Arduino, hx711, about hx711

[electrokean]

@bneedhamia nice writeup there, and good to see you doing some statistical analysis. I actually have calculations built into my sketch to calculate noise and ENOB (effective number of bits) on command.
I think you biggest issue you have (per your blog) is that the "load cells" you're using just aren't particularly good, thus leading to mismatches between cells, which totall screws things up when you wire them together. You don't need expensive load cells, but it isn't difficult or costly to find some better than those.
Get some proper 4-wire (or 6-wire) beam style ones and mount them solidly, wire them in parallel (although test them individually first) and you should get significantly better results.
FYI, 6-wire ones just separates the excitation and excitation sense, to remove voltage drop effect from long or thin wiring. The HX711 doesn't support 6 wire load cells, but I guess you could use channel B to measure any error, and I suspect that is why the channel B input exists. Otherwise you just connect E+ and Sen+, and E- and Sen-. Note that Sen+/Sen- is different from S+/S- (signal).
Hope this is of help.

[bogde]

did you follow the calibration procedure? what weight did you use?

[electrokean]

As bogde mentions, this sounds like a calibration issue.
You effectively need to reduce your value passed to set_scale() by a factor of 100.
Are you using get_units() which returns the scaled weight, or get_value() which returns the raw value?

[dmurphydrtc]

Kean/Bodge,

I obviously stuffed that up, will work through my calibration again and
report back.

Thanks.

On 28 March 2016 at 17:28, Kean Maizels [email protected] wrote:

As bogde mentions, this sounds like a calibration issue.
You effectively need to reduce your value passed to set_scale() by a
factor of 100.
Are you using get_units() which returns the scaled weight, or get_value()
which returns the raw value?

—
You are receiving this because you authored the thread.
Reply to this email directly or view it on GitHub
#33 (comment)

Regards,

Dave Murphy

[bogde]

any news?

[dmurphydrtc]

Bodge,

Just have not had time to get back to it...Hopefully in the next night or
2. Thanks again.

On 6 April 2016 at 15:17, bogde [email protected] wrote:

any news?

—
You are receiving this because you authored the thread.
Reply to this email directly or view it on GitHub
#33 (comment)

Regards,

Dave Murphy

[bneedhamia]

If you aren't aware of it, check out Nate Seidle's beehive scale project, at http://makezine.com/projects/bees-sensors-monitor-hive-health/ although he doesn't go into detail about calibration, he describes a lot of other details.

I just recently blogged about how to calibrate an "old school math" way, for my Dog Bed Weight Scale project. See https://needhamia.com/?p=720 for a step-by-step of how I calibrated it. The raw numbers I read and formulas I used are in a Libre Office spreadsheet is at the github repo of that project, at https://github.com/bneedhamia/CurieBLEWeightMonitor