Author Topic: Composite video issue  (Read 1454 times)

0 Members and 1 Guest are viewing this topic.

Offline schmitt trigger

  • Super Contributor
  • ***
  • Posts: 1364
  • Country: mx
Re: Composite video issue
« Reply #25 on: September 09, 2019, 12:22:55 am »
Line rate is exactly what you have done: set the timebase to display a waveform from the start of a horizontal sync to the start of the next horizontal sync.
In other words, you are displaying a single video line.

And as you can see from the cursor readings (which I may add, you also have set correctly), the frequency reading is now very close to the actual NTSC horizontal frequency of 15734 Hz.

A terminated video waveform should be 1V, from sync tip to peak white, of which about 0.28V belong to the sync (blacker than black), and the remaining is the active video.


Your waveforms look fine, other than the levels appear to be a little different from the NTSC specs, but that may be the cause that your black background is not as black as it should be.
The color burst appears OK, too.
« Last Edit: September 09, 2019, 12:27:37 am by schmitt trigger »
 

Offline mayor

  • Regular Contributor
  • *
  • Posts: 142
  • Country: ca
Re: Composite video issue
« Reply #26 on: September 09, 2019, 08:46:38 am »
Cool, thanks for the explanation. I'll bring the DC voltage down and see where that leads!
 

Offline mikerj

  • Super Contributor
  • ***
  • Posts: 2151
  • Country: gb
Re: Composite video issue
« Reply #27 on: September 09, 2019, 09:06:41 am »
The output waveform looks horribly messy, assuming you were displaying the same screen as the input (though presumably not the same line).  The input looks exactly as I'd expect from low resolution computer generated video with a limited colour palette; there are four discrete luminance levels clearly visible.

Displaying the same video line on input and output simultaneously, with the timebase speeded up to see some detail would be very helpful in diagnosing this.  I'm wondering if your emitter follower is oscillating, if so a ferrite bead on base wire (assuming through hole components) or a lowish value resistor (typically in the range 47-470 ohm) in series with the base may fix this.
 

Offline Renate

  • Regular Contributor
  • *
  • Posts: 103
  • Country: us
Re: Composite video issue
« Reply #28 on: September 09, 2019, 03:37:47 pm »
Off hand, I don't think that it's oscillating.
If you were using my circuit that has no collector resistor, oscillation would be impossible anyway.

Are those last scope shots with the monitor connected or not?
 

Online vk6zgo

  • Super Contributor
  • ***
  • Posts: 4946
  • Country: au
Re: Composite video issue
« Reply #29 on: September 09, 2019, 04:25:08 pm »
The output waveform looks horribly messy, assuming you were displaying the same screen as the input (though presumably not the same line).  The input looks exactly as I'd expect from low resolution computer generated video with a limited colour palette; there are four discrete luminance levels clearly visible.

Displaying the same video line on input and output simultaneously, with the timebase speeded up to see some detail would be very helpful in diagnosing this.  I'm wondering if your emitter follower is oscillating, if so a ferrite bead on base wire (assuming through hole components) or a lowish value resistor (typically in the range 47-470 ohm) in series with the base may fix this.

The video waveform does look better than that in the previous postings, but it is still a bit "rough around the edges"

A good composite video waveform should look pretty much like the one in this link.

https://hackaday.com/2018/01/18/know-your-video-waveform/
Note, however,that the "Raspberry Pi" video in this screenshot doesn't show "setup"between blanking & black level,which a standard NTSC signal should include.

The burst peak to peak amplitude should be about the same as the sync amplitude--- anything less points to loss of high frequency response in the transmission system.

With the waveform shown in the OP's most recent posting, the burst amplitude is low, & the noise level is high, both at blanking,& sync level  either of these conditions would be seriously "out of spec" in TV Broadcasting.
 

Offline StillTrying

  • Super Contributor
  • ***
  • Posts: 2221
  • Country: se
  • Country: Broken Britain
Re: Composite video issue
« Reply #30 on: September 09, 2019, 04:29:25 pm »
Amplitude, timing? The composite signal is about 700mV p/p, with DC of around 5V.

As your video signal is still riding on top of 5V DC, I'd try just AC coupling it to the monitor with a 10 to 100uF, +ve end to your TR's emitter.
CML+  That took much longer than I thought it would.
 

Offline mayor

  • Regular Contributor
  • *
  • Posts: 142
  • Country: ca
Re: Composite video issue
« Reply #31 on: September 09, 2019, 05:22:17 pm »
The output waveform looks horribly messy, assuming you were displaying the same screen as the input (though presumably not the same line).  The input looks exactly as I'd expect from low resolution computer generated video with a limited colour palette; there are four discrete luminance levels clearly visible.

Displaying the same video line on input and output simultaneously, with the timebase speeded up to see some detail would be very helpful in diagnosing this.  I'm wondering if your emitter follower is oscillating, if so a ferrite bead on base wire (assuming through hole components) or a lowish value resistor (typically in the range 47-470 ohm) in series with the base may fix this.

Alright, I'll capture the same line and post back! I think it was the same screen, but can't be 100% sure.
 

Offline mikerj

  • Super Contributor
  • ***
  • Posts: 2151
  • Country: gb
Re: Composite video issue
« Reply #32 on: September 09, 2019, 07:07:49 pm »
Off hand, I don't think that it's oscillating.
If you were using my circuit that has no collector resistor, oscillation would be impossible anyway.

No true, emitter followers are notorious for oscillating.
 

Offline Renate

  • Regular Contributor
  • *
  • Posts: 103
  • Country: us
Re: Composite video issue
« Reply #33 on: September 09, 2019, 07:28:18 pm »
Not true, emitter followers are notorious for oscillating.
Thanks, I learn something new every day.
I was thinking of collector-base feedback.
This is lead inductance/base capacitance resonance.
The frequencies involved would be pretty high.

Still, although the scope view still looks cruddy, I'm not convinced that it's oscillating.
You'd have to crank up the scope timebase and see a nice ringing on an edge.
 

Offline mayor

  • Regular Contributor
  • *
  • Posts: 142
  • Country: ca
Re: Composite video issue
« Reply #34 on: September 09, 2019, 07:46:52 pm »

Still, although the scope view still looks cruddy, I'm not convinced that it's oscillating.
You'd have to crank up the scope timebase and see a nice ringing on an edge.

Will do that too.
 

Offline mayor

  • Regular Contributor
  • *
  • Posts: 142
  • Country: ca
Re: Composite video issue
« Reply #35 on: September 10, 2019, 08:41:01 am »
Off hand, I don't think that it's oscillating.
If you were using my circuit that has no collector resistor, oscillation would be impossible anyway.

Are those last scope shots with the monitor connected or not?

None of my captures are with tv attached. Haven't had time to experiment yesterday, but hopefully will tonight.

I will def try the DC blocking cap, and if there are suggestions on how to limit noise, that'd be great. Feels like this might be a little more involved, though...
 

Offline mayor

  • Regular Contributor
  • *
  • Posts: 142
  • Country: ca
Re: Composite video issue
« Reply #36 on: September 11, 2019, 12:29:14 am »
Some more pics from input & output for same line + close ups of source. Output doesn't really seem worse than input, but input looks bad -- at least, to my untrained eye! (input on ch1, out on ch3)

« Last Edit: September 11, 2019, 12:31:01 am by mayor »
 

Offline Renate

  • Regular Contributor
  • *
  • Posts: 103
  • Country: us
Re: Composite video issue
« Reply #37 on: September 11, 2019, 08:33:17 pm »
All the scope waveforms look halfway decent for that era.
They don't correspond with the smear that you showed on the TV.
You have to either scope it with the TV connected or try out a more serious buffer.
 

Offline nali

  • Regular Contributor
  • *
  • Posts: 237
  • Country: gb
Re: Composite video issue
« Reply #38 on: September 12, 2019, 05:50:48 am »
I'm not so sure,  the screen shot image-data-closeup.png looks like the white levels might be driven into saturation somewhere along the line?

That exponential decay from the white levels is the ghosting effect seen on screen.
 

Online vk6zgo

  • Super Contributor
  • ***
  • Posts: 4946
  • Country: au
Re: Composite video issue
« Reply #39 on: September 12, 2019, 06:12:17 am »
All the scope waveforms look halfway decent for that era.
They don't correspond with the smear that you showed on the TV.
You have to either scope it with the TV connected or try out a more serious buffer.

Coming from a TV Broadcasting background, I would say it looks unacceptable.
Were video signals from Games machines always that bloody awful?
 

Offline maginnovision

  • Super Contributor
  • ***
  • Posts: 1185
  • Country: us
Re: Composite video issue
« Reply #40 on: September 12, 2019, 06:51:09 am »
All the scope waveforms look halfway decent for that era.
They don't correspond with the smear that you showed on the TV.
You have to either scope it with the TV connected or try out a more serious buffer.

Coming from a TV Broadcasting background, I would say it looks unacceptable.
Were video signals from Games machines always that bloody awful?

That era? Absolutely. It's ok because the screens were also garbage.
 

Online vk6zgo

  • Super Contributor
  • ***
  • Posts: 4946
  • Country: au
Re: Composite video issue
« Reply #41 on: September 12, 2019, 10:16:50 am »


All the scope waveforms look halfway decent for that era.
They don't correspond with the smear that you showed on the TV.
You have to either scope it with the TV connected or try out a more serious buffer.

Coming from a TV Broadcasting background, I would say it looks unacceptable.
Were video signals from Games machines always that bloody awful?

That era? Absolutely. It's ok because the screens were also garbage.

Try to tell that to the licencing Authority back in the day-------they could pull your station licence if you consistently did not make specs.
Somehow I don't think "The screens are crap, so we can transmit non standard signals" would have flown as an answer to ACMA(or their predecessor), or the FCC!

