Load cells

You should connect your load cells in parallel, i.e. all plus excitation with plus, all minus excitations with minus, the same approach with sig outputs. It's not recommended to connect in parallel more than two of load cells.

It's recommended that all load cells have the same impedance parameters.

I assuming that distribution of mechanical loading for load cells are equal, though it'll be never absolutely equal.

This way of load cells connection has obvious disadvantage as measurement system a whole is getting worse in term of sensitivity and linearity especially for low level loads. Sensitivity goes worse by reason of decrease of input/output impedance. Linearity goes worse by reason of non-linear character of output signals summarize.

The only (non obvious) advantage --- is saving of ADC channel and number of power supply devices. But is it worth it to compare with accuracy?

Four load cells connected in parallel behave very "stupidity" and there is so hard to avoid a noise problems, though I've met such design and I was compelled to apply one for reasons that I couldn't change in no way.

I would suggest you to apply two pairs load cells connected in parallel. You need two excitation and two ADC/Instrumental amplifiers.

Hope it helps.

It's not recommended to connect in parallel more than two of load cells

Hi caramba, I don't know if we're using the same load cells (Sartorius/Global Weighing/Philips) but I've used up to 6 load cells in parallel (for really big truck scales). Typical applications use 3 to 4 load cells.

I guess it depends on your transmitter or load cell power supply.

Their manuals all say that you can parallel up to 6 and gives instructions on how to connect them. Here's one that I redrew from one of their manuals (4 load cells):

There is an issue with using four load cells and that is "balance". It's hard to get four load cells balanced (equal distribution of weight). Careful attention to installation is key.

Three load cells, however, are much more forgiving.

Lately, however, I've taken to using just one load cell. I find it to be more accurate and simple to install. The scale structure needs to be done perfectly though.

HI Vulcan,
I said before as rule when you used to post your comments, there is nothing to add as you're so competent and I truly respect you. But I would disagree with you to this particular case in next reasons. Yes I'd used also up to 4 and 6 load cells as well. Sartorius and Philips and Metler Tolledo have very good balanced sensors and special methods how to replace and install ones. Though even with all recommendations it's so hard after even one sensor replacement to achieve desired perfomances. Scale platform rested on more than 3 sensors is statically undetermined and it's nightmare too. It's so hard from position of diagnostic issues also to deal with 4-6 sensors connected in parallel: zero balancing, repeatability (no matter where you place a load) and etc.

1) As I could understand madness is a novice (he asked about serial or parallel connection!);

2) It's not clear for me, but I was assuming he'd been charged to design new device for his plant.

It's not a big problem today 16bit, 24bit ADC. But there you get advantage do not mind about electric disbalance (sensors can be with absolutely different characteristics) and very easy diagnostic if something goes wrong during maintenance.

Regards.

Don't get me wrong, caramba. I agree that 4 or more load cells are difficult to work with. But if you're careful with the design and installation, you shouldn't have any trouble with them. The only time I've had problems with them is if someone else did the design and wasn't careful with the installation.

A lot of people seem to think that load cell installation is so simple. It's not, and if the mechanical guys who fabricate the scale don't have any idea (or don't care) about precise center-of-gravity and elimination of side loads, then you will have a nightmare.

I learned a long time ago that using one load cell is better in terms of accuracy but makes designing the scale that much more difficult. The results, however, are worth it.

I did not dare get you wrong Vulcan! And we were learned the same. And mechanical problems concerning load cells aligning is crucial issue. Sure!

I had a case in my experience when three load cells had a huge gap at its response when fourth of them had no any response till platform has been acted a significant loading. I connected and re-connected every cell separately to analog transducer for adjusting their level micron by micron. It was a hard work and it would be much more easy if scales were equipped by separated measurement channels at once. Nowadays weighing systems have "goodness" to adjust zero, span, non-linearity by means of PLC programming, but it is a trap as if platform is ill-designed in mechanic no one PLC,ADC won't make scales' readings reliable.

But do you agree that separate connected load cells better at least for trouble fixing, especially for non-trained personnel?

All my best regards to you.

Ah! I get what you're saying!

Yes, I've had to check the output of each load cell one at a time, but this is usually just once (during the initial installation). If we need to replace one of them or if a modification is done to the mechanical structure, we need to do it again. It's not very often but, yes, it is tedious.

Recently, our project team tried direct current-output load cells in our new installations. These are the types that have 4 to 20mA output instead of the usual millivolt output. These were fed directly into our PLC so that we had a reading for each load cell. We only use it for tank level measurement, not for weighing purposes since the accuracy is not as high.

An alternative is to use individual indicators for each load cell and feed the current outputs to a current-summing module so that only one 4 to 20mA signal goes to the PLC instead of four.

Many thanks for the advice

Good time of the day or night, Guest,

Your request is figured out as pretty googable on my opinion. Either way, ok look over here:

These mentioned are neither 4-20mA nor digital. Seemed you're keeping at hand 6-wire LC.

Vexc+ Vexc- : Excitation voltage which is likely DC. Might be 5-10 V (depends on particular spec).

Vsen- Vsen+: feedback or excitation voltage reference points for further instrument's performance improvements. Actually you could apply LC with Vsen+ not connected to. 'cos if your PWS is good, cable is excellent and distance is short to LC installed.

Vsig+ Vsig-: LC's output in mVs. There might be up some 5-20 mV for average LCs.

Hope it helps.

caramba

pls i really need ur help. can u pls show me how to connect a 6-wire cable load cell to pic16f877a. thx

Thanks for the advice

I forget one else "win-win' technique. Your load cells can share the same excitation voltage supply and have separated ADC/Amplifier channels.

pls i really need ur help. can u pls show me how to connect a 6-wire cable load cell to pic16f877a. thx