OT: How to convert composite video into TTL?

898 views
Skip to first unread message

jb-electronics

unread,
Apr 9, 2021, 11:28:41 PM4/9/21
to neoni...@googlegroups.com
Hi all,

This is a bit of an off-topic question, but I hope there will be
somebody here that can help. I have bought a new-old stock monochrome
CRT for a HP Agilent 8920A, basically this unit here:
https://www.ebay.com/itm/273930914548 .

It looks to me that it is just a rebranded OmniVision 6" display:
http://www.omnivisionusa.com/Industrial-LCD-CRT-Monitors/replacement-crt/crt-monitors/6-inch-kit-.html

Now unfortunately it takes TTL video as input. But I have a composite
source (a Raspberry Pi). How can I convert composite into TTL? I
basically need to extract Hsync and Vsync and feed it separately to the
unit, okay. There is an old circuit here that does just that:
https://www.elektormagazine.com/magazine/elektor-198812/47485

But I know there are also dedicated chips, like the GS1881:
https://www.mouser.com/datasheet/2/761/GS1881_GS4881_GS4981_Datasheet-769183.pdf

So basically here is my question: before I dive into this any further,
is there any chance of success? What is the optimal choice? Or is it
quite unlikely that I will be able to convert the signals? I mean, even
if I manage to extract Hsync and Vsync, is it likely to work?

I am sorry for the naive question, but I am no expert on video signals,
and it would be nice if anybody more experienced could chime in with a
few words of caution/experience.

Best wishes
Jens

Adrian Godwin

unread,
Apr 10, 2021, 12:29:09 AM4/10/21
to neonixie-l
It's easy enough to extract the sync signals, as you've seen. It may not be particularly difficult to modify the TTL input to analog, or to create a TTL level signal if you don't need a grey-scale.  On many general-purpose monitors like that they could often be built for either standard - the microvitec Cub  popular with the BBC Microcomputer had that option.

But the critical thing is that it runs at the right speed. CRT circuits are built around the line oscillator which generates the horizontal scan AND the EHT voltage. It was only when multisync monitors came along that the optimisation was split to reduce the dependency. And if the frame frequency is wrong, you will often have problems getting a full frame scan when you force it into sync.

So the first thing is that you need to make sure the 8920 monitor ran at either PAL or NTSC rates so you can choose a raspberry pi format to match. There's a good chance it's NTSC but later monitors like the ones in the 54 series of digital scopes were more like the IBM standards MDA and EGA (CGA was NTSC). 

--
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/e0f398e8-24b1-952c-c998-0ddfbed6849e%40jb-electronics.de.

Alex

unread,
Apr 10, 2021, 7:03:36 AM4/10/21
to neonixie-l
I actually have a panasonic M-5001A CRT module sitting on a shelf waiting for a project that looks very similar to that which I scavenged out of a dead bit of custom test equipment a decade or more back - it indeed uses TTL level inputs and back along I got it working from a PIC using, I think, this project here (all my bookmarks are 404 now but I think this was it : https://web.archive.org/web/20121106102606/www.rickard.gunee.com/projects/video/pic/howto.php)

I think you may have a better project using the PIs GPIO to run the display, its nice then being TTL as it becomes a timing and frame store programming exercise more than anything else, though that would be a proper project (though probably simpler hardware with just level transition) than running it from composite.

Found a photo of it running from the PIC in my archive, love a mono phosphor CRT:

2014-06-07_20-54-30_164.jpg

jb-electronics

unread,
Apr 10, 2021, 9:04:10 AM4/10/21
to neoni...@googlegroups.com
Thank you! I did some more diggin and found the service manual for the display (Panasonic TR-60S1A, see here: https://www.opweb.de/english/company/Panasonic/TR-60S1A)

There is is a timing chart (see below). It looks an awful lot alike NTSC to me, can somebody confirm?

Best wishes
Jens

jb-electronics

unread,
Apr 10, 2021, 9:04:15 AM4/10/21
to neoni...@googlegroups.com
Thank you! I did some more digging and found the service manual for the display (Panasonic TR-60S1A, see here: https://www.opweb.de/english/company/Panasonic/TR-60S1A)

There is is a timing chart (see below). It looks an awful lot alike NTSC to me, can somebody confirm?

Best wishes
Jens




On 2021-04-10 12:28 a.m., Adrian Godwin wrote:

David Forbes

unread,
Apr 10, 2021, 3:54:09 PM4/10/21
to NeoNixie
I built a couple LED TV sets about ten years ago. I had to solve this problem. I bought a sync separator chip for the first one, and a digitizer chip for the other.
See the bottom of the page for schematics.



To view this discussion on the web, visit https://groups.google.com/d/msgid/neonixie-l/0b6843bf-8882-8a17-7c62-437ffc98a5ff%40jb-electronics.de.
mphoilcjefeebdjc.png

chuckrr

unread,
Apr 10, 2021, 6:13:33 PM4/10/21
to NeoNixie

David,  How come you never answer any of my questions about the rubidium oscillator?   I have asked you about it many, many times and have never

received any response from you.    -Chuck
 

To view this discussion on the web, visit https://groups.google.com/d/msgid/neonixie-l/CAPbqtvdWTmhDEjMoQQvcS%2BBROe64k3S%3Dy9_g2%2Bn%2B_ACQDVyHJg%40mail.gmail.com.

5-ht

unread,
Apr 10, 2021, 10:52:47 PM4/10/21
to neonixie-l
Jens,
The horizontal sync requirement of this monitor (52 uS / 19.2Khz) is somewhat higher than a standard NTSC signal which is about 63 uS / 15.7 Khz.

Mark

Mac Doktor

unread,
Apr 11, 2021, 5:25:10 PM4/11/21
to neonixie-l

On Apr 10, 2021, at 3:53 PM, David Forbes <nixie...@gmail.com> wrote:

I built a couple LED TV sets about ten years ago. I had to solve this problem. I bought a sync separator chip for the first one, and a digitizer chip for the other.
See the bottom of the page for schematics.

Are you going to take it to the next level, i.e., a mechanical television? I always wanted one of those.

This is assuming that there aren't a dozen of them on Hackaday or a crowdfund site right now.


Terry Bowman, KA4HJH
"The Mac Doctor"


“...the book said something astonishing, a very big thought. The stars, it said, were suns but very far away. The Sun was a star but close up.”—Carl Sagan, "The Backbone Of Night", Cosmos, 1980


jb-electronics

unread,
Apr 11, 2021, 5:51:06 PM4/11/21
to neoni...@googlegroups.com
Thank  you! I think it should be possible to modify the horizontal and vertical deflection circuitry on the board to make it a proper NTSC frequency, yes?

Jens

Adrian Godwin

unread,
Apr 11, 2021, 7:19:20 PM4/11/21
to neonixie-l
It sounds a bit of a stretch - 15kHz to 19kHz. You could try running the monitor without an input signal and adjusting horizontal frequency (if there is one) to see what range it can manage.

Another possibility is to fix it in software by creating a custom monitor signal. Some information is at
https://github.com/raspberrypi/documentation/blob/master/configuration/config-txt/video.md



To view this discussion on the web, visit https://groups.google.com/d/msgid/neonixie-l/37bf74f3-842f-2d17-c5b7-0606f09d9d4e%40jb-electronics.de.

gregebert

unread,
Apr 11, 2021, 9:26:36 PM4/11/21
to neonixie-l
At least it's just monochrome. NTSC composite video with color is rather complex, though I must say it is also ingenious in that it is backward-compatible to monochrome and packs so much video information into a 3.58Mhz bandwidth. Similar for PAL, though higher bandwidth.

jb-electronics

unread,
Apr 11, 2021, 10:22:56 PM4/11/21
to neoni...@googlegroups.com
Still new to the whole composite video thing---could you clarify what is the stretch? Do you think the driving electronics might not work at the lower NTSC frequency? Or is there a mechanical impediment, like the windings of the CRT's yoke? Thank you!

