Pages: [1]   Go Down
Author Topic: No Serial communication (not 10k resistor problem)  (Read 1019 times)
0 Members and 1 Guest are viewing this topic.
0
Offline Offline
Full Member
***
Karma: 0
Posts: 121
View Profile
 Bigger Bigger  Smaller Smaller  Reset Reset

I have pretty simple accelerometer display setup.  It consists of an arduino, a serial display and a 2 axis accelerometer.  The arduino is powered (I used an inline fuse for safety) by a car battery (12V).  

When I installed this setup in car it initially didn't work unless I had it plugged in to the computer for a couple of seconds.  Accelerometer data was coming in, everything was powered just fine, etc.  Then I read about the 10k ohm resistor fix to clean up the serial tx signal (10k ohm resistor between ground and rx pins).  

This worked fine for a couple of days of racing (2 hot days in a black car driving in an autocross).  Then it stopped working.  I figured it was the resistor which I installed quickly and carelessly.  So, I replaced it with another...no dice.  The serial tx/rx lights do not come on/blink and I checked all my connections for shorts and such.   I will see if serial communications can be "jump started" by reading from usb.  Unfortunately I can't do that for a couple of hours.

Where should I look next?  I'm thinking of pulling out the multimeter and making sure I didn't break of a header pin or something/Find out where the missing pin is.  What could I have broken?  
Logged

Daniel
Guest
 Bigger Bigger  Smaller Smaller  Reset Reset

hi

yes you are faced with the standard troubleshooting problem: where to begin? I usually go at it from two perspectives: a general overview and a detailed analysis.

Overview: look at things as a whole and think about what might be going wrong. Make individual tests to confirm that power is present etc.

Detailed: if things just won't work, then see what will work, in increasing levels of detail:
- will the atmega blink a led?
- will the atmega display things on the LCD?
- can the accelerometer data de sent back to the Arduino serial monitor?
- etc
- etc

The goal of course is to look for working and non-working parts of the circuit that will help to isolate the small thing that is causing the problem...

It's much like the way a doctor has to think these days: a general overview and then specific tests.
(Of course the difference is we make a lot less money than doctors.) smiley

D

« Last Edit: April 23, 2007, 02:51:49 pm by Daniel » Logged

0
Offline Offline
Full Member
***
Karma: 0
Posts: 121
View Profile
 Bigger Bigger  Smaller Smaller  Reset Reset

I owe you all follow up so here goes:

I tried connection the usb to the arduino and reading off of serial...no dice.

I tried powering from usb and the worked, but wouldn't send serial.

I re-uploaded the sketch and that worked.  Then I ran the sketch as it should be run with power off the 12V battery and it works!

This raises questions:
What could have caused this?  Did something short out/corrupt the sketch on the chip?  To this end I've taped all the wires and such down so they don't move around at high speeds?

What else could I do to prevent this?
Logged

0
Offline Offline
Full Member
***
Karma: 0
Posts: 121
View Profile
 Bigger Bigger  Smaller Smaller  Reset Reset

Bumping the thread back again from the dead.

I took the car autocrossing (for the curious: http://en.wikipedia.org/wiki/Autocross) and the same issue cropped up.  I've narrowed it down to two causes.

1) Wiring is shorting something out - The current setup is a rats nest of wires and is set up to be easily taken apart/put back together etc.  I'm using a lot of screw terminals, etc. (There are pictures on the my project thread (http://www.iwsti.com/forums/showthread.php?t=84794).)

2) Heat - I hope this is not the cause because this will make the device almost unusable.  I also think that this is not the cause due to the specs of the devices I am using.  Heat outside the car during the event was in the 80's.  So figure up to 100 degrees F where the arduino was installed in the car.

Questions: Why would a short cause such a specific problem?  What kind of short could cause this problem (in other words...what should I look to keep wires and stuff away from? Current thinking is I actually shorted to the chip and crapped out some of the program space...if so...I risk losing the bootloader...ick)?  Has anyone else had a problem with heat and the arduino?  I'm guessing no, since I haven't seen any report of a problem from those involved in the making of weather data loggers.

Conclusion:  I'm going to rewire a neater, cleaner, soldiered in, etc. setup and see what happens during the next event.
Logged

Austin, TX
Offline Offline
Jr. Member
**
Karma: 1
Posts: 94
Shh! I am hunting wabbit. Killlll the wabbit!
View Profile
WWW
 Bigger Bigger  Smaller Smaller  Reset Reset

There are a couple of things to keep in mind when developing electronics for an automotive environment.

The power supply and long signal lines are subject to a great deal of electrical noise--which can cause intermittent hard to diagnose problems.

Second there is a great deal of vibration which requires that all devices have a solid mechanical as well as electrical connections.

If your project works well on the bench but problems appear in the car then one (or both) of these two areas are the likely source of the problem.
Logged

0
Offline Offline
Full Member
***
Karma: 0
Posts: 121
View Profile
 Bigger Bigger  Smaller Smaller  Reset Reset

According to the manufacturer the ATMega168 operates up to 185F (85C).  I can see it reaching that temperature range where I mounted it.  I mounted the arduino in the dash on top of the ducting for the AC.  The AC was not on at the time (windows down for safety, etc. at driving even).

If this is the case.  What exactly happens when the temperature gets that high?  Bit start flipping?  

As far as a solution I'm thinking of hooking up some temp sensors (I'll be developing for them anyway) to that area.  I could mount it in another location (easy), I could add a heat sink to the chip and a fan (also relatively easy).

http://adsabs.harvard.edu/abs/2001APS..SES.KE016C
« Last Edit: June 11, 2007, 11:58:27 am by sti_robot » Logged

0
Offline Offline
Full Member
***
Karma: 0
Posts: 121
View Profile
 Bigger Bigger  Smaller Smaller  Reset Reset

I think I may have figured it out!!

Being that I'm a software guy I never am really concerned for these EE type issues.  I got in all my sensors and was thinking of hooking them up when I started to think about the power they draw.  

I believe I may be drawing too much power out of the 5V power onboard the arduino causing it to loose memory/crash. (see: http://www.arduino.cc/cgi-bin/yabb2/YaBB.pl?num=1178565105/4#4)

Apparently it has a ~400mA current limit (http://www.arduino.cc/cgi-bin/yabb2/YaBB.pl?num=1183268436/2#2)

In the current setup I am powering a serial LCD (I believe it draws ~60mA with the backlight and an accelerometer which should be using only 0.6 mA).  So I'm not sure how this is the problem.  It could also be a noisy power supply from the car.  Cars have a problem delivering a nice even 12V.  Sometimes its 13something and sometimes under load it can be lower.

All the sensors I'd like to power off the 5V:

1 accelerometer drawing 0.6mA
2 buttons drawing 2mA each (using 2.2kOhm resistors...set to HIGH normally open)
2 thermistors in voltage dividers using 2.5kohm resistors providing 790-18680Ohm's themselves ???draw each??
1 pressure sensor drawing 0.5mA

This should all work correctly right (we're way far from 500mA)?
What about the voltage dividers?? How much do they draw/how do i calculate that?
Logged

0
Offline Offline
Full Member
***
Karma: 0
Posts: 121
View Profile
 Bigger Bigger  Smaller Smaller  Reset Reset

Quote
There are a couple of things to keep in mind when developing electronics for an automotive environment.

The power supply and long signal lines are subject to a great deal of electrical noise--which can cause intermittent hard to diagnose problems.

Second there is a great deal of vibration which requires that all devices have a solid mechanical as well as electrical connections.

If your project works well on the bench but problems appear in the car then one (or both) of these two areas are the likely source of the problem.


If there are power supply problems from the 12V coming from the car how do I fix them?
Logged

Greenwood, Indiana
Offline Offline
God Member
*****
Karma: 0
Posts: 508
Arduino rocks
View Profile
 Bigger Bigger  Smaller Smaller  Reset Reset

Make an additional power supply that takes the automobile power and cleans it up.

http://www.elecfree.com/electronic/power-filter-regulator-circuit-with-lm1084/

That's a pretty good circuit, you can leave out the relay and it's control circuitry unless you have a need for it. As well as some of the other connectors and wires.


http://cache.national.com/ds/LM/LM1084.pdf

I really like the specifications of that part but you can use about any three terminal regulator fixed or adjustable that's within the Aruduino's recommended voltage and current rating. Just setup an adjustable one properly and you're set.

Logged

If it was designed by man it can be repaired by man.

0
Offline Offline
Full Member
***
Karma: 0
Posts: 121
View Profile
 Bigger Bigger  Smaller Smaller  Reset Reset

^ I'm beginning to agree with your theory.  Power output from a car's 12V system really isn't 12V but something the range of 12V to 13.5V normally.

Should a sustained 13.5V cause the issue I am seeing (flipped bits on the chip)?  I am driving the car hard so the power steering pump, iginition, and other engine related electrics are getting a workout.  
Logged

0
Offline Offline
Faraday Member
**
Karma: 8
Posts: 2526
View Profile
 Bigger Bigger  Smaller Smaller  Reset Reset

The 78xx regulators can typically deal with voltages up to 35V.  A typical car is actually running around 14.5V when the alternator is charging.

The problem with automotive power is usually the noise.

-j
Logged

0
Offline Offline
Full Member
***
Karma: 0
Posts: 121
View Profile
 Bigger Bigger  Smaller Smaller  Reset Reset

...and noise could cause my problem on a semi consistent basis.  (typical situation is like this.  On and regular drive on highway for an hour-On and drive hard for 3 x 40 second intervals...with 5-10 minutes of idling in between each run-Off for hours-try to turn on but doesn't run my code.
Logged

Pages: [1]   Go Up
Jump to: