Thread Tools
Jan 26, 2018, 04:06 AM
OlliW
some update.
I did two further flight tests with my copter and the "old" ESCs and the "new" FuriBee ESCs, where in both cases I was using the UC4H ESC node but with PWM signals, that is the ESCs were connected to the node and run with ordinary PWM, and the node was controlled per UAVCAN. The result is: with the "old" ESCs everything worked perfect, with the FuriBee's I got exactly the same problems as before.
So, the list of results is this

C1) "old" ESCs directly connected to the PWM ports on the flight controller: OK
C2) FuriBees connected to UC4H ESC node using DSHOT: not OK
C3) FuriBees directly connected to the PWM ports on the flight controller: not OK
C4) FuriBees connected to UC4H ESC node using PWM: not OK
C5) "old" ESCs connected to UC4H ESC node using PWM: OK

"not OK" means the troubles described in the above.
So, all possible combinations have been tested now, and the picture is unambiguous:

Whenever I use the "old" ESCs things are just fine, and whenever I use the FuriBee's I get exactly the same sort of problem.

This, and especially test C3, demonstrates that this is an issue of using the FuriBees, and has nothing to do with the UC4H node, which from test C5 appears to work fine, at least as much as testable in these tests.
As a BTW: Knowing the firmware of the UC4H nodes would be of exactly zero help to sort out the issue I have with the FuriBees, in much contrast to Leaflyer's pleas.

Now, obviously, I need to sort out what's going on when using the FuriBee's.

I suspected that this might be an effect of improper tuning of the copter, and the first idea would be to run autotune. However, ArduPilot's autotune would do exactly this kind of tilt maneuvers, which in the above tests made the copter to get confused and fall down, so I'm not sure if running autotune is really a good idea. This is a general ArduPilot question, and not placed well here, yet, maybe someone has an idea how I could tune such a misbehaving copter?
Last edited by OlliW; Jan 26, 2018 at 07:01 AM.
Sign up now
to remove ads between posts
Jan 26, 2018, 11:33 PM
Wisconsin
Quote:
Originally Posted by OlliW
some update.
I did two further flight tests with my copter and the "old" ESCs and the "new" FuriBee ESCs, where in both cases I was using the UC4H ESC node but with PWM signals, that is the ESCs were connected to the node and run with ordinary PWM, and the node was controlled per UAVCAN. The result is: with the "old" ESCs everything worked perfect, with the FuriBee's I got exactly the same problems as before.
So, the list of results is this

C1) "old" ESCs directly connected to the PWM ports on the flight controller: OK
C2) FuriBees connected to UC4H ESC node using DSHOT: not OK
C3) FuriBees directly connected to the PWM ports on the flight controller: not OK
C4) FuriBees connected to UC4H ESC node using PWM: not OK
C5) "old" ESCs connected to UC4H ESC node using PWM: OK

"not OK" means the troubles described in the above.
So, all possible combinations have been tested now, and the picture is unambiguous:

Whenever I use the "old" ESCs things are just fine, and whenever I use the FuriBee's I get exactly the same sort of problem.

This, and especially test C3, demonstrates that this is an issue of using the FuriBees, and has nothing to do with the UC4H node, which from test C5 appears to work fine, at least as much as testable in these tests.
As a BTW: Knowing the firmware of the UC4H nodes would be of exactly zero help to sort out the issue I have with the FuriBees, in much contrast to Leaflyer's pleas.

Now, obviously, I need to sort out what's going on when using the FuriBee's.

I suspected that this might be an effect of improper tuning of the copter, and the first idea would be to run autotune. However, ArduPilot's autotune would do exactly this kind of tilt maneuvers, which in the above tests made the copter to get confused and fall down, so I'm not sure if running autotune is really a good idea. This is a general ArduPilot question, and not placed well here, yet, maybe someone has an idea how I could tune such a misbehaving copter?
Unless I am misunderstanding there is a parameter to reverse the direction of the motor from UAVCAN GUI. Unless you mean bi-directional 3D type motors? It would be a useful test to confirm that the Furibee work PWM sans UAVCAN. Just to make sure they work properly.

Any possibility the CANbus is being saturated?
Jan 27, 2018, 02:07 AM
OlliW
it's not about reverse or not, all motors do spin in the correct direction (otherwise the copter would never get in the air, right)(BTW, a reverse parameter exists only for your Sapog, it's not valid to generalize)

it's certainly also not about bi-directional, uni-directional as usual

I tried to stress that, but I obviously need to stress that stronger: Please note case C3): Here the FuriBee are connected to the flight controller exactly as anyone does with all the ESCs out there, and run from the PWM of the flight controller. Absolutely no CANbus or UC4H in the way.

pl note also case C5), it shows no issue with CANbus being saturated

Last edited by OlliW; Jan 27, 2018 at 02:23 AM.
Jan 27, 2018, 04:38 AM
OlliW
I btw have further looked at the logs, comparing C3 and C1, and I'm getting more and more convinced that this is indeed a tuning problem, i.e., that the PID controllers self-excite to final lose of control ... maybe I just need to reduce RATE_ROLL/PITCH let's see

EDIT: dammed ... in trying lowered rateP in configuration C1 the copter flipped over and fell down (just 1.5m, but) ... one of the motors obviously broke, one of it's phases is now dead !!! ... dammed, no progress this WE

EDIT2: HAHA ... no, it's "just" that the solder in one of the gold contacts apparently went bad and has now infinite resistance !!! That's "funny", since it's a E300 DJI set, and it of course came pre-soldered and configured ... that's how that piece looks like, really, infinite Ohm ... everyday a new experience LOL
Last edited by OlliW; Jan 27, 2018 at 05:46 AM.
Jan 27, 2018, 08:11 AM
OlliW
ok, I now definitely give up on these stupid FuriBee 30A BLHeli_32 ESCs ...
I did now quite a number of different tests with different PIDs, ranging quite a bit, varying all Staibilze_P, Rate_P, Rate_D ... the longer I played the less it was possible to get the copter even up into the air, it would tilt over already at startup, even the original PIDs didn't allow me to get the copter up (no, the motor assignment and spin directions are all good) ... as a control I just went back to the "old" ESCs, and the thing just flies with them exactly as before ... I don't know what's up with these FuriBees, maybe there is one dude which I don't identify ... anyways, these FuriBees just cut some dents in my props, that's all they are good for ...
(if anyone wants to analyze the logs, you're welcome )

so, anyone around here who would borrow me a set of KISS 32A ESC for few weeks?

(they are quite pricey and I don't feel like investing in them unless I know they'll work for me)
Jan 27, 2018, 01:56 PM
OlliW
UC4H ESC mini has arrived

at least this works LOL. Today the pcbs for the UC4H ESC mini arrived, and I could not resist to build four of them ... each node can support 2 motors, so that's plenty.

bench-tests were also successfully passed

initially I wanted to combine them with the FuriBees and install them in my test copter. I then thought that they+FuriBees also would fit nicely into the Solo, and that this would be a cute conversion of the stock Solo to allow using ArduPilot master with all it's advantages. However, obviously these plans are obsolete (see above ). So, I guess, I will install the nodes themselves in the Solo as UAVCAN-PWM converters, i.e., with using the stock Solo ESCs, which however also would overcome the Solo motorpod issue and allow using ArduPilot master ... let's see
Jan 27, 2018, 03:49 PM
Wisconsin
Does not really surprise me. I have had more problems with bad ESC's than anything else in my builds. I have tried a lot of them and the only ones that I really trust are Hobbywing Xrotor. It would be interesting to check out the new Hobbywing 30amp BLheli_32 but it is not clear exactly what features they support from their US website.

Since the ground offset appears to come from high current in the ESC how did you decide that the UC4H node PWM output would not eperience a problem with the ground shift, in the Solo?
Jan 27, 2018, 03:56 PM
OlliW
the ground offset does NOT come from high currents in the ESC. For an offset to occur some resistance is needed, and there is no such resistance in the ESC. It's due to resistance of the - bat wire and on the main board.
Jan 27, 2018, 10:59 PM
Wisconsin
Quote:
Originally Posted by OlliW
the ground offset does NOT come from high currents in the ESC. For an offset to occur some resistance is needed, and there is no such resistance in the ESC. It's due to resistance of the - bat wire and on the main board.
It could be but it could be an inadequate conductor in the ESC. The power distribution through the main brd was certainly a bad idea though.
Jan 28, 2018, 02:21 AM
OlliW
Quote:
Originally Posted by mike_kelly
it could be an inadequate conductor in the ESC.
it could, but it is not
this is the difference between speculation, and having investigated it
(and it's wasn't the likely cause anyways)
Quote:
Originally Posted by mike_kelly
The power distribution through the main brd was certainly a bad idea though.
no, this can be and often is an excellent idea

1 V at 50 A max implies a resistance of 20 mOhm, which is common. I haven't investigated the main board, since that's not what I needed to know, but from the available info there is no fundamental issue in the Solo with this. Also the motorpod by itself has no design flaw. It's the overall combination which causes the problem. It's actually a trap many stumbled into at the time then the never 3.3V electronics met the ATmega (BKonze/SimonK) ESCs.

the main problem is the Chinese wispers effect, too many people distribute to many could be's, and in this matter it's all clearly colored to one side ... a mistake has clearly happened, but many aspects of the design are not as bad as it is wispered
Jan 28, 2018, 10:31 AM
Wisconsin
Quote:
Originally Posted by OlliW
it could, but it is not
this is the difference between speculation, and having investigated it
(and it's wasn't the likely cause anyways)
no, this can be and often is an excellent idea

1 V at 50 A max implies a resistance of 20 mOhm, which is common. I haven't investigated the main board, since that's not what I needed to know, but from the available info there is no fundamental issue in the Solo with this. Also the motorpod by itself has no design flaw. It's the overall combination which causes the problem. It's actually a trap many stumbled into at the time then the never 3.3V electronics met the ATmega (BKonze/SimonK) ESCs.

the main problem is the Chinese wispers effect, too many people distribute to many could be's, and in this matter it's all clearly colored to one side ... a mistake has clearly happened, but many aspects of the design are not as bad as it is wispered
Urban electronic design myth?
Jan 28, 2018, 10:38 AM
OlliW
I don't understand this phrase, obviously because not being native English speaker, but it is most likely very funny, so it made me laugh anyways
Jan 28, 2018, 10:31 PM
Wisconsin
Quote:
Originally Posted by OlliW
I don't understand this phrase, obviously because not being native English speaker, but it is most likely very funny, so it made me laugh anyways
The term "Urban Myth" is used to describe stories passed around extensively that are just not true but commonly believed.

"Urban legend
An urban legend, urban myth, urban tale, or contemporary legend is a form of modern folklore. It usually consists of fictional stories, often presented as true, with macabre or humorous elements, rooted in local popular culture. More at Wikipedia "
Jan 30, 2018, 01:16 PM
OlliW

UC4H "Solution" to Solo's Motorpod Issue


Hey Folks,

since my original plans didn't work out as described in the above, I decided to install the UC4H ESC mini modules in the Solo in order to overcome the stock Solo's "motorpod issue" and allow me to flash BetaCopter. The goal obviously is to get, thanks to BetaCopter, perfectly working smart shots with a STorM32 gimbal.

The "motorpod issue" prevents us from flashing the stock Solo (which I do have) with the latest ArduCopter code. Various "mods" such as installing the green cube or the infamous level-shifter mod are know. My approach here is to use the UC4H ESC mini modules as converters from UAVCAN to PWM, which when installed close to the motorpods, feed them with nice and clean 3.3V PWM signals, overcoming the issue.

The conversion of the Solo unfortunately wasn't as simple as I initially thought, because the mini ESC pcbs are just not designed for that. This is kind of sad, since with pcbs designed for the purpose the conversion would be nearly plug&play (just one wire would have to be soldered). Anyways. I had these pcbs, so I went with them.

The conversion is shown in the pics below.

And, most importantly, the converted Solo indeed FLIES !!!!!!

At this point I really have to thank mike_kelly for advice with parameters and such stuff, which really helped me to get it up. Thx, sir !

Obviously more flight testing needs to be done. Especially, I would like to know what kind of flight maneuvers did trigger the "motor pod" issue. I would like to do exactly these problematic maneuvers, to test, and (hopefully) demonstrate that the concept works. It's clear that it are maneuvers which lead to high current, that's known and understood, but which maneuvers should I do?
EDIT: not really needed, it's just clear that this is a working solution overcoming the issues
EDIT2: I tuned up the Solo a bit, but the largest current load I could get so far, according to the logs, is 23 A ...
EDIT3: see post #511 for a real stress test, speeds of up to 82 km/h and currents up to 41 A ... mission accomplished I'd say

Have fun,
Olli
Last edited by OlliW; Feb 13, 2018 at 10:55 AM.
Jan 30, 2018, 04:11 PM
Registered User
OlliW you can install Arducopter 3.5 directly on stock Solo w/o any mods as long as you keep the max motor current and max PWM out settings limited to the original 3DR values in the 3.5 parameter file. I have it running on 2 Solos for the last 9 months w/o any failure or issues over a countless number of flights under all sorts of conditions.
Root cause of the reported failures are PWM signal voltage drops <3.3V which may occur under high peak loads if current is not limited appropriately.
I tested all sorts of extreme maneuvers and was not able to reproduce any ESC failure with my settings.

As you said, it’s “the Chinese whisper effect “ :-)
Jan 30, 2018, 04:20 PM
OlliW
thx
I think I have analyzed the situation carefully and understand the issue well, but thanks for summarizing it here again
I'm glad to hear that your approach works for you, but, to be honest, I don't consider it a real solution, and wouldn't recommend it generally
it's as you said, the issue are high currents in peak loads, but these happen mostly and especially in transients and not static situations. Thus, any static limiters such as max PWM out won't solve the problem.

I didn't meant "Chinese whisper" to mean that there is no issue. There is.

I'd like to add that with the default parameters the Solo is a lame duck, and it is surprising it consumes any current at all given that it flies like that. This might be an additional reason for why AC3.5 might work on stock Solo's in some/several cases.
Last edited by OlliW; Jan 30, 2018 at 04:26 PM.
Feb 03, 2018, 02:10 PM
OlliW

UC4H ESC Nodes work!


Hey Folks,

next big step forward, the UC4H ESC node works perfectly!!!

This time it's with my test copter, which gave me such a hard time, see few posts up.

First, I have to deeply thank Flyudino and especially Felix. Flydunio offered me a reduction of 50% on a set of 4 KISS 32A ESCs, which I of course grabbed, and without which this here would not have happened. Felix was really very helpful with a number of discussion points. THANKS, guys!

With the KISS 32A ESCs installed and hooked up to the DIY variant of the UC4H ESC node, some pics are below, and configured to DSHOT, the test copter just flew out of the box. Not a single issue or anything (the setup of course also passed all my bench tests). I pushed the copter a bit around from left to right, as you can see from the log screenshot, no issue.

As it seems, I do have now two copters successfully flying with UC4H ESC nodes.

Mike: I caught up, you are not the only one anymore with a only-three-wire setup. Sorry, sir

I guess it's time for me to get back to betacopter, to make use of the ESC telemetry data. I guess I also want to design a pcb which snugly fits to the KISS, to make this all more clean and convenient.

Cheers,
Olli
Last edited by OlliW; Feb 04, 2018 at 09:25 AM.
Feb 03, 2018, 11:35 PM
Wisconsin
Quote:
Originally Posted by OlliW
Hey Folks,

next big step forward, the UC4H ESC node works perfectly!!!

This time it's with my test copter, which gave me such a hard time, see few posts up.

First, I have to deeply thank Flyudino and especially Felix. Flydunio offered me a reduction of 50% on a set of 4 KISS 32A ESCs, which I of course grabbed, and without which this here would not have happened. Felix was really very helpful with a number of discussion points. THANKS, guys!

With the KISS 32A ESCs installed and hooked up to the DIY variant of the UC4H ESC node, some pics are below, and configured to DSHOT600, the test copter just flew out of the box. Not a single issue or anything (the setup of course also passed all my bench tests). I pushed the copter a bit around from left to right, as you can see from the log screenshot, no issue.

As it seems, I do have now two copters successfully flying with UC4H ESC nodes.

Mike: I caught up, you are not the only one anymore with a only-three-wire setup. Sorry, sir

I guess it's time for me to get back to betacopter, to make use of the ESC telemetry data. I guess I also want to design a pcb which snugly fits to the KISS, to make this all more clean and convenient.

Cheers,
Olli
Now the race is on for an all UC4H with STorM32NT CANbus gimbal. Vroooomm.
Feb 04, 2018, 12:09 AM
OlliW
Quote:
Originally Posted by mike_kelly
Now the race is on for an all UC4H with STorM32NT CANbus gimbal. Vroooomm.
ähm ... I assume, I should not be admitted to this race ... since then the race would be over:

test copter:
- UC4H GPS magnetometer
- UC4H power brick
- UC4H ESCs
- T-STorM32 gimbal with CANbus via UAVCAN

looks quite complete to me ... only gadgets come to my mind, such as a Notifier node

but, I agree, and herewith offer another badge to win: The first all UC4H with STorM32 CANbus gimbal User.

this one might be even harder to win than the other badge. Will be interesting to see.
Last edited by OlliW; Feb 04, 2018 at 12:17 AM.
Feb 04, 2018, 09:52 AM
Wisconsin
Quote:
Originally Posted by OlliW
ähm ... I assume, I should not be admitted to this race ... since then the race would be over:

test copter:
- UC4H GPS magnetometer
- UC4H power brick
- UC4H ESCs
- T-STorM32 gimbal with CANbus via UAVCAN

looks quite complete to me ... only gadgets come to my mind, such as a Notifier node

but, I agree, and herewith offer another badge to win: The first all UC4H with STorM32 CANbus gimbal User.

this one might be even harder to win than the other badge. Will be interesting to see.
I want a UC4H Pixhawk with built-in SLCan node so you can view Canbus messages with mission planner during flight.
Feb 04, 2018, 12:06 PM
OlliW
I was actually thinking about how to integrate a SLCAN adapter ... in my STorM32 code this would be relatively easy since I this "tunnel" mode quite extensively for a variety of cases ... but in ArduPilot I'm not sure how one would do that, ArduPilot has such a thing e.g. for GPS, I think, so it should be possible
this would be kind of cool
it would however never work through telemetry, i.e., during flight, just because of bandwidth
Feb 04, 2018, 12:38 PM
Registered User
Well I am going on this. I truly hate soldering 10 AWG cables to anything but it is done. The other side of the power brick is has the wires and solder joints entombed in epoxy and likewise for the PDB. My work bench is the usual catastrophe.

This will not win any badges as the power system is a DJI E1200 Pro. This is an extraordinarily rigid frame that is rated for 15KG total weight but will be no more than 6KG even with DLSR, batteries and gimbal. It has an RTK module as well that goes on the front with a hard mounted Nex 5 camera -- this was a prototype mapping quad supposed to fly around Alberta, Canada. I have the cable on order to connect it to the Pixhawk and on to the GPS/Mag node.

I should get it ready soon as I am rolling. Probably will have a 1.5W Dragonlink V3.

I have scratched my head as to what to do with it since it never carried the FLIR that was envisioned (Matrice system came along). I have not flown Arducopter much since I acquired a Phantom 4/ 4 Pro but I think this will carry a 2 or 3 Axis axis gimbal (probably Storm32) and a good DLSR and I will use it for coastal photography. I expect this to be untroubled by wind. Probably will drive around Lake Michigan and do some landscapes.

I will post better pics as I progress, including flight pics.
Last edited by Marc Dornan; Feb 04, 2018 at 02:41 PM.
Feb 04, 2018, 01:42 PM
Wisconsin
Quote:
Originally Posted by Marc Dornan
Well I am going on this. I truly hate soldering 10 AWG cables to anything but is done. The other side of the power brick is has the wires and solder joints entombed in epoxy and likewise for the PDB. My work bench is the usual catastrophe.
Come on Marc you can see your bench top, that means it is not a mess yet! I hate soldering 10g wires too and OlliW can attest to my really bad results.
Feb 04, 2018, 02:14 PM
Wisconsin
Quote:
Originally Posted by OlliW
I was actually thinking about how to integrate a SLCAN adapter ... in my STorM32 code this would be relatively easy since I this "tunnel" mode quite extensively for a variety of cases ... but in ArduPilot I'm not sure how one would do that, ArduPilot has such a thing e.g. for GPS, I think, so it should be possible
this would be kind of cool
it would however never work through telemetry, i.e., during flight, just because of bandwidth
Sure but I was thinking about sampling at a low rate for telemetry. It would be very powerful to know the temp of your ESC, for instance, even if you got one sample per second.
Feb 04, 2018, 02:39 PM
Registered User
Quote:
Originally Posted by mike_kelly
Come on Marc you can see your bench top, that means it is not a mess yet! I hate soldering 10g wires too and OlliW can attest to my really bad results.
Well, I made an effort. I moved my empty whisky glass from last night out of frame. This build is sponsored by modest drams of Lagavulin...
Feb 04, 2018, 02:51 PM
OlliW
@Marc:
GREAT, looks like a cool re-purposing of this frame

pl let me mention, what are the expected currents? the power brick has a 100A Hall sensor, and isn't really designed for large currents.

did you had a chance to subject the power brick to a "cold" test, by connecting it to a SLCAN adapter, just to see if it is alive, and e.g. reports the voltage

I think the workbench looks like it has to . One should e.g. always have a fragment of a micro gimbal with 1806 motors lying around ...

@Mike:
ah, I see

the temp of the ESC btw is already obtained with the UC4H ESC node when used with a DSHOT ESC, such as the KISS
it is just a matter to make ArduPilot digest that message already available on the CANbus

I guess it wouldn't be easy for me to get it to MissionPlanner, since this would need adding a Mavlink message (or maybe there is already one???), adding it to ArduPilot's streams, and adding it to all GCSs.
Maybe one could add it easily to the FrsSky telemetry? For those using OpenTx or such.

That's useful discussion, because my plan so far was to make ArduPilot to log the telemetry data in the DataFlash, so that it could be read after flight by MissionPlanner. But maybe there is an even simpler to implement realtime solution.
Feb 04, 2018, 03:44 PM
Registered User
Quote:
Originally Posted by OlliW
@Marc:
GREAT, looks like a cool re-purposing of this frame

pl let me mention, what are the expected currents? the power brick has a 100A Hall sensor, and isn't really designed for large currents.

did you had a chance to subject the power brick to a "cold" test, by connecting it to a SLCAN adapter, just to see if it is alive, and e.g. reports the voltage

I think the workbench looks like it has to . One should e.g. always have a fragment of a micro gimbal with 1806 motors lying around ...

@Mike:
ah, I see

the temp of the ESC btw is already obtained with the UC4H ESC node when used with a DSHOT ESC, such as the KISS
it is just a matter to make ArduPilot digest that message already available on the CANbus

I guess it wouldn't be easy for me to get it to MissionPlanner, since this would need adding a Mavlink message (or maybe there is already one???), adding it to ArduPilot's streams, and adding it to all GCSs.
Maybe one could add it easily to the FrsSky telemetry? For those using OpenTx or such.

That's useful discussion, because my plan so far was to make ArduPilot to log the telemetry data in the DataFlash, so that it could be read after flight by MissionPlanner. But maybe there is an even simpler to implement realtime solution.
I thought about that. I have a Mauch 200A module that I intended to use for this machine. I checked out out the specs of the sensor and my system. The absolute max rating of this system would be 160A! In reality I am almost certain it will not hit 100A. It would need to be on a heavily laden machine at the limits of its weight in heavy wind and even then I think each esc would not hit 40A. It is a theoretical bench figure. The power module this replaced was a joke (a 3DR style shunt sensor) and it survived testing. So while I would not use this for a professional application, I think there is really is no risk as I have deployed it. I will check it carefully as I test though. This has hovered at 28A with mapping payload. Even with a camera and 2 axis Nex gimbal I doubt it will hover much over 35A.

If I have totally screwed up (doubt it) I could swap it but I could also shunt the sensor and change the mv per Amp, right? I made sure there was a lot of solder on the back so the tracks will not be an issue. A bit unorthodox but I think it would work. Again, I do not think this will be an issue.

Yes, it is alive and it tested ok.
Last edited by Marc Dornan; Feb 04, 2018 at 03:51 PM.
Feb 04, 2018, 03:48 PM
OlliW
I was sure you had thought about it, just wanted to double check
the copter is just bigger than anything I would ever had in my mind (my mind stops at 450 )
maybe next time I order pcbs, I'll order them with 70u Cu layer, and if it's just to get a feel for how these would solder

Quote:
I made sure there was a lot of solder on the back
it's a bit of an urban myth (hey, Mike, I'm learning ) that putting on a lot of solder helps ... solder has roughly 10 times lower conductance than Cu ... but I assume you certainly have kept the distances of the plus wires to the hall sensor pins short, as well as have no or a very short gap in the minus wire, so I'm actually not that much concerned about the tracks ... it was really just the max amp of the Hall sensor

the 70u Cu comment was not meant to point to an issue with the tracks, it's just an additional level of just doing it a bit better on my side

Quote:
Yes, it is alive and it tested ok.
great, I'm happy to hear that

looking forward to see how this continues
Last edited by OlliW; Feb 04, 2018 at 05:29 PM.
Feb 05, 2018, 04:26 AM
Registered User
Olli, based on your approach, do you think someone could develop a "drop-in" or "plug&play" solution which requires no soldering for the stock Solo's ESC issue? Would it be cheaper than a PxH 2.1 GreenCube?
Feb 05, 2018, 08:37 AM
Wisconsin
Quote:
Originally Posted by Marc Dornan
Well, I made an effort. I moved my empty whisky glass from last night out of frame. This build is sponsored by modest drams of Lagavulin...
My family history is from Inverness so I must be loyal to the distillers on the river Spey.
Feb 05, 2018, 09:41 AM
OlliW
Quote:
Originally Posted by WaldoPepper99
Olli, based on your approach, do you think someone could develop a "drop-in" or "plug&play" solution which requires no soldering for the stock Solo's ESC issue? Would it be cheaper than a PxH 2.1 GreenCube?
not fully
if you would however accept one solder joint per ESC, then: yes, I would think so
(it would have to be tested, but I think that's possible)
it certainly would be much cheaper than a green cube, just to give a number out of the blue, depending on how mass produced this would be it could be below 25$ or so for all 4 ESCs I think
there could be even much cheaper and easier solutions, not based on my approach, could come down to few bucks per ESC
Last edited by OlliW; Feb 05, 2018 at 09:55 AM.
Feb 05, 2018, 10:00 AM
Registered User
Nothing wrong with Speyside malts.

So I can make a can I simply splice a CAN cable from Pixhawk to the power prick and GPS/Mag node? Or does it need to daisy chain using the connectors? A bit of a CAN noob question.

Olli, you may not have seen my addition to my post as I think I edited it after you replied. My fall back plan if my module is not within safe limits is to add some 12 AWG wire in parallel with the sensor and then recalibrate the mv per amp. I was careful to make sure I have added solder so the copper tracks are not a limiting factor. The system should still be linear. Of course I am not expecting you to certify the safety of this. Just the general principal. I stress test all things on the ground extensively before I fly.
Feb 05, 2018, 10:10 AM
OlliW
Quote:
Originally Posted by Marc Dornan
So I can make a can I simply splice a CAN cable from Pixhawk to the power prick and GPS/Mag node? Or does it need to daisy chain using the connectors? A bit of a CAN noob question.
the correct answer I guess depends strongly on who you ask
a technical perfectionist would likely say, no way other than to the specs is allowed hence must be daisy chained and correctly terminated
if you ask me, given that we are talking about short wiring distances, don't care and wire as you like ... that's at least what I do, I don't even place two terminations, just one (on the pixhawk), and I have no indication that this is a problem (I gladly chose CAN transceivers which help quite a bit in making the bus more robust)
your copter is somewhat bigger but if it would be mine I would not care and wire as its convennient
all this is of course just what I would do, I might be wrong, and of course may not to be held responsible for anything

Quote:
Originally Posted by Marc Dornan
Olli, you may not have seen my addition to my post as I think I edited it after you replied.
I've noticed it and edited my response accordingly
Feb 05, 2018, 12:25 PM
Registered User
Quote:
Originally Posted by OlliW

it's a bit of an urban myth (hey, Mike, I'm learning ) that putting on a lot of solder helps ... solder has roughly 10 times lower conductance than Cu ...


Well slap my empty whisky bottle! I have been slathering solder for years!

Well you can also call it an "Old wives' tale" or Alte Ehefrauengeschichte, but in these days that may be regarded as sexist, and I do not know any old wives that solder, so maybe stick with Mike's vernacular.
Feb 05, 2018, 05:33 PM
OlliW
just did and ordered this:
Feb 05, 2018, 05:46 PM
Registered User
Kiss ESC Can node adaptor?
Feb 05, 2018, 06:05 PM
OlliW
correct
I'd like to clean up my copter, this DIY esc node is great, but the way I've build it it's quite messy ... with the carrier it should be nice and clean
Feb 05, 2018, 11:16 PM
Wisconsin
Very cool! Is it a common profile or only good for the Kiss BLHeli_32? Not that I will use them after your report but as an example would the furibee 32 esc fit?

Like the wraith look like the same form factor but are 33% less expensive.
https://www.amazon.com/Wraith32-Plus.../dp/B075JSHP3W
Last edited by mike_kelly; Feb 05, 2018 at 11:24 PM.
Feb 06, 2018, 02:26 AM
OlliW
there is no such thing as a "Kiss BLHeli_32"

this I've designed to (hopefully) fit the KISS 32A. It likely should also fit the KISS 24A, I believe it is at least supposed to. It may happen that it fits to other ESCs, but this would be accidentally.

The FuriBee's certainly do not fit. I don't have the Wraith, so can't answer to this question. When I was looking for BLHeli ESCs I was also looking at the Wraith, but didn't considered them because they do not have the telemetry pin easily accessible.

Yes, I too consider the KISS 32A quite expensive. However, even with a uc4h esc kiss carrier it is well below 40Eur, and thus still ca half the price of the next UAVCAN ESC, which I think is the Orel20/Sapog. From that perspