Hi all.

I am currently working on building a dyno for cars. A have the rollers made and they are ready to be balanced.

I have two new load cells 0-100kg (the roller weighs approx. 130kg).
I have two PR load cell amplifier units 0-10v
I have one 10-bit absolute encoder
I have a siemens s7-1200 plc.

The only problem with this setup, is that its too slow.

The amplifiers only update output every 60ms. If I want to have a measurement every 10 degree of rotation then I would only be able to spin the roller at about 27 RPM. That’s not good enough.

The perfect solution would be to rotate the roller to 2000rpm and have a measurement every 10 degree, that means a measurement rate every = 1000ms/((2000rpm/60sek)*36) == 0.833ms

So I need other hardware that are fast enough. I have looked at the HX711 load cell amplifier but that has an update rate at 80hz == 12,5ms So that’s one fast enough either.

the Siemens PLC can maybe update every 10ms, so that also not good enough.

I am lost here, what too do??

Do you have any suggestions?


Why are you weighing the roller?
You should be measuring torque.
Check and talk to the Seimens rep or seller, they should have a High Speed input module.
I would be discussing the whole project with them as to its feasibility with the speed of the PLC.

Seimens charge enough for their product, they should be able to answer your questions.

What are you going to use for a retarder/brake to absorb and measure the energy produced.

Tom.... :slight_smile:

This is only for balancing the rollers, not measuring torque. The rollers need to be perfect balanced before I can place a car on them and spin them to max 5000rpm.

This post is NOT for the dyno itself, only on how to balance them, and what hardware to use.

I use load cells to determine the weight difference when they are spinning.

Shouldn't you be using accelerometers?

Hmm maybe. I haven't considered that.

I just saw the logic by using load cells.

How can I use accelerometers? I don't know that method.



Okay, dynamic balancing.

The load cells will be fine, they will give you an AC output due to any imbalance plus a DC offset due to the mass of the rollers.

I used to service wheel balancers, they used piezo load cells.

Strain-guage type should be fine, as you say its the response of the load cell amplifiers and the input of the PLC that is the problem.

Have you got an oscilloscope to see the waveforms of the cell and out of the amplifier.

Tom..... :slight_smile:

Exatly Tom...

Yes I have an oscilloscope an old 2 channel. I could connect the load cells directly to the oscilloscope and see the waveform, but how to locate were the unbalance is???


There were a couple of different methods to get position.
One was a peak detector, it would produce a sync pulse at the peak of the waveform, then knowing the speed of the shaft you could delay that pulse to the required phase shift to align to a reference point indicating where the imbalance is or where to add mass.
Two, was to detect a positive going zero crossing to produce a sync pulse, then phase shift etc etc.
That was using analog/digital circuitry.
Using a controller I don't have any info, sorry.
I have some circuit info for the old analog stuff.

Tom... :slight_smile:

Balancing a boat propellor can be done using two low friction bearings. No electronics required.


I dont understand what you mean by phase shift??

I could make a weel with 4 edeges with one beeing longer than the others and mount it on the rollershaft. I could then take a standart PNP inductive sensor (ex from SICK) then i will be able to see approx were the unbalance is, or were to add mass. With this setup i can only balance one end at a time, but thats ok.

I think I will try this with the oscilliscope. It should work??


Yes that is correct, but what if there were more blades on the shaft making the shaft 1 meter long. then you woldnt be able to se the unbalance from one end to the other.


If you use a scope then your method would be fine, one channel on the loadcell amp and the other on the inductive pickup.
You can see the relationship between unbalanced sinewave and pickup pulse to workout where to put or remove mass.

Just make sure loadcell is in line with plane of roller where you are going to adjust mass for best and quickest results.

When I was talking about phase shift, the balancers concerned used either
a flashtube and numbered index wheel on the shaft to show where to place mass,
or displayed on a digital readout where to place the mass using an index wheel on the shaft.

Tom.... :slight_smile:

The load cell produce 2mv/v. if I supply the load cell with a car battery I will have about 24mv @ 100kg. The roller weighs about 130kg that's 65kg per roller end and about 15,5mv. The oscilloscope is "Phillips PM 3208" and has a selector to adjust the measured voltage and that can go down to 5mv.

Is it fine to through the load cell amplifier away and measure directly on the output on the load cell?

Keep the loadcell amps, they will provide good, low noise, STABLE signal.

Tom.... :slight_smile:

Year but they only output signal every 60ms and thats too slow. Then i need another amplifier? Mabe a circuit that provide a constant signal?

Okay, then use a regulated power supply, not a car battery, to excite the load cell.

Tom..... :slight_smile:


Yesterday I tried to test the oscilloscope with the load cell. First I used til amplifier module to exite the cell, that resulted in a very weird curve on the oscilloscope. Then I used the car battery and that did the trick, a nice flat curve. But, when I place myself on the load cell and bounce very little the curve hardly moves. So that tiny change in mv is not to be seen on the oscilloscope. So load cell can not be used in this case when I want to balance down to + - 1gram.

Then I found this one on spark fun.

What do you think of this sensor?

Draw a circuit diagram showing the load cell, battery and oscilloscope connections.

I have connected the load cell according to its data sheet.

Supply directly to the battery and output directly to the oscilloscope. 2 wires to battery and 2 wires to oscilloscope.

Nothing complex about that.

When I step up on the load cell I can clearly see a change in the curve, but not when bouncing very little. So I don't think it can measure down to 1g with load cell.

Check out the link I posted and tell me what you think about it

Provide a link to the data sheet.


You can use the load cell, just you need an instrumentation amplifier, it has differential input and single ended output that is suitable for your scope.

google instrumentation amp
google diffrential amp
google loadcell amp

Tom... :slight_smile: