help on Open bathroom scale project

Go To Last Post
25 posts / 0 new
Author
Message
#1
  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Hello :-)

I am fat and I am being looking for a way helping controlling myself. In recent past I hacked 2 cheap bathroom scales for log my weight on SDCard, and I am being making a graph of time VS weight, me and my girlfriend, we are using it.

Now I found a cheap (20€) scale on local shop that also measure body fat, water and muscle :-) -- I started hacked it but is far from the end... and I though that maybe my energies would be better to invest on building a scale from the start, as an Open Source project.

I want to make simple and cheap as possible, DIY bathroom scale. It will need electronics + mechanical construction.
Mechanical may be difficult for me but now that I have a RepRap 3D printer, I can print plastic objects and use nuts and bolts in conjunction just like RepRap Mendel.

So I am looking for guidance on some cheap hardware/sensor for making the scale. Please help me start this project, because later I can/know how to use some cheap ARM + cheap LCD + SDCard.

I am looking to build something modular, for later try to add some kind of wireless and user be able to get the data on his mobile PC/android. Then some other hackers can develop application to show the graph in real time. Also other sensor as for body fat can be added to the system.

I had looking at google, there is none Open Hardware bathroom scale yet, and many hackers are fat :-) -- people like me needs a smart bathroom scale to help us getting better shape and health.

Thanks in advance.

JPCasainho,
www.Casainho.net
.Portugal

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

I built something similar for my wife a while back (she's not fat)!!

I used a scale I bought and took out the controller and the lcd...
Only thing left was the glass platform and the 4 sensor legs which were surprisingly linear in response.
I'd guess whatever mechanical lever translated the force to the sensor was the reason for the linearity...

Anyhoo, I replaced the electronics with a Mega32, SD Mem, DS1307 RTC, FT232 USB, blue 128x64 LCD, Boost SMPS and an AD623 instrm. amp...

The code identified the user by weight, stored it to txt file on SD and then graphed the result over the last month or so...
The 512mb SD card is not likely going to fill up anytime soon...

I thought about removing some of the paint under the glass and adding a solar cell and a supercap.. Perhaps an RF PC link would be a nice feature as well..

I can post the sch and code if you like...

Michael

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

krazatchu wrote:
I built something similar for my wife a while back (she's not fat)!!

I used a scale I bought and took out the controller and the lcd...
Only thing left was the glass platform and the 4 sensor legs which were surprisingly linear in response.
I'd guess whatever mechanical lever translated the force to the sensor was the reason for the linearity...


So that's why I would prefer to try build an Open Source project, so we can go ahead instead of "build the wheel" everytime :-)

I appreciate your offer for the schematic and code, however I am looking for a way to build the sensors and all the mechanical, to not be dependent of some model of commercial scale.

JPCasainho,
www.Casainho.net
.Portugal

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

I looked into building sensors before I found a cheap scale at the store...

I couldn't have bought the sensors for the price of the scale ($15)...

You can however, build the sensors using a long zigzag trace on a pcb glued to a flexible piece of steel..

The sensors are wired in series to a bridge and an amp...
Using 4 sensors for each foot, the summing is analog...

My project used only the glass platform and the sensors, the electronics were completely rebuilt..

Michael

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Here's a pic...

Attachment(s): 

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

The sensors, I'm pretty sure, are called strain gauges...
http://en.wikipedia.org/wiki/Strain_gauge

They are bonded to a flexible metal arm that is tapered.
It is the taper that should make output linear to input.
A LUT can be used if non linear.

Michael

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Here's a good vid from nerd kits that lays out the general idea...

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

Michael

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Do you have any idea on how I could build the sensor on a construction based on M8 bar bolt, nuts and plastic pieces? -- like Mendel: http://objects.reprap.org/wiki/M...

JPCasainho,
www.Casainho.net
.Portugal

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

I wish I had time to jump into this. For years I've wanted to design a scale that would send the weight wireless to my PC but I've never gotten around to it.

Let me repeat what was said above, it is unlikely that you will ever be able to build a sensor cheaper than you can buy a scale and hack it. I understand that you want to do your own mechanical work so that it will be fully open source, but in my experience (I tore apart several scales a few years ago) the internal designs are very similar. All you have to do is tie into the strain gauge and then figure out a way to calibrate it. One company I visited that made scales had a way of piling on gallon water bottles as their weight standard - clever and cheap.

Anyway, good luck and I'll be trying to keep up with your progress.

Smiley

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Exactly what I used to check the sensors on the scale I used...
Measured amounts of water...

