A2320 - banding / colour artifacts

[system11]

Has anyone seen this before?

Apologies for the poor capture quality, bunch of scaling going on - but you can clearly see heavy banding in the VGA one, it almost looks like certain shades are missing or weak in the blue/green channels, or too high in the red, in character wide columns.  The banding is visible all the time and especially noticeable on mid greys or scrolling multi coloured graphics.  If you look in the top left you can see red artifacts which should be areas of crosshatch.

I've performed the fine and coarse calibration procedures, I've checked the ground is good.  Unfortunately it's been over 10 years since I used the machine and I don't recall if it had this issue when I got it.

As a separate question, are Amigas known for high RGB output levels?  I've got a slight red tint from RGB on the capture card and a slight green one on the TV - yet they're actually being fed from a splitter and getting an identical signal.  I bought a scart cable on ebay....  I've seen this issue before on arcade PCBs which are known to have levels outside SCART spec.

[magnetic]

Bro you have a lot of signal degrading going on there with all your attachments and splitting...

[system11]

The VGA one is actually directly into the capture card with a 15cm cable.  As a sanity test I plugged it into an actual monitor directly too and it does exactly the same thing with banding and artifacts.

I didn't realise the forum would resize the photos so much!  Here are ones you can actually see:

Monitor port RGB feed (via splitter):



Monitor RGB feed:  

[wmaciv]

Something I just remembered...  If this is in an A2000, make sure VSYNC is based on the power supply reference, not the video section vertical sync pulse.  J200 between the CIA chips, I believe.  It will totally throw off a 2320 if set incorrectly.

[system11]

Yep, checked that - was already timing on the power supply.  Even tried it the other way out of curiosity and it made no difference.  This really reminds me of something I've seen in arcade hardware faults, where a single bit of a colour is stuck somewhere.  I'll check it out with a logic probe at the weekend from the colour hybrid backwards.

[BLTCON0]

@system11
Is the A2320 displaying on an LCD screen?

This kind of banding is typical with old days scandoublers used on modern LCD screens.
As LCDs are inherently digital, they must sample the analogue input signal and typically they will only perfectly recognise some standardised VESA modes.
For "uknown" (yet still displayable) modes the likely mismatch between the engaged sampling period and the actual pixel clock will create a beat-frequency effect manifested as (perfectly periodic) banding.

Toying around with the LCD's  "clock" setting may sometimes yield slightly improved results.