Jens

jb-electronics

unread,
Apr 11, 2021, 10:23:27 PM4/11/21
to neoni...@googlegroups.com
Oh, yes, the entire discussion is about monochrome video. Sorry, should have mentioned it in the follow up email as well.

Adrian Godwin

unread,
Apr 11, 2021, 10:41:33 PM4/11/21
to neonixie-l
The line oscillator has to run close to the line frequency and is then phase-locked to it by the horizontal hold mechanism. If the oscillator is tuned for 19kHz it's a big stretch to get it to sync to a 15kHz input signal. 

Yes, you might be able to modify some components to slow it down, or even just change the horizontal preset frequency to get it closer  but that line oscillator is the most complex thing in as monochrome monitor. It generates a linear ramp for the horizontal coils and is also wound up with line output transformer which generates the high voltage for the tube. I guess this makes sense as during retrace the line coils are getting a big pulse to shove the spot to the other side of the screen. But drastically changing the line frequency is likely to have an effect on the line linearity and the EHT voltage.

I'm not saying don't try it - it would be very educational, though best if you can get hold of a schematic. But a lot easier to get hold of an NTSC monitor, I think. The people who maintain arcade machines have been scraping up all the large colour monitors recently but I doubt they'd have much use for a small monochrome.

If the stored energy in the line coils can be dumped into the line output transformer at the right phase of the cycle, it's saving a lot of power that would otherwise require a separate circuit. In the old days. the more use you could get out of any one vacuum tube the better, and the style persisted until multisync monitors needed to be more flexible (and even the simple earlier ones of those had more complex LOPTs to deal with the variation).
 

To view this discussion on the web, visit https://groups.google.com/d/msgid/neonixie-l/e49bca6f-fc9e-e4ee-9003-ccd950ce1d82%40jb-electronics.de.

Robert G. Schaffrath

unread,
Apr 12, 2021, 8:34:31 AM4/12/21
to neonixie-l
Yes the work done to get color into the 3.58MHz (well something like 3.579545MHz) was incredible. The math involved in transmitting the RGB info is truly amazing. And it was all decoded through analog circuitry! No digital processing involved. Quite a feat for its day. In fact, NTSC TV's were pretty much Analog Computers that received a complex signal and processed it with discrete components into a viewable picture. The real complexity of the system was in the TV transmitter that had to provide a perfectly timed and on frequency signal that the TV would receive. When turned on, the TV had to be somewhere in the ballpark in order to latch on to the signal. But upon receiving that broadcast signal, the circuitry would lock its oscillators onto the horizontal and vertical timing pulses and be in sync (of course you had to adjust the horizontal and vertical hold to give the TV a fighting chance of locking on). NTSC (Never Twice the Same Color) was a bit of an issue with needing Color and Hue controls that adjusted the TV's reception of the 3.58MHz color carrier. PAL eliminated the need for those controls.

Of course now with digital TV, the idea of a vertical and horizontal hold controls, along with hue and color, seems so quaint!

Adrian Godwin

unread,
Apr 12, 2021, 8:59:47 AM4/12/21
to neonixie-l
On Mon, Apr 12, 2021 at 1:34 PM Robert G. Schaffrath <robert.s...@gmail.com> wrote:
locking on). NTSC (Never Twice the Same Color) was a bit of an issue with needing Color and Hue controls that adjusted the TV's 

To complete the trio : SECAM is Something Essential Contrary to the American Method, and PAL is Perfection At Last.

Charles MacDonald

unread,
Apr 12, 2021, 11:06:07 PM4/12/21
to neoni...@googlegroups.com, Adrian Godwin
On 2021-04-12 8:59 a.m., Adrian Godwin wrote:

> To complete the trio : SECAM is Something Essentialy Contrary to the
> American Method, and PAL is Perfection At Last.

