Strain gauge load cell circuit?

Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
I have a 6kg load cell from a crappy digital scale. I assume it is some sort of strain gauge arrangement - it is an aluminum bar with a figure 8 shaped hole through the side with a strain gauge (presumably) on top of that area.

It has 4 wires, black, white, green and red; measuring ohms between pins:
b-r 985Ω
b-g 745Ω
b-w 745Ω
r-w 745Ω
g-w 997Ω
r-g 745Ω

Does anyone know the circuit, or can guess it?
...just had a forehead-slapping thought, was measuring it in circuit still, but unpowered... There is also what appears to be another area on the side of the bar that has "glue" on it and the wires come from there, so perhaps a half bridge with temp compensation?
 
Last edited:
I have a 6kg load cell from a crappy digital scale. I assume it is some sort of strain gauge arrangement - it is an aluminum bar with a figure 8 shaped hole through the side with a strain gauge (presumably) on top of that area.

It has 4 wires, black, white, green and red; measuring ohms between pins:
b-r 985Ω
b-g 745Ω
b-w 745Ω
r-w 745Ω
g-w 997Ω
r-g 745Ω

Does anyone know the circuit, or can guess it?
...just had a forehead-slapping thought, was measuring it in circuit still, but unpowered... There is also what appears to be another area on the side of the bar that has "glue" on it and the wires come from there, so perhaps a half bridge with temp compensation?

You probably have two wires for excitation and two for sense the extra resistances are probably in the excitation branches for TC comp etc. Applying an excitation and looking at the outputs should get you to the correct pin out.
 

PRR

Member
Joined 2003
Paid Member
The four "745r" suggest a bridge. A hasty thought suggested the corners should be 745r; duh! The "745" is really one in parallel with three series! So each arm is 993r, which is indeed what the corners say (1% error).

It appears to be full symmetrical, though we do not know the mechanical orientation. Power up two corners with say 1.5V, meter across the other corners, and lean on the cell. If no real change, try swapping corners.

Dratted scanner quit. Sorry for the smear.
 

Attachments

  • RonE-loadcell.gif
    RonE-loadcell.gif
    10.3 KB · Views: 71

PRR

Member
Joined 2003
Paid Member
Tempco correction is not likely.

A 4-arm bridge has inherent 1st-order correction.

This is said to be a "bathroom" scale. When the power is out, my potty may get chilly, but if my butt freezes to the seat (or the bathwater is solid), then my weight is not an urgent question.

This is said to be "cheap". It won't have 2nd-order frills (unless there is a standard cell/chip package also used in commercial scales).

If this were a commercial (billing) truck scale selling to Siberia and the Sahara, the designer would have to study wide temperature range errors. Not a cheap bathroom scale.

Googling load cell tempco, it seems that a tempco spec applies to the electronics used to recover the cell voltage, not the cell. I believe there must be a cell (and beam!) tempco but small relative to low-level electronics.
 
Last edited:
The elements are above and below the figure 8 cutout, which makes more sense.
from board silkscreens hidden before I cut off the wires, the pinout appears to be E+ red, E- black, S+ green and S- white. Presumably E is power and S is signal.

Cut loose from the board the resistances measure substantially the same, except both of the higher resistances are 997 and the rest are 748. I'll have to pull out my alligator clips and have a play this weekend.
 

PRR

Member
Joined 2003
Paid Member
> except both of the higher resistances are 997 and the rest are 748.

Remember that in-bridge, 748 = 997||(997+997+997)

Or in round numbers: 1K||3K (the three other resistors series) makes 750.

They are four "1K" resistors. (997=1K for practical purpose.) A ring of four will be 1K over the diagonals and 750 at each arm.
 
Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.