New.. and confused about load cells & pressure transmitters ohm's

[paulmac]

Hi Ian,

Sounds interesting, but how would you set the scale? If say the transmitter (0-5v) gives 4v at a pressure of 2000psi when you have 3000kg in the bucket, wouldn't you need something a bit more complicated to do the calculation to get the voltage to Kg's ?

Checked out your site and products. All look very nice

Regards

Paul

 

[Ian Rogers]

Its not that hard really.. I don't care to much about pressure all you need to know is what the digital input is at zero and span at a known weight. then there is a simple scaling routine to sort it.

digital_value_reading =  ((digital_value_span * ( sensor_reading - sensor_zero))  / sensor_span) + digital_value_zero

Cheers Ian

 

[paulmac]

Ian,

So what would you recommend to use for the display?

 

[Reloadron]

This link provides a basic overview of the calibration of a pressure transducer feeding a display.

I believe it was mentioned earlier but if for example the output of a 0 to 5,000 PSI (I only use PSI because it is familiar for me) is 0 to 5 volts the output could be measured with a voltmeter capable of measuring 0 to 5 volts and the scaling is obvious. However, things get a little more complex. Note what is going on in the link. Consideration at zero is given to the empty fork (or bucket) weight. Then a calibrated load is applied and the meter is basically told when you see this much, display this much.

There are meters designed for this purpose. Some simple metering is relatively inexpensive while the more features the more the cost. For example a pretty nice panel meter designed for doing this type of work is one like this one. which at about $600 USD is not exactly inexpensive. I have one simple application where I use several of these units. I have a flow sensor that transmits 4 to 20 mA for 0 to 4 GPM of flow. They cost about $80 USD, aren't fancy, no features, loop powered and do the trick.

Another option and a popular one in the forums (I believe Bryan will go this route) is the use of a PIC micro-controller, writing their own programming code and using a small LCD display of their choosing. That would be a home brew example.

What it all comes down to is taking a physical force like force or pressure, converting it to an electrical signal, conditioning that signal, scaling that signal and finally displaying that signal.

Ron

 

[bryan1]

Hi Paul,
In a few weeks I'll be able to spare some cash and I'll grab one of those sensors and fittings. Then hopefully the bobcat will be finished and I'll have time to devote to making a pic controlled device for it. i'll install it on my MF50B frontend loader bucket and get it working then after that I'll be able to do one for you and workout the cost of the parts. My idea is to have the ability to add bucket weights for a daily total etc. I am thinking one of my maximite boards might do the trick then it be be an easy case just to download the data straight onto a sd card so I can save the data in an excel file. Soon I'll be going into the landscaping/ earth moving business and keeping an accurate record of the bluestone rocks and roadbase will be a huge bonus.

Regards Bryan

 

[Ian Rogers]

Well as you've seen on my web site, I have a complete system in place to read process and display, but as you already said these are expensive.

You can get a volt module (similar to this) Anders Electronics | Process Control | Panel Meters | Digital Meters | LCD 5.5 to 15mm Digit Height, OEM Applications |OEM24LS and a small ina125p instrument amplifier (a small pot on the gain and one on the -vref) and you have a little display.

Cheers Ian

 

[paulmac]

Bryan,
sounds great. I like the idea of recording the weights on a SD card, though I wouldn't need that for myself, but it's a nice option.

Ian,

Had a look at the meter and ina125p instrument amplifier datasheet and see what you mean, but putting it together is beyond me

I guess something like this; Mini-Type 4-20mA Panel Meter 24VDC Powered would work, as it allows setting an offset and scaling, although perhaps not easy to zero an empty bucket. Both of these, 5 Digit Process Indicator(24X48mm) 4-20mA Input and **broken link removed** would be better as they have buttons on the front for programming, but I'm not sure how easily you could zero them either. Also the displays are a bit small on all of them. Need to be able to read the thing easily. This, **broken link removed** is by far the cheapest prebuild option I've found, but scale and offset have to be programmed by plugging it into a computer and there's no way to zero it once your using it. Thought for $23 it's almost worth playing with.

I guess the above would work fine if you didn't need to zero it , but I think as the hydraulic oil and rams themselves warm up with use, the zero point will drift and you will need to be able to reset the zero. I guess if oil/ram temp affects the zero point then it will also affect the calibration weight point? Perhaps if you took the offset amount of the zero point and adjusted the calibration weight point by the same amount you would maintain a better accuracy?

As you can see I'm good at taking something simple and making it more complicated than it needs to be!

Regards

Paul

 

[Ian Rogers]

We use a position in the ram (forklifts) where we zero and read. A standard settle time of 3 seconds is needed to get quite accurate results.

