) These drive the cathode(s). I vaguely recall the transistors are typically driven in common-base arrangement, since that gives the best voltage-gain/bandwidth.

If your CRT is from a TV, you'll find it very hard to drive the deflection yoke with high enough frequency of current variation to generate much text on the screen. TV yokes are optimized for operation at the H and V scan frequencies, and in the case of the H deflection coil, typically their inductance forms a part of the overall flyback oscillator system. But maybe, on that small a tube the yoke will be fairly small and low inductance too.

Hmm... Somewhere I have the schematics (and a working display) from an old Atari Asteroids arcade game. That basically does what you're trying to do, but with a larger B&W TV screen with deflection yoke. Want me to find it and scan it? The deflection drive circuits are non-trivial, I warn you. And I think they used a custom yoke.

Oh and one more thing. A TV CRT electron beam is quite capable of drilling a hole in the back of the screen glass, if you leave it in one spot while bright. I killed a screen that way once. The Atari system had a spot-killer circuit, that turned the beam off if there was no deflection happening. This is _why_ CRT TVs used the yoke as part of the EHT oscillator circuit. No yoke, no EHT, no screen burn.

[AlfBaz]

I know precious little about TV's or CRT but when I was younger I got my hands on an "luggable" PC with a small CRT. It worked but it was some kind of custom unit whos intended function I knew nothing about. I started playing around with its innards and could clearly see the two connectors for the X and Y deflection coils. I disconnected them and could see the dot in the middle of the screen. I then applied a voltage from a 9V battery to the x deflection coil via a potentiometer and behold, the dot moved...

So yeah a small CRT might be easier to drive

[c4757p]

Hmm... Somewhere I have the schematics (and a working display) from an old Atari Asteroids arcade game. That basically does what you're trying to do, but with a larger B&W TV screen with deflection yoke. Want me to find it and scan it? The deflection drive circuits are non-trivial, I warn you. And I think they used a custom yoke.

That would be very nice, if you don't mind. I figured the deflection circuits would be non-trivial, considering the fairly large inductive load, but I would like to see how it's done. It has actually been a bit hard to find a decent schematic to look at for a monochrome, electromagnetic CRT. They all seem to be color magnetic or mono electrostatic, and I don't really want to have to dig through the complexity of a color CRT driver to do my very first monochrome one.

A TV CRT electron beam is quite capable of drilling a hole in the back of the screen glass, if you leave it in one spot while bright.

OK, so I was totally wrong about the beam being "decently bright" then. There's no way it was anywhere near that strong. At some point I did something and got it really freaking bright, but I haven't been able to reproduce that  Didn't know it was so capable of damaging the screen though - perhaps I'll hold off on trying to reproduce that until I have at least one deflection coil running.


Another question that I think I hinted at but didn't ask directly: how are the transformers usually driven? (Or will that be in your schematic too?) I've looked around for driver circuits for CRT flybacks, but most of the ones on the Internet seem to be geared towards "Big voltage and short circuit current for impressive arcs, who cares if you burn out a few transistors in the process?" rather than a reliable EHT source for long-term CRT drive... Most of the decent circuits have a feedback winding, but this originally did not! There are no more windings on the transformer (I checked every pair of pins for continuity - most seem to be internally unconnected and just there for physical stability), and while there is an exposed ferrite arm to wind one on, there wasn't anything on it. I suppose I could always add one.

[SeanB]

If I remember those Asteroids games used 90 degree deflection tubes, as the longer electron path made it easier to get the deflection linear without needing complex compensation. 110 degree guns have to have s correction on both directions with saturable inductors to get an acceptable linear scan. You cannot run a linear current through them and get consistent deflection. The long tube was not a problem in the vertical case it was used in.

I remember playing that when I was young, when there were still arcades............ lost a lot of money to Pac Man at the local tearoom.

[c4757p]

110 degree guns have to have s correction on both directions with saturable inductors to get an acceptable linear scan. You cannot run a linear current through them and get consistent deflection.

Forgive me if it's a stupid question - this sort of stuff is not an area where I have any experience at all - but instead of an actual analog oscillator circuit with compensation, could this be driven by a DAC-controlled circuit with the compensation in a lookup table, or would it be nearly impossible to match the coil current to the DAC value because of the inductance? (4.1mH, by the way, on the longer coil. 410µH on the shorter one.)

