Featherweight Blue Raven Development Thread: Ground tests

[ihbarddx]

Although there are a number of minor bugs we're cleaning up and the automatic download of time series data is not available yet, I want to look ahead to the last major phone interface feature to be added, which is ground testing of deployment charges.

For people who have used a remote ground test feature for deployment charge sizing with other wireless devices, what features did you like and what would you like to see done differently. Here's my own wish list:

• Record a short slow-motion video
• One button starts the countdown and cancels it
• Show live continuity voltages for each channel
• Use on-board sensors to record data during the firing (accelerometer, baro sensor for pressure leak, etc.) as if it were a short flight
• Add an easy way for the user to enter the charge size that was used
• Keep the slo-mo video, recorded data, and associated charge size together for future reference.

Err... Will time series download involve a hardware change?

 

[kjhambrick]

Adrian --

It's even easier to kill the 16g Accelerometer than running the Blue Raven on the Charger.

All I had to do was recharge the battery and the 16g Accelerometer appears dead.

Reloading the latest firmware restored it.

Attached are two images:

1. After charging the battery -- 16g is dead

2. after reloading the firmware -- 16g is restored.

Ideas ?

Thanks

-- kjh

 

[Adrian A]

Adrian --

It's even easier to kill the 16g Accelerometer than running the Blue Raven on the Charger.

All I had to do was recharge the battery and the 16g Accelerometer appears dead.

Reloading the latest firmware restored it.

Attached are two images:

1. After charging the battery -- 16g is dead

2. after reloading the firmware -- 16g is restored.

Ideas ?

Thanks

-- kjh

It sounds like there is a new interaction with the initialization timing during the power-up that I haven't seen before. I'll be following up with my scope and further testing. Since this hasn't been an issue on previous builds, there is likely some unlucky initialization timing difference that is making the initialization of the accel/ less reliable with the new build. I expect I'll be able to adjust the initialization timing to get back to the reliable start-up of previous builds. Thanks for finding this and narrowing down the conditions where it was happening for you.

 

[Adrian A]

Err... Will time series download involve a hardware change?

This will be available with a new phone app version and an over-the-air update to the firmware. In the meantime, the time series download is still available through the USB serial port, which is how I have been doing the flight testing over the last year.

 

[ihbarddx]

This will be available with a new phone app version and an over-the-air update to the firmware. In the meantime, the time series download is still available through the USB serial port, which is how I have been doing the flight testing over the last year.

Great!

 

[Adrian A]

An update on the issue that @kjhambrick saw.

I have isolated the problem and I have a new build that corrects the issue. In build 7d4488f, I had added some code to give the baro sensor a chance to recover if it does not initialize properly, but it turned out that in some cases it would interfere with the accel/gyro sensor initialization. I didn't find the problem before software release because it doesn't show up when the board is mounted on my software development fixture, and it was intermittent when the board is powered up in a normal configuration. Yesterday I was attempting a 2-stage test flight, and the Blue Raven in my sustainer powered up cleanly on Saturday during prep and testing, and on Sunday morning in the prep area. But then when I powered it on again at the pad, the accel/gyro initialization was unsuccessful as long as I had that firmware build installed. While the rocket still the rocket was in the tower, I was able to use the phone app to revert the firmware back to an earlier file that I had stored on the phone and continue.

The most recent phone app build automatically starts an update to the latest firmware, and so I wasn't able to use the app again after reverting the firmware, and had to rely on the beeps. The beeps indicated continuity on my airstart channels after I closed the screw switch I use to isolate them, but I forgot that channels assigned to be an airstart function come up in a software-disarmed state for additional safety. I launched and the airstart did not light because of the lack of software arming.

The actions I'm taking based on this are:
1. Find and fix the source of the problem (done)
2. Add a cancel button to make it possible to avoid the automatic firmware update from the phone.
3. Make it more obvious in the live data screen when there is a sensor problem that makes the altimeter no-go for launch. At the moment it gives you an indication that there's a problem, but it doesn't jump out and grab you.
4. Include a software-disarmed channel in the new go-for-launch summary and in the channel ready beeps
5. Add testing the build with the Blue Raven powered outside of my development test fixture to my list of regression tests that I need to do before releasing a build.

A new app release with the corrected binary should be available tomorrow.

 

[kjhambrick]

That is awesome, Adrian, great debugging there.

Sorry about the sustainer failure.

Q1: Did you still get drogue -and-or- main deployment from the Blue Raven ?

Last Wednesday or so, I requested a firmware update via the Featherweight Contact Page and I received an email from the Firebase App Distribution with an update link and your release notes.

