Electronics > Projects, Designs, and Technical Stuff
Home Brew Analog Computer System
<< < (15/66) > >>
GK:
If I had a spare LOPT/trippler that is something that I would consider, but how long before these things are unobtanium and what would I do if the thing went poof a couple of years from now? They are a highly stressed part and failures are not uncommon. There is still heaps of this stuff around right now, but people (down here at least) are now throwing away their CRT televisions and computer monitors as if they were something one could catch rabies from! 

I was considering winding my own HV step-up transformer on a large ETD core (I have a kit of selected samples from the entire range) and making my own voltage multiplier. The multiplier could have its HV diodes and ceramic caps laid out on a spacious PCB. Spray half a can of conformable coating over it and it should last forever.

 

SeanB:
I can still buy LOPT's and triplers for Trinitron TV sets, the bloody things just keep on trucking, horrible bodges to replace GTO thyristors notwithstanding. If you are worried just use the LOPT and grab a few spares from similar sets, they all run off pretty much the same voltage (108-130V) and have near identical charactaristics, so can just be substituted when needed, just make a note ( place in plastic baggie and cable tie to LOPT EHT lead) of pins used for HT and GND along with the drive pin, and the values of the tuning caps ( put the caps themselves and the transistor in the baggie as well) needed.

3 Dead Trolls in a Baggie...........

http://youtu.be/nL24aNugo_4

GK:
Well I guess I could get my hands on a few spares if I really wanted to, but I still prefer to roll my own, just because I can.



-----------------------------------------------------------------------------------------------------------------------------------------------------------------------------

I've completed the design for the horizontal coil driver amplifier. The yokes horizontal coil has an inductance of 2mH shunted with 70pF winding capacitance and a series resistance of 1.2 ohms and requires a smidge under 1.4A (accurately measured this time) peak for full scale deflection.

The composite current driver amplifier I have come up with (schematic attached) forces -Vin across resistor Rsense (lower right hand side of the schematic). The current flowing through Rsense is the deflection coil current. The small signal bandwidth is 35kHz and the large signal bandwidth (a coil current of 2.8Ap-p) is a fraction over 2kHz. The large signal bandwidth and slew rate is constrained by how much "compliance" voltage the I-out driver amplifier has. I settled on +/-40V rails as a sane compromise between bandwidth and worse case power dissipation, which gives a clipping threshold of approximately 35V.

The small signal bandwidth is constrained by how much feedback can be wrapped around the deflection coil whilst maintaining stability. The first attached plot shows the loop gain and phase of the current sensing feedback loop enclosing the deflection coil. The second plot shows the voltage and current waveforms for the coil driven to full scale deflection (2.8Ap-p) by a 1kHz squarewave input.
   
I'm currently laying out the PCB for this circuit. Should have it knocked off in a couple of evenings. Then I will start considering the design of the current driver for the vertical deflection coil. It will be a fair bit different as the vertical coil has completely different characteristics (140mH inductance, 62 ohms Rs, 180mA peak for full scale deflection).

   


   
GK:
I'm making steady progress with the deign and construction of the electronics for this magnetic deflection CRT XY oscilloscope, but right now I'm still playing around with the crappy/bodged/experimental/test bread-boarded circuits on the bench.

I'm having an issue with the beam convergence when wishing to display additive colours. Displaying red, green and blue traces independently works fine, but I'm having an issue when wanting to mix red with any of the other colours (eg red+green for a yellow trace). At the moment, the green and blue beams are, for all sakes and purposes, perfectly converged, but the red beam will only accurately converge with the others when the focus control is right at one end of its travel - but display focus is optimal at roughly mid position.

So, obviously, I need to adjust/calibrate the convergence for red beam. My CRT here is of the modern type in which the individual beam convergences are adjusted by means of rotatable pairs of magnets on the neck of the CRT only. There are no coils nor electronic compensatory circuits at all. I've attached a picture of the neck of the CRT and the rotatable magnet assemblies.

However, I'm not sure what the correct procedure is and I'm a bit paranoid about screwing things up! Any ex-TV service technicians out there who can offer advice? To the best of my knowledge, for this type of CRT, the convergences are adjusted for a single point in the middle of the screen with no currents applied to the deflection coils. However if someone could kindly outline the magnet adjustment procedure before I take a crack of it I'd be quite appreciative!

I've also attached screen pictures of the 7 binary (a beam either on or off) colour combinations as they are currently generated, with the less than optimal focus setting required to keep the red gun converged with the green and blue. The colors don't appear as vivid in photos as they do in real life, unfortunately.




 
GK:

--- Quote from: GK on March 27, 2013, 11:55:51 am ---Some major progress has been made in the construction of this computer, but I won't have much further fully functional to show until a bunch of back-order bits arrive.

Here is a little play in LTspice in the meanwhile   :D:

www.users.on.net/~glenk/bouncingball.asc

--- End quote ---


Some more playing around............ Just for fun I used the ".wave" LTspice directive in a modified version of that simulation to record and save the simulated X and Y signals into the left and right channels of an audio file. This enabled me to connect the headphone socket on my PC to an XY oscilloscope for a visual display, playing the audio file with windows media player.

I modified the simulation for a much quicker solution time (10mS instead of 1s) and used alternate-write multiplexing to simulate and plot the trajectories of two bouncing balls in unison; one having a slightly higher mass than the other and subjected to a slightly stronger constant for gravity. I simply couldn't do a slow simulation such that the "balls" (CRT beam dots) could be observed slowly bouncing about on the screen (which I could have videoed), due to the high-pass, low frequency response of my PC's audio and there is some significant waveform distortion due to the AC coupling, but it makes a nice demonstration of a PC emulating an analogue computer nonetheless  ;D.   

 

 
Navigation
Message Index
Next page
Previous page
There was an error while thanking
Thanking...

Go to full version
Powered by SMFPacks Advanced Attachments Uploader Mod