Michael

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

I hacked four scales to weigh the corner weights on my go-kart. As previously mention it is hard to beat the cost of the mass produced scales in regards to the platform and the strain gauges. For the micro, I usedan Analog Devices AduC834 which has an inbuilt 24bit a/d converter that basically wires directly to theuV output strain gauges. There's also a number of 24bit a/d devices that can be interfaced to the AVR to do the same job.

Can't go too far wrong with water - 1L = 1kg at standard temp and pressure.

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

I would be pretty embarrassed using an Open Bathroom...

John Samperi

Ampertronics Pty. Ltd.

www.ampertronics.com.au

* Electronic Design * Custom Products * Contract Assembly

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

js wrote:
I would be pretty embarrassed using an Open Bathroom...
... not to mention the trauma done to the passing observer(s) ...

Ross McKenzie ValuSoft Melbourne Australia

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

smileymicros wrote:
I wish I had time to jump into this. For years I've wanted to design a scale that would send the weight wireless to my PC but I've never gotten around to it.

Maybe I will start creating a group on code.google and another on Facebook. I need to get help on ideas to start building a reliable and cheap sensor/electronics. The idea is to use easy local materials. We need to be able to build using local and cheap technological.

smileymicros wrote:
Let me repeat what was said above, it is unlikely that you will ever be able to build a sensor cheaper than you can buy a scale and hack it.

I understand and I think is always like that. But there are a lot of projects that are more expensive but people prefers them anyway.

This problem of fat is a big one for people like me, I had spent much money on few devices. There is a scale on market that log weight to a web page, but is not Open and is a expensive monthly service $$.

JPCasainho,
www.Casainho.net
.Portugal

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

I think I'd start on that bathroom scale with a little thermite, just 'cause I'm not allowed to play with C4.

The largest known prime number: 282589933-1

It's easy to stop breaking the 10th commandment! Break the 8th instead. 

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

I have just taken my scales apart and there are 4 sensors, one in each corner, each with 3 wires. It looks like they are wired together somehow - anyone any idea what the circuit diagram would look like ? All the stuff on the internet has 4 wire sensors. It all goes into the single black blob (ASIC I assume).

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

I dare say they're half gauges. If you look at the pcb how they're wired together, you'll find they form a bridge (wheastone bridge). From that you'll need a high resolution analog to digital converter or a high-gain op-amp stage and a 'normal' ADC.

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Just got the meter out and tested it, you are spot on. Each of the strain gauges in each corner has 3 wires, red, blue and white. Red to blue and red to white measure 1000 ohms. Red to blue increases resistance when pressure is on the scale, so must be tension, and red to white decreases, so must be compression.

They form a bridge with the reds being the exciter and outputs, and the whites and blues connected to form the bridge. So I need to build an op amp amplifier and get my ADC to read the output.

BTW the scales are the cheapest available on ebay - branded Fitness Pro.

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

When you are done with this hacking, I need one that displays only 0 and 140 lbs.

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Beermaker, its probably easier to get a high resolution adc rather than using an opamp and fiddling the gain. Something like a LTC2440 or similar. All it needs is a few wires for the spi interface to the AVR and the rest is done in software. The strain gauge connects directly to the chip and the excitation is got from the 5V power.

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Most bathroom type scales do well to accurately measure in 1/2 pound increments. At 10-bits that would let you weigh up to 500 pounds in 1/2 pound increments. I suspect you'd have to use some pretty outstanding mechanics to get anything more accurate in the lever system so I'm not sure why you'd need more than the 10-bits that come with the AVR. The 12-bits on the Xmega would theoretically give you 1/4 accuracy, but I doubt if a reasonably inexpensive lever mechanism would actually support that accuracy.

Smiley

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

I had a play with an old op-amp and soon gave up. Just got a TI ADS1232 demo board and it is fantastic, plugs right into the strain guage and with my £10 scales gives sub-1g accuracy. I have a ADS1230 sample (free from TI) which I will try next, this is really the way to go.

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

With weigh scale applications, the dynamic range comes into play as you usually have the zero offset to contend with. No doubt with some analog work and you can feed it into an AVR but just getting a fairly cheap 24bit adc makes the hardware easier and the zero offset and scaling are done in software. You probably only end up with around 14bits of actual reading as the lower bits just become noise. Of course , with the AVR, you have eeprom to store the offset and scale values.

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

hello,every one,anybody have the RCT DS1307+12864LCD source code,i'm a beginner of AVR.

waitting reply
hqjin87

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

hqjin87 - you've just hijacked a thread. so you might be waiting a while. Since you're a beginner, try a search of the projects to see if anything resembles what you want.