Need Advice - DIY Sea Turtle Nest Temperature Logger

I just finished building a temperature monitor system that modified to log instead of communicate via the LCD and wireless...and then it would be less than $40 (of course the battery power and waterproof case would have to be figured out)

It uses this temperature sensor which is as accurate as you requested..and you could save even more by assembling the temp sensor yourself

Here's the project

The Complete Channel One Temperature Monitor and Alarm Project - The Ultimate Raspberry Pi Bundle

The OP won't be able to buy enough battery to power an RPi for 2 months for $40 !

@diverdan - what electronic skills do you have - can you solder parts to a pcb and write software ? (if not can you get a fellow student who can to help ?)

How many loggers do you need ?

If you can the cost of the bits for a temperature logger can be less than £5 ($10). How many do you need to make ?

The cheap way of waterproofing is to put he board and battery in a little plastic box like these from Empty Film Canisters For Sale. Buy Film Canister Rockets Bulk

(sorry *%$£?!!! E14's editor has made the link look silly but it still works)

If you think you dare go the home build route then let me know and I'll make more specific component suggestions.

MK

You'll have to do some calculations on current consumption vs. battery needed, and wat the temperature is you want to measure. The photo film canisters are a great idea (also for later data extraction!), but you might want to stick the sensor to the outside wall to get faster temperature transients.

Some hints on components if you want to design yourself:

- Use FRAM; although a bit more expensive than EEPROM, you have very short access time, and thus longer battery life. Although it's called FRAM, it's non-volatile.

- If you're going to do microcontroller development take a good look at all power off modes. In general, you'll want to do a measurement, store the result, and power off until the next measurement. Choose a microcontroller that has features for this. MSP430, or EnergyMicro are specializing at this and offer many feautures / documentation to accomplish your goals.

On the cost issue: how many would you need? And will your hours be for free? If you need a great lot of them, assembling will cost you a lot of time....

You can edit the way a link looks by mousing over the link. It defaults to using the title of the page you're linking to

Thanks, I hate disguised links !

MK

Thanks for all the support guys. I really appreciate the contribution y'all are making by at least with regards to brainstorming.

Great concept Tim, but it might be a little bit more advance to what I would like to accomplish.

With direct response to Michael, I do have some experience with breadboarding and soldering. So that in itself wouldn't be an issue. I've never really had a hand in coding, but I do have some potential resources around me that might be willing to help out.

I'm still working out the kinks to the whole project design, but my projected budget that the university is willing to provide is US$500. That severely limits my options, but I am willing to personally contribute to this research to ensure it gets completed. If I could get 50 data loggers at US$10 a pop, I would be ecstatic. Realistically, I'm thinking that 20-25 data loggers will suffice, and I will have to utilise each data logger twice to get a decent enough sample size.

The film canisters are a great idea, as I don't want to disrupt the sea turtle nest in itself while collecting these measurements.

vsluiter:

From what I understand, most of these devices use a coin cell battery, which should be able to power the device at least 6 months, using 1 minute intervals. I would be collecting these measurements at 1-4 hour intervals; it hasn't been decided yet. This is in turn would increase the length of the battery life.

The temperature collected would be around 28-30 degrees C, which is 82-86 degrees F.

Hi Diverdan,

Nice project... If you want it to succeed, think about what you need first (sample rate, needed accuracy, how to get data off the device), and then start developing. Changing specs along the way is a surefire way to go outside space / time / money budget.

By the looks of your numbers you want to have useful data between 28 and 30 degrees; with 0.5° resolution, that would only give 5 values. Do you need a larger range, or a more precise sensor? I've once made a circuit that was measuring within 0.1° Celsius over a range of -5 to 40°C, and in hindsight I think I could better have bought one of these sensors: http://www.ist-usadivision.com/sensors/temperature/tsic-sensors.php

The reason I warn about battery usage is that it's very hard to get to low power usage if you don't take it into consideration from the start of your product design. You can easily get this started while not paying attention to it, but it'll be much harder to get to a lower consumption from that point, than it is to first start at very low power consumption, and gradually add features. Then you'll also be able to see what feature is affecting the current consumption.

Read this: Making the Arduino sleep the long sleep and this: https://www.sparkfun.com/tutorials/309 for more info on making Atmels sleep.

Good luck with your design, it would be nice if you kept us posted!

Here are some component ideas:

Temperature sensor: TI TMP112AIDRLT (Farnell 1735574 @ £1.41) - 10uA active, 1uA shutdown, 0.0525C resolution, typical accuracy 0.1C worst case 0.5C.

Micro: LPC811M001 (Farnell 2320693 @ £0.56), ARM M0, 16 pin, low power modes, cheap

Real time clock: Microchip MCP7940N (Farnell 2361116 @ £0.49)

EEPROM: Microchip 24AA128 (Farnell 1331285 @ £0.43)

Battery: CR2450, about £1.50 (don't buy junk)

Battery holder, about £0.25.

Each time the recorder wakes up to measure it must be on for about 100mS to measure temperature and will use max 2mA, then write to EEPROM for 5mS using 5mA = 0.225mAS.

If it does once per minute the average current drain is 3.75uA.

Allow the RTC to use 2uA (way more than spec) gives an average current drain of 5.75uA

The battery is good for about 500mAH (depends on which you get) so the battery will do 86,000 hours.

You might be tempted to use a smaller battery but aim for 10x the calculated capacity - you really don't want to run flat for the sake of 25p.

You can get the data out of the uP by a uart interface using one of those FTDI logic level uart to USB cable thingies which work with almost any Windows PC.

By the time you get the pcb, crystal etc I reckon the total materials cost will be about £9, about $15.

Sounds such fun I quite fancy making one myself but I'd better get on with some real work

MK

Nice breakdown!

Since they've got such good documentation I'm advocating the use of the Zero Gecko with internal RTC (EFM32ZG108F160.9 μA @ 3 V Deep Sleep Mode, including RTC with 32.768 kHz) and replacing EEPROM by FRAM. But of course that's personal favourism and some overoptimization, Michaels setup seems very good too.

The Gecko has the RTC on chip so would save the external RTC and so the BOM would reduce a bit - which is good. I picked the NXP LPC800 part because it's in an easy to solder package (and I didn't know about the Geckos - never used them).

I think the Gecko is a better choice than the NXP part if the package is OK - I don't mind QFNs but I know that some hand solderers can't manage them.

I chose the EEPROM because I wanted everything to work down to 2V (you can't trust coin cells to give you any more over their life and a reasonable temperature range) - have you a part number for a FRAM ?

MK