Scope Clock - small display issue

[Max DN]

Hello,

I'm building a scope clock, same board as another one I built before with no issues (after some help I got on this group, thank you!) and I'm getting a flat line on the display.

I have tried replacing the CRT, the DAC, the Op Amps and the MPSW42 deflection amplifiers but made no difference. CRT is 3RP1A. Code and circuitry have been fully tested on another working board.

I do recall having a similar issue before with a DH3/91 CRT and I seem to recall that I just needed to adjust the pot for Astigmatism that controls the voltage on A2. Schematics as in here:  http://www.cathodecorner.com/sctv/sctv-schem-2.png

I'm a bit at a loss here, although I'm sure it's something simple that I'm missing.

Pic attached.

Any suggestions welcome,

Max

[Max DN]

Tried to replace a few more parts, including potentiometers for Focus and Intensity, made no difference. So, back to basics: using an oscilloscope I see that the X and Y signals straight out of the DAC look exactly the same as in the picture I had attached. Maybe a faulty DAC? Tried to replace the DAC twice with new ones and made no difference. Looks like I'm hitting some sort of physical limits... More debugging to do...

[Dekatron42]

Could it be the that any of the power lines sags as the signal tries to increase to the right (in the picture right where the line is) or that any of the resistors in the amplifier chain has an incorrect value?

/Martin

[Max DN]

Martin, that's what I am thinking... however I don't think it's the amplifier chain as I get the same issue if I just display X and Y signals straight out of the DAC on the oscilloscope. That brings it down to a few components only: 2 potentiometers for centerting X and Y, their two bypass capacitors, the DAC and its 1 bypass capacitors, the +3.3V Linear regulator, the +5V supply to the DAC. I'm going to try and replace everything once more by one by one... Not the smartest way of doing it but let's see what it brings...

[Dekatron42]

Can you remove the processor and feed the amplifiers directly from a potentiometer connected to NEG5 and POS5 and see how the dot moves on the screen? You need one potentiometer for X & Y so that the dot will appear on screen. This way you will test out the amplifier chains without the DAC involved.

/Martin

[gregebert]

Is there a Z-axis ??? That would be my first suspect; if the trace is still visible, then the appropriate grid needs to be driven a bit more negative.

If the CRT uses rapid movement for dark (technically dim) regions, and slower movement for bright regions, then my best guess is some kind of software tweak would be needed to get the beam to move faster, or off the phosphor-coated area (or both).

[Max DN]

@Martin, if I try to bypass the DAC using POTs for NEG5 and POS5 it all seems fine (the dot goes across the whole X axis and Y axis).

@gregebert, no Z-axis.

I am probing X and Y just out of the DAC and not connected to anything else. I can see the image on the oscilloscope as a full round circle, but as I try to center it to what would be the center of the CRT then thr circle flatlines. 

It feels as though past a certain point on the vertical axis, the dots are just collapsing into the same line and cannot go past (on one end of the Y vertical axis, X axis is fine).

Very odd as I have already used same code and same pcb for another scope clock. After 10 hours on it today, I'm going to let go until I have any other ideas. I have tried replacing the processor, the DAC (3 times), checked all resistors and bypass caps and triple checked all soldering points. Maybe it's time to bin this pcb, but I do like a challenge!

[Dekatron42]

So what happens if you put this Teensy in the other working board, does the error follow the Teensy?

If so, have you tried to reprogram the Teensy or even erase it to set  it to its original setup condition?

/Martin

[gregebert]

So another possibility is that one of the amplifiers is clipping, usually because the input signals are trying to drive the output beyond the supply rails. BTW, most op-amps are not rail-to-rail, but they can get pretty close. It will show up quickly on a scope.

[Max DN]

So, it seems it is now working as expected, although I'm not counting my chickens yet. In the SPI routine from the Teensy to the DAC, I changed the gain for the DAC from two to one, loaded the code, edited the code back to a gain of two in the DAC, loaded it again and voila'. Odd!

It's working as expected. Although I don't understand why!

[gregebert]

Weird indeed, and needs to be investigated. Have you done a few resets and power-cycles just to be sure ?

Any chance you are using revision-control with your source-code ? Sometimes there are other changes you made, but forgot. I know it sounds stupid, but I've done it many times myself and the evidence was right there, thanks to git (or whatever source-control you use).

Years ago, I have also seen bugs in CAD tools. Anyone here remember CUPL ? It was a tool used for generating PLA code from logic equations. Sometimes as you got close to filling-up the device, the tool would report you ran out of logic elements. But, if you were lucky, simply re-arranging lines of code got it to fit-in. It seems unlikely that Teensy would have this problem, but you never know.

[Terry S]

Yep. CUPL and ABEL (from Data I/O).

[Max Di Noi]

Yes, tried everything I could think of. Moved the same teensy to another board and works fine, tried new teensy, tried to erase teensy eeprom, tried to reset teensy to factory (15 secs long press). Nothing.

--
You received this message because you are subscribed to the Google Groups "neonixie-l" group.
To unsubscribe from this group and stop receiving emails from it, send an email to neonixie-l+...@googlegroups.com.
To view this discussion on the web, visit https://groups.google.com/d/msgid/neonixie-l/11475335-9f02-4341-b732-f34db7359f97n%40googlegroups.com.