Normal TV CRTs were designed to display proper analog video signals, not video games, & were very good in their intended usage.

One of my activities in a past life was fixing, replacing CRTs & calibrating Studio picture monitors to specs.
The Studio units were capable of displaying the whole resolution capacity of the PAL system, when fed with Composite video.
They could do even better fed with RGB.

Even domestic TVs were not much worse.

The pix in the original posting is really pretty dire.
I seem to recall the Commodore 64 doing considerably better on a little 13" Sanyo ------RF modulator & all!

I wish I still had the C64, so I could look at its video signal with a 'scope.
« Last Edit: September 12, 2019, 10:19:55 am by vk6zgo »
 

Offline mikerj

  • Super Contributor
  • ***
  • Posts: 2151
  • Country: gb
Re: Composite video issue
« Reply #42 on: September 12, 2019, 11:26:29 am »
I'm not so sure,  the screen shot image-data-closeup.png looks like the white levels might be driven into saturation somewhere along the line?

That exponential decay from the white levels is the ghosting effect seen on screen.

I'd agree with this, it's taking around 3us to fully return to black level which is an awful long time when a whole line is only ~64us.  OP: is that scope trace taken with or without your buffer connected?  If with, does it look the same when it's disconnected?
 

Offline StillTrying

  • Super Contributor
  • ***
  • Posts: 2221
  • Country: se
  • Country: Broken Britain
Re: Composite video issue
« Reply #43 on: September 12, 2019, 11:43:32 am »
I'm not so sure,  the screen shot image-data-closeup.png looks like the white levels might be driven into saturation somewhere along the line?

The rigol draws straight lines when the signal reaches the top or bottom of the screen, those completely flat tops will be due to the waveform hitting the top of the scope's screen and then being brought down 1/3 div. after being STOPed.
CML+  That took much longer than I thought it would.
 

Offline mayor

  • Regular Contributor
  • *
  • Posts: 142
  • Country: ca
Re: Composite video issue
« Reply #44 on: September 13, 2019, 09:27:41 am »
I'd agree with this, it's taking around 3us to fully return to black level which is an awful long time when a whole line is only ~64us.  OP: is that scope trace taken with or without your buffer connected?  If with, does it look the same when it's disconnected?

Hey mikerj, I'll do a comparison, but iirc, those were without the buffer.
 

Online vk6zgo

  • Super Contributor
  • ***
  • Posts: 4946
  • Country: au
Re: Composite video issue
« Reply #45 on: September 13, 2019, 03:12:15 pm »
I'm not so sure,  the screen shot image-data-closeup.png looks like the white levels might be driven into saturation somewhere along the line?

That exponential decay from the white levels is the ghosting effect seen on screen.

I'd agree with this, it's taking around 3us to fully return to black level which is an awful long time when a whole line is only ~64us.  OP: is that scope trace taken with or without your buffer connected?  If with, does it look the same when it's disconnected?

The OP has "zoomed in", in the process changing the time/div to 1us/div, so that return to black level is very much less than 3us.

This is obvious if you compare the previous pictures which are 10us/ div.
The H.A.D. of a horizontal sync pulse is around 5us, so you have a known reference to work from.
 

Offline mayor

  • Regular Contributor
  • *
  • Posts: 142
  • Country: ca
Re: Composite video issue
« Reply #46 on: September 14, 2019, 07:36:59 pm »
Hi,

finally tried lowering the DC bias. I am sure it could be further improved, but I'm happy with it. The actual image is better than the attached picture shows, the bleeding is subtle, colors are saturated and the background is truly black.

Thanks to all, this has been fun and educational!
 

Offline mayor

  • Regular Contributor
  • *
  • Posts: 142
  • Country: ca
Re: Composite video issue
« Reply #47 on: September 15, 2019, 10:55:53 am »
Hi, one thing I want to go back to. The clock on the video chip looked jittery to me. Did you consider it to be normal?

 

Offline StillTrying

  • Super Contributor
  • ***
  • Posts: 2221
  • Country: se
  • Country: Broken Britain
Re: Composite video issue
« Reply #48 on: September 15, 2019, 11:42:47 am »
"The clock on the video chip looked jittery to me."

Yes, maybe the 3.5795Mhz is getting mixed in with it somehow, I'd try to see if there are only 3 traces of jitter.

"Did you consider it to be normal?"

Well I dunno. :)

Years ago when I was using video emitter followers I used 220R for the emitter resistor with 12-15mA flowing though it.
CML+  That took much longer than I thought it would.
 

Offline mayor

  • Regular Contributor
  • *
  • Posts: 142
  • Country: ca
Re: Composite video issue
« Reply #49 on: September 15, 2019, 01:57:43 pm »
"The clock on the video chip looked jittery to me."

Yes, maybe the 3.5795Mhz is getting mixed in with it somehow, I'd try to see if there are only 3 traces of jitter.


Interesting, not too sure how to go about that?
 


Share me

Digg  Facebook  SlashDot  Delicious  Technorati  Twitter  Google  Yahoo
Smf