VGA Phase acting strange

[MrClump]

Is it normal for the phase on an A1200's VGA output to drift?

My 1200 is connected via an OSSC, and looks amazing. However whenever it's started up cold the phase is always out, you get the characteristic wavy lines with the chequerboard wallpaper. After the Amiga has been on for 15 minutes or so and has warmed up the phase is spot on as set on the OSSC.

I first noticed this as I was tweaking the phase, and then having to set it back again after a while as it drifted out again!

The Amiga was recapped about 9 months ago by Retropassion so no concerns there, and I am using it with a brand new Electroware PSU.

Can anyone offer any suggestions? Perhaps an issue with the video DAC?

[mark_k]

Maybe the Amiga's clock crystal frequency drifts as temperature changes?

[MrClump]

Originally Posted by mark_k

Maybe the Amiga's clock crystal frequency drifts as temperature changes?

Good point, I have a scope I can hook it up to.

[Stedy]

Hi,

It's likely caused bye either the Amiga 28.37516MHz oscillator drifting, as mentioned by mark_k or the OSSC master 27 MHz oscillator drifting.
Neither the Amiga or OSSC appear to have tight tolerance oscillators, they are both listed as 100ppm stability. For broadcast quality video you need <10ppm drift, for most video designs, never use a part with >30ppm tolerance for consistent results.

A simple test, power up the OSSC, leave it on for 15 minutes, then turn on the Amiga, does it look Ok straight away?
If it does, its likely due to the oscillator on the OSSC, if not it's the Amiga oscillator adding to it.

[MrClump]

Originally Posted by Stedy

Hi,

It's likely caused bye either the Amiga 28.37516MHz oscillator drifting, as mentioned by mark_k or the OSSC master 27 MHz oscillator drifting.
Neither the Amiga or OSSC appear to have tight tolerance oscillators, they are both listed as 100ppm stability. For broadcast quality video you need <10ppm drift, for most video designs, never use a part with >30ppm tolerance for consistent results.

A simple test, power up the OSSC, leave it on for 15 minutes, then turn on the Amiga, does it look Ok straight away?
If it does, its likely due to the oscillator on the OSSC, if not it's the Amiga oscillator adding to it.

I never mentioned I’d tried that. It’s definitely the miggy. I left the OSSC on all day yesterday, dodgy phase as soon as I turned the Amiga on. Left the Amiga on all day today with the OSSC off and it was perfect as soon as I switched the OSSC on.

The problem is the OSSC puts out such a bleedin great image all these little flaws suddenly become very obvious! I think I’ll put the scope on the craystal and see what it looks like cold vs warm.

[MrClump]

I've now completed some more investigation, though am not any closer to finding the cause.

It's not my OSSC or anything to do with my cabling. I have another Amiga 1200 that I've just finished refurbishing and that has a pin sharp image immediately after every device in the chain is powered up from cold.

With my scope I've checked the clock signal on pin7 of U30 and on both Amigas it reads a steady 28.37Mhz, both immediately after power-up and after leaving the computer on for half an hour. Now I'm not sure whether that reading is accurate enough, I can only get the frequency to two decimal places on my scope (I think!). It's just occurred to me that while I had the Amiga open and the scope out I should have probably checked the 5v supply to U30 to see if there was any ripple after power on.

I also tried using a different PSU, just in case my (brand new) PSU was dumping a load of noise into the Amiga when cold.

It's seems to be something to do with heat, seeing as the issue vanishes after the Amiga has been powered up for 20-30 minutes.

So I am now wondering either caps or U30 may have an issue? As I said it was recapped by Steve at Retropassion about 9 months ago, so I would be surprised if any were at fault, but I guess it's possible.

[MrClump]

Rule number 1, never forget the most likely cause of a sudden issue is your endless tinkering. :-)

Problem solved! I had obviously been messing with the settings on my OSSC, because switching Allow upsample2x on (or rather back on!) completely cures the problem! I remember messing with it wondering what it did, must have left it set to off and then didn't notice an issue with my display until sometime later. Enough time to forget I was tinkering. Again.

[Daedalus]

Good that you've sorted it It might just be that you're working around the issue (which is probably a perfectly good solution), but a little note on what was already said above: the precision of the oscillators that Stedy quoted of 100ppm is quite loose as crystal tolerances go, but the ppm unit stands for Parts Per Million. A ~28MHz crystal that's 100ppm out will have an error of around 2.8kHz, or 0.0028MHz. If you can only measure to 2 decimal places / 4 significant figures on your 'scope, it's not enough to detect that error, let alone a 30ppm one.

[MrClump]

I’m going to be stripping it down very soon to transplant it into its new A1200.net case, so I might as well swap out the 28Mhz Crystal while I’m at it. I’ve ordered one on eBay, I’ll post my findings when I’ve done it.

[MrClump]

The issue remains, despite me thinking I had solved it with the change of setting on my OSSC (it just improves it slightly).

I've now swapped out the crystal, no change so it's not that.

I had the 1200 stripped down today, so did some experimentation. It seems the cause may be Alice and is related to the temperature. From a cold switch on if I warm up Alice with some hot air from my desoldering station the issue goes. Then cool Alice down with some freeze spray the issue return.

Seeing as the issue goes away once the computer is warmed up I'm not inclined to rip my Alice chip off and replace it. Not really sure why Alice would/could cause this? At first I thought it would most likely be the video DAC, but no matter whether the DAC is cold or warm it's the same.