Consumer grade power meters accuracy

[quest for silence]

Hi all.

As far as you know, how much accurate consumer-grade power meters can be, I mean for for standby power measurements?

I know (but I can't recall the source, atm) that very often the claimed precision (1%, 5%, etc.) is referred to a certain power range, while outside of that the mileage may vary, greatly, and that the minimum error can't ever be lower than the absolute value referred to that range.

So, for instance, whether a power meter is "calibrated" (inaccurate terminology, sorry) to do measures for a 500W load within a 1% precision range, the minimum error can't be lower than 5W, whichever the device power draw will be.

Given that, first of all is what I recall true? Whether it were true, what about home made standby power measurements on contemporary rigs? I mean those dual core Haswells which show an apparent idle power draw of about 10-15W at the wall (with 1 SSD, keyboard, mouse and network on, but no activity).
Are those figures to be rejected as actually unreliable, or not?

Thanks in advance for any insight.

[Philipus II]

I really need to start publishing in English: https://www.computerbase.de/2014-05/...berblick-test/

There are some decent cheap powermeters and some really crappy ones. The price doesn't exactly help.

[quest for silence]

Originally Posted by Philipus II

I really need to start publishing in English

Is that sarcasm?

Originally Posted by Philipus II

There are some decent cheap powermeters and some really crappy ones. The price doesn't exactly help.

Unfortunately Google Translator is really bad at german-to-english.

So, summarizing, are all my expressed doubts groundless, and providing that a specific meter works correctly (something like the PM231E or the Volt Craft 4000), even just a single measure will be quite accurate (I mean a 13W reading can't be actually 8W as well as 18W), whichever the load figure is (I mean measuring 4W, 40W or 400W doesn't matter)?

[jonnyGURU]

Brand Electronics is the best for the money.

[Philipus II]

Originally Posted by quest for silence

Is that sarcasm?

No, actually I'd like to do that. We'll see.

[CM Phaedrus]

Brand is good, but keep in mind it's only really accurate when measuring values of the same order of magnitude as the point it was calibrated at.

So Brand usually calibrates them (using a big electric heater) at I think 800-1300W, you have a bit of choice in the matter. However if you calibrate it at that point and you try to measure something at 50W you're going to get results off by several percent. Without knowing this, I calibrated ours to an 81W load and I get excellent accuracy between 25W and 250W. But above 250W results diverge more and more from the Extech and Yokogawa meters, as well as from expected.

I spoke to Brand's owner/engineer, and he says that he built a semi-custom improved version with multi-point calibration and refined circuitry that gets much better accuracy. I think he offered it for around $400. If you're interested in that you should call in person, he doesn't list it on the site.

[quest for silence]

Originally Posted by CM Phaedrus

Brand is good, but keep in mind it's only really accurate when measuring values of the same order of magnitude as the point it was calibrated at.

So how the tiny equipment tested by Philipus II can be really accurate at low load, given that he used a 500W load from a Chroma (if I'm not wrong)? Is the "calibration procedure" you're talking about anything different?

[c_hegge]

Originally Posted by CM Phaedrus

I spoke to Brand's owner/engineer, and he says that he built a semi-custom improved version with multi-point calibration and refined circuitry that gets much better accuracy. I think he offered it for around $400. If you're interested in that you should call in person, he doesn't list it on the site.

$400? I think I'd sooner try this one - http://www.ebay.com.au/itm/RF9901-DI...item2a09d1238e

The meter Wolf uses is only about $200pp from that seller. Not that it exactly offers $2000 Tektronics or Yokogawa type accuracy either, but it doesn't seem to have that issue with accuracy deteriorating outside certain load ranges.

[Oklahoma Wolf]

Originally Posted by c_hegge

it doesn't seem to have that issue with accuracy deteriorating outside certain load ranges.

I've just been impressed with how infrequently units are able to fool it. Only the Leadex platform has done it, and only at really low loads with the main power online.

The Brand and KaW were constantly fooled.

That's actually the seller I got mine from. Had to ask for the 120V version, and it was shipped direct from China.

[c_hegge]

My Cabac power meter gets fooled too sometimes. I opted not to publish a review on a low end FSP unit recently just because I didn't believe its readings. The power supply (an FSP350-60HCC) is not 80plus certified at all, but I was getting efficiency numbers of around 87% at 50% load - and that meter cost me about $150 (and is supposedly accurate to within 1%).

I've been thinking about getting the Everfine 9901 (which I suspect is very similar to the Rek 9901, albeit without the universal power socket on it). The trouble is, though, that when I still have to purchase nearly every PSU I review and HWI still costs me money to run, I've usually got better things to spend another $150 on.