PDA

View Full Version : 50 quid Cell Checker



notdeadyet
12-05-2013, 07:49 PM
I promised a few people I would do some pics and a write-up.

Since the demise of Teledyne cells in the rebreather market I've been toying with the idea of buying a cell checker. However I've got oddball cells in my unit which don't fit any of the connection options on commercially made units, they are fatter than regular cells so wasn't sure if they would fit the chamber and as they are high output then they'd be off the scale of a standard 0-200mv DPM LCD display. I also didn't fancy paying 300-500 for a pot either.

So after a year of apathy after my last Teledyne cells died I finally got round to building one. The design is nothing fancy, it's not a million miles different to many of the units shown on CCRX or RBW. The basic chamber is a 10" water filter houseing which is rated to 135psi, plenty for what I needed. It has a built in purge button and brass 3/4" NPT ports.

First step was to reduce the ports to a more manageable 1/4" NPT (I've got shitloads of 1/4" stuff in my spares box as a lot of KISS and Mk15 connections are based on it). Next I added a 0-2.5bar gauge to one side to monitor the pressure in the chamber. On the other side I added a t-piece with an oxygen add port and a gland for the cable to the cells. The oxygen port is just a push on hose barb with an old piece of LP hose cut down and connected to a spare first stage. I may swap that to a quick release in the future to save tying a first stage up permanently. I should really dial the IP of the reg down to below the rating of the chamber.

The cable gland is just a 1/4" Swagelok tube connector as used on Hammerhead handsets and other bits of kit. It's basically just a compression fitting with a piece of cable rather than a pipe. The cable itself is 4-core telecom cable from Homebase. Once in place the whole fitting was potted with a bit of hot melt glue.

Connection to the cells is via some mini crocodile clips on the cell connectors. As it's a 4 core cable they share a common negative (as they do in the rebreather anyway). I was having trouble with gas leaking through the sheath so I sealed it with some more hot melt, it's ugly but does the job.

Cell monitors are just some cheap multimeters on the 2000mv setting. The test leads are spliced on to the cell cable. It isn't as pretty as a NarkedAt90 but it is a 6th of the price:

http://i113.photobucket.com/albums/n233/lizardland/2013-05-12194650.jpg

Operation is dead simple. Connect the cells, let them sit in air and take the reading. With these they were in the region of 20-22mv. A bit of maths would say they should read 95-105mv in a pure O2 environment. Screw the housing on to the lid and pressurise to 1bar and then purge back to ambient. I thought about adding a needle valve but the cylinder valve gives fine enough control. Obviously the atmosphere in there is now half air and half oxygen so I repeated the flush and purge cycle until I started getting the predicted value on the multimeters. It took about seven flushes to get the readings right. The cells were showing 102-112mv with the pressure at ambient which suggested the atmosphere inside was now more or less pure oxygen.

A bit more maths and I should get 190-210mv when I up the pressure to 1bar (10m depth). I brought it up, let it settle and the cells were showing 195-216mv which I'm happy with. It isn't dead on but equally there is a bit of tolerance on the gauge so it probably isn't dead on 1bar.

It follows that 2bar (or 20m depth) should give 285-315mv. Bring the pressure up again, let it settle and the cells are showing 298-320mv. Again, fine by me.

Purge the chamber back down to ambient and let it settle just as a final check in oxygen at 0bar (surface pressure). Final final check is to make sure they settle back to their original readings in air with the chamber open (which they did).

So...

Filter housing: 13 from Ebay
2x3/4"NPT reducing adapters: 6 from Ebay
1/4" NPT Cable gland/compression fitting: 3 from Ebay
T-piece: I already had but I think is in the region of 7 from Swagelok
Cable: 1 from Homebase
Crocodile clips: 2, Ebay
3xMultimeters: 9, Ebay
Pressure gauge: 15 from a pneumatics supplier
BSP/NPT adapter for the gauge: 3 from same place as gauge
Hose barb: already had but about 3 from Swagelok

OK, I've gone over my 50 quid budget (and doesn't include the first stage) but not by much. It does the job surprisingly well and easily. I think my next step might be to knock up a spreadsheet to track the results and predict the values in oxygen at ambient, 1 & 2 bar. Might also include something for gases other than O2.

It is a very easy project, easy enough that everyone with a rebreather should have a cell checker.

And if anyone mentions the Little House on the Prairie table cloth then you can get fooked.

Major Clanger
12-05-2013, 07:55 PM
Lots of useful stuff.

Can't green you but would if I could...

Capt Morgan
12-05-2013, 08:12 PM
Thanks NDY, I have been following your building of this
and have soruced most of the parts to copy it.
Just one thing I was thinking was if it would be better
to actually read the pO2 as opposed to the mv ?

notdeadyet
12-05-2013, 09:04 PM
Thanks NDY, I have been following your building of this
and have soruced most of the parts to copy it.
Just one thing I was thinking was if it would be better
to actually read the pO2 as opposed to the mv ?

You could do either I suppose, the numbers ultimately have the same source. I used mV as I can compare it to the data sheet of the cell and it's just easier to use a multimeter. Probably cheaper too :D

BTS
12-05-2013, 09:07 PM
Can you put up a bigger picture... I like pictures....

Ta very much like...

DiverMike
13-05-2013, 07:10 AM
I wonder if with a bit of know how the 3 multi meters could be replaced with a raspberry pi (25) and then you could watch the traces respond to the pressure. Even better still if there is a pressure gauge with digital out. Nice bit of kit - I'm tempted to make one.

LearnerDiver
13-05-2013, 08:26 AM
I wonder if with a bit of know how the 3 multi meters could be replaced with a raspberry pi (25) and then you could watch the traces respond to the pressure. Even better still if there is a pressure gauge with digital out. Nice bit of kit - I'm tempted to make one.

Take the three cell outputs, run them through op-amps to raise them to the range handled, fire them through an Arduino, some calcs and output all 3 on an LCD/OLED screen; just like the prototype rebreather HUD I built that works in this fashion but outputs the values to RGB LED's then via fibre optics to the HUD rather than a screen but the principle is the same

You could add a pressure sensor inside the unit like a BMP085 Barometric Pressure Sensor and read the pressure and temp and use them for calcs or in the display

I'll dig out some bits and post info on them later on

nickb
13-05-2013, 09:31 AM
This is great stuff

LearnerDiver
13-05-2013, 06:49 PM
Take the three cell outputs, run them through op-amps to raise them to the range handled, fire them through an Arduino, some calcs and output all 3 on an LCD/OLED screen; just like the prototype rebreather HUD I built that works in this fashion but outputs the values to RGB LED's then via fibre optics to the HUD rather than a screen but the principle is the same


You could add a pressure sensor inside the unit like a BMP085 Barometric Pressure Sensor and read the pressure and temp and use them for calcs or in the display


I'll dig out some bits and post info on them later on


I'll assume you want to use an Arduino or clone to measure and display the voltages, and then optionally convert to pO2, mainly because I use Arduino's & clones and know nothing about doing this on a Raspberry Pi but assume you will come up against the same issues


I'll also assume that the cells you are using are the same as the ones used above as different cells output different milli-voltage ranges so you may have to play with op-amps if their output is very low


There are some issues with measuring milli-voltages on a Arduino;


Arduino measures analogue voltages in the range 0 to 5v in 1024 steps so that means we can only easily measure to accuracy of 5/1024 = 4.8mv so we can only measure in 4.8mv steps so thats no good for our cells as we want better accuracy than that


We could run the cell voltage through an op-amp and boost it say 10 times so our 0-400mv input becomes 0 to 4volts which gives us steps of 3.9mv after amplification which is 3.9mv/10 = 0.39mv resolution which is ideal and then we can divide the value read by 10 to get our actual reading, downside to this is we need to build an op-amp circuit


Other option is to use the internal 1.1v reference which means no op-amp to build just a bit more maths in the code and it gives us a resolution of 1.0742mv which is near enough 1mv as makes no difference


So;


Take the three cell outputs and wire into three Arduino inputs

Wire a pressure sensor into the Arduino


Measure the voltages and pressures and using the reference voltage calculate exactly what the voltages are

Store the calibrated values and pressures


Display them


Using the voltages calculate the three pO2 readings;


100 / (Calibrated mv in 100% O2 at atmospheric * Pressure in bar) / mv measured


This assumes voltage output is absolutely linear, if it isn't then you could have a calibrate routine and do the maths better by using a lookup table of the actual values measured during calibration or extrapolate between calibration values


Display them


Bear in mind my rebreather experience so far is a try dive on a Meg and a day out with Simon TW, my electronics is mainly self-taught, and my maths is crap.....

I think some cells may need a 10k ohm resistor across them to create a load

If anyone has a spare cell and is happy to lend it to me for a couple of weeks I'll knock up a board with display capable of doing the above or I'll pinch the cell out of my Analox Analyser and build one based on that

trebor
13-05-2013, 06:52 PM
LOve it stuart, Nice one.

ebt
15-05-2013, 10:55 AM
Its amazing what you can achieve when someone removes the distractions.......

http://s3-ec.buzzfed.com/static/imagebuzz/web04/2009/6/3/15/buckfast-wine-17258-1244057745-17.jpg

notdeadyet
15-05-2013, 11:38 AM
Its amazing what you can achieve when someone removes the distractions.......

http://s3-ec.buzzfed.com/static/imagebuzz/web04/2009/6/3/15/buckfast-wine-17258-1244057745-17.jpg

Most of the parts came from my home distillation project.

Sent from my GT-I9100 using Tapatalk 2

ebt
15-05-2013, 11:50 AM
looking at the pictures, i wonder if you could fit a modified mark munro board inside the chamber.....

Major Clanger
15-05-2013, 12:20 PM
If anyone has a spare cell and is happy to lend it to me for a couple of weeks I'll knock up a board with display capable of doing the above or I'll pinch the cell out of my Analox Analyser and build one based on that

I have one you could have. Just swopped it out but still read over 1.7 on last lim check . PM me your address.

scubadogct
15-05-2013, 12:31 PM
I'll assume you want to use an Arduino or clone to measure and display the voltages, and then optionally convert to pO2, mainly because I use Arduino's & clones and know nothing about doing this on a Raspberry Pi but assume you will come up against the same issues


I'll also assume that the cells you are using are the same as the ones used above as different cells output different milli-voltage ranges so you may have to play with op-amps if their output is very low


There are some issues with measuring milli-voltages on a Arduino;


Arduino measures analogue voltages in the range 0 to 5v in 1024 steps so that means we can only easily measure to accuracy of 5/1024 = 4.8mv so we can only measure in 4.8mv steps so thats no good for our cells as we want better accuracy than that


We could run the cell voltage through an op-amp and boost it say 10 times so our 0-400mv input becomes 0 to 4volts which gives us steps of 3.9mv after amplification which is 3.9mv/10 = 0.39mv resolution which is ideal and then we can divide the value read by 10 to get our actual reading, downside to this is we need to build an op-amp circuit



Actullay audrions AD's are 10 bit. (2^10) -1 = 1023 bits of resolution =4.8mV step = 2.44mV max error
Over a range of 0 -5V I think that is more than enough.
over a range of 0 - 100 % o2 that is a maximum scaled error of 0.04 % 02

An op amp will add slew - best avoided.

If you feel you do need a higher resolution then use a better chip - audrinos are for education.

An arm core or a pic can offer 16 bit resolution = 65535. = 0.076mV steps = 0.038mV max error = 0.00077 % o2
32 bit is way over kill for this application.

Of course I have assumed that the cells output is 0-5v.
If it is less just use the an external or internal reference voltage.

ebt
15-05-2013, 12:56 PM
http://www.mipex.eu/blog/wp-content/uploads/2011/10/MG_0045-by-h3nr0-on-Flickr.jpg

:D


doc, your cell range is a wee bit off mate. usually 7-13mv in air (or 24mv for high output jobbies).

gobfish1
15-05-2013, 03:50 PM
Don,t know much about printed circuit boards and the like , but my first analyzer was just a cheap multimeter with a resistor swap . feek i payed 100 quid for a 5 quid multimete and a 50p resistor , Mr Gurr seen me coming lol ,

thanks for posting , nice work ,

ps if you had a shearwater you could just plug that in and bingo , but seeing your a cheap tw&t its not a bad job .

scubadogct
15-05-2013, 09:40 PM
http://www.mipex.eu/blog/wp-content/uploads/2011/10/MG_0045-by-h3nr0-on-Flickr.jpg

:D


doc, your cell range is a wee bit off mate. usually 7-13mv in air (or 24mv for high output jobbies).

Then you will need to use the internal or external vref.

Not a rebreather man - just an electronics geek.

Mike_123
16-05-2013, 10:33 PM
Good job - definitely a gadget worth having imo.

Here's my spreadsheet (http://www.filedropper.com/o2sensorchecker) if you want something to work from rather than starting from scratch.
Just type the mV values into the yellow box.

http://i44.tinypic.com/723dqp.jpg

_cys
17-05-2013, 08:09 AM
Nice cell checker. I like the multimeters. I was going to say something about the tablecloth...

One thing I like doing with my home-brewed cell checker is running the cell outputs from the checker into the rebreather electronics. I can check the mV and ppO2 readings using both my Shearwater secondary and HH primary. I picked up a few extra Winchester connectors to make it happen for the R10DS equiv cells. I like getting info on the rig and the cells at the same time.

notdeadyet
17-05-2013, 08:59 AM
Nice cell checker. I like the multimeters. I was going to say something about the tablecloth...

One thing I like doing with my home-brewed cell checker is running the cell outputs from the checker into the rebreather electronics. I can check the mV and ppO2 readings using both my Shearwater secondary and HH primary. I picked up a few extra Winchester connectors to make it happen for the R10DS equiv cells. I like getting info on the rig and the cells at the same time.

My next step was to rig some leads into the centre section so that I could also calibrate the HH and the secondary at the same time.

The tablecloth is not my choice, agreeing to it makes life easier :)

Sent from my GT-I9100 using Tapatalk 2

Major Clanger
17-05-2013, 12:05 PM
.

The tablecloth is not my choice, agreeing to it makes life easier :)

Sent from my GT-I9100 using Tapatalk 2

What's the carpet like? My brutal honesty in these matters may smart a bit ;)

notdeadyet
17-05-2013, 12:07 PM
What's the carpet like?

Are you speaking euphemistically?

Sent from my GT-I9100 using Tapatalk 2

Major Clanger
17-05-2013, 12:11 PM
Are you speaking euphemistically?

Sent from my GT-I9100 using Tapatalk 2

Lol, shagpile or laminate...