YARD SALE: SkyRC Watt Meter

[iloveelectronics]

When I first found these from a supplier in China I thought I struck gold. With its nice compact form factor, ability to measure temperature in addition to all the other measurements and resolution to match the Watt's Up meter ($50+), at only a fraction of the price, I thought there was little risk so I went ahead and order a small lot of them in order to get the wholesale price. How I wish I knew its shortcomings then... Anyway, I don't think they are worth their regular price so here I am "giving them away" in this yard sale!

See Martin's review of the meter here:

If you search on eBay you will find that these normally go for about $26 shipped. And that's the price WITHOUT the temperature probe. I haven't found anyone on eBay selling the temperature probes.

I am offering 3 sets of the following here on this forum to members (I'm offering another 4 sets over at the MJLorton.com forum):

1. New in-the-box SkyRC Watt Meter WM010, WITH the temperature probe. Details at http://www.skyrc.com/index.php?route=product/product&path=24&product_id=46 (please watch Martin's video for its issues). I'm selling each set for $9 shipped worldwide. That's basically just the price of the temperature probe (they might be handy for other uses), shipping, PayPal fees, plus some rounding up.

Please post in this thread to claim. PM me your PayPal and shipping address too so I can send Paypal invoice. First come first serve. 3 sets available here.

Here are some pictures of them. The inside shots are taken with my own unit. I removed the gunk on the watch crystal so I could see the chip's number, it is an ATMEGA32A. WARNING to you if you want to crack open your own: the retention clips are rather fragile. I broke 2 of the 5 clips while I opened mine up!

[TerraHertz]

Have you been able to identify what the board uses for a reference, and AtoD converter?
I agree with the guy doing the review - 200mV inaccuracy on the voltage, and the non-responsiveness of the current measurement seem like a real problem. But maybe this can be 'twiddled', if the reference circuit is identifiable. In the review video it appeared that the V & A under-reporting was by roughly the same percentage error, and that the voltage error was predominantly an offset rather than a scaling error. So maybe...

It would have been interesting to see a comparison of the temp reading to true value, to see if it had the same percentage offset.

I didn't look up the micro part number yet. If that contains both the reference and AtoD, then I guess that idea is out.

Edit: http://www.atmel.com/Images/doc8155.pdf
OK, 10 bit ADC, with an external reference input pin. CPU substrate ground is the other reference - which immediately suggests one potential problem.

Also 10 bit ADC means the resolution is 1 in 1024, or about 0.1%. So if the measurements are scaled for 60V full scale, the resolution _should_ be around +/- 0.06V.
Notice the web page says "Accurate & precise – 0.01 A current and 0.01V voltage resolutions" which is bullshit.

Though, I don't know why a device designed for RC system use would have a max voltage range of 60V. I'm not an RC enthusiast, but isn't 60V ridiculous in that context? Should have been much lower, giving better measurement precision.

[GK]

But it says "Accurate and precise" on the packet, so what's the problem?

Fly your dream!

LOL.

Does that model Atmel uC have in in-built ADC? Probably uses that. uC Supply rail is likely the reference.

[iloveelectronics]

Well, if anyone wants to have a go at "fixing" these here's your chance. Otherwise they will continue to serve the purpose of dust collection quite well here 

[peter.mitchell]

Have you been able to identify what the board uses for a reference, and AtoD converter?
I agree with the guy doing the review - 200mV inaccuracy on the voltage, and the non-responsiveness of the current measurement seem like a real problem. But maybe this can be 'twiddled', if the reference circuit is identifiable. In the review video it appeared that the V & A under-reporting was by roughly the same percentage error, and that the voltage error was predominantly an offset rather than a scaling error. So maybe...

It would have been interesting to see a comparison of the temp reading to true value, to see if it had the same percentage offset.

I didn't look up the micro part number yet. If that contains both the reference and AtoD, then I guess that idea is out.

Edit: http://www.atmel.com/Images/doc8155.pdf
OK, 10 bit ADC, with an external reference input pin. CPU substrate ground is the other reference - which immediately suggests one potential problem.

Also 10 bit ADC means the resolution is 1 in 1024, or about 0.1%. So if the measurements are scaled for 60V full scale, the resolution _should_ be around +/- 0.06V.
Notice the web page says "Accurate & precise – 0.01 A current and 0.01V voltage resolutions" which is bullshit.

Though, I don't know why a device designed for RC system use would have a max voltage range of 60V. I'm not an RC enthusiast, but isn't 60V ridiculous in that context? Should have been much lower, giving better measurement precision.

This things are generally used in the RC plane world, where models run many many cells in series (10-12s lipoly)

[T4P]

Hang on a sec ... 10-12 series cells of Lipos are so much that most chargers don't do more than 6 cells ...

[peter.mitchell]

They usually are charged in 2 halves of 5 or 6 cells, but the ESCs and everything use the full 10-12s