Windows is not in the habit of changing video modes on the fly unless switching full screen mode on a games etc. Is there any unusual software installed on them?
Do the two machines you tried differ much? i.e. same type of video card etc?
What about power supply? A reasonable spec PSU will ride out a couple of hundred ms of power glith, but beyond that its possible you could see a power interruption cause a flicker / momentary loss of sync. Have you got a UPS or at least a filtered lead you could try them on?
First I should say that I believe that there are - or were - two or more faults. I'm detecting features of the problem that I'm sure I'd have detected earlier had they been the same all along.
I'm testing this evening with the set-up as it's been for some time: PC, short VGA lead, VGA-RJ45 adapter, short patch lead, wall plate, about
10m of fixed wiring, wall plate, short patch lead, RJ45-VGA adapter, short VGA lead, monitor.
The signal interruption occurs on a regular schedule. Every ten minutes and two seconds, the signal will (probably) disappear momentarily, resulting in the monitor blanking the screen for several seconds.
If I switch the PC off/on, or adjust its clock, the schedule is unaffected.
19:24:20
19:34:22
19:44... no interruption
19:54:26
20:04:28 (switched PC off/on)
20:14:24
20:24... forgot to check
20:34:28 (PC clock +1 minute)
20:45:28 (PC clock corrected)
20:54:27 ...
22:14:45
I've tried exiting all my installed applications and stopping all the installed services: no difference.
I've looked at the Intel software for the video card with a view to disabling anything that looked suspicious. So I've disabled 'image enhancement auto adjust' and selected power plan 'maximum battery life' which is said to give 'minimum visual impact'. I've disabled 'optimum resolution notification'.
That's it for tonight. Tomorrow I'll try disconnecting the PC from the network and then stopping all the services (binary-chop-wise) to narrow the problem down further. Then there's the chipset setup to investigate.
Regarding the two PCs: they're very different. I'm not pursuing that angle while software seems to be the likely culprit.
ok, try booting the machine and entering the BIOS setup page and leaving it there. I.e. before it loads windows or anything. Do you still get the glitch. (if you do, chances are we are looking for a hardware problem). If not, then its not conclusive since we are in a lower scan rate text mode.
Next try in windows "safe mode" (F8 at startup, select from menu). Or you could even try a linux "live" CD or Bart PE boot disk.
Don't suppose you can get your hands on a digital storage scope by any chance? I would be inclined to capture the video and sync waveforms around the expected glitch time, and see what sort of disturbance we are actually looking at.
There are six drop-out times per hour, at approximately nn:04:30, nn:14:30, nn:24:30 etc, sometimes up to about a minute later. Drop-outs occasionally occur at other times too.
With the monitor directly connected to the PC, the probability of a drop-out is small. But if the monitor is connected via CAT5 with adapters, the probability is high. I suspect, but don't have the time or motivation to find out, that the probability increases with the length of the link.
So the CAT5 and its adapters are not the cause of the problem, but they do magnify it considerably.
The problem remains if the PC and the monitor (with its CAT5 link) are moved to another room in the house, with nothing else connected.
I've eliminated all plausible software causes, and I'm drawn to the conclusion that's it's externally-generated interference that's causing this.
One interesting aspect of the setup is the earthing. Neither the PC nor the monitor are themselves earthed. The PC is powered by a laptop-style brick and the monitor is powered by a wall wart. Normally (but often not during my testing this week) the PC has a network cable and a USB DAC plugged into it (sorry about all the TLAs). No monitor, keyboard, or mouse: it's a media server. The network cable doesn't provide any earth AFAIK. Which leaves the DAC, which receives mains through a three-pin IEC connector.
Usually the DAC had three connections: mains, USB to the PC, and twin photo leads to the amplifiers. But during testing I plugged the DAC into the USB port on the PC *without* any connection to the amplifiers, and the drop-outs came fast and furious, even with the DAC switched off at the wall mains socket. Obviously there's more investigation to be done, but I've had (more than) enough for this week.
Did you verify the problem remains even if looking at the BIOS screen or windows safe mode?
If so, then yup, that only leaves hardware.
Not close to an airport are you?
(then again, radar sweeps tend to be far more frequent).
Have you got a UPS? You could try the whole kit running from battery for
15 mins.
