Monday, September 15, 2008

Test Results

We did one tethered test this weekend.We intended to do a 10 to 14 sec stable hover at 2 meters in altitude. It seems like the throttle gains were a little bit aggressive as the vehicle was very lightly loaded with fuel. It jumped up to 3m  and started to swerve  in pitch .as the throttle over compensated and interacted badly with the pitch loop. To protect the vehicle I had set aggressive abort limits (20deg in pitch and roll) It hit the abort limits and shut down. The vehicle coasted up to 4m then began to fall  while slowly continuing to roll in pitch. When the tethers came taught at the bottom the vehicle was almost inverted. I had used PVC tether covers like Armadillo does, but I had not secured them in position on the vehicle. they slid up the tether ropes and formed a loop that caught around the peroxide vent valve and the primary flight control battery. When the lines snapped taught it removed both the vent valve and battery.


The Tethers absorbed the shock and prevented the vehicle from contacting anything other than the tethers. Other than the vent valve and battery mount the vehicle is undamaged. In reviewing the video the I am very pleased with dynamics of the tether system.  I inspected the tether mounts on the vehicle and saw no signs of damage or distress from a fairly severe usage.


As soon as I saw the vehicle abort I reached for the command abort RC transmitter to depressurize the vehicle. The removal of the main peroxide vent started the de pressurization, but that vent is VERY small, so when I activated  the emergency vent one heard a distinct POP and  large woosh. Lasting less than a second.



The 900Mhz primary telemetry channel is used for both data status download and command upload. To insure reliable command upload I've reduced the data coming down from the vehicle. The largest part of this data is the IMU data and the control loop responses.  I record 100% of all vehicle parameters at full rate on board   on a 2Gbyte micro SD flash card on the vehicle.  this recording process has between an 1 and 2 second delay. As the tether loop removed the vehicle battery we lost the last 2 seconds of the 2.5 second flight.


Anomalies and corrective actions:

Emergency Vent:

The emergency vent uses off the shelf COTS RC equipment. the servo arm and linkage controlling the vent used to be plastic.  It looks like chemical and weather exposure made the linkage brittle and we broke it when preparing the vehicle for flight. We repaired it in the field with some safety wire. (Its in tension to operate the vent) and it operated  correctly as indicated above. We are going to replace this with a proper  all metal linkage before we fly again. The present wire linkage will actuate  100% reliably, but it might also allow it to activate without being commanded.


The Tether ropes:

The dynamic 9.5Kn climbing rope worked very well. The PVC pipe covers did not work correctly.We have already modified these to extend them and retain them on the vehicle end to prevent rope loops.


Pressurization Quick disconnect:

The dual pressurization quick disconnects were actuated twice, the first time only one of the two disconnected. After de-pressurizing the vehicle and trying again both actuated.  As the vehicle was up on blow out stands we could not pull on the end of the hose as we could if the vehicle was on the ground. So in a "real" flight its likely that  we could have recovered from this failure.  The disconnects were designed to be rugged as they will fall to the pavement after every actuation. The one that failed to deploy  had been "Adjusted" by prior test impacts.



Control laws:

We have already modified the throttle control loop to change the behavior and we intend to try again on Wednesday.


Hardware:

We repaired the valve and attempted to re-fly on Sunday, but  we had intermittent computer hardware issues and choose to return to San Diego to evaluate.  We evaluated this on Sunday night and it looks like an epoxy blob we used to pot the wires coming off the PCB lifted a an IC lead.  The cause and effect is unclear here as another lead of this IC (Not under the epoxy blob),could be displaced with a probe and appeared not to be  soldered. We are in the process of swapping out the primary carrier board that holds the peripherals for the flight control computer. We will get that assembled today and I will probe every single IC lead solder joint under the microscope before installing it in the vehicle.

 

I will try to post the vido in the next 24 hours.

5 comments:

Anonymous said...

Have you considered RTV silicone instead of epoxy for PCB encapsulant?

Paul Breed said...

We have already decided to swithch to RTV.

Daniel said...

I'm sure you've done so already, but for those hobbyists following you - do check the type of solvent the RTV uses - avoid the acetic acid based kind. Look for the alcohol based RTV to avoid corrosive damage to electronics.

Paul Breed said...

Electronic Grade RTV (Also called low corosion RTV) is good, I also like hot melt glue as a way to retain wires.

Paul Breed said...

Hot melt is my usual method, but my Hot melt gun died and I used epoxy instead as I was finishing up.