UK Vintage Radio Repair and Restoration Discussion Forum

UK Vintage Radio Repair and Restoration Discussion Forum (https://www.vintage-radio.net/forum/index.php)
-   Television Standards Converters, Modulators etc (https://www.vintage-radio.net/forum/forumdisplay.php?f=95)
-   -   Modeline problem (https://www.vintage-radio.net/forum/showthread.php?t=144356)

Keith 28th Feb 2018 9:49 pm

Modeline problem
 
Having now got my Bush TV22 running (https://www.vintage-radio.net/forum/...d.php?t=143583) I'm trying to generate some 405 line video for it.

Inspired by rambo1152 (https://www.vintage-radio.net/forum/...d.php?t=130716), I fitted an NVIDIA NVS 285 card to a WinXP machine, updated to the lastest driver from the NVIDIA website and tried putting the suggested modeline into WinModelines i.e:

Modeline "405i50" 8.10 664 680 752 800 377 378 385 405 -hsync -vsync interlace

As I was typing it in it was clear that there was a problem as I got red error messages. Trying to run it returned the error window "Mode not supported" with "-2" in the header bar.

The preset modes work fine - I can see the video timing changing on a scope connected to the VGA output.

Any suggestions would be gratefully received.

Graham G3ZVT 1st Mar 2018 2:45 pm

Re: Modeline problem
 
You won't regret embarking on this project, I promise!
I got the Nvidia card working under XP without difficulty, but as you may know, after some investigation by Peter Scott, the ATI Radeon HD 2400 Pro became the card of choice, because the drivers compatible with Windows 7 ( and AFAIK later) were suitable.

I will look back through my notes and see if I can find which version ATI driver I was using. It is almost certain to be a graphics driver version issue.


Try this in the meantime, it's what I'm using.
Modeline "664x377_25 10.1kHz 50.0Hz" 8.100 664 680 752 800 377 378 385 405 interlace -hsync -vsync

Keith 1st Mar 2018 3:15 pm

Re: Modeline problem
 
Hi Graham,

I was rather hoping you would answer. The PC I'm using can boot into either XP or XP Pro. Both have problems, however. I currently can't use XP Pro as it's not recognising the keyboard for some unexplained reason. I may have to try and sort that in due course.

The problem with XP is that its IE is corrupted. This doesn't normally bother me as I use Firefox but it was possibly the reason I had difficulty updating the video driver using the NVIDIA installer. I eventually managed it manually through Device Manager and it seems to work OK. I am able to set up dual outputs and monitor the second one on a scope. When I can see sensible 405x50 video on it I'll feed it to the grid of the last video IF for now.

I'd be interested to know which NVIDIA driver version you used - maybe there has been update since your installation and possibly this could be my problem (/clutching at straws mode).

Haven't got round to building a modulator yet but I have bought a couple of 1496 multipliers and obtained 45 and 41.5MHz crystals.

Keith 1st Mar 2018 3:54 pm

Re: Modeline problem
 
1 Attachment(s)
Well - I'll be blowed! Today it works. I now have what looks like 405 line video coming out of the PC (scope is on 20uS/div). Maybe it just needed a restart. Anyway, I think it's time to get the TV22 connected up and see what I get...

Graham G3ZVT 1st Mar 2018 6:38 pm

Re: Modeline problem
 
The more I thought about the fact the card could output 625 & 525 the more convinced I was that your setup mwas viable for 405.

Anyway, well done.

Here are some "rules" that I stick to for a successful outcome.
Always make sure the PC boots up with VGA screens on both ports showing the extended desktop.

Then switch port 2 to the modulator

Only then activate the modeline ensuring you have the secondary or Video1 destination set on the dropdown menu

Also I found that if several modelines are listed in the application, things didn't always behave as expected, so I only have the 405 one in the list.

What modulator are you using?
I'm using this one
http://www.earlytelevision.org/405_modulator.html
I'm really impressed with the performance, both sound and vision, in fact I am building a second one now, which will have a tidier construction.

You may like to consider downloading Vidblaster X Trail

https://www.youtube.com/watch?v=ahzdx4KgWBg

Graham G3ZVT 1st Mar 2018 9:00 pm

Re: Modeline problem
 
1 Attachment(s)
Apologies Keith, I just re read post#3 and you told me your modulator plans.

Personally, I feel that xtal control is an unnecessary extravagance at Band I
Likewise, a custom PCB. I have various bits and pieces for making simple boards, but I always seem to chicken out and reach for the drawer full of little project boards.
Attachment 158476
That's the sound part finished, it just needs an evening's work to complete the rest. If it works as well as the prototype I'll be happy.

Yes, those coils have too many turns, I forgot to shorten them before mounting them.

Argus25 2nd Mar 2018 1:47 am

Re: Modeline problem
 
Keith,
I have built many video modulators over the years. For your application I would recommend:

Use crystal control 41.5MHz and 45 MHz with two separate amplitude modulators and mix them in the correct ratio with the sound one 6dB down. Then a buffer amplifier and level control is helpful.

Don't forget, in the case of the video one, to sync tip or black level clamp the video at your modulator's IC's input (if you are using an IC) or otherwise on many sets (eg Bush TV22 and others) you will have retrace lines appearing at times.

Also, make a board with a ground plane, or you will find any measurements you make with the scope will be confounded by where you connect the scope earth along with other issues of not using a ground plane at these frequencies.

Hugo.

Keith 2nd Mar 2018 10:58 am

Re: Modeline problem
 
Thanks for your replies and advice. Much appreciated. I've now tried inputing the video to the grid of the TV22's video amp but not getting anything sensible as yet. Horiontal sync very ragged and no vertical lock. Probably not surprising as all I've done to the TV so far is replace the waxies and a few high value resistors. I need to get in there with a scope and have a look at the sync separator. I will continue the saga on https://www.vintage-radio.net/forum/...d.php?t=143583 .

peter_scott 2nd Mar 2018 3:02 pm

Re: Modeline problem
 
Hi Keith,

I've just come across your thread and I'm away from home for a few weeks.
I think if you view your video waveform at frame rate you'll see why your set is not locking.

The composite sync mode from the graphics card doesn't give totally correct line pulses within the frame pulse. Although rather crude my far from professional piece of design can improve the syncs with a couple of logic devices.

See: http://www.nostalgiatech.co.uk/Baird%20240%20lines.htm

Peter

Keith 2nd Mar 2018 5:56 pm

Re: Modeline problem
 
Hi Peter,

What a wonderful website - don't know how I haven't found it before!

Thanks for the advice on the sync pulses. I may get to the fine details later but right now I'm just trying to get something recognisable on the screen. No progress today as other household items have needed attention first.

Graham G3ZVT 2nd Mar 2018 8:07 pm

Re: Modeline problem
 
Its interesting to note how different set types respond to the non-standard waveform. My Cossor 916 has no problem with vertical lock, it's rock steady over 70% of the travel of the vertical hold pot.

What I don't get however is any interlace.
I don't regard that as a major problem, it might even be considered an advantage when staring at Testcard C at close range, and even with a full spec BBC signal, good interlace was a challenge, due to understandable compromises made by Alan Blumlein et al.

Notwithstanding the above, I did build Peter's logic circuit in order to improve the interlace situation, but didn't have much success, I will revisit it soon, particularly as I now have a cheap PC oscilloscope.

peter_scott 2nd Mar 2018 8:42 pm

Re: Modeline problem
 
I must admit that the correct adjustment of the dreaded monostables does require a 'scope and I don't think Rambo currently has access to one.

Hands up, anything with monostables is crappy design and requires pot adjustment to work. (But it does give a simple solution and I can get good stable results with a 1937 set without the advantages of flywheel sync.)

Peter :-]

Keith 2nd Mar 2018 9:42 pm

Re: Modeline problem
 
Going backwards, I'm afraid. Eventually I took a closer look at the frame sync as suggested by Peter. No subtle error involving the inclusion or otherwise of line pulses though, but a massive error - the vsync is upside down (i.e. +ve) and spaced at 40mS!
The modeline I pasted was as stated by Graham, i.e. "Modeline "664x377_25 10.1kHz 50.0Hz" 8.100 664 680 752 800 377 378 385 405 interlace -hsync -vsync", so I'm guessing I'm back at a driver compatibility problem.

The driver version as reported by Device Manager is:

NVIDIA QUADRO 285
06/11/2012
Version 6.14.13.475

Graham, I wonder if you can recall which NVIDIA driver you used successfully and how you found it?

peter_scott 2nd Mar 2018 10:25 pm

Re: Modeline problem
 
Just put a +ve in the vsync. It won't be perfect because the composite sync from that mode line is far from perfect. You can't rely on the composite sync. for the reasons I mentioned earlier but it might give you enough to see that there are other things that make it worth pursuing. (like better video bandwidth for example.)

The images in my website are genuine.

Peter

Argus25 3rd Mar 2018 5:09 am

Re: Modeline problem
 
Since it appears that the modeline system produces good video picture data, but has unsatisfactory sync, then a reasonable solution would be to simply make a new sync generator circuit, probably just synchronised with resets, and re-insert a good sync to the output video signal.

There is really no good reason why the interlace should be problematic, provided the sync signal has the timing that there is an exact a half line difference at the leading edge of the vertical sync . If it does have issues on the actual set even with a correct signal and there is poor interlace, this merely suggests that the frame osc in the TV is being inadvertently triggered by line pulses which should not happen if the interlace filter has the correct proportions/time constant.

Keith 3rd Mar 2018 1:14 pm

Re: Modeline problem
 
Getting nowhere today. Winmodelines seems to have a mind of it's own. With the same modeline that worked (at least for hsync) two days ago, I now get lines of about 60uS. Time to give up maybe - anyone got any 405 line video tapes?

peter_scott 3rd Mar 2018 4:18 pm

Re: Modeline problem
 
Hi Keith,

Are you using Jeroni Paul's Modeline Editor ?

Can you give us a screen shot of what your settings are?

Peter

Keith 3rd Mar 2018 4:20 pm

Re: Modeline problem
 
Yes - Version 1.10. It worked yesterday (apart from the incorrect frame sync) but not today:angry:

Keith 3rd Mar 2018 4:27 pm

Re: Modeline problem
 
1 Attachment(s)
Here is screenshot.

peter_scott 3rd Mar 2018 5:34 pm

Re: Modeline problem
 
Hi Keith,

In your modeline try changing the 404 number to 405.

Peter


All times are GMT +1. The time now is 9:50 am.

Powered by vBulletin®
Copyright ©2000 - 2024, vBulletin Solutions, Inc.
Copyright ©2002 - 2023, Paul Stenning.