Good Morning! I am trying to implement the project highlighted here:

http://www.nerdkits.com/videos/weighscale/

It is simply taking apart a cheap digital scale so that the strain gauge can be read by an MCU. In the video he uses an AD620 instrumentation amplifier to amplify the signal before it goes into the MCU. I went on Mouser and looked up the AD620 and they cost about $10 a piece. I did a search for "instrumentation amplifier" and got results as low as $2 a piece. I'm curious, what are the parameters that are most important in an instrumentation amplifier? My circuit will be running on 5V and according to the video he has the gain set to 150. I am completely ignorant of amplifier circuits so I apologize for my confusion. If the difference between the $10 amp and $2 amp is irrelevant to me, obviously I would prefer to spend $2 or at least <$10.

 

It is many things to consider like supply voltage. Most Instrumentation Amplifiers are made for dual supply voltage. For the sake of a hobbyist project you could try to build one your self. But you will find it is some problems with such amplifier. This is outlined in section 3. (see link)
http://www.analog.com/static/imported-files/design_handbooks/5812756674312778737Complete_In_Amp.pdf

 

I use the AD623 IA for load cell work, it designed to work from a single supply, from 3V thru 12V.

What weighing accuracy and range are you designing for.?
E

 

It is many things to consider like supply voltage. Most Instrumentation Amplifiers are made for dual supply voltage. For the sake of a hobbyist project you could try to build one your self. But you will find it is some problems with such amplifier. This is outlined in section 3. (see link)
http://www.analog.com/static/imported-files/design_handbooks/5812756674312778737Complete_In_Amp.pdf

That looks like a great link. Thanks!

 

I use the AD623 IA for load cell work, it designed to work from a single supply, from 3V thru 12V.

What weighing accuracy and range are you designing for.?
E

It is part of my "Aquarium Automation" project. Basically I want the major steps in a water change to be automated. The only step that I need to be involved in is adding the dry salt mix to the saltwater mixing reservoir (35 gallon trash can). Everything else will be handled by a system of pumps, float switches, and an Arduino Mega. I currently have the Arduino handling temp control, water topoffs, and the freshwater filter (using a solenoid valve).

For adding salt, I could add/check/add/check/etc. until I get the right salt concentration…..or….I was thinking that I could just place the entire saltwater reservoir on a scale. Weigh the reservoir before I add the freshwater, weigh it after I add the freshwater, and then perform a calculation that will tell me how much salt to add by volume to get the proper concentration. Taking it one step further, the Arduino could continue to weigh the reservoir as I add the salt and it could just tell me to keep adding more salt until the proper amount (by weight) has been added. That is the thought at least. Obviously I would check the concentration using a refractometer at the end.

 

A US gallon of fresh water weighs 8.34 lbs, so a 35 gallon would be 291 lbs.

What typical weight of salt would you add to say 35 gallon, reason for asking is the weight resolution you are requiring to measure too.?
ie: what would the Gross weight be of salt+water+tank.?

 

A US gallon of fresh water weighs 8.34 lbs, so a 35 gallon would be 291 lbs.

What typical weight of salt would you add to say 35 gallon, reason for asking is the weight resolution you are requiring to measure too.?
ie: what would the Gross weight be of salt+water+tank.?

35 gallons would be roughly 10.5 lbs of salt mix. Resolution down to an ounce or two, if possible, would be perfect.

The weight of the reservoir is negligible (it's just a cheap rubbermaid trash can). The water is 291 lbs like you said, but it would actually be slightly less because it isn't filled completely to the top. I would say just over 300 lbs all in.

I think I see where you are going with this. If I want 1 ounce resolution, I would need to set it up for the micro controller to be able to distinguish 4800 steps (300 *16).

Given that the ADC on the Atmega micro controllers is 10 bit, the max we can get is 1024 steps. So really assuming we are using the absolute 0 - 5V range, the max resolution that we could get would be about 5 oz. I could live with that.

Am I thinking this through correctly?

 

Why would you be adding salt to water replacing loss thru evaporation?

 

Why would you be adding salt to water replacing loss thru evaporation?

This isn't part of the water-topoff system. This is the water-change system. It is for mixing up new saltwater for when I perform a water change. Old salt water will be pumped out of the aquarium sump and new saltwater will be pumped in from the saltwater mixing reservoir. The water-topoff system pumps water in to the sump from a freshwater reservoir to replace water that is lost through evaporation.

 

I think I see where you are going with this. If I want 1 ounce resolution, I would need to set it up for the micro controller to be able to distinguish 4800 steps (300 *16).

Given that the ADC on the Atmega micro controllers is 10 bit, the max we can get is 1024 steps. So really assuming we are using the absolute 0 - 5V range, the max resolution that we could get would be about 5 oz. I could live with that.

Am I thinking this through correctly?

hi,
Your above quote confirms my doubts over the resolution you were expecting, the 8 ADC will give only 1024 values within the PIC.

The way around this is to 'Tare Off' the 291 lbs of fresh water, this could be done by injecting an offset voltage into the bridge, so that the 291 would give a zero voltage output.

Then you would have 1024 stepped values for the increase in salinity.

 

hi,
Your above quote confirms my doubts over the resolution you were expecting, the 8 ADC will give only 1024 values within the PIC.

The way around this is to 'Tare Off' the 291 lbs of fresh water, this could be done by injecting an offset voltage into the bridge, so that the 291 would give a zero voltage output.

Then you would have 1024 stepped values for the increase in salinity.

Ugh…..that sounds complicated. So do I just include a variable resistor on one leg of the bridge to zero it out before I add the salt or is there some super snappy way to do it with the micro controller? How does the tare feature work on commercial digital scales?

 

Ugh…..that sounds complicated. So do I just include a variable resistor on one leg of the bridge to zero it out before I add the salt or is there some super snappy way to do it with the micro controller? How does the tare feature work on commercial digital scales?

Its not really complicated, a 20 turn Cermet pot and a couple of 0.1% resistor will get the job done.

Fill the tank with fresh water, say 35 gall, adjust the trim pot to cancel out the Tank+water weight, so that the ADC reads zero.

You add salt, I guess you already know the correct water to salt ratio for the required salinity, As you add salt, stir the mix and observe the weight until reaches the desired weight.

You could calibrate the tank volume/gall.

The 'Tare Off' does not have to be full to 35 gall, you can Tare off any weight.

Do you follow OK.?

How does the tare feature work on commercial digital scales?

Its usually done in software, pressing a Tare Button, but its also possible use the method we are discussing.