Spectrum Analyser CRT drive problem
 

Apologies in advance if this is considered OT.

I am having a minor problem with the display on my spectrum analyser Mk2
project, see
http://www.hanssummers.com/electroni...ser2/index.htm
.. The display for this analyser is an 4.5-inch ex-TV tube. I successfully
dismantled that TV and rebuilt the necessary drivers etc as a miniature
monitor having TTL-level video and composite sync inputs.

Naturally the type of display this tube is now being subjected to is very
different to a TV picture. There seem to be some problems with the video
drive circuits, see pictures at the bottom of the above web page. The
graticule fades out in the top eighth of the screen. Also see the "noise
floor" photo which shows significant horizontal "bulge".

Also there is a resolution problem - one spectrum analyser sweep consists of
640 measurements, and I have a complicated interpolation circuit which
"joins the dots" i.e. draws lines between the individual measurements, and
also reduces the horizontal resolution to 320 pixels which I hoped would be
displayable on the monitor. Problem is I can only get 160 horizontal pixels
to display reliably - at 320 pixels a vertical line is very faint or
invisible (depending on how much white there was in the preceeding
microsecond or so's worth of scan line.

The circuit diagram is also on that page
(http://www.hanssummers.com/electroni...yser2/circuitm
..pdf). I'm no TV expert. Can anyone suggest what modifications should be
made to the video driver circuit to correct these problems? Alternatively
what newsgroup might be populated with TV gurus?

73

Hans G0UPL

In article ,
says...
Very nice construction on the little TV conversion to display on your
website! :-)

No kidding. I've bugged Hans a lot about how he really should be doing
all that work with a Mini-ITX PC motherboard, but all kidding aside,
that project is what "homebrewing" is really all about.

-- jm

------------------------------------------------------
http://www.qsl.net/ke5fx
Note: My E-mail address has been altered to avoid spam
------------------------------------------------------

In article ,
says...
Very nice construction on the little TV conversion to display on your
website! :-)

No kidding. I've bugged Hans a lot about how he really should be doing
all that work with a Mini-ITX PC motherboard, but all kidding aside,
that project is what "homebrewing" is really all about.

-- jm

------------------------------------------------------
http://www.qsl.net/ke5fx
Note: My E-mail address has been altered to avoid spam
------------------------------------------------------

"John Miles" wrote in message
...
In article ,
says...
Very nice construction on the little TV conversion to display on your
website! :-)

No kidding. I've bugged Hans a lot about how he really should be doing
all that work with a Mini-ITX PC motherboard, but all kidding aside,
that project is what "homebrewing" is really all about.

Thanks guys.

John, the sci.electronics.repair folks didn't have anything to say about it
:-(

Someone else thought the series LC from the driver transistor's emiter to
ground might be a "colour killer" and putting a notch in the frequency
response of the driver at about the same frequency my 320 pixels are at
(4MHz). I tried shorting the L altogether and there was no visible
difference on the display. That same person also mentioned gamma of the
tube, but I don't know much about that. I would find it hard to believe that
the tube itself was the limitation here, it must be the video driver.

Interestingly the fading at the top of the screen dissappears when the
"brightness" control is turned up, but that makes the bulging noise floor
worse and also dims thin vertical lines even more. Arrgh.

73

Hans G0UPL

"John Miles" wrote in message
...
In article ,
says...
Very nice construction on the little TV conversion to display on your
website! :-)

No kidding. I've bugged Hans a lot about how he really should be doing
all that work with a Mini-ITX PC motherboard, but all kidding aside,
that project is what "homebrewing" is really all about.

Thanks guys.

John, the sci.electronics.repair folks didn't have anything to say about it
:-(

Someone else thought the series LC from the driver transistor's emiter to
ground might be a "colour killer" and putting a notch in the frequency
response of the driver at about the same frequency my 320 pixels are at
(4MHz). I tried shorting the L altogether and there was no visible
difference on the display. That same person also mentioned gamma of the
tube, but I don't know much about that. I would find it hard to believe that
the tube itself was the limitation here, it must be the video driver.

Interestingly the fading at the top of the screen dissappears when the
"brightness" control is turned up, but that makes the bulging noise floor
worse and also dims thin vertical lines even more. Arrgh.

73

Hans G0UPL

Hans Summers wrote:

: Naturally the type of display this tube is now being subjected to is very
: different to a TV picture. There seem to be some problems with the video
: drive circuits, see pictures at the bottom of the above web page. The
: graticule fades out in the top eighth of the screen.

That looks like a video DC-restoration problem. Try mixing your TTL video
signal with the composite sync signal in order to obtain a more TV-like
video signal. Note that the video amplifier is AC-coupled to the picture
tube and DC-restoration is done in that case by the CRT itself, the cathode
and grid electrodes work like a diode that clamps the video signal to the
sync level (no sync pulses = no proper DC-restoration, brightness will
fluctuate with video contents). The time constant of the RC coupling to the
CRT should be high enough (in your case it is).

: Also see the "noise floor" photo which shows significant horizontal "bulge".

That could be due to poor HV regulation (check by reducing the brightness,
if the image looks OK when you have just a faint image, then the problem is
the HV). That's is a problem of using a normal TV for digital signals, and
there is not much that can be done, other than redesigning the whole
horizontal sweep circuitry (needing perhaps a different flyback). Check in
any case the DC supply to the horiz output for fluctuations. Computer
monitors are much better in that respect.

: Also there is a resolution problem - one spectrum analyser sweep consists of
: 640 measurements, and I have a complicated interpolation circuit which
: "joins the dots" i.e. draws lines between the individual measurements, and
: also reduces the horizontal resolution to 320 pixels which I hoped would be
: displayable on the monitor. Problem is I can only get 160 horizontal pixels
: to display reliably - at 320 pixels a vertical line is very faint or
: invisible (depending on how much white there was in the preceeding
: microsecond or so's worth of scan line.

That's a bandwidth problem in the video amplifier. You may begin by removing
the audio IF trap (the L-C connected to the emitter of the video amplifier),
but I don't think that would help that much. Try using a higher frequency
transistor for the video amplifier, you should choose one with low b-c
capacitance, a cascode amplifier will do a better job since it does not
suffer from Miller effect problems (you might need a bit higher voltage
supply for the amplifier, though). A small inductance in series with the 8K2
resistor in the collector ("peaking coil") should improve high-frequency
response by compensating for the CRT input capacitance.

Hope that helps a bit.

Regards,
Hector.

Hans Summers wrote:

: Naturally the type of display this tube is now being subjected to is very
: different to a TV picture. There seem to be some problems with the video
: drive circuits, see pictures at the bottom of the above web page. The
: graticule fades out in the top eighth of the screen.

That looks like a video DC-restoration problem. Try mixing your TTL video
signal with the composite sync signal in order to obtain a more TV-like
video signal. Note that the video amplifier is AC-coupled to the picture
tube and DC-restoration is done in that case by the CRT itself, the cathode
and grid electrodes work like a diode that clamps the video signal to the
sync level (no sync pulses = no proper DC-restoration, brightness will
fluctuate with video contents). The time constant of the RC coupling to the
CRT should be high enough (in your case it is).

: Also see the "noise floor" photo which shows significant horizontal "bulge".

That could be due to poor HV regulation (check by reducing the brightness,
if the image looks OK when you have just a faint image, then the problem is
the HV). That's is a problem of using a normal TV for digital signals, and
there is not much that can be done, other than redesigning the whole
horizontal sweep circuitry (needing perhaps a different flyback). Check in
any case the DC supply to the horiz output for fluctuations. Computer
monitors are much better in that respect.

: Also there is a resolution problem - one spectrum analyser sweep consists of
: 640 measurements, and I have a complicated interpolation circuit which
: "joins the dots" i.e. draws lines between the individual measurements, and
: also reduces the horizontal resolution to 320 pixels which I hoped would be
: displayable on the monitor. Problem is I can only get 160 horizontal pixels
: to display reliably - at 320 pixels a vertical line is very faint or
: invisible (depending on how much white there was in the preceeding
: microsecond or so's worth of scan line.

That's a bandwidth problem in the video amplifier. You may begin by removing
the audio IF trap (the L-C connected to the emitter of the video amplifier),
but I don't think that would help that much. Try using a higher frequency
transistor for the video amplifier, you should choose one with low b-c
capacitance, a cascode amplifier will do a better job since it does not
suffer from Miller effect problems (you might need a bit higher voltage
supply for the amplifier, though). A small inductance in series with the 8K2
resistor in the collector ("peaking coil") should improve high-frequency
response by compensating for the CRT input capacitance.

Hope that helps a bit.

Regards,
Hector.