Notifications
Clear all
Topic starter
2019-06-21 12:52 pm
With thousands of cheap multimeters leaving China every day, how would I determine if the one I just bought is producing reliable results?
I could send it to a lab for testing, but that would probably cost more than I paid for the equipment itself.
Can I test it myself? The answer is YES! And you can make an acceptable evaluation.
But you need 2 multimeters, but if you only have one, perhaps borrow one for testing. You will also need a load of discrete components. Make selection of resistors (10Ω - 1MΩ), transistors, capacitors, diodes, etc.. A known voltage source and constant current source is also required. Put all these components in a box and only use for testing. This way, the actual values of the components are pretty much irrelevant.
The principle!
A long time ago I went to witness pressure testing on some parts for a ship. The first thing I asked was whether the manometers had valid documented calibrations, something that in the oil and gas industry is compulsory. They said they didn't but their test procedure called for two manometers for testing. The principle is that if both manometers show the same value then they are showing the correct readings because no two manometers will have exactly the same error.
Test all the discrete components mentioned above with two meters and tabularise the results. This way you can see immediately if there are any discrepancies. This is where a third multimeter comes in to play to resolve discrepancies.
So if you buy another meter, test it with the same components, add the results to your table, so you will always know which meter produces reliable results, and which produces unreliable results for each type of component.
I already own three meters of varying cost and am now looking to buy with automatic ranges, and this is how I test them before use.
Steve
2019-06-23 2:57 am
I was taught the same process, especially for home consumer and hobby use. I am fortunate enough to have a few digital meters, and have done this myself, producing very acceptable results and confirming their accuracy.
Another thing I do is to use the same set of test leads, so the only variation is the multimeters, not the leads.
Thanks for sharing this tip!
2019-06-23 5:45 am
I have several VOMs. After a career of 35+ years doing HVAC repair my analog Simpson sits unused in the shop along with several Fluke, with and without ammeters.
My favorite meter is the one (of many) that I got free from Harbor Freight Tools. And guess what? When they fail or the battery goes dead, I have 5 more in the cabinet to fall back on.
If you wanna spend a bunch of money on your VOM feel free but in my opinion, your wasting it over a couple of 10th of a volt.
Gee, wondering who's gonna jump in my poop about this, or, waytogo Bob, everyone hates you already.
?
This message was approved by Recycled.Roadkill. May it find you in good health and humor.
2019-06-23 7:30 am
It all depends on how accurate you need your instruments to be.
For the needs of us; hobbyists, makers and tinkerers, even low priced instruments are accurate enough.
It is another story if you need super high resolutions (microVolts, microAmps) or have to measure high energy sources which require high levels of protection to be engineered and built into the instruments.
As for myself I have two multimeters. This is handy for measuring voltage and current at the same time. I also sometimes compare their readings to check that there are no large deviations, which would indicate that there is a problem with one of them.
hj
Topic starter
2019-06-23 9:42 am
I don't really agree with using a dedicated probe set for testing MMs unless you only use one set of probes all the time.
The reason is that most people are using the stuff that came in the box and the probes are part of the whole system, therefore the comparison must be of the whole system.
I have seen the idea of testing MMs with one dedicated probe pair elsewhere on YouTube, with the same given reasons.
Having worked in QA for the last thirty years, at factory acceptance testing, use of foreign components i.e. components that are shop supplied and not being delivered with the equipment is NOT allowed. Exceptions (just as an example) may be, for instance, using shop electric motors with 50Hz on pumps being delivered to areas where 60Hz is used, but then the power rating and speed must be the same.
That is why I don't think it is a good method for product comparison.
Topic starter
2019-06-23 10:11 am
The question of accuracy came up last year when I was trying to build a simple clock and a strobe with a 555 timer. I was using 5V, but not getting the correct frequency according to my RC calculation.
The first thing I discovered was that despite the datasheet value range of 4.5 - 16V recommended supply voltage. The frequency at 5V was less than that at 12V, I was expecting the same frequency across the whole range. I plotted frequency against voltage and discovered that frequency rises from 4.5V up to about 8V where the graph flattens out after that there is no considerable rise in frequency up to 16V.
I set the voltage to 12V, but my RC calculation was still slightly off. I knew the resistor could have +/- 5% error, but that was still not enough to compensate for the deviation I was getting. I then tried 10 capacitors of the same nominal value, only to find a variation of +/- 15% capacitance.
This the point where I started to consider the accuracy of my MM and my mental state. ?
Steve
2019-06-24 3:51 am
It's quite possible that my lifetime lack of need for accuracy is due to my lack of understanding of electronics. But it seems that with the arduino and all the little goodies premounted on tiny circuit boards change everything.
Heck, hopefully I'll learn something here and actually retain some of it.
This message was approved by Recycled.Roadkill. May it find you in good health and humor.
Topic starter
2019-06-24 2:45 pm
To supplement my earlier post about the 555 timer.
Here is a plot of the voltage against frequency that I measured.
The nominal frequency according to the RC*0.693 calculation should have been 218Hz, so I was still getting a 5% error.
If you are not just making LEDs blink, this should be taken into account.
hstaam reacted
2019-06-24 4:15 pm
Thank you for sharing this. It is very interesting and instructive.
hj
hj