I recall hearing it as "Pay (for) Additional Luxury".

the early sets would have needed extra circuitry to decode the
alternating lines. BUT the explanation at th etime was that microwave
inter station transmission in Europe had not been built with colour in
mind and PAL had sufficient error correction built in that it did not
matter.


--
Charles MacDonald VA3CPY Stittsville Ontario
cm...@zeusprune.ca Just Beyond the Fringe
No Microsoft Products were used in sending this e-mail.

John Rehwinkel

unread,
Apr 13, 2021, 8:41:58 AM4/13/21
to 'David Weiner' via neonixie-l
> I'm not saying don't try it - it would be very educational, though best if you can get hold of a schematic. But a lot easier to get hold of an NTSC monitor, I think. The people who maintain arcade machines have been scraping up all the large colour monitors recently but I doubt they'd have much use for a small monochrome.

Small monochrome monitors used to be really common as viewfinders, security monitors, etc. They're less so today, but they're still out there. Let me know if you have a need for one.

- John

Robert G. Schaffrath

unread,
Apr 13, 2021, 2:09:54 PM4/13/21
to neonixie-l
Back in the 1990's I purchased new surplus CRT based viewfinder module designed for a camcorder from a surplus electronics vendor. Was rather inexpensive. It has a tiny 1" B&W CRT and takes composite NTSC video. The only issue with it was the image was mirrored due to it being designed to be viewed indirectly via a mirror that would cause the image to display correctly. It was a simple matter to swap two wires to make it work correctly for direct view. IIRC, it would run on a common 9V battery. This post now has me wondering where I put it. I know I never sold it or threw it out. Big problem these days is where to get an NTSC video source. The only things I currently have left are an old 8mm camcorder and a digital to NTSC converter box that were common when the US went all digital.

gregebert

unread,
Apr 13, 2021, 2:39:22 PM4/13/21
to neonixie-l
VCRs and older game consoles (like the original XBOX) will provide NTSC via their AV output.  Some of the newer TVs dont even have composite video jacks, so the standard is rapidly dying. 

Too bad, because studio-quality NTSC is actually very crisp; it's the transmission medium and receivers that make it look so horrible.

jb-electronics

unread,
Apr 13, 2021, 5:57:16 PM4/13/21
to neoni...@googlegroups.com
Hey all,

I am looking for a 5"-6" monochrome CRT, preferable green or amber, with
composite video input. If anybody has one for sale, please let me know!

Best wishes
Jens

Mac Doktor

unread,
Apr 13, 2021, 7:54:40 PM4/13/21
to neonixie-l

On Apr 13, 2021, at 2:09 PM, Robert G. Schaffrath <robert.s...@gmail.com> wrote:

Big problem these days is where to get an NTSC video source. The only things I currently have left are an old 8mm camcorder and a digital to NTSC converter box that were common when the US went all digital.

FYI: if you have a bunch of Video8 or Hi8 tapes you'd like to dub off (I have quite a pile) the best thing to use is a Digital8 camcorder. It can play (not record) both formats. It can also convert them, as well as both composite and S-video, to a DV stream over FireWire. I have an older iMac with El Capitan that can still import DV streams via Apple's old iMovie HD app (iMovie X is a no go). I just used it a few weeks ago.

The Sony Digital8 camcorders were the last to have really fast optics. Their miniDV camcorders (I have one of those as well for daylight video) have sharper video but poor low light and infrared performance. 

SD video is a thing of the past but the Digital8 is still my go-to for infrared. It comes in handy every Halloween.

I have two of them, BTW. Just in case one breaks.


Terry Bowman, KA4HJH
"The Mac Doctor"

"If only you could see what I've seen with your eyes."—Roy Batty, Blade Runner

Adrian Godwin

unread,
Apr 13, 2021, 9:05:32 PM4/13/21
to neonixie-l
Could you use an old 'scope ? Or would it be too big ?