I suppose the next trick is to attempt to work out the vector for the interference - i.e. is it RF, or mains carried. Powering kit from a filtered mains lead may help - even just a metal oxide varistor style peak suppressor. Perhaps looping the supply cables through ferrite rings would be worth trying. Failing that, sticking some kit in a metal box might be worth trying.
I've been doing that for literally years (on and off). I blitz the problem for a week or so, always thinking that the next test will identify the problem, but it never does, so I give up until the next time it really starts bugging me.
No. There's a British Aerospace factory a few miles away. There's a phone-type mast (actually for emergency services, allegedly) about 300 metres away.
Yes I have got a UPS, and it would power the kit for rather more than 15 minutes. So I will try that.
But often the problem ducks out of sight for an hour or two. It's the unpredictability of it that makes it such a bugger to test. The ten minute intervals have made things a bit easier, but it doesn't always happen on schedule and sometimes it happens off-schedule.
Before putting everything back where it was (in the interests of domestic harmony over the weekend) I advanced the PC clock by five minutes. I was interested to see whether the glitch time stayed the same or shifted by five minutes. It shifted by two and a half minutes. Bloody typical. And for the last couple of hours it's not happened at all.
At that range you can forget interference to this sort of system. If it were TETRA it might interfere with the TV if that was in the lower channel numbers but not this..
The problem remained when looking at the BIOS setup screen. Which left me thinking it *had* to be interference.
Moving the UPS is a big job so before doing that I switched off some equipment which I thought might be the cause of carefully timed interference - two desktop PCs and a pair of Homeplugs (the Homeplugs were clear favourites). Problem remained.
So I heaved the UPS in place and bingo! It solved the problem (apparently: it's all very probabilistic), even when still on mains power. So it seems the filtering effect of the UPS does the trick.
I've since narrowed it down to the VGA/UTP transmitter that's sensitive.
Whew! Thanks to everyone who's supported me in this unrewarding task.
Now, rather than trying to track down the source of the interference (which might be something unfixable like the smart meter or the intruder alarm monitoring) I'll go down the filtering route.
I've no experience of mains filters. Any tips?
Oh, and there are still a couple more jobs to do. I need to buy a new UTP/VGA receiver, to replace the one I accidentally fried by sticking the monitor's 12V plug into its 5V socket (perfect fit, physically). And I need to buy a new NAS drive to replace the one that didn't survive being powered off/on. :-(
What sort of UPS is it? (i.e. backup, online, or line interactive)
Backup units filter the mains and run the load directly from that when available. Then switch in the inverter quick enough on a power fail that stuff does not notice.
Line interactive (some of the APC range), are like the above, but can also do more conditioning to boost and trim in the incoming mains.
Online ones are pricey and fairly rare, they run the load from the inverter all the time and hence show no interruption on switching to battery.
Ferrite on its power lead? Tinfoil hat for it?
You can get extension leads that are supposed to filter - however these are of variable quality. You could make you own with an off the shelf mains input filter and a 250V transient peak suppressor.
It's an APC, don't know what class exactly. The conditioning is "Full time multi-pole noise filtering : 0.7% IEEE surge let-through : zero clamping response time : meets UL 1449". I've no idea what that means but it sounds OK and certainly did the trick. I'm pretty sure the inverter isn't running full-time because I'd hear it.
Speaking of APC, that's one company that will never earn another penny from me. Their software knackered my browser by installing a JavaScript engine whose security certificate expired some years later. It took me nearly a week of work to discover that, reinstalling Windows and my applications one by one until I found out what was doing it. People all round the world had the same problem and were up in arms about it, but were APC apologetic? Fat chance.
Doesn't sound to me like external radio based interference more likely to be a fault somewhere with the hardware. You could use a simple MW radio nearby and see if there are any large disturbances in the RF background not that precise but better then nothing...
DEC used to supply them as standard with their StorageWorks SCSI arrays, never had a problem those, but not used them lately.
Didn't Eaton hoover-up the cheapo MGE kit? I've seen heaps of those fail in bad ways (a quick look at their web page seems likely the "Ellipse" range), I think the good MGE kit went to Schneider/APCC?
I had a couple of versions of APC software and both were a pita. For years now, I've just used the inbuilt WinXP UPS monitoring which works well enough without the bells and whistles.
HomeOwnersHub website is not affiliated with any of the manufacturers or service providers discussed here.
All logos and trade names are the property of their respective owners.