VOGONS


Reply 20 of 34, by relo999

User metadata
Rank Newbie
Rank
Newbie

Made one with all separate ground lines, but no change. Maybe there is some way to filter it on the PSU side?

Have you read the comment above your last Deunan?

Reply 21 of 34, by Deunan

User metadata
Rank l33t
Rank
l33t

Yes, I've read it, but you said connecting the ground to your socket (that does not have true earth wire) didn't do much. So I don't think it can be easily solved with the PSU setup you have.

There is one more thing to try maybe. Is the power supply that's conncted to the mains using 2-prong plug? If so, have you tried rotating the plug 180 degrees? To swap live and neutral wires that go to the PSU. There is usually a capacitor between primary and secondary side in SMPS, to reduce noise on the secondary. But with just 2-prong plug the PSU can't know which wire is N and which is L. So the cap is randomly connected to either N or L, depending how you rotate the plug. And the optimal connection is to N rather than L, which is more noisy.

Rotating the plug alone might not help, you might need to also have the extra grounding wire connected. But if it is connected, and rotating the plug does nothing, then I would say you need a true earth wire in the socket. A different PSU might help but you will have to experiment in that case, sometimes even decent, efficient PSUs can be quite noisy.

I might have another idea to try - I assume you have a 12V PSU connected to wall socket, and then Pico PSU adapted to Towns that generates all the other voltages? The problem might be the first PSU, the Pico, or even both. If the above doesn't help then try connecting the socket "earth" wire not to the screw terminal on the back of the unit, but rather to the GND wire from the 12V PSU output. In theory it should be the same thing but as you saw with the ripple between video connector shield and GND pins, for AC noise the routing of the power planes matters a lot too. So maybe connecting the earth between the PSUs will have better results if the noise is mostly coming from the first one.

Reply 22 of 34, by relo999

User metadata
Rank Newbie
Rank
Newbie

Yea, it's an F-plug. I did turn the plug and the first time it worked, though I should note I had it fully unplugged on both ends of the 12V supply so it also could be a "solve" by simply waiting some unknown time instead of the twist. As after power cycling it again and trying it again I could flip it either way and the bars were back. I have it fully unplugged now so I can test tomorrow.

I have a different Pico and 12V supply I have for my X68k I could try. And I'll try the earthing to the ground wire of the PSU tomorrow.

I have the 12V in the wall which goes to the pico.

Reply 23 of 34, by Deunan

User metadata
Rank l33t
Rank
l33t

The problem randomly went away with some time? That's odd, it shouldn't happen like that. It reminds me of something, many years ago when I was still using CRT monitor with my PC, it had similar issue to yours - but not exactly the same. The upper/left corner of the screen would get a bit darker, and "pulsing" fast, even the edge of the screen was moving a little bit. The culprit was electric heater in my kitchen stove. It's a gas stove but uses electric heater for the grill function. I didn't use it much, but every time I did the montior would show this issue. No other load (washing machine, iron, vacuum cleaner, fridge) ever caused me any such problems. So in my case this was a particular heavy power load, in particular location that caused it - I guess it somehow affected the power wires it this area of the house? I mentioned this is an old building, the wires are good copper but who knows how exactly connected. There are no detailed diagrams.

My problem went away after I switched to LCD monitors, so I never did anything about the stove. That's why I said your problem might be difficult to solve - it can be not just the PSU(s) but a combination of several factors.

Reply 24 of 34, by relo999

User metadata
Rank Newbie
Rank
Newbie

No, the problem went away after extended disconnection from the electrical net. Tried it again this morning, first boot had no issue, but turning it off an on again gave me the issue again.

Also tried connecting it to a different part of the house (on a different electrical group) without grounding and it went away and didn't come back after turning it off an on again.

Will try the other PSU and Pico tonight probably. But before that will just remove as much as possible from the electrical group.

Reply 25 of 34, by Deunan

User metadata
Rank l33t
Rank
l33t

How long does it need to be disconnected to work OK? What about the opposite situation, you disconnect it, then connect again, but not power on right away. Wait some time before powering on. Is it good or bad now?

Reply 26 of 34, by relo999

User metadata
Rank Newbie
Rank
Newbie

No clue about the wait time, at least 6 hours probably less. It doesn't get the issue on first boot after being connected for half an hour.

But I found out what causes it. I was disconnecting things to see what's going wrong and followed the power strip I use to access the plug, found out I plugged the strip into another strip (with nothing in it) for no good reason beyond what likely was laziness years ago. Removed that intermediate strip and the issue was significantly reduced. Plugged in the PSU directly into the wall and the issue was mostly gone, only if you're really looking for it on a big single color image it's still visible on only on the NEC.
So yea, power strips are what's causing it it seems and multiple connected together seem to make it worse.

Reply 27 of 34, by Deunan

User metadata
Rank l33t
Rank
l33t

It gets more interesting with each new discovery. Are these power strips plain cable and socket(s) or the "computer" kind? These ones have noise suppression filters in them - some older strips actually had pretty beefy iron chokes for that purpose, the new ones are just a small filter and surge protector. The problem with these is they don't work nearly as well if you have the kind of electric setup where there is no true earth wire.

So on second power-on the problem comes back? What if you turn it off, but do not unplug from socket, and leave it be for few hours? Then when you turn on, is it good or bad? All these timing-related issues suggest the problem is the PSU, the grounding and shorter cables (without filters) help, but in the end the true culprit seems to be the PSU. One of them at least.

Reply 28 of 34, by relo999

User metadata
Rank Newbie
Rank
Newbie

One is a big one with some fuse in it and a power switch with LED, the other is one with a power switch with LED. I'm guessing both are the "computer" kind, seeing as they have LED and I doubt there are LED's that run on 240V AC directly. And I don't hear some kind of relay turning them off.

And I'll tell you when I know later today, though I booted it again with having the 12V PSU disconnected at both end for hours and the bars were back. So full disconnection for extended periods isn't consistent to solve it on first boot.

Reply 29 of 34, by relo999

User metadata
Rank Newbie
Rank
Newbie

Tested letting it be connected after a boot, didn't solve it. Just got bars on the first time booting after 8 hours.

Did try a different power strip without grounding pins on it, worked without issue.

Reply 30 of 34, by relo999

User metadata
Rank Newbie
Rank
Newbie

Sorry for the double posting.

Found the issue. It's the 12V PSU I was using. I tried using the X68k Pico with it's PSU in my Fresh, no issue. I tried using the X68k Pico with the PSU for my Fresh, the bars are back. Only haven't tried using the Fresh's pico with the 12V PSU from my X68k.
This mean I need to find a new PSU that doesn't have this issue. Seeing as the X68k's PSU is only 60W while the Fresh says it wants 90 with peaks of 120w.

Reply 31 of 34, by Deunan

User metadata
Rank l33t
Rank
l33t

I suspected that PSU, I've seen many that cause problems. That's why I stay away from any power supplies with only 2-prong plugs, except for devices that are not connected via cables to other stuff. So it's OK for phone chargers or Raspberry Pis that only use WiFi. But everything else can cause problems, switching mode PSUs are a known problem in measurements/metrology, there are small but noisy currents going through ground leads that affect readings even on 4.5 digit instruments.

What brand was that PSU? If you can't find anything good, consider 12V "brick" PSU from old Xbox. Not the very first ones, you don't need that much power (and the brick is huge) but in general these PSUs, if original, were pretty good quality. The plug is useless but you can cut it and solder your own.

Reply 32 of 34, by relo999

User metadata
Rank Newbie
Rank
Newbie

It's a LEICKE.

You mean the PSU's from the 360?

Reply 33 of 34, by Deunan

User metadata
Rank l33t
Rank
l33t
relo999 wrote on 2025-06-30, 11:14:

You mean the PSU's from the 360?

Yes, but you need to pick carefully. I see a lot of them now come with just 2-prong cable. My old one was 175W and had 3 prongs, in fact the socket had a notch that required a special cable - or some "modding" on typical computer type cable - to fit properly.

Perhaps a lot of these new "original" PSUs are not in fact original at all. So unless you trust the source/seller, maybe it's not such a good idea after all. Maybe a laptop charger then? I know some older IBM laptops had these smaller plugs with 3 prongs, but I'm not sure if any of them was 12V.

Reply 34 of 34, by relo999

User metadata
Rank Newbie
Rank
Newbie

Checked my own stash, I got an extra 360 PSU from a butchered system but it's a 2 pronged one. Though would the 2 pronged one not be better in my case as one with earthing would have possibly the same issues as my LEICKE one?

Edit 2:
For anyone interested, the 175W 360 PSU works great.

Edit:
Also seems the early models are the only ones with earthing.
s-l1600.jpg