Very easy to bump the Blue Raven firmware revision via my Android Phone !

Q2: Do I need to send an email to receive an upgrade link for the latest firmware when it's ready or is the email automatically sent to owners of the Blue Raven ?

I am in no rush -- my next opportunity to fly will be June 3 and I still don't have a rocket ready yet ( need to glass and finish a new 48mm main chute tube for a good, 25 year old Fin Can I've got laying around ).

But the Blue Raven is still fun to play with on the ground

Thanks for the update and good luck !

-- kjh

 

[Adrian A]

That is awesome, Adrian, great debugging there.

Sorry about the sustainer failure.

Q1: Did you still get drogue -and-or- main deployment from the Blue Raven ?

Yes. All standard functions except for airstarts come up in the software armed state at power-up. So the airstart didn't ignite, but the apo and main fired.

Last Wednesday or so, I requested a firmware update via the Featherweight Contact Page and I received an email from the Firebase App Distribution with an update link and your release notes.

Very easy to bump the Blue Raven firmware revision via my Android Phone !

Q2: Do I need to send an email to receive an upgrade link for the latest firmware when it's ready or is the email automatically sent to owners of the Blue Raven ?

I think it's sent automatically.

 

[Adrian A]

I just released the new app build 171 for Android and iPhone, which includes new firmware that corrects the problem discussed above for build 167 (iOS) and build 168 (Android). As a rule I don't make previous releases unavailable for people who have installed them, but I'm making an exception for builds 167 and 168 because that problem can affect critical functions of the Blue Raven.

We are getting close to having the ground test functionality working; it could be available this week. It may get done ahead of the data download capability because it's a simpler interface between the phone and the Blue Raven.

 

[kjhambrick]

Thanks Adrian.

Downloading now.

-- kjh

p.s. I received an email from the Firebase App Distribution that notified me of the new Firmware. Very nice !

 

[Adrian A]

The new release (178) includes the ground test feature now!

Here's a screenshot:


Here is a slow-motion video that was recorded during a ground test:

View attachment 1684442185246.MP4

And a screen recording video showing the process:


View attachment RPReplay_Final1684442248.mov

 

[watheyak]

Maaaaannnnn..... Such a cool feature.

 

[kjhambrick]

Very nice Adrian !

You've been busy !!

I tried it yesterday afternoon -- Version 1.0.0 178 ( c22abd1 ) runs great !

I've got Flight Sim and Ground Test Data to review on my Linux Box now ( BlueTooth -> Phone -> email -> Linux File System )

Will test ( i.e. try to 'break it' ) tonight after work

Thanks.

-- kjh

 

[Neutronium95]

I used Blue Ravens for both of my BALLS flights this year, and found myself very unhappy with the ground test functionality.

The first issue I ran into was when doing some initial tests with my bulkhead Blue Raven when powering it on for the first time. I was testing with bare ematches, and a preliminary version of my avbay. Twice, when powering it on I accidentally nudged it with the magnet and knocked the avbay over. Both times it immediately fired the ematch that was attached to the apogee channel. I was taking reasonable precautions and nothing bad happened, but it definitely spooked me.


The other problem was that the ground test functionality in the app just doesn't seem to work. When I was running those initial tests, I'm not sure if I set off a single ematch via the ground test menu. I did validate functionality by running a simulated flight and having that set off the ematch, but the ground test didn't work.

I tried again at BALLS, to test charges for my H13 flight, and again the ground test didn't work. I have several slow mo videos on my phone of ejection charges not firing. It did eventually go off on maybe the fifth attempt, but the app was lagging enough that it surprised the hell put of me when it eventually went off.

My phone setup is slightly unusual (Pixel 6A running CalyxOs) but I haven't had any issues with any other apps.

I've npticed the app lagging a ton when ground testing, maybe the video is causing problems.

The flight sim is also a bit awkward, with lots of lag that left me unsure what state the altimeter was in. Have you considered just curating a list of possible flight sims for people to run, something like "Nominal flight to 2000 feet" or "Two stage flight at an angle" to test staging logic?

 

[Adrian A]

I used Blue Ravens for both of my BALLS flights this year, and found myself very unhappy with the ground test functionality.

The first issue I ran into was when doing some initial tests with my bulkhead Blue Raven when powering it on for the first time. I was testing with bare ematches, and a preliminary version of my avbay. Twice, when powering it on I accidentally nudged it with the magnet and knocked the avbay over. Both times it immediately fired the ematch that was attached to the apogee channel. I was taking reasonable precautions and nothing bad happened, but it definitely spooked me.


The other problem was that the ground test functionality in the app just doesn't seem to work. When I was running those initial tests, I'm not sure if I set off a single ematch via the ground test menu. I did validate functionality by running a simulated flight and having that set off the ematch, but the ground test didn't work.

I tried again at BALLS, to test charges for my H13 flight, and again the ground test didn't work. I have several slow mo videos on my phone of ejection charges not firing. It did eventually go off on maybe the fifth attempt, but the app was lagging enough that it surprised the hell put of me when it eventually went off.

I'm surprised by the problems you're describing; I would not release an app that behaved like that in my testing. One thing that might tie the issues together is if the accelerometer is so far out of calibration that it is falsely detecting extra flights. That can cause additional downloads that could make the app non-responsive for the ground test. The app should show a status of "Downloading and clearing flash memory" if that is happening. Another thing to be aware of with the ground testing is that you need to select the channel that you want to fire by clicking on the channel button at the top to change its status from "Disarmed" to "Armed"

My phone setup is slightly unusual (Pixel 6A running CalyxOs) but I haven't had any issues with any other apps.

Ah, this seems like the more likely explanation. I had never heard of CalyxOs. The Blue Raven app is only designed to support Android and iOS. Do the apps that run fine with CalyxOS have a lot of Bluetooth interaction? The details of the timing and interaction between the Blue Raven and the phone could easily cause issues if CalyxOS doesn't behave quite the same way. Do you have another Android or iOS phone that you could try?

I've npticed the app lagging a ton when ground testing, maybe the video is causing problems.

The flight sim is also a bit awkward, with lots of lag that left me unsure what state the altimeter was in. Have you considered just curating a list of possible flight sims for people to run, something like "Nominal flight to 2000 feet" or "Two stage flight at an angle" to test staging logic?

 

[Neutronium95]

I'm surprised by the problems you're describing; I would not release an app that behaved like that in my testing. One thing that might the issues together is if the accelerometer is so far out of calibration that it is falsely detecting extra flights. That can cause additional downloads that could make the app non-responsive for the ground test. The app should show a status of "Downloading and clearing flash memory" if that is happening. Another thing to be aware of with the ground testing is that you need to select the channel that you want to fire by clicking on the channel button at the top to change its status from "Disarmed" to "Armed"

I'll try recalibrating it and doing more tests this weekend.

Ah, this seems like the more likely explanation. I had never heard of CalyxOs. The Blue Raven app is only designed to support Android and iOS. Do the apps that run fine with CalyxOS have a lot of Bluetooth interaction? The details of the timing and interaction between the Blue Raven and the phone could easily cause issues if CalyxOS doesn't behave quite the same way. Do you have another Android or iOS phone that you could try?

CalyxOS is a privacy focused Android fork. It's pretty much just stock Android with a bunch of Google integration stripped out. The Bluetooth stuff should in theory work exactly the same. I noticed it took a really long time to download the flight data, but just assumed that was a limitation of the chip on the Raven. I'll try with some other Android phones this weekend.

 

[Adrian A]

I'll try recalibrating it and doing more tests this weekend.


CalyxOS is a privacy focused Android fork. It's pretty much just stock Android with a bunch of Google integration stripped out. The Bluetooth stuff should in theory work exactly the same. I noticed it took a really long time to download the flight data, but just assumed that was a limitation of the chip on the Raven. I'll try with some other Android phones this weekend.

Downloading does take a long time, and about twice as long on Android as iOS. If the flight is particularly long, resulting in a file over 16 MB, there is an issue with the currently-released build 206 that we will fix with the next release we are working on now.

 

[thzero]

The other problem was that the ground test functionality in the app just doesn't seem to work. When I was running those initial tests, I'm not sure if I set off a single ematch via the ground test menu.

I had similar. Stock Pixel 6 Pro with latest Android.

 

[Neutronium95]

I just had a Blue Raven fire the main charge on powerup. I had some issues prior with my bukhead Blue Raven firing charges on powerup when knocked over by the magnet I was using to switch it on.

 

[Adrian A]

I agree that's not acceptable. I would be ticked off too. Could you send it to me to investigate? I suspect some kind of one-off hardware problem because the firmware can't detect launch immediately. I'll DM my address.

 

[Neutronium95]

In slightly better news, I used the ground test feature a few times while trying to figure out the failure at the launch site, and it worked much better than in the past. I'm not sure what was changed, but it fired the ematch reliably and my phone didn't become super laggy when the countdown hit 0. In general the newer version of the app just feels a lot better.