It also depends on which ram you're reading from if its the bucket's ram there will be a lot of non lineararity as the bucket is never at the same angle reference to the floor.
Gravity will mess you up a bit.

Some people don't try and read kg, just moment if you place a transducer on the main ram (get the 4 to 20mA type) and connect a variable resistor to an LED, when there is say 300kg in the bucket turn the pot to light the LED. (This is for moment, tipping etc...)

Cheers Ian

 

[SABorn]

I have been watching this thread from the first post, and was going to throw my 20c worth in but Bryan1 covered most of what i was to say.

Now i will commit a sin here and offer another route simular to Bryan1.

I have been building a circuit for use with load cells in a platform to weigh sheep and have that completed now, so watched with intrest to this thread.

A very simple Picaxe (thats the sin) can handle all that is needed here as it has ADC inputs and will connect direct to a 0 to 5 volt input.

As for calibration... well that can simply be done in program as i did first for my load cell circuit or via a trimpot on a second ADC input as i changed my circuit to.

All you do is read the adc input from the sensor and divide the reading by the adc value of the trimpot input.
This allows for the calibration to be set in the field without a computer needed.

The display, i used a 4 digit 7 segment led display that is I2C, which drives directly off the picaxe on a 4 wire bus (5v+, neg, clock, data.) (led is easy to see in sunlight ...lcd are not so easy)

A few lines of program and the project is complete (and a 5 volt supply)

All up about $20.00 worth of parts.

No need for fancy pic programmers or a degree in programming and half the fluffle that has been covered thus far.

If using I2C then the 18m2 picaxe is the smallest chip they have with I2C, and with I2C you could add eeprom for data storage if required but also agree it is not needed.

All up you could say it is a very simple task with a 1 chip answer.

Pete.

 

[paulmac]

Pete,

Sounds good. I think that would be more than adequate for what I need. When will you start?

 

[Reloadron]

Pete, somewhat of a question. I am also guilty of hands on a Picaxe. Sorry, I am not a programmer type and just dabble with programming.

Anyway, the chips I have worked with all have been 10 bit resolution.

Now I assume down under you guys weigh your sheep in kg where up here in my world they weigh the sheep in lb. So what? 1 lb is about .453 kg or so?

Now if I want to weigh sheep I figure to what resolution? I have no clue but what I figure .5 kg? Using Google as my friend I did manage to find out depending on breed an adult sheep can range from 100 to 200 lbs or what about 45 to 90 kg?

OK, if I use a load cell that is 0 to 500 kg and scale it so o to 500 kg = 0 to 5 volts and use a 10 bit PIC I get:

Full Scale Measurement Range = 0 to 5 volts.
ADC Resolution is 10 bit or 2^10=1024 quantization levels.
ADC Voltage Resolution = 5 / 1024 = 4.882 mV

Now the load cell is let's say 0 to 500 kg. so we get:

5 volts / 500 kg = 10 mV per kg. We can easily resolve 1 kg.

Now what when we use lets say a pressure gauge of 0 to 5000 PSIG = 0 to 5 volts.

5 / 5,000 = .001 or 1 mV per PSI. The best we could resolve with 10 bit resolution would be about 5 PSIG increments.

Not sure how that would figure into things overall as to how many PSI aquaint how to the weight in the bucket? Would 10 bit A to D be good enough?

Ron

 

[SABorn]

Hi Paul,

Hmmm, as though i dont have enough projects on the go already.

What are your electronic skills like as i could be tempted to help you with this project, and dont see it all that hard to do.

BUT... i dont have a loader in my back yard or a sensor cell to test with, so would be some engineering guess work involved and require you to make changes as needed.

To quote $20.00 in parts is about what it would cost me for the basic bits without any assembly and things like a enclosure to mount it all in etc.
Also i have access to some surplus stuff that makes projects cheaper for me but for you to source simular it might cost more.

I would be happy to write the program for the Picaxe and design a pcb but are you able to make a pcb and assemble the board.

To give a basic idea of how simple a picaxe is to program the basic code would be something like this..........
*********************************
main:

Readadc10 1, w1 '......read the pressure sensor connected to input 1

Readadc10 2, w2 '..... read the trimpot calibration reference... input 2

let w3 = w1 / w2 '......divide pressure reading by calibration reading

'Display the data here

goto main
********************************************
W1,w2,w3 are word registers that the value of the adc reading is put into.

Now that dont include the code to write to the display and any other functions that might be included like tare etc, and is just an example of how you would read the data from the sensor and calibration.

This will give 10 bit resolution or a count equal to 0 to 1024 for 0 to 5 volt input, which should be good enough for a simple scale of KGs or 00.0 kgs.

What voltage is the loader? (24v i guess)

If you were happy with a lcd display than the project gets easier (but smaller numbers to read) as the picaxe has a inbuilt Ascii converter that will write data direct to a lcd.

A led display is something that will need to be built up and is harder to write to in code, but can be done.

Are you still interested?

Pete.

 

[SABorn]

Hi Ron,

You posted as i was typing so sorry for the double post here.

Now with the sheep scales i used a dedicated chip that is a ADC to 24 bit, which is so bloody touchy i only used the upper 16 bits, as the first 8 bits was not stable enough without some fancy circuit work, and found 16 bits was more than enough.

The chip i used was from TI and is ADS-1231, well worth a look at and is a wonderful chip for the application (designed for load cells)

Now if the 10 bit resolution is not enough, than we could use simular device to gain the high resolution.
I think 10 bits would be enough for an average indication of bucket weight, as it will never be a actual true weight reading anyway only a close average.

I also retrofited an old supermarket set of dead scales (15kg) with a picaxe and a ADS-1231 chip, which i can read down to 0.1 grams very easy and up to the 15 kg.
By blowing on the scale it will record a 25 gram weight, to give some idea of the sensitivity.

Now who would want to weigh their sheep to 0.1 grams unless you wanted to know if they had passed gas or not.

Hope that help explain how i get the weight readings required.

Pete.

 

[Reloadron]

Hey Pete What's Up

No problem at all, never knew you weighed sheep. Interesting stuff. I mean I never had a clue what goes into it like what is the tolerance? Do they get weighed before or after clipping? Before and after for wool content? Seriously, I love this place as the learning curve never ends.

Ron

 

[SABorn]

Hi Ron,

Hope Paul dont mind us getting a little off topic here.

I dont have any sheep myself, this is a project for someone else and is a part of a larger system that will read a RFID ear tag containing a registered animal number (national data base for livestock), log the data (and any comment entered) and draft the animals according to parameters set in the computer.

Sheep weights has become more important over the latter years and is used for application of treatments like drench, lice treatments, also to determine pregnancy, body weight, and overall performance, as the ones that dont produce well are culled and on the truck to market, as you might consider each animal eats the same amount but some have a greater return in body weight and are the ones you want to keep, so the skinny ones are best to cull as their return over time is lower.
This also shows how rfid fits well into this scheme of management, as with a flock of 100s is hard to track each animal but with personal earring's this becomes a easy task.

These systems are available and cost $1000.00'sssssss, but we think it is not so hard to develop our own for a few $100.00's, and it is the challenge i enjoy. (my part is for the love of electronics and is not charged for).

The rfid is something that we done some years back and made hand held wands for reading cattle ear tags, and now it is to take it a step further and increase the reading distance to around 2-3 feet and integrate that data back to a PC to store the data and do the number crunching. (the hand held wands was downloaded to a PC after the job)

I roped a few other friends in on this project to help out, as i will do most of the electronic design and testing/development/construction, another will do the VB program software for the PC as i am not that fluent on VB, and the person who will use this system will construct all the needed hardware, (weigh bridge, auto drafting gates, etc,)

Its no small project but so far 90% of the basic electronics is sorted and now needs to be all intergrated as one package.
The last problem to sort out is rfid reading range, and from research it would indicate the antenna voltage needs to be increased from around 200 volts that we presently have up to 700 to 1000 volts to gain the distance needed. (although 700 volt is the Australian max allowed limit )

The weight of fleece can be an important factor to record too, but the price of wool is poor compaired to body weight so wool is not a major concern.
If it is a matter of scanning the animals on the way into the shed and on exit after shearing then this might also be of interest in the scheme of data but not a intended use at present.

The rfid also allows auto drafting of male/female, age, selected rejects from previous data, and a one man operation in many cases. (very handy for the smaller farmer)

Now how many sheep do you have Ron? (and i dont mean the ones in the fridge)

Pete.

 

[paulmac]

Pete,

When they first bought the RFID tag system in, I was still running about 2000 sheep and remember going to a field day where they had a drafting race with scales setup and were showing what it could do. All looked great and saved alot of time, but the cost was ridicules . I remember thinking then that someone with a bit of electronics knowledge would be able to build their own for a fraction of the cost. Glad to see someone's doing just that. I don't have any sheep now, but I do have alot more time to do other things besides chasing after flyblown sheep!

For $20 or so I'd be willing to have a go myself. I guess I could just use a breadboard for the circuit board, if there's only a few components. I have built kits before and can usually solder things together without killing them. I just need the circuit and components layout so I know what to solder to what. A 10bit resolution would be about 3kg of a 3000kg range, correct? This would be fine as I only need a 50kg accuracy and don't think a pressure transmitter would give any better that 2% anyway.

LCD is fine. If we could use potentiometers instead of trimpots then it would make it easy to adjust with a knob. If you used only 1 trimpot would you still be able to zero the display? Wouldn't you need to be able to set an offset value as with the bucket empty there will still be a load on the hydraulics, so to get it to show zero you would need to subtract this from the reading. Then when weighing a load, you would again have to subtract the offset before dividing the trimpot value for scale? That's just how I see it anyway. Also you would have to calibrate for each different machine as offset and scale values will be different.

Yes the loader is a 24v system, though my forklift is 12v.

Bryan,

I'm still interested in you idea too!

Regards

Paul

 

[SABorn]

Hi Paul,

What i am thinking for tare is you lift the bucket (empty) to the reading position and press a tare button, this would write the current ADC reading to eeprom in the picaxe.

The program would simply subtract the value stored in eeprom from every ADC reading taken there after, giving a adc value for the load in the bucket.

The load adc value is then divided by the adc value from the trimpot giving a display value in kgs.

The trimpot is needed as apposed to a standard pot because it would require to be a precision multi turn trimpot for the accuracy needed to calibrate the display reading.
It would be a set once and leave alone setting anyway.
If it was to be used on more than one machine you could have several trimpots, each set for a particular machine and select in program which trimmer value the picaxe read for that machine, but doubt that would be needed if the same sensor was used as it would just be a different tare required per machine. (still an adc reading divided by X to = kgs.)

The advantage with writing the tare value to eeprom is it will be retained after power off and be read back into memory on power up.

I2C displays are not so easy to get and i create my own using a pic for a I2C driver interface, but there is no reason a lcd can not be driven direct from the picaxe and needs something like 8 wires to do this (ribbon cable) for this reason i would move up to a 20x2 picaxe for the higher number I/O pins to accomadate the requirements and it is a better chip overall (about $2.00 dearer)

For this project i see little advantage using a pic as the adc reading is the same because picaxe uses the internal adc of the pic the same as a standard pic would anyway.
Processing speed is a none issue here so no advantage there.

There is certainly no reason you could not vero board this project other than messy and larger in size.

I will put together a schematic for you to ponder over, but dont wish to spend hours desiging and testing things if you are likely to throw this in the too hard basket, as many often do here after the rest of us spend our time on their projects.

I do think you will get great plaesure out of having a go at this and much saturfaction using something you built.

There is a few hidden costs for a one off project like this that will push the budget up a little, like the need for a usb to serial cable to program the picaxe with, should your computer not have a serial port on it, (most later PC's dont have serial anymore)
Otherwise its a simple 3 wire cable needed to a serial port.

You will need to download a 40 meg program to use for picaxe programming but this is free software.

I too would not have maggot factorys (sheep) if i owned property, and it is highly likely we will post all out development of the rfid/weight system public on a different forum for others to follow as we have done already for many of our projects.
My view is to share information and designs free for others to follow as i have done with projects from other people.

The only design data for a home build rfid of animal tag reader on the net that i am aware of is the basic early design we created, and i have looked long and hard for information, and found nothing in 137khz half duplex range, just getting a copy of the animal tag protocol to work from with decoding the data was a difficult proceedure.

All up it was about $30.00 worth of parts and these things sell for well over $1000.00 for a hand held wand.

Pete.

 

[SABorn]

Hi Paul,

I made a few quick changes to an exsisting schematic that will show the overall basic layout required.

The display shown is not the one i would use as it is rather small but the wiring is the same other than the pinouts on a different display will be in a different order.

The voltage reg shown i would change to a LM317 as the 7805 has a 30 volt max input and is a little close to the 28 volt of a 24 volt system for my liking and the 317 has a 40 volt max input so more head room.

Otherwise its the basic system required, (subject to change as needed)

As you can see most of the circuit wiring is to the lcd in this case, but even that is simple enough.

Pete.

 

[paulmac]

Pete,

Yep, I think I can just about understand that. Certainly not a lot required to make it work. You only show 1 wire going to the Picaxe at B2 from the transducer. Is the second wire common ground, or does that go somewhere too?

Paul

 

[SABorn]

Hi Paul,

From the data sheet i viewed it is only a 3 wire sensor V+, com, output, so V+ would go to 24 volt+, common would go to Negative, and output to the picaxe input.

Data sheet here....... **broken link removed**


Pete.

PS:- check your private messages


Specs