I have a non-functioning and disconnected Belkin UPS rated at 550VA (330 watts ?) which is about 10 years old - suspect that an input relay has become stuck in the "Mains only" position . At the time of purchase, it was a good idea because despite being in suburban NW London, I was getting regular large variations in voltage supplied and occasional outages. Since EDF replaced my local electricity substation (70 years old) with brand-spanking new one about 3 years ago, the outages have dried-up and I suspect that the voltages may be a bit more stable.

How can I tell what the mains voltages supplied to the computer PSUs are like at the moment. Does anybody know what's the best modern Windows software that will do the job reliably (I seen some terrible reports about the accuracy of these plug-in mains socket testers) and what are acceptable limits ? - the power supply unit on my main desktop is a Corsair CX750 80 plus bronze with Active Power Correction (750 Watt), so that, apart from the fact that this PSU probably wouldn't benefit from the power conditioning effect of an upstream UPS, the Belkin UPS would be incompatible being underpowered on battery and non-APFC compatible (Simulated Sine wave ?).

However, the Belkin UPS could be used, if repaired, to provide smoothing and battery back-up to an older non-APFC computer PSU (350 Watts) in my Media Centre mini tower.

But, the way I see it, its only worth attempting to repair the Belkin UPS if the current mains voltage is still variable and outside limits and if the repair is cheaper than buying a brand new APFC PSU for the media centre.

I have to say that, at first sight, repairing the Belkin UPS doesn't look promising - they used 6 "One-way" "Z" screws to secure the case (These would have to be drilled out) and, if that's their attitude regarding customer repairs, what chance getting proprietry spares or generic parts fitting for something this old ? - The battery in the Belkin UPS is only 3 years old - replaced by me on the premise that it was that causing the problem. Do Belkin have a "Public facing" repair or parts service ?

Any advice appreciated.

Thanks.

Nick

Edited 24 April, 201313 yr by Clunkclick

Do Belkin have a "Public facing" repair or parts service ?

Probably not. HTH

It'd likely be cheaper to replace it after 10 years. The batteries are likely shot anyway

IMO UPSs are cheaper than losing any sort of equipment, and anything you may have stored on them (photos etc)

just buy a new one, stick with APC, Leibert or a company that does power products as their core product.

Most UPSs will have a USB or Ethernet port for monitoring. USB for home products and small business. Older ones will have serial interfaces.

If it's the relay, it'll probably be available online somwhere (doubt Belkin would use a proprietrey one). But, if it's not the relay, or the battery has went duff, you'd have wasted money and time - I'd reckon a new or eBay sourced one would be better value.

The one-way screws are probably a good idea for their own legal protection, they don't want customers changing anything but the battery. If a typical home-office customer seen ordinary screws, they might assume that customer serviceable parts were inside - only to shockinly find out that UPS' can still bite with the cable out!

I've ordered one of those 3 pin socket plug-in power analysers (<£20) just to see what's happening with the voltage at the moment.

If its stable over a given period, then really, the only other justification for a UPS for a home system (Voltage smoothing and surge/spike suppression) will be gone - I've already got good surge/spike suppressors in the power supply line at present, in the form of trailling sockets and they seem to adequately deal with the worst local offender, the central heating pump (Max 720W draw).

Frankly, even if there is a voltage variation still present, it would appear, from what I read on the net, as if the PSU in my main desktop is better at smoothing than any UPS (Corsair quote a Power Correction Factor of .99, whereas most UPS's can only manage .60). Plus the fact, that the literature tells us that only "Pure Sine wave" output from a UPS is compatible with PSUs with Automatic Power Correction. The Corsair PSU in my desktop has APFC and the Pure Sine Wave output UPS's are very expensive for the level of power output I need i.e 750W, 1300/1350VA - APC want minimum of £350-£430, whilst Cyberpower (Dell) want £225-240. Ouch ! On the otherhand my main desktop principal components ( Main board Asus Sabretooth, processor AMD Bulldozer six core, 8GB Corsair memory and Corsair 250GB SSD) has a shorter life and easily tops the cost of a UPS in depreciated cost per annum terms, so there may be a case for having a UPS.

Having in-line, trailing sockets with surge/spike suppression, followed by a UPS with voltage smoothing and then a PSU with APFC would, IMHO, be over-icing the cake for an urban area supply `with minimal power outages now and improved voltage stabilisation . May be different for country areas, with flown cables and frequent outages and consequential voltage variations arising from weather effects and the actions taken to manage power system outages or in premises which have a lot of "Noisy" appliances.

If there is a slight variation in the incoming mains voltage, then I suspect, that I'll just be able to get away with a new APFC PSU for the Media mini tower and the existing trailling sockets for Surge/spike suppression.

Waiting to see what the plug-in multimeter reveals.

Nick

It's only worth repairing the big UPS units and they'll normally be on maintenance contracts.

+1 for APC they seem to make good kit.

You won't get software for windows to monitor your power quality the power analyser you've got might be better but I've no experience of them.

A UPS is a good investment if your PC is important to you in terms of your work or you've invested heavily in it in other ways. If you're just looking to protect hardware then a good surge protector might do the trick for less money.

