AltAcc2C repair

The Rocketry Forum

Help Support The Rocketry Forum:

Joined
Feb 8, 2009
Messages
19
Reaction score
0
I'm having a problem with my AltAcc2C. On its last two flights, the AltAcc did not fire either of the charges. Luckily, I had an ARTS as a backup.

When I run the Quick Functional Check, all indications are correct except the bulbs never flash, or take over 70 seconds to flash. The bulbs are good.

Also, on Power Up, the altitude indication flashes 16 times, then once, as if I flew it to 16,100 feet, but it's never been much over 10,000 feet. And it goes through the cycle three times, not twice like it says in the manual. In Standby, the LED blinks once every 7 seconds instead of every two seconds.

But when I Power up with the Serial cable connected, it goes through the altitude report faster (still does it 3 times) and Standby flashes every 2 seconds like it should.

Has anyone seen this? Is it something I can fix? Any help would be greatly appreciated. I stopped using it after the second failure, but I just lost my ARTS in a spectacular crash, so I'm trying to revive the AltAcc.

Joe Cox
Jolly Roger Rocketry
TRA 10686
 

UhClem

Well-Known Member
Joined
Feb 6, 2009
Messages
1,782
Reaction score
261
My manual indicates that the altitude is flashed three times at power up.

Not sure what your problem could be. I repaired my old AltAcc and a friends by replacing the ceramic resonator. But their symptoms were very simple: they didn't work at all.

Two questions:

1) The altimeter will flash the type 194 bulbs? (If yes then the output FETs are good.)

2) Have you performed the calibration? If not, do it as this will test the accelerometer. If that is dead then all hope is pretty much lost as that part is obsolete. Include a copy of the .cal file so I can look it over.

Your symptoms are very odd and I can't think of anything offhand but I will think about.
 
Joined
Feb 8, 2009
Messages
19
Reaction score
0
The altitude is flashed three times at power up. It's flashing three sets of 16 then one, which indicates the last flight was to 16,100 feet. I have never flown it nearly that high.

I do not have type 194 bulbs, but I use the small Christmas lights, which I believe work the same. During the Quick Functional Test, the AltAcc will occasionally light the bulbs after 70 seconds, but most of the time they don't flash at all.

Here's the calibration file. I had to zip it so TRF would accept the file type.

View attachment prodata.zip
 

UhClem

Well-Known Member
Joined
Feb 6, 2009
Messages
1,782
Reaction score
261
The calibration looks like the accel is working but I think it isn't a very good calibration. Your values are:

avg-1g 118.9844 # Minus One Gee Avg
std-1g 0.1526 # Minus One Gee Std Dev
fid(0) 3.0742 # Finite Difference on [ -1, 0 ]
avg0g 122.0586 # Zero Gee Avg
std0g 0.2514 # Zero Gee Std Dev
fid(1) 4.1836 # Finite Difference on [ 0, 1 ]
avg+1g 126.2422 # Plus One Gee Avg
std+1g 0.4383 # Plus One Gee Std Dev
Here is what my cal file shows:

avg-1g 115.1523 # Minus One Gee Avg
std-1g 0.3601 # Minus One Gee Std Dev
fid(0) 3.8945 # Finite Difference on [ -1, 0 ]
avg0g 119.0469 # Zero Gee Avg
std0g 0.2118 # Zero Gee Std Dev
fid(1) 3.9688 # Finite Difference on [ 0, 1 ]
avg+1g 123.0156 # Plus One Gee Avg
std+1g 0.1243 # Plus One Gee Std Dev
The difference in between your fid(0) and fid(1) values is about 1 while my data shows a difference of less than 0.1. They should be nearly the same.

If your values are accurate rather than the result of not holding the altimeter in the proper orientations, then the accel has a lot of non-linearity error. Much more than it should have and way too much for the altimeter to work well. But it is far more likely that you held it wrong.

The only idea I have right now is that maybe a PC trace is broken/intermittent and the stress from connecting the serial cable completes the connection. But it isn't much of an idea.
 
Joined
Feb 8, 2009
Messages
19
Reaction score
0
Thanks for checking. I honestly don't understand the data from the calibration file, but I'm pretty sure I was holding the altimeter in the correct positions for the calibration. At least I followed the instructions on the screen and in the manual.

Your idea about a bad trace may be a good one to follow up on, as the problem I mentioned about the LED flashes on start-up has another caveat. When the cable is not connected, the start-up flashes are very slow. When the cable is connected, the flashes are at normal speed.
 

UhClem

Well-Known Member
Joined
Feb 6, 2009
Messages
1,782
Reaction score
261
The calibration data for the accel is really pretty simple.

It takes a bunch of samples of the data using the ADC. From those samples it computes an average value and standard deviation. Because it uses a 8 bit ADC and the values are close to mid scale, the ADC returns numbers near 128. It isn't exactly 128 at zero G because of offset error. (Calibration uses the sensor noise to good advantage to get better than 8 bit resolution from the 8 bit ADC.)

Once data is collected for the three orientations (1, 0, and -1G) then more processing can be done. This is where the finite differences come from. The difference between the average zero G reading and +/1G are the two finite differences. Because the sensor is linear, they should be equal. Random noise and small sample size will result in them being slightly unequal.

Finally the software uses linear regression to find the best straight line fit to the three points. This is the final calibration value. Only the slope is used as the offset varies with temperature and is measured just prior to launch.

The barometric calibration information in that file is very confusing especially it you look at the code that uses it. (Or doesn't use it. Source code for POSIX versions of prodata were published once long ago. It takes an odd approach to processing the data.)
 

Len_Lekx

Active Member
Joined
Jan 19, 2009
Messages
42
Reaction score
0
Joe...

If it turns out that your AltAcc is toast, PM me and I can send you mine. I don't use it anymore, so I'm willing to let you have it for the cost of sending it to you.

(Last time I tested it was about two years ago... but it worked fine.)
 
Top