Lesson Working with Load-Cells

Hello!
Here are some information about load-cells for future reference:

2 common types of load-cells
- you have 3 wires load-cells

a single resistor is variable under load​

- or and 4 wires load-cells

2 resistors are variable under load​

Some Theory
Here you can find some theory nicely explained :

- http://www.ni.com/white-paper/3642/en/

- https://learn.sparkfun.com/tutorial...-hx711-breakout-hookup-guide/hardware-hookup-

- detailed 3 wires load-cell tutorial http://www.fut-electronics.com/wp-content/plugins/fe_downloads/Uploads/Tutorial_for_Load_Cell_weight_sensor 50KG.pdf thanks @tombo​

As far as I understood the point,
- 4 wires load-cells are in Wheatstone bridge configuration and that's the best for measuring the small variation of resistance.

that's what I used for my DIY load cell hydraulic hand brake
​

- 3 wires load-cells, like Fanatec brake pedal or any low cost chinese load-cells, are only half bridge...

but you could increase sensitivity by using
- a pair of them back to back : like I did for DIY brake pedal


- or by four : http://www.neuroproductions.be/experiments/6-axis-robot-arm/


- or if you want to use a 4 wires PCB amplifier with your 3 wires load-cell: you can proceed like Jet Badger did

​

Reading the signal

you cannot replace a potentiometer by a load-cell : the resistor variation values are too small.
That's why it's mandatory to use an amplification board : INA122, HX711​

INA122 PCB board is perfect for turning a laod-cell into a "potentiometer":

the output is an analog voltage
you can replace any pot by the INA amplifier PCB
the price is around 10€ (chip, PCB...)
http://www.ti.com/lit/ds/symlink/ina122.pdf
Operation supply voltage range: 2.2V to 36V
​

HX711 is really low cost

0,6€!
but the output is serial communication: it requires an arduino or a microchip establishing a serial communication

Operation supply voltage range: 2.6 ~ 5.5V
https://cdn.sparkfun.com/datasheets/Sensors/ForceFlex/hx711_english.pdf​

Examples

4 wires load-cells -> my hydraulic hand-brake converted to load-cell hand-brake
http://www.xsimulator.net/community/threads/hydraulic-handbrake-converted-to-load-cell.7326

3 wires load-cells -> brake pedal with personnal scale hack
http://www.xsimulator.net/community/threads/diy-load-cell-brake-pedal-short-tuto.6042/

​

Any further hints and information are welcome!
yours
Mat

Thank you for that detail Information.
I think i will try the Jet Badger method first because i have all the parts laying around.

Just one point to add about using two separate load cells for a full bridge is that they need to be matched exactly in resistance otherwise there will be an inherent difference that gets amplified via the comparator. This is not a resistance that can be measured with a multi-meter as it is too small. I noticed this on one pedal that I made where it looked like the accelerator was being pressed ever so slightly.

Great write-up!

Thanks guys!

Yes @Alexey, indeed the pair of load-cells should be exactly the same!

@tombo, try to use low tolerance resistors and keep us informed

Hello,
so i made some further testing on my Handbrake with the Fanatec loadcells. I found a great Info sheat how to use only one of them.
http://www.fut-electronics.com/wp-c...Tutorial_for_Load_Cell_weight_sensor 50KG.pdf

At the Moment I get some readings rom 0.6mv to 2.6mv. I can't press with full pressure because the handbrake is not mountet yet. I hope I'll recive the in122 this weekend and try to amplify the signal.

Here is a Picture of the current test setup. I have only plexiglas as building material at home. There is also a little sketch so the hole concept is explained.

thanks @tombo! nice document about 3 wires load-cell

about your design, here are my thougts :
- it's good to have a mechanical stop behind the load-cell to prevent over-pressure (that could destroy it)
- what is the purpose of your spring? in the configuration you drawn, your spring does nothing...

Hello,
I have some kind of mechanical stop but you can't see it. I milled the form of the Loadcell in the plexiglas so it can move and at the end stop at the plexiglas.
For the Spring, maybe I'm a little bit stupid but it should bring the lever back to start position.
If the spring would be between the nut and the tube at the loadcell it will compressed when you pull the lever. I'm anyways not shure about position of the spring maybe in front at joint and plexiglass.
I get my notice today that my parcel with ina122 will arrive in next 2 days so I will keep you updated with the final result at weekend hopefully.

Here are some pictures. You can see its broke at the upper hole. Plexigas isn't the right material for this. I will try aluminium.

I would put the spring here...
or here but much larger: it should not touch the load cell itself

Hello,
thanks for your explentation.
I'm woking on the design anyways and added a plexiglas plate which holds the tube. Now the pressure of the spring is against this plate.

Today I managed to get the qater Bridge loadcell design working. It semms that my loadcells are very cheap, because all of them, output a little voltage when no pressure is applyed. It's the same at my fanatec break pedal where I used them as replacement.
Maybe i change the resistors again. As RacingMat mentioned low toleraqnce ones will be good. I'm currently using 2,2Kohm instead of 1K in the above tutorial.

Here is a little Video.

@tombo, you have to use resistors that are the closest to the internal resistors in your load-cell!

measure between the 3 wires:
you should find something like 1 kohm, 1kohm, 2 kohm
or X kohm, X kohm, 2X kohm

therefore you have to complete the bridge with X kohm resistors
not other values! if not, the bridge would be unbalanced

It's not so much using the same resistors but having the same ratio. What you are looking at is called a voltage divider. When the load cell is strained or compressed its internal structure changes and thus creates more or less resistance, almost like becoming more dense (less resistance) as it is compressed. The change in resistance causes an imbalance in the voltage divider and that is what you measure. Unfortunately this change cannot be detected using a multimeter which is why we use a comparator to detect the change and then amplify it to something we can measure.

Using plain old resistors on one side of the full bridge is setting the 0 point of detection, so it should match the ratio of resistance from the other side. The comparator looks at the voltage output of your two passive resistors and compares the voltage to the voltage coming out of the load cell voltage divider. The comparator will then amplify the difference in voltage.

Using passive resistors is never going to be perfect because of component tolerances.
When a load cell changes resistance due to strain it changes at a rate of about 0.04 ohms per kg of load (different cells will have different values but in the same ballpark). Your standard component tolerances on a 1k ohm resistor even at +/- 1% tolerance is a swing of 20 ohms which is 500 times the size of one 1kg change in the load cell.

Basically what I am saying is that it is virtually impossible to match a load cell with passive resistors.
One way you might be able to do it is with potentiometers. Set up the full circuit, have no load on the load cell and adjust the potentiometers until there is no voltage on the output. Sounds simple but it can be quite tedious.

I'm terrible at explaining things but I hope it helps.

Hello,
for me it helps, I've done it like you said and I'm able to use my handbrake with quaterbridge design.
I've had some ussues but finally find out that these are caused by my gear indicator. If it is plugged into usb the handbrake readings go crazy.

you used adjustable potentiometers to balance your bridge?

Hello,
I used two trimmer pots, adjusted them to nearly the resistance of the Loadcll before inserting into pcb. Than I measured the output signal of Ina122 and finetune the pots until I get a full 0-5V a arduino and 0-3,3V at my fanatec pedals.

It's the pcb from this thread http://www.xsimulator.net/community/threads/make-your-own-pcb-with-toner-transfer.7343/

Great!

About Amplifier, I just found an equivalent pin-to-pin AD627
only 4€ with shipping
http://fr.aliexpress.com/item/1PCS-...est=201526_5,201527_4_71_72_73_74_75,201409_3


it could be worth to have a try!

Ok,
the upper circuit worked properply, but was a pain to set up.
I found a much easyer way. I used one cell with two 1k resistors like before. Now there is some little offset Voltage. This can be adjusted by the second trimmer. Just connect the output to a multimeter and set it to 0v with the offset pot. Than you can adjust gain normaly.
Testet with 3.3V on fanatec Pedals and 5V on breadboard.


Edit: today I see that in Assetto Corsa there is a graphic slider for input devices. On this the Handbrake output doesn't look so good. There are some pums in the signal when pressed. You can't notice it while playing, but it is there. I will take a little depper look at this.

just looking at loadcells and i am sure i read somewhere that th t500 pedals can run it as a straight swop out from the standard pot 0-5v im just looking into adapting somethings on my rig, but as it was 2 years ago i was looking ijust have that memory of a straight 3 wire loadcell swop out on t500 pedal sets.
is this wrong ? or am i missing it is still resisted.
t500 pedals are 0-5v as far as i know atm.

You're missing something, the circuit above in Tombo's post would have rj12 socket to it removing brake pedal connect then second rj12 out plug to go to wheel with the new connection wired, I will be making a few shortly will document on here when I do & drop link on this post.

Super post (and thread). I am glad I found it.
I am starting to get frustrated with my potentiometer break pedal, hopefully I will be abble to convert it to load cell.
It is a bit overwhelming, the amount of info and knowledge needed.
Fingers crossed I'll be able to sort it out.
I'll make sure to be back with any stupid question
Thanks everybody.