Edit: Probably not worth it just because that'd be one really fast DAC.

[SeanB]

The corrections are not linear, and depend on where the beam was before, and the direction. Big CRT units have up to 100 little magnets on the yoke to do adjustment for the errors. Some are rings, some are little plastic ferrites on plastic tabs pushed up inside the yoke and others are stuck on the CRT outside during final test. This just gets the linearity to a point where a line is mostly straight all the way as it moves from top to bottom, from side to side and most importantly stays in focus from the middle to the far corners. The line must stay the same width as well or characters look distorted as they scroll.

[c4757p]

This one has some magnets stuck to the yoke - would that be it?

[SeanB]

I have seen some with only a few, and some with literally a forest of platic tags. Biggest I saw was a 87 inch Panasonic CRT, where they were sharing the neck of the CRT with a massive deflection array with literally dozens of wires connecting it to the main board, along with more connecting an auxillary deflection controller board with a whole lot of power transistors and such all on it. that set came with handles to carry it, you needed 4 people minimum.

[c4757p]

I think I'll just wire it up tonight and test the deflection and see if I can get anything going. I'm not going to worry too much about the linearity of the deflection if it's not too bad.

Oh, and to address something I missed earlier:

If you look at the small PCB typically mounted on the tube rear socket, it will have either one HV transistor (monochrome tubes) or three (RGB.) These drive the cathode(s).

Nope, it's just wires. I think I'll also try properly driving the cathode tonight - might as well try as much as possible at once, since setting it up with all the HV stuff going all over the bench is a massive pain. Would an MPSA92 do a decent job for this? I think I'll also attach two neon lamps in series between the cathode and ground to avoid building up too much voltage on that transistor (300V max). (Is this something that's actually done? I remember that my old oscilloscope had a neon lamp on the PCB near the CRT driver that was clearly being used as a proper component, not for indication.)

Also, it was mentioned earlier that the "grid" is usually connected to a fixed voltage - does anyone know about what this voltage usually is? I don't want to damage anything by doing it wrong. Positive or negative - and does it need a series resistor if positive, like a triode would?

[c4757p]

Actually I think I did burn the screen! I just cleaned the dust off the front (there was a good bit) and found a small black spot dead center in the phosphor coating that blended in with the dust...  Oh well, it's not noticeable. Guess I'll be more careful with it and try not to make it bigger.

[SeanB]

Grid is normally  a little negative, around -100 to -200V below the cathodes. You might need higher current drivers for the cathode, or ones with a TO220 housing, as they are required to give a 100V plus swing at around 30mA to drive the cathodes, with also having to discharge the whole tube capacitance of around 2nF as well while doing so.

[c4757p]

Thank you! I will see if I can find something. In the meantime maybe I will parallel a few MPSA92 with current-sharing resistors and just swear a lot if I blow them up... 

Is it reasonable to get the -100 to -200 voltage by adding another winding to the transformer on the external ferrite arm to save the space of a separate DC-DC converter, or will this cause some sort of problems?

[c4757p]

One more question just to clarify - just in case I have the connections totally wrong. The one I have been referring to as the 'grid' goes to a metal cup that appears to surround the cathode (as a grid would). There is another metal cup in front of that one, connected to another wire. What is this one - what should I do with it?

[TerraHertz]

Darn it! I can't find my old Asteroids doco.  But I _did_ find the manual for another arcade game called Rip Off, that was another vector graphics thing (tank warfare) using an ordinary magnetic yoke B&W CRT.
Scanned the deflection schematic for that, and also a couple of small B&W TV schematics I found. (Man, those are old! Almost childhood relics.)

They're at http://everist.org/archives/scans/crt_xy_circuits/

Heh, no difficult archive format this time, just png images. Three of them, about 4 MB each.


I did tell you there are no 'grids' in TV CRTs, just cups and rings. All of which have to be right voltages to form a clean focused beam.
Does your tube have a model number on it? You can probably find the spec data for it online if you search. That would give you the recommended voltages for the pins.

[c4757p]

Yeah, I know there are no true "grids", but having a fair amount of vacuum tube experience the term feels familiar. Also I continued to use it because SeanB was using it and I don't know what they're really called. (What are they really called?)

Well, I feel like a bit of a fool: I was having some trouble finding the specs online - being from the "Tian Jin Jin-Jing Glass Bulb Co.", it wasn't exactly a standard type - but I finally did notice that the specs are within the first few results for the tube number (14SX3Y4), just under a link that didn't look the least bit relevant. I guess I should know better than to expect relevant-looking links for random cheap Chinese parts.

The heater voltage is confirmed to be 12. Anode voltage is 8 kV, E_g2 is listed as both 100V and 300V. (I'm guessing E_g2 here refers to the voltages for the "grids"?) Is that positive or negative? And E_kco is 35 ± 10 - I'm guessing that's the cathode cutoff voltage?

[c4757p]

Thanks for the schematics! They look very helpful. Not too confusing at all - I just wish they weren't crammed together the way they are. I just had to print them out and draw boxes around subcircuits with a colored pen.

The deflection circuits actually looks similar to what I was planning to try, so that's good, I was at least barking up the right tree.

[TerraHertz]

Huh. It's surprising to see people advertising this CRT on ebay.
http://www.ebay.com/itm/5-Inch-B-W-CRT-14SX3Y4-Yoke-Coils-QPH-20-70-N57C-mounting-holder-and-tube-/310530272611?_trksid=p2047675.m1850&_trkparms=aid%3D222002%26algo%3DSIC.FIT%26ao%3D1%26asc%3D13534%26meid%3D4096836877334669473%26pid%3D100011%26prg%3D1080%26rk%3D1%26sd%3D251162969551%26

Sorry, your guess about the meanings of those few parameters is as good as mine.
But it might help to think about what the electrons are doing as they head down the gun, to get an idea of the relative voltages needed.
First, it's all relative. The successive focusing rings are making E field lenses, to adjust the beam divergence.
Starting at the cathode tip, electrons leave it when the cathode is negative enough relative to the first cup - the one with the small hole. Electrons that happen to go through the hole are a fairly wide and diverging beam, and need to be brought back to a fine, collimated beam. Now they're in the E field between first cup, and the next ring. So what V does that have to be relative to the cup? It needs to bend the electron paths inwards... so it must be negative relative to the cup.

So now the electrons are converging, and once they get down to a fine beam they need to be bent 'outward' again a bit so they are all traveling in parallel. So, next focusing ring is more positive than the previous ring. After which they'll start to see the HV gradient towards the anode metalization, and get going faster.

One of those circuits I sent has voltages for each CRT pin. Will give  a general idea.

[amyk]

http://shiximm.com/baogaofanwen/261729.html
Chinese article but has a PCB layout (!) and schematic for a TV that uses this tube.

[c4757p]

Nice find on that schematic, amyk! Thank you.

I do have to whine a bit though! WHY do people draw schematics that look like that?? It's not like they pay per square inch of schematic sheet - how about some organization??

[Mr Smiley]

You MUST include a X-Ray killer circuit in your design if your going to roll you own.

 

[Paul Moir]

I wouldn't think that would be much of an issue on a small monochrome CRT?

[Mr Smiley]

All CRT tubes are X-ray machines, that's why the glass is ultra leaded and very heavy and there is a X-ray killer circuit there by law.

 

[TerraHertz]

About 'X-ray killer' circuits - they monitor the EHT to the CRT anode, and turn the system off if the voltage goes above some limit. So far as I know this is only done in large CRTs (typically colour) since their EHT voltages are in the 20KV range.
For small CRTs with HV under 10KV, I don't think X-rays are a problem. But... designing your own EHT circuit could accidentally take the EHT too high. So probably a safety cutout circuit tripping at say 12KV would be a good idea.

Incidentally, the X-rays are generated when the electron beam hits the glass. It's the sudden stop of the beam electrons, and the excitation of electrons in the glass, that produces the X-rays.

[c4757p]

How do these circuits typically go about sensing the voltage? 12kV is a lot. I suppose you could do it by ratio from another voltage on the flyback transformer, but I cannot find any data at all on this one.

I thought X-ray emission was usually only a problem at significantly higher voltages - like 20kV plus? I don't have an HV probe, but would it be safe to say that if I am operating an 8kV-rated tube perfectly well without damaging it I am not in the 20kV range? Can they really tolerate over double rated voltage without even acting funny at all?