AES 3-3 RGB bypass video not perfect

Pahnda

n00b
Joined
Feb 14, 2020
Posts
8
Hi,

I recently got an AES 3-3.

The last days I did the RGB bypass mod following the guide on retroRGB (https://www.retrorgb.com/aes-rgb-bypass.html)

https://imgur.com/a/Qa2avf5
Following the link, you find the pictures of the result. The 1. picture is from the AES with my OSSC on my LG Oled. 3. picture is on my BVM 20F1.

The picture is quite fine, but I notice this "bleeding" or shifting of the picture. On the OSSC this effect is quite noticable.
I thought this might be the OSSC having problems with the signal, but when connecting the AES to the BVM I also saw it.

Playing with diffrent settings in the OSSC didn't help. The pictures on imgur show the best results i could get.

I'm using a self made scart-cable which uses the csync on pin 7 of the AES.
I have also a scart-cable which uses the composite/Luma signal on pin 3. Using composite/Luma for sync helps a bit, but the effect is still there.

The cap for the sync was replaced with a 470uf cap on the AES 3-3. Else the sync route on the AES board is untouched/stock.

As I'm not an expert on video signals, I don't know how to fix this problem or where it comes from. Is it a problem with the sync signal?
But why is composite/luma sync also showing the same symptoms?

I'm thinking about getting an oscilloscope to analyse the signal more and see if they are in spec or if there is some interference.

But first I wanted to ask if someone here knows a solution for this.

Thank you in advance.

PS.: sorry for the link to imgur, I can't figure out how to upload pictures to the forum
 

Neodogg

Dogg-Father,
20 Year Member
Joined
Nov 27, 2002
Posts
5,588
you're no Panda! I know Panda and you are not him...fake panda! Unless you're the German Panda? Do you have a baker trapped in your basement?
 

Pahnda

n00b
Joined
Feb 14, 2020
Posts
8
WTF? :D I have no idea what you're talking about ...

EDIT: Anyway ... would be great if someone could help me with my problem, or give me a hint where to look at.
 
Last edited:

DaisyAge

Galford's Armourer
Joined
Jun 10, 2018
Posts
457
It could be luma bloom but I'm thinking it isn't. Also, if you did the RGB mod then your composite signal should be cut and you shouldn't be using your 2nd scart cable with composite sync. Can you test other boards or have any pictures of other test screens? your BVM might need a convergence tuning. All else considered you should look for any other threads relating to bvm/pvms and scart cables. You may need a different value capacitor or there may be a way to fix it in the service menu. I had a similar issue with a console using svideo/composite but the source of the problem was a switch box I was using. It had two outputs so I cut traces to one of them and it solved the problem. Now that I think of it, do your cables have proper termination impedance? Should be 75 ohm for a bvm but check anyway.
 

Pahnda

n00b
Joined
Feb 14, 2020
Posts
8
It could be luma bloom but I'm thinking it isn't. Also, if you did the RGB mod then your composite signal should be cut and you shouldn't be using your 2nd scart cable with composite sync. Can you test other boards or have any pictures of other test screens? your BVM might need a convergence tuning. All else considered you should look for any other threads relating to bvm/pvms and scart cables. You may need a different value capacitor or there may be a way to fix it in the service menu. I had a similar issue with a console using svideo/composite but the source of the problem was a switch box I was using. It had two outputs so I cut traces to one of them and it solved the problem. Now that I think of it, do your cables have proper termination impedance? Should be 75 ohm for a bvm but check anyway.

But luma bloom wouldn't be on pin 7 as there is csync on that pin.
Acording to the guide from Firebrandx you will get Luma instead of composite after this mod on pin 3.
Unfortunately i only have this one board, but i will post some new pictures later this day.
I think the bvm isn't the problem here as other consoles like the snes have no problem at all and convergence is fine.
I justed showed the picture of the bvm because it shows the similar effect as seen with the ossc.
I got 75 ohm termination on the outputs of the bvm, so yes, if you mean that.
Thanks for giving feedback :).
 

DaisyAge

Galford's Armourer
Joined
Jun 10, 2018
Posts
457
Run this by Xian Xi from jamma nation. He's the guru on this shit.
 

Xian Xi

JammaNationX,
15 Year Member
Joined
Dec 1, 2005
Posts
27,748
Why the fuck would they tell you to use pin 11 for sync, put it on 10.

If using an OSSC, I'd run everything through a 4 channel buffer like the THS7374. But ya, don't take sync from pin 11.
 
Last edited:

Pahnda

n00b
Joined
Feb 14, 2020
Posts
8
Are you sure with pin 10? The specsheet for the CXA1145P/M says "C Sync out" is on pin 11, "C sync in" is on pin 10.
https://console5.com/techwiki/images/f/fb/CXA1145PM.pdf
I didn't do the sync part of the guide, as the AES 3-3 has stock traces from pin 11 to the pin 7 of the AV-out (470uf cap and 68 ohm resistor inbetween).
I will take a look at the THS7374. Thank you.
 
Last edited:

Xian Xi

JammaNationX,
15 Year Member
Joined
Dec 1, 2005
Posts
27,748
3-3 also has the worst RGB output of all revisions.
 

Pahnda

n00b
Joined
Feb 14, 2020
Posts
8
3-3 also has the worst RGB output of all revisions.

Yes I know that it has bad RGB output, that‘s why i‘m here and want to improve it. The bypass improved the the picture a lot. However I think it could be better. I will buy an oscilloscope and check the signal and then try a ths7374.
 

Pahnda

n00b
Joined
Feb 14, 2020
Posts
8
So I got my oscilloscope today.
The voltages:

Red: 720 mV
Green: 720 mV
Red: 720 mV

Sync: 380 mV

The following picture shows the red signal (red) and the sync signal (yellow) on a full white 100 IRE picture.

aes3-3.JPG

The two peaks of the sync signal are kinda wierd... also the voltage of the sync signal might be a bit high with 380 mV?

What do you think? Any suggestions?
 

Pahnda

n00b
Joined
Feb 14, 2020
Posts
8
Isn't the LM1881 for composite sync signals?
As I don't have a LM1881, I'm wondering if I should buy a LM1881 or a THS7374.
Which will be better for cleaning up the CSYNC signal?
 

DaisyAge

Galford's Armourer
Joined
Jun 10, 2018
Posts
457
I'd do as Xian Xi said and go with pin 10 and the THS7374, he knows best
 

Pahnda

n00b
Joined
Feb 14, 2020
Posts
8
Ok so if you use pin 10 you get TTL csync. Today I tried pin 10 with a 470 ohm resistor and a 470 uF cap to to bring the signal down to about 300 mV.
It works, but the problem still persist. The sync signal is a bit cleaner now (the two spikes before and after the signal are gone), so I will leave it like that.

I will see if i try that THS7374...

Could it be, that the problem comes from the RGB lines and not the sync?
 
Last edited:

DaisyAge

Galford's Armourer
Joined
Jun 10, 2018
Posts
457
I'm not totally sure. My guess would tell me no because the RGB lines are just the lines for color. Their position on screen are dependent on sync and the timing circuits involved in that. Keep us updated
 
Top