I want to buy a CG scale, but price seem to be too big for my budget. I decide to build my self one. I choose a variant of Olav Kallhovd and Josh Fint, but with some adaptations. Electronic schema is very simple. I choose power up electronic from UBEC 3A/5V with Lipo accumulator 2-3S. Sensors are 3 kg.

2 - weight sensor 3 Kg
2 - adaptor HX711
1 - lcd i2c
1 - arduino mini pro 5V
1 - UBEC 3A/5V

I find a company for 3D printed parts, and from 3 online store I bought electronic components. A lot af money for shipping .

Now is ready to be used. I'm very happy for that.

Calibration was a nightmare. I wasted a day reading everything I found about calibration, then another day to do the actual calibration. But now the device is ready for use. Was more simple to build and mount all, but calibration is necessary.

Originally Posted by popad

Calibration was a nightmare. I wasted a day reading everything I found about calibration, then another day to do the actual calibration. But now the device is ready for use. Was more simple to build and mount all, but calibration is necessary.

Calibration has fortunately been made a lot simpler in the new version.

Cheap CG scale ... https://medium.com/rc-soaring-digest...p-25f50c3d16e8 !

I will try this or similar variant. I have bougt two cheap scales. More, soon ...

Materials: one obechi board 1000x100x3mm, one obechi stick 1000x10x10mm, a rest of 3mm balsa board, and a rest of 10x10mm balsa stick, a rest of 2mm foam board. Pictures speak for me .

Ready for use ... No problem with calibration in this variant, and really cheap.

First real application. Find CG of my current "Introduction F5J".

I set the CG of this glider at the airfield, based on my experience with the two previous fuselages. The glider flew very, very well. And I kept thinking what is the reason, that it flies better than the previous two. Today thanks to this device I found out.

The precision depends on the precision of the scales, which are very cheap, so not very precise. I made several measurements and made an average.

L = 250 mm

0. W1 = 639 gr, W2 = 368 gr. CG = 91.4 mm
1. W1 = 635 gr, W2 = 372 gr. CG = 92.4 mm
2. W1 = 636 gr, W2 = 374 gr. CG = 92.6 mm
3. W1 = 640 gr, W2 = 373 gr. CG = 92.1 mm
4. W1 = 637 gr, W2 = 368 gr. CG = 91.5 mm
5. W1 = 639 gr, W2 = 372 gr. CG = 92.0 mm
6. W1 = 640 gr, W2 = 370 gr. CG = 91.6 mm
7. W1 = 636 gr, W2 = 371 gr. CG = 92.1 mm

CG beteen 91.4 - 92.6 mm. CG = 92 mm ! ( in the manufacturer's manual it is recommended CG = 87 - 90 mm ).

I changed the scales, with ones that behave better. I had to set the CG to the 3 "Introduction F5J", and before this operation I checked the scales about that I was writing previously, and found a difference of about 10 grams. So I decided to get other two scales from another company. The two new scales are not perfect, but at least they show the same thing, and the difference between them, when it occurs, is not greater than 1 gram. In fact, I noticed that I have to leave them under voltage for a few minutes, after which they work well. A few photos ...

Hello,
can someone give me a hint whats wrong ?
I build this cg scale too,but have some issues with calibration.
I have tested the loadcells each and they show me weight. The calibration (c1) shows invalid value, weight is NAN .
What is wrong ?

Originally Posted by jochen594

Hello,
can someone give me a hint whats wrong ?
I build this cg scale too,but have some issues with calibration.
I have tested the loadcells each and they show me weight. The calibration (c1) shows invalid value, weight is NAN .
What is wrong ?

No images, no schema, no program. How can I tell you something about that. Seem that one cell do not transmit data on calibration.

Everything worked well for me, but during calibration I noticed that the weight measuring cells show different results over time, thus making calibration almost impossible.

That's why I switched to the version with two scales. And here only on the third pair of scales bought did we get the desired result. So I stayed with this solution. Not so elegant, but work very well !