Sunday, October 19, 2014

Nova Repaired

I called it "repaired" and screwed body back together. It was only on the bench for one week.

I upgraded to MissionPlanner 1.3.10. I re-did all calibrations (Radio, Compass, Accelerometer, and ESC) and everything looks fine in MissionPlanner, DroidPlanner, and with Taranis control. First flight will be an AutoTrim.

CH-7 is already assigned to a 3-Position switch at Taranis. In MissionPlanner, changed CH-7 to be Super-Simple Mode. Middle position (or first click) ends up being Simple-Mode (non-GPS but directional) and Down should be Super-Simple-Mode (GPS required but even works "behind" pilot). When bench testing radio, there is no indication in MP or DP that SM or SSM is engaged. Even though it's just a Mode-Modifier, I find it strange that the only way to know if it's setup properly is to fly quad and see what happens (especially for a Mode that changes what the sticks do).

I really liked the Prop Guards (for the short time I had them) so I installed a new set. If you drill a small hole in them, you can leave them installed while opening the quad. This way, you don't have to remove the 8 prop-guard/motor screws. With them only being Phillips-head who wants to be removing them just to get inside the quad (not to mention the fact that they hold the motors securely in place). Be sure to only use these 8 special-longer screws with the prop-guards. If used incorrectly, them will touch/scratch the internal motor windings and kill the motors (and possibly the ESCs).

I already have a set of genuine DJI Phantom (P330/SPN9) 8045 props that I picked-up at LHS. Best I can tell, the stock props were clones of these (and these will continue to be my "spare-set"). I ended up just getting a new set of stock white 8045 props from along with my Prop Guards.

If I do anything else to it before next flight, I'll document it here.

Just got back from 2 flights (batteries) on small WLT v262 quad. Good to go flying again.

The Texas daytime temps are now more reasonable. It's been 3 years, so I am starting the rather time consuming process of cleaning and staining my back wooden deck. I'd rather it didn't, but most of my free time will be spent on this project (instead of quadcopters) so I can finish before winter comes.

Deck is 12 year-old white-pine. This pic after 30 man-hours of cleaning prep
and staining with 6 gallons of Defy Extreme stain (Light Walnut) ... 2 coats.

Wednesday, October 15, 2014

Nova ESC and body Repairs

Repaired (red led) ESC and it tests good. Saved some down time and $35. I am using (normal) leaded solder and have iron set to 360c.

Notice PWM signal from chip, thru resistor and then to top of board where it uses PCB Via to jump to other side.

I verified that the 5v trace (with usable solder-pad) is indeed not connected to anything else (notice a 5v component is intentionally not soldered in). All the Vias are the same size and very tiny. As other blog pics show ... the only PCB traces (on this side) that still carry the PWM signal is the PWM's Via's circle. So that it will accept solder, I used finger-nail buffing-pad to carefully buff-away the green paint from the Via circle pad AND a 2mm section on adjacent 5v trace. After fluxing this small area, created a solder bridge between the two. The tiny blob of solder pictured here (where "W" used to be) is covering the whole via top surface as well as the whole 5v trace cross-section. Checked work with ohm-meter.
So, now the previously un-used 5v pad becomes the PWM pad. Since we now have a solder-pad large and sturdy enough to handle a 26awg wire, soldered the PWM wire to it.
The "PWM detection for LowVoltageAlarm & LED control circuit" wire gets a proper connector (so whole ESC can be installed or removed without soldering). Soldered it to it's own pad further along this new hi-jacked trace. I suppose it should be noted that it was this short red hard-soldered wire (with no slack) that lead to the pad being ripped from the board in the first place.
The repaired version was tested in the quad with a battery and works fine.
Three strips of Plastruct (added in a second phase) 
A new frame/body/shell is about $40 (shipped) but currently Back-Ordered at HK. Repaired cracked styrene body with Plastruct Plastic Weld Solvent-Cement and then some thin strips of Plastruct (# 90739), were layered inside for reinforcement (on side-edges and middle). They also sell "dowel-shaped" pieces that might work. Over the years, I have gained experience with various glues and materials. I know that using the right glue for each material is important. The Plastruct Solvent "wicks-into cracks" and melts the plastic back together. It the end, the cracked pieces (and any reinforcement pieces you add) all become very close to "one piece" again. I was impressed how well it worked.

Basic Testing
IIRC, you can test motors and ESCs via MissionPlanner option and/or Terminal without arming motors (or pre-arm checks). Be sure props are removed first.

Troubleshooting "All ESCs are beeping"
  • ESCs are getting main (hard-wired) run voltage from main LiPo battery, but ESCs can tell they have no PWM control (control signal that tells them to stay Off or stop, Arm motors and start them, speed-control, etc.)
    • Check in this order
    • This can be caused by no BEC (+5v) voltage being connected to RX radio (and therefore, also no power being passed to FlightController).
    • FC might not be getting power (usually via CH-1 Aileron connection) 
    • FC might be damaged or malfunctioning
      • FC might have a corrupted firmware and needs to be reset and Firmware re-loaded with MissionPlanner.
        • Followed by loading of Default params & complete calibration.

Trouble-Shooting visibly fried or damaged PowerBoard, Motors, or ESC after a crash or LiPo battery problem.
I've never had to do this, but I have wrote this procedure to help a forum member. I think it might come in handy in the future. Seems to fit on this post.
  • Unplug all ESCs and RX.
  • Connect LiPo and check PowerBoard for BEC +5
  • Check PowerBoard for +11v output at each corner (ESC Input motor-run voltage)
  • Connect one (non-fried) ESC
  • Connect one (non-fried) Motor
    • You can test the motor's 3 phases with a LCR or LC meter. It measures the inductance of the coil-sets. All 3 should be similar. The motor should not be connected to ESC or anything else during test.
  • Use MissionPlanner Motor Test and try to get this set working.
  • Keep testing at this location the rest of motors and/or ESCs.
    • Remember that testing a fried motor might destroy a previously good ESC and vice-versa. I would just write-off any obviously blown parts and try to check and/or save the rest.
ESC Replacement Notes:

I've never had to do it, but according to forum reports, different version ESCs can be mixed. Be warned that the orientation of the solder-pads is reversed on some models. If in doubt, use a meter to check volts and/or ohms before connecting. As I stated before, there doesn't appear to be a 5v generation circuit (aka +5_BEC) on these ESCs.

There are 3 heavy-gauge (thick) Power-Input wires:
Red - 12v +
Black - GND
White - 5v+ (for LEDs)

There are 2 thin PWM motor-control wires (black Dupont/RC-Type plug):
White - Signal
Black - GND

There are 3 wires for brushless motor connection:
Red, Black, and Yellow

The PWM signal only needs White (Signal) and Black (GND) wires ... for each ESC. If there is also a red wire in that set, it should not even be there. But, it seems that at the Cheerson plant, pre-made 3-wire Dupont cables (RC servo type) are plentiful. So if it is there, either heat-shrink it or leave it soldered down to the ESC with the other two (actually used) wires. If you trace the ESC's PWM red-wire solder pad, you will see it actually goes no-where. You can also just remove it (back-out pin, etc.) on each end.

The 5-volt power for the LEDs comes from the PowerBoard, but is only sent to the LEDs when the quad wants each of them to light up. If you think about it, it must be this way. Otherwise, you would have separate control lines going to each ESCs ... to control the LED lighting.

There is one other thin (usually red) wire, but only on the Forward-Port ESC. That is something different.

Motor Troubleshooting:

Haven't done much with Nova motors, so no dedicated page yet. Seems to fit here for now. You can easily open motors and check inside (glued-on magnets, etc). If you ever mounted motors with wrong screw (too long) windings are likely permanently damaged underneath (just takes a scratch). Electrically, they can be checked with LCR Meter (aka LC or Inductance Meter). A basic Volt/Ohm meter is usually not sensitive enough.
Stock Radio Notes:

Since I purchased a PnF Nova (without radio), I have never owned the stock white radio set transmitter or receiver (TX or RX). However, I know many of you visiting this Blog might have them, so here are some notes from the RCGroups forums.
Here is the basic wiring, showing where the Ground, Positive-5v, and Signal rails are located on RX. In a stock config, power comes from BEC on PDB to Battery (B) socket on RX. Then (since all the "power-rails" are connected) the power gets to FC through standard 3-wire Dupont (servo-like) cable connected to Aileron (A) channel.

Please note that (like any electronic machine or vehicle) through-out the whole aircraft ... proper power polarity must be followed (for all standard voltages ... like +5, +12, etc.). If any component's power is connected with reversed polarity, it will short-out when power is applied. It might destroy several components when the "magic smoke" escapes (as components fry and burn-up). If unsure of voltage or polarity of a "power supply point", check first with voltmeter before connecting. Verify plug and socket polarity (with docs or pics), and power requirements before connecting anything.
The stock white RX/TX radios work as a matched set. Only this Cheerson RX model will work with this TX model. It might be FlySky protocol (not to be confused with FrSky) but I don't know it's ever been proven. If you need a new one of either, you can get them at HobbyKing (for Nova) or BangGood (for CX-20) ... or if serious about hobby ... use a better, more hobby-grade radio.

Monday, October 13, 2014

Flight # 6 - Crash Landing - Damage Evaluation

The good news is that all electronics seem to still be working. On the bench, I connected a LiPo and used the Taranis and DroidPlanner. I can Arm and spin motors, instrumentation sensors are all working, etc.

Lower shell only 60% cracked (but mounting point is damaged). Notice oval copper solder-pad ripped from ESC.

Upper shell 100% cracked. Screw mounting point compromised.

Top and bottom body-shells are pretty much trashed unless I try to glue them back together. I would like final repair to look acceptable, but more worried about structural integrity since it is a motor arm and shell is also the frame.

ESC damage. Not sure what PCB Revision this is, but I think it's older (ie v1.0).

Scratched SMT micro-trace and soldered on jumper wire for testing. It worked but didn't hold. Notice trace lifted from board but stayed connected long enough for testing.

The PWM solder-pad oval got ripped right off the ElectronicSpeedController (ESC) PrintedCircuitBoard (PCB). I did a very-temporary micro-solder-hack to make sure ESC was still working, and it is. However, all the PWM traces are SMT form-factor, so re-acquiring that trace (on either PCB side) will be tricky. There appears to be several different "MFL1218" Nova/CX-20 ESC PCB Revisions and but the word in the forums is that they can be mixed in the same aircraft and work (calibrations, actual operation, etc.). Appears to be a simple 2-layer PCB.

PWM detail. Under the letter "W" signal uses a hole-thru Via to other PCB side. All the way back, PWM trace is about 50% width of a normal trace like the 5v one pictured here.
Not sure what is risky-er ... trying to fix this one and use it or trying to integrate a different revision one. Also, a new ESC is $35 shipped (from Hong Kong and the wait). One possible way to fix this one (and it be durable enough) might be to cut the 5v trace toward middle of PCB, jump to it near Via hole, and use it's pad instead (5v trace/pad isn't being used). Actually, as I examine under the QC-1 sticker ... this version doesn't appear to have 5v capability, so the 5v trace isn't connected to anything on either end.

It came to rest like this. It was still on and knew it had landed. I don't think landing in the tree caused much damage, rather it was the Nova hitting the ground under the tree. In addition to the Body and ESC, one prop-guard was broken and twisted beyond repair. Three of the props were nicked but none had broken blades. Prop-guards seem to do their jobs ... absorbed crash energy, and likely prevented even more damage to delicate parts (ie motors).

The repair starts here.

Wednesday, October 1, 2014

DroidPlanner (GCS) on Android tablet

I lucked out and just managed to get IceCreamSandwich 4.x installed on my old (circa 2012 with GingerBread 2.x) Chinese Android 7" tablet (iView 760TPC). It is dual-core with 512mb RAM, and supports USB Host Mode (for 3G cell dongles, etc.). The tablet came with a little OTG USB adapter cable that allows you to plug in normal-sized USB flash drives (while I don't use it much, it goes work). The tablet's native USB port is Mini (not Micro) so I can't use the OTG cable supplied with HK 3DR v2 Telemetry radios. Therefore, for now, I will use the iView OTG cable, and the normal USB cable that I use with the radio on the laptop.

At first, DroidPlanner didn't seem to work (radio lights weren't flashing quite right, and DroidPlanner would not Connect). I checked the page at DroidPlanner GitHub, and it mentioned trying the USB Host Check app. Sure enough, the checks all failed (which I thought was strange ... since old OTG USB Adapter was already working for Flash drives). I went ahead and clicked the "Fix It" button and rebooted. I still had the 915mhz radio connected. After it booted up, it detected the radio and ICS asked me if I wanted to use it with DroidPlanner. This was a sure sign that things were looking better. Sure enough, after starting DroidPlanner, it connected with no problems. HUD is working nicely.

I was then getting a "Unfortunately, TTS has stopped" (crashed) on every startup. I clicked OK, and the app worked fine. The error went away completely after the Google Play Store offered a "TTS Update" and I let it install. Now, TTS (Speech Prompts) work fine in DroidPlanner.

I used this mount from HobbyKing. I also upgraded to this 5inch USB Mini-Micro male OTG cable for a much cleaner look.

Edit: In March 2015, I upgraded to a better Android tablet.


USB-OTG cables are asymmetrical (only properly communicate when connected in proper direction). It has more to do with internal wiring than connectors that might be soldered on each end. USB-OTG port should at least work with USB flash drives or "wired" computer direct-connected storage (as a starting point).

DroidPlanner (even DP v1) requires ICS v4.x or higher on Android device. Test with DP v1 first (it has smaller memory and lower resource footprint). If that works, you can uninstall it and try a higher version. DP v2 and v3 might crash or malfunction if your device is not up-to running them. DP v3 (Tower) might require a dual-core or better and 1gb ram (depending on what background apps you are running).

With the Nova's main LiPo battery connected, get Nova working with 915mhz v2 radios in MissionPlanner for Windows first. This checks telemetry radios (important step) and Nova. Always properly Disconnect MP with button (before disconnecting physical cable). Leave radio on Nova, but move laptop's radio to Android device, now using USB-OTG cable.

To verify the your Android device even has a functioning USB port, I like to first test it with a flash-memory device.

You can really only have one version (either v1, 2, or 3) installed at a time (to avoid conflicts). Be sure radio-dongle and/or USB-OTG cable is disconnected at first. Install Droid Planner and reboot device. After full boot, connect radio (it should be detected). Select radio as Default device for DP. Start DP and try to connect.

If radio is not detected, run USB Host Check App from TaurusLabs. If device abilities fail, click Fix button and run Check again. After Connect, move quad and see if Hud responds (this will check basic communications). Forum member Jester1964 reports that they had to enable "USB Debugging" in Developer options before it would work for them.