Denial

Well let us know what you find. In the meantime here's something else to try: a clip-on common-mode choke on the DC cable itself, even better if you can chop up the cable and only choke the 5V and ground lines.

You can buy a variety of clip-on ferrites but there's rarely any specs attached to the cheaper ones but it might be worth a try. I have a few of these clip-ons somewhere in my stuff.

I only recently started to use ferrite clips. I wanted to believe it is not a placebo effect, but I really like the results. I probably have a really cheap kind, but I put them on everything now.

It occurred to me that you can maybe reduce the noise by adjusting the R10 potentiometer indicated above by the circle. Turning it to the left increases the output voltage and could (in theory) reduce noise.
Ideally you want this so that the output becomes more-or-less compliant (i.e. 1V between sync and white.)

The power supply I built for my Vic uses a Mean Well RS-15-5 and doesn't exhibit this fault. (I definitely would notice if it did, as I spend hours staring at the screen looking for artefacts that shouldn't be there for the FPGA work I'm doing). There's jail bars, sure, but not the rolling bar.

interesting. What's the schematic, and what else is plugged to the VIC-20? Monitor, RF? 1541?

There is no schematic to it really - Meanwell switch mode power supply is a single unit and there's also a bog standard 9V AC transformer both wired up to a single fuse and switch to live, the other side is a straight connection to neutral, and earth (ground) goes to both the metal case of the switch mode and to the frame of the transformer. Lots of chunky solder joints where needed, heavily tinned wires tightly clamped into the screw terminals of the Meanwell. And the whole lot wrapped up in a nice plastic project box.

Would a small plastic box with a 7-pin DIN socket on each end and a couple of capacitors and an inductor inside, with values picked to filter out this particular noise, be a useful "product" do you think? If we know the frequency, we can build a suitable filter to squash it. The problem is, of course, that I don't know what frequency it is at because I don't have this problem. Next time I'm near my VIC, I'll disconnect the ground from the Meanwell and see if it starts rolling noise bars down my screen. Note to self: slap a bit of tape around the exposed end of the ground wire.

Sorry if I sound ignorant (which I probably am) but I understand this PSU has basically two sections with transformers in it. Would it be possible that the noise from the 9V section gets coupled into the 5V section or vice versa?

It could be that noise from the 5V switched mode supply gets into the 9V AC output of the supply - however, the 9V side doesn't go anywhere interesting in the VIC-20, just to the user port and the cassette port (the same is not true of a C64, where it is used to generate a "clean" 5V supply for the VIC II chip.)

The 9V AC side is just a transformer with no active components, so it won't be introducing electrical noise at anything other than 60Hz (50Hz), which wouldn't produce the kind of rolling interference pattern you are seeing.

I finally got around to catching up with my uncle who is a technician. I'm not hugely experienced in electronics.

As you guys know, the PSU's from Electroware.pl on Ebay consist of a standard transformer for the 9v AC and a Meanwell switch mode for the 5v DC.

The Meanwell 5V switch mode was where all the problems lay. We measured the ripple and it was horrendous. I can't remember the measurements, but the saw-tooth on that signal was enough to cut through concrete like butter, and was entirely responsible for the video noise. We did put a 2200uF Electrolytic across the output of the 5v switch mode, and whilst it did reduce the ripple, it did nothing for the video signal.

We then disconnected the 5v wires of the cable from the PSU and connected them to a bench power supply, whilst leaving the 9v connected to the Electroware PSU, and voila! Completely clean video signal.

I can't speak for other Meanwell switch modes, but this one is complete garbage and we were both astonished at how terrible it was. I did speak to the Ebay seller about this, but typically he did not respond. All I can say is avoid this seller and his products.

If anyone has had success in building a 9v AC/5v DC PSU with high quality parts and can point me to an easy to follow step by step guide, I would very grateful. Anyway, thanks for listening guys!

Thanks for this.

I guess there's only one Poland based Commodore PSU supplier on E-Bay, so it's worth spreading the word that these are not suitable for the Vic-20.

Assuming it works okay on the C64?

I purchased two of them a couple of years ago, one for the 128, and used the C64's also on my VIC-20s. I wasn't aware of this ripple, nor did I notice the video noise (not more than with other PSUs). Now I doubt about those products quality, and unfortunately the Polish one is my only PSU for the 128 .

Should I throw them in the bin? What about other PSUs available in EU? Suggestions? TIA

I would have thought those Meanwell thingies would work fine, but apparently not then.

One thing you could do is get a separate 5VDC supply from somewhere and then combine that with the 9VAC from an old VIC-20/C64 PSU. There are Y-cables that let you combine these. Or at least there used to be, I have one of these. But I forgot where I bought it from. But then you would still need a low-ripple/low-noise supply ideally.

You could buy a linear supply but I think these can get quite expensive for some reason.

If you already have a couple of these PSU's and they work fine with the C64 and 128 then I don't think there is too much reason to be concerned. I could be wrong, but I think the C64 has 3 large electrolytic capacitors for filtering the voltage rails versus the VIC-20's one cap which may explain why these issues don't seem to manifest themselves as much on the C64.

Having a look at the PSU, the design is basic with little effort to ensure power filtering and unfortunately we weren't able to see what was inside the 5v switching supply due to it being filled with potting resin. The thing that really annoyed me was the zero support I got from the seller and that's enough for me to never deal with that person again.

Anyway, some suggestions here in this thread about ensuring the 5v supply is separate from the 9v are good ideas and I'm currently investigating putting together my own. It may still involve using a Meanwell 5v supply but one with better filtering. It will take some time, but will post my results.

This product bought through Amazon worked for me without interference - make sure you give it a good ground connection. This is either 110 or 240V compatible: https://www.amazon.com/gp/product/B005T ... UTF8&psc=1

And this is the 9V AC transformer I used *** this is 110V only ***:
https://www.amazon.com/gp/product/B00B8 ... UTF8&psc=1

Here's the box I crammed it all into:
https://www.amazon.com/gp/product/B07TS ... UTF8&psc=1

Fuse holder & fuses:
https://www.amazon.com/gp/product/B00AQ ... UTF8&psc=1
https://www.amazon.com/gp/product/B083R ... UTF8&psc=1

I chopped one of these in half for the cable. I'm not all that confident that the conductors are really man enough to carry 2A though:
https://www.amazon.com/gp/product/B07VT ... UTF8&psc=1

You'll also need an illuminated rocket switch and some mains power cable. And some rubber feet.