Matrix user forums

Hello all,

I would like to read the analogue to digital converter in the link below.


https://cdn.sparkfun.com/datasheets/Sen ... nglish.pdf


Would it be possible to know how to get the data from HX711 chip?

Is there any FC6 example ?


Best Regards
George

Hi George,

Attached is a macro is created in fc6, it will also work in fc7. It reads 24 bits of data from the hx711 chip....the project I used this for was a 4 load cell scale connected in a Wheatstone bridge config. Import the macro into your project and call it as a routine.

Note - given the sensitivity of the hx711 you will need to average the output .

Let me know how you go

Best

Pete

Hi Pete! Thanks for your reply and your time!

I tried to make up a program including your macro. Unfortunately i ended up with no good results although i wont be disappointed at all.

In order to avoid a lot of texting here, i will upload the FC6 program with this post so you all can have a look and/or improve the file.

I will also include the wiring i used between load cell and HX711 module .

The weight value in LCD is not stable and seems to fluctuate up and down a lot.


Thanks again!
George

http://www.ebay.co.uk/itm/291644547509? ... EBIDX%3AIT

Hi George,

The fluctuation is pretty normal to a degree - hence my reason to suggest an averaging methodology.....Also check your PSU - do you have filter caps close to the HX711 Chip/board?... this can be a source of noise...

The other thing you can do is reduce the gain and scale up... this may also assist.

When I get moment I will have a look at your project

best

Pete

ok... had a look at your project....

when I created the macro, I left all of the config options disabled - you will need to enable the icons in the red circle and then it should work...the data sheet will give you a good insight as to the config options vs function

let me know how you go

best

Pete

Hi Pete,

I tried your suggestion but again the results were not very good.

I get at least a number on the LCD and i assume that the Load Cell may be a crappy one...therefore i will look for a better load cell to make my test.

Is there any load cell which operates well and been tested with this module ?


Thank you
George

Hi George,

I used 4 load cells from a bathroom scale I purchased for $25...very cheap and cheerful

4 cells wire in a Wheatstone bridge config is good for 150kg with a resolution of 1 gram (ish)

Can you describe what not very good means?..do you know the scaling factor?

Best

Pete

Hi Pete / George

First off, thanks for posts on this subject. I have found them to be very helpful.

I recently obtained one of the HX711 modules from eBay so armed with your posts and a lot of gung-ho I set about experimenting.

I see that the modules come in a variety of flavours, mine is the XFW-HX711, a green board with headers for the four Load cell connections, nothing for the screen. I also have quite a few load cells (each is its own wheatstone bridge) rated 5Kg and 20Kg. Each of these cells come with a calibration certificate stating that their accuracy is class C3.

Ages ago, using these and an Instrumentation Amp, I could get quite accurate and repeatable readings just by feeding the output of the amp into an ADC port on the uC directly. I could differentiate between the addition and removal of a "few tens of grams" at a time, and readings were quite stable in that each "read cycle" returned the same values more or less (I would like to stress that absolute accuracy was not a consideration in this application, just a means to know when certain criteria had been met, and I did have to average out readings per cycle) so I was keen to see what this module would offer.

Hooking up the HX711 to an EB-006 (PIC18F4520) via a plug-in expander on Port C and an EB083 connected for its LCD, running Pete's code with mods made to connect over Port C gave immediate results. Scaling and calibration variables in the macro I didn't care about for testing, so they were set to 1. Loop delay was set at 10ms as per Georges posted Flowchart and I varied the average, counting 5, 10, 20, 100 and 200 times.

In each instance the results returned after averaging, had a variation of approx +/- 7% between readings. By that I mean I took 20, separate "averaged" readings, and each reading varied by around 7% from each other. Not good. Repeating with other load cells gave similar results and clearly something was amiss and I initially thought the module to be at fault. Perhaps this is what George saw too?

Reading the datasheet (which is always a last resort <s>) suggested that I might be accessing too quickly, therefore I increased the loop delay from 10ms to 400ms which gave an immediate improvement. Averaging 5 times gave a variation between readings of approx +/- 0.6%, averaging 10 times gave a variation of +/- 0.3% whilst averaging 100 times returned a variation of +/- 0.12%, although admittedly it does take "ages" to update

As this module is far cheaper than the Instrumentation Amplifier I used, and is capable of repeatable accuracy (for my needs anyway) I think it will be included in my armoury..

I hope this may help.
Regards

Hi Pete,

Which variable in the macro should I use to display in an LCD?

Best

Chris

Hi

I don't have access to FC just now (traveling) but from memory I believe Pete's macro, that he kindly shared, "built" a value. By that it was created taking other variables into account (such as scaling) etc, therefore the "final "weight" was if I remember correctly calculated near the very end of the macro, after averaging etc. I really can't remember it's name but I would guess it would be something like "weight" or an abbreviation of such.

Try displaying whatever variable(s) you find near the end, after all you can't break anything <s>

Hope this helps

Hi,

Ok i will check this out and see what i can do!

Thank you for your help.

Chris