Someone correct me if I am wrong here, however, my understanding of power factor correction is that its related to kva on the grid and a high power factor makes kwh = kva. If you are billed in kwh it does not really provide any direct benefit or benefit to the electronics.

I'm not particularly up on the science or applicability of the terminology, apart from what's on wiki and in my battered copy of Morley and Hughes.

As far as I understand the PFC ratio represents the phase lag of current behind voltage in a circuit having resistance in series with a capacitance. Presumably that can apply to other components downstream from a UPS e.g. a computer PSU as well as a power supply grid and, may be of relevance to fine electrical components in terms of shortened life caused by receiving power outwith specifications.

In the kit specifications, most manufacturers state the battery output of their UPS's in the following form e.g. 1000VA/600W, and that the ratio of watts to VA is referred to as the Power Factor Correction. I've also noticed that manufacturers state that the UPS's that are set up to produce either square wave or stepped sine wave output when operating from battery usually have a lower wattage output per unit of VA typically 0.6. Whereas units outputting Pure Sine on battery typically have a ratio of 0.9.

I'm not really appraised what the technical implications of that are or whether they have any significance other than for marketeers trying to distinguish their products, but some of the literature claims that to connect a square wave or stepped sine wave unit to a computer PSU equipped with APFC can damage the computer power supply or, at least degrade its components on an accelerated basis. Leibert, for instance, don't seem to produce any pure sine wave UPSs, whereas APC, Cyberpower do.

Postscript

To the extent that Power factor is a measure of efficiency in usuage of electricity, the following reference explains why this may be relevant to a UPS:-

http://www.kwsaving.co.uk/Business/pfc/pfc-simple.htm

Presumably low PFC UPS will deplete the battery quicker ?

Nick

Edited 25 April, 201313 yr by Clunkclick

Plug-in multimeter arrived.

Tested the MM against a rated load and calibration was spot-on.

This morning, with no major loads operating in the house, plugged-into a wall socket with light loads, it shows a consistent slight over voltage at the wall sockets (236-240v), frequency spot-on @50cps and power-factor varying between .70 and 0.80 depending on load.

These figures stay the same when the tester was plugged into the line supply of the Media centre (350w non-APFC PSU) and the latter booted-up. At start-up wattage peaks @ 130w and then settles down to 70-80w when the system is fully-loaded and at idle. Playing a 480p You tube video sends it up to 90-100. Very strange, all other settings remaining equal, there was a consistent 10w difference in consumption between different 480p videos ??? Presume the playing of one required more CPU processing than another (Processor is AMD Athlon II 630 x 4, 2.8Ghz, rated @ 95W + Corsair SSD, 10W ? on an Asus M4A88TD-M Evo Mboard).

So, at first sight, it all looks OK. Supply voltage is within post 2008 UK (Euro harmonisation limits) i.e. 230v +10% / -6% and wattage seems to be within what would normally be expected from this device .

However, the PFC is 20-30% off of unity.

Does anybody know whether this is within limits for a UK domestic supply ?

Nick

Edited 26 April, 201313 yr by Clunkclick

More extensive testing throughout the house has shown some highly variable results which seem to differ according to device tested:-

A 0.9-1.8K Delonghi ceramic heater showed a PCF of .99.

A Chinese 10 w Ozone generator showed a PCF of .40.

I'm confused now. If PCF is an objective measure of the effective power delivered by the grid, why the difference between devices ?

Nick

As I said my understanding is power factor is a factor of the load not the grid.

We install power factor correction on all the sites we do (water treatment) as electric motors are used on the pumps which generally have a poor power factor. The client get charged extra for a poor power factor as the grid has to transfer more power, domestic customers don't incur the extra charge.

We did a site with extensive ground dewstering and tunneling. Our power factor was 0.65.

As I said my understanding is power factor is a factor of the load not the grid.

That is my understanding too.

Plus a lot of modern switch mode power supplies may cause cheapo current monitoring devices to return inaccurate figures.

Understood.

I had read various articles on the web that sort of implied that bad PCF could be "Imported" from nearby equipment on the same circuit (Distortion of the sine wave on the grid by other heavy loads ?), especially in industrial situations. It just so happens that I am 250 yards from my High street where there are the usual selection of shops (Including a small Sainsbury), offices and restuarants, some presumably with three-phase supplies and they may share the same supply circuit as me, 'cause when my power supply has been down, they are also down.

It will be interesting to see the results for the main desktop when I get that up and running again - that's got a two-year old Corsair 750w power supply with active PCF in it. So I would expect the PCF reading to be better on that.

Plug-in multimeter specification states limits accuracy of measurement on AC +/- 0.2% ! Whereas my 30 year old Fluke multimeter has a stated accuracy of +/- 2.5% on AC ! - However, I did notice a slight difference between the value I calculated for PCF and that displayed on the plug-in meter whilst testing the older 350 watt supply, so it may be worth double checking.

I'm coming round to the idea that there's probably no need, as long as I stay in my current location, to provide for outages as they are not occurring now and that perhaps, the only action required is replacement of the existing 12 year old 350watt power supply in the media centre with an active PFC (assumption being that the plug-in multimeter gave a reasonably accurate reading on this -I can always test again with the Fluke).

Nick

Edited 28 April, 201313 yr by Clunkclick