--
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/26bb430c-025f-f3c4-34d2-42db16b2dbf1%40jb-electronics.de.

jb-electronics

unread,
Apr 14, 2021, 9:44:38 AM4/14/21
to neoni...@googlegroups.com
It's a good idea, but these scope tubes are too deep and don't have a composite input usually (but x/y deflection setup). I am trying to build a luggable computer (think Osborne I).

Jens

Adrian Godwin

unread,
Apr 14, 2021, 10:02:30 AM4/14/21
to neonixie-l
No, they often have a Z modulation input and many have a TV sync, so you'd only need to add a vertical scan. But tube depth is a problem. Green isn't, though !

To view this discussion on the web, visit https://groups.google.com/d/msgid/neonixie-l/abc536e3-c2f5-7a6a-0696-b70153f0106b%40jb-electronics.de.

Toby Thain

unread,
Apr 14, 2021, 10:18:57 AM4/14/21
to neoni...@googlegroups.com
On 2021-04-14 10:02 a.m., Adrian Godwin wrote:
> No, they often have a Z modulation input and many have a TV sync, so
> you'd only need to add a vertical scan. But tube depth is a problem.
> Green isn't, though !
>
> On Wed, Apr 14, 2021 at 2:44 PM jb-electronics
> <webm...@jb-electronics.de <mailto:webm...@jb-electronics.de>> wrote:
>
> It's a good idea, but these scope tubes are too deep and don't have
> a composite input usually (but x/y deflection setup). I am trying to
> build a luggable computer (think Osborne I).
>
> Jens
>


Portable TV tubes are still available if you're willing to build the
supporting circuitry. This one is 6 x 4.5cm, white phosphor, magnetic
deflection, and some are sold with the TV frequency yoke.

https://www.ebay.ca/itm/8LK3B-vintage-CRT-tube-3-6-cm-x-4-5cm-USSR-Soviet-Oscilloscope-NEW/254462273192

(A friend and I are trying to integrate this tube at the moment.)

--Toby


jb-electronics

unread,
Apr 15, 2021, 8:41:32 PM4/15/21
to neoni...@googlegroups.com
Hi all,

I got it working... somewhat! I hooked the Raspberry Pi's NTSC composite up to the LM1881 sync separator, and after fiddling with the horizontal hold it snapped into place. A few problems:
  • As mentioned before, the screen coverage is a bit poor (lots of space to the left and right).
  • The brightness is rather low (what is full brightness at TTL? Probably 5V? Composite maxes out way below that, so maybe I should hook up an amplifier?)

So here is the schematic of a very similar module: http://oz1db.dk/hp/omnivision_man_lp06xx_rev0_0611.pdf . Can't I just change the timing capacitors for the horizontal and vertical sync? (I.e. mess with R17 at U1 and R35 at U2)? Based on what Adrian wrote below it seems it might be possible, within reason? Since it is already working (kind of, limited screen coverage) I probably don't ahve to tweak much?

So if anybody can help on the brightness issue as well as on the frequency modification I would be very grateful!

Best wishes
Jens

Adrian Godwin

unread,
Apr 16, 2021, 5:45:04 AM4/16/21
to neonixie-l
That's very encouraging - if it can go far enough to sync it's probably usable.

By all means try adding some capacitance to the timing caps. Just turn it off quickly if it refuses to oscillate - it might be warming up the line driver transistor. You shouldn't need to adjust the vertical stuff at all. 

There may be a width control - either a normal preset or an inductor. Try getting more width with that.

TTL can be as low as 2.4V but is probably more like 4, and the entire signal is video. Composite is only 1V, and the sync makes up the bottom 0.3V so you will certinaly be short of contrast. An amplifier is possible but it's amplifying to 60V or so already so you can probably get away with just tweaking the input circuit.
  

To view this discussion on the web, visit https://groups.google.com/d/msgid/neonixie-l/5a7d7063-5c00-2fc2-3fd7-eaeba8333662%40jb-electronics.de.
Reply all
Reply to author
Forward
0 new messages