ArduPilot Mega, Operational temperature range defined by the cheapest component?

Hello fellow enthusiast,

I'm currently working on a project where a UAV will be operating in a relatively quiet cold environment: less then -40 degrees Celsius. (possibly down to -60C or even worse)

We are currently considering using the ArduPilot Mega as the main unit for the flight control system, there are however some issues with the temperature range of some of components on the board.

First of all the most expensive components on the board, the ATMega1280 and ATMega328 both have a operational temperature down to -40C, as stated here:

http://www.atmel.com/dyn/resources/prod_documents/2549S.pdf

http://www.atmel.com/dyn/resources/prod_documents/doc8271.pdf

However, the multiplexer, 74LS157D, has a minimum operating temperature of only 0C! This could be significantly increased using a replacement IC from the 4000 series, specifically the 40257 that has a temperature range down to -55C. I am not sure if these are 100% compatible and was wondering if anyone could please confirm that. Alternatively one could use the military grade version of the component (54LS157) instead, although only some suppliers stock it.

Ti datasheet here: http://www.ti.com/lit/gpn/sn74ls157

and here: http://focus.ti.com/lit/ds/symlink/cd40257b.pdf

The final component I'm a bit unsure about, the LM317 for the voltage regulator. Some datasheets define the 317 having a range down to 0C, the 217 having down to -25C and 117 down to -40C. But there are also other documents stating that the temperature range is only dependent on the component packaging. I can't find anything similar to what is written on the back of the component on the board (92AG NO3B?), but there are several parts with the same dimensions that can handle temperature down to -40C, for instance the LM317AEM. So which one is on the board and could it be changed?

Is this really the case, or is it simply an other manufacturer that has other temperature ranges? Or using other packaging for the component?

I do realize that the board is not meant to be used in extreme environments, but the change of these components might still be a good idea. Especially the 74LS157, since the replacement is cheap.

Oh, and btw, we are going to have a lot of insulation and possibly some heaters, but to keep weight and power usage down we would like the components to handle a good temperature range (like the extended industrial standard).

Best regards,
David

Forget the logic chips, its the gyros that have to worry about. As log as they are initialized at the temperature where it will be flying you will be ok. If the imu is at not so cold when the flight starts and changes in flight, the gain and more so the offset or zero settings will change. I put an imu in my freezer with the USB calble thruogh the door seal, and initialized it. Then I took the unit out and watched it with Happy's IMU tester. You can watch the uncorrected yaw axis start rotating faster as the unit heats up. The APmega has enough extra power and inputs it could easily run a PID temperature controller in its' spare time. Put the unit in a foam box and use some high wattage resistors as heaters. Could easily be done with a little software and home made heater board.

Replies

ZHAO Zichao March 22, 2014 at 12:58am

Hi David S

Does the APM drifts with temperature? If so , do you know how can it fly at -22C?

/zzc
Greg Fletcher January 24, 2011 at 8:42pm

Forget the logic chips, its the gyros that have to worry about. As log as they are initialized at the temperature where it will be flying you will be ok. If the imu is at not so cold when the flight starts and changes in flight, the gain and more so the offset or zero settings will change. I put an imu in my freezer with the USB calble thruogh the door seal, and initialized it. Then I took the unit out and watched it with Happy's IMU tester. You can watch the uncorrected yaw axis start rotating faster as the unit heats up. The APmega has enough extra power and inputs it could easily run a PID temperature controller in its' spare time. Put the unit in a foam box and use some high wattage resistors as heaters. Could easily be done with a little software and home made heater board.
Bill Sanford January 24, 2011 at 8:05pm

I flew my quad last weekend at -7F with no issues, but I wasn't out for a long period of time.

I'd say a lot of testing would be in order for sustained low temp operation...
Developer

Randy January 23, 2011 at 4:41pm

Slightly anecdotal evidence but I know of one person who flew last weekend in -22C without any problems.
- ZHAO Zichao > Randy March 22, 2014 at 1:05am
  
  Hi Randy
  
  Is there temperature control in APM? If there is, what is the heater? If there is not, why the copter doesnot drift at -22C?
Uwe Gartmann January 23, 2011 at 1:59pm

Hi David,

Its not only the electronic component. Its the PCB, cabling, housing and many other things as well. Have a look at the industrial electronic system suppliers. You can't compare such stuff with ArduPilot hardware.
David S January 23, 2011 at 3:14am

Might have found the answer to this myself when I bought the hardware. Seems like at least the multiplexor is of a different type then the old 7400-series one, the news one with a temp range down to -55C. Going to do some tests soon. talk back later.

This reply was deleted.

Discussions

ArduPilot Mega, Operational temperature range defined by the cheapest component?

Replies

Activity