Mystery Throttle Ceiling - Worked Fine Yesterday!

Pixhawk - DiscoPro - EZUhf - TBS 900kv + Bulletproof TBS ESC - Futaba T8FG

It’s been in the air every day for a week now, tuning and adjusting. AltHold was rock solid (hovering at about 50% throttle), Land was impressive. Stabilize worked well if I took off in Stabilize.

Ran AutoTune successfully yesterday (Saved changes as per instructions), and then flew another battery in AltHold. I didn’t like the way it fell from the air when I switched from AltHold to Stabilize as the last thing I did to it yesterday, so today’s mission was to set the throttle midpoint. I attempted to take off in Stabilize with the goal of hovering in one spot for 30 seconds.

The Disco got six inches off the ground and skittered around in ground wash. Repeatedly.

What I have tried so far:
Checked the endpoints in the Futaba.
Re-Calibrated the ESCs.
Checked all the throttle related parameters in MP.
[attachment=2]Throt Tree.PNG[/attachment]
Compared the RCin and RCout in the logs.
[attachment=0]RC.PNG[/attachment]

Log:
[attachment=1]2014-09-03 09-56-27.log[/attachment]

Help for this bumbling old man would be greatly appreciated!
Joe

@JoeBob,
What version of ArduCopter are you using?
Regards,
TCIII GM

Arducopter V3.1.5 Quad
MP 1.3.9

Thanks!

Other possible clues:

I have all 3 wires from my ESCs attached, including power. One of the Wiki pages said leave the power attached with Opto ESCs to power the Opto isolater. ESCs are 30A Opto flashed with custom SimonK.

Throttle in Radio Calibration looks the same as before: PWM = 1098 to 1938

No weight added.

When I switched from AltHold to Stabilize at the end of yesterday’s last flight, it dropped straight down from 4 meters with slow turning props. I hammered the throttle up, but no response. I thought I was too slow, but that may have been the start of my problem? PWM value for Stabilize Mode is 1260 (1231 to 1360 is the range.) That’s why I went out to set Throttle Midpoint today.

I don’t understand the difference between ‘Refresh Parameters’ and ‘Write Parameters’. Could I have messed them up?

Appreciate the help. I promise not to be a long term bother.
Thanks,
Joe

Okay, it’s been a couple of days now and it looks like everybody online is just as stumped as I am. So I’m going back to ‘start’ and begin again.

Uninstalling MP and re-installing MP I can handle. How do I make sure to erase all the basic files in the Pixhawk to make sure that I start with a clean slate?

Is one build/firmware for a quad recommended over others?

Thank you,
Joe

Reinstalled MP after completely uninstalling and removing directories.
Faied when attempting to clear the Pixhawk memory, the ‘terminal/settings/command/erase’ option is no linger an option.
Couldn’t get it to re-install firmware as ‘no need to upload; already onboard’ .

No improvement. The Disco just sits on the ground sounding like an under-powered window fan.
Yaw, Pitch, and Roll inputs result in higher RPMs; but it doesn’t even make it off the ground. At all.

Would a video help diagnosis?

Working at it,
Joe

I don’t think it’s a Pixhawk problem. Your RCOUTs were pretty much all maxed out but you didn’t lift off. So why if the ESCs are being told to deliver full throttle isn’t your quad lifting off?

My hunch is that there’s something wrong with your power feed from the battery that’s causing a voltage drop. Since you don’t have any logging going on of your voltage or current there’s no way to see what your operating voltage is (Your Vcc looks OK though).

You didn’t say if you attempted flights with different batteries all with the same result. If this has all been on the same battery it’s possible you lost a cell in the battery and are running at essentially 3S instead of 4S.

If you have used different batteries with the same results I’d look toward something failing on the Disco’s PDB that is causing a voltage drop. You didn’t say if you were using a Core, but if so I think there’s a shunt resistor the Core uses to do its current measurement thing. If there’s an issue with that a voltage drop could be created.

Are you using a BEC to power your Pixhawk? Your Vcc is around 5.2 volts.

At this point what I would do is hook up a voltmeter to the input leads of one of the ESCs, tether the Disco down and run up the throttle. My guess is you’ll see a lower voltage than what should be being fed to the ESCs. If so it’s just a matter of tracking the power feeds back to the battery looking for the problem.

Oh, one thing I forgot to mention, your question about Refresh versus Write in Mission Planner.

The Refresh selection reads the parameter values from the Pixhawk and displays their values in the various Mission Planner screens.

The Write selection takes whatever values are currently shown in Mission Planner and writes them into the Pixhawk. So after you make any changes to the parameters in Mission Planner you need to do a Write to get them into the Pixhawk. In some MP pages the Writing gets done automatically and there’s no Write button to select. But if you see a Write button be sure and use it otherwise your changes won’t be saved.

[quote=“OtherHand”]
My hunch is that there’s something wrong with your power feed from the battery that’s causing a voltage drop. Since you don’t have any logging going on of your voltage or current there’s no way to see what your operating voltage is (Your Vcc looks OK though). [/quote]
Is voltage logging something that can be turned on?
Using the 3DR power module. Is there something that might be burned out? (This quad flew well for a week.)

[quote=“OtherHand”] You didn’t say if you attempted flights with different batteries all with the same result. If this has all been on the same battery it’s possible you lost a cell in the battery and are running at essentially 3S instead of 4S. [/quote] Multiple batteries over many days now.

I’ll check on that resistor.
When I tied it down and watch the telemetry while trying to lift a table, voltage numbers from the TBS core (via OSD in the FPV camera) and the numbers on the MP Flight Data screen matched. They max out around 18 A, but don’t hold there. They seem to move up and down (along with motor RPMs) on their own.

[quote=“OtherHand”] Are you using a BEC to power your Pixhawk? Your Vcc is around 5.2 volts.[/quote] Pixhawk is powered by the 3DR power module that came with it, and the red power leads from the 4 ESCs.

[quote=“OtherHand”] At this point what I would do is hook up a voltmeter to the input leads of one of the ESCs, tether the Disco down and run up the throttle. My guess is you’ll see a lower voltage than what should be being fed to the ESCs. If so it’s just a matter of tracking the power feeds back to the battery looking for the problem.[/quote] What range should I look for going to an ESC?

Almost ready to put the Naza back in just to see if it flies without the Pixhawk and narrow down the possibilities.

BTW; thanks for clearing up Write/Refresh.

Joe

If you are using a 3DR Power Module, you have your current and voltage monitoring disabled. You should have that turned on. I’m not sure if having that off could be causing weirdness for you.

In Mission Planner it’s under Initial Setup/Optional Hardware/Battery Monitor. Set the “Monitor” field to “4: Voltage and Current”, the “Sensor” to 3DR Power Module and “APM Version” to Pixhawk.

Once you make those changes you could try a short hover and it will log some current and voltage data. That might be enough info to determine where your problem lies.

In terms of what sort of voltage you should be looking for at the ESCs, it should be just maybe a couple of tenths of a volt less than what you would measure at the battery.

I don’t think switching out flight controllers would help you at this point. The Pixhawk was clearly telling the ESCs to provide full power, but you weren’t getting it out at the props. There’s no reason to think a Naza telling your ESCs full power would be any different.

I really appreciate the help -

Made the suggested changes to the Battery monitor. (I’m using the TBS CORE OSD, so I hadn’t bothered with Pixhawk battery info. But TBS doesn’t log…)

Flew a short hop tied to a table. Here’s that log.
[attachment=0]2014-09-08 00-22-35 26.log[/attachment]

Also, here’s a video that I uploaded to the Disco Pro forum to see if they have any clues:
vimeo.com/105516498
Sorta reminds me of an old girlfriend.

Thanks,
Joe

Are you running a 5S battery on this and was it fully charged for the test? I note that after 1 minute the voltage recorded was only about 18.1 volts (3.6 volts per cell). That’s pretty low.

I like how you’re thinking outside the box; but my batteries for the Disco are all 4S. They read around 16.8 when fully charged.

The Disco CORE has 2 50A capacitors that have caused issues in other MRs. I’m pursuing that line of researcher, too.

Thank you,
Joe

Well then something is definitely off with the Power Module calibration. I thought 5S was a very strange setup. FYI, your initial, no load voltage was recorded as 19 volts. I’ve never seen a voltage measurement that far off from what it’s supposed to be for 4S.

Found 2 possibilities:

In searching for ‘Motors enable logging’ (Old Option in the CLI) I found this:

I’m wondering if the ‘thrust curve max max thrust percentage’ should be set at less than 80 to get max thrust at a lower throttle point?

Finally, this log graphed under ‘Power Issues’ got me to thinking:
[attachment=0]RC in RC out 800.png[/attachment]
The red line is throttle inputs (Channel 3 - RCIn). The green line is throttle outputs (Channel 3 - RCOut)
This flight started from cold. No warm up flights.
The throttle Out never matches the highs of throttle In. That mirrors the observed action of the Disco.
The throttle Outs bounce erratically during the first minute of flight. But then start to hold the highest points reached after warming up.
Hmmm.
Something is retarding current flow, but less after getting hot.
Did that hard bounce break a solder joint where the power module connects to the board? The battery connections are coated with Liquid Tape for waterproofing, which would hold the joint together if it were cracked, allowing some current to pass by contact, but not enough for full throttle.
Another interpretation would be that something in the Power Module is toast.

Any other ideas?

Thanks,
Joe

Your THR_MAX is probably fine at 1000. I wouldn’t focus on that. The other motor parameters look good too.

From the data so far I’d focus on a bad electrical joint between the PM and the battery. Specifically I’d look more toward the ground connection than the positive lead. The reason I’d look at the ground is the way the PM measures voltage. It’s a simple voltage divider tapping between the + and - leads of the battery (through the connectors of course). Since the PM is reading higher significantly than actual it suggests the resistance on the ground side of the voltage divider is higher than it should be. The most obvious way for this to happen is some sort of increased resistance in the ground lead. But I’d look at the entire wiring between the battery and PM, both + and -.

Firmware 3.1.4 Flies just damn fine.
Hoooo Yaaah!!

Under Terminal, Select Pixhawk, connect; then in the command prompt area,

Type ’ Help ’ [enter]
Type ’ Setup ’ [enter]
Type ’ Erase ’ [enter]

It should say “Erasing eeprom”

Go to Initial Config.
Install Firmware
In the bottom right corner select "Install previous firmeware."
3.1.5 grounded my Disco, so I selected 3.1.4.
I could tell that the motors were sounding better, so I put the props back on and flew that table!

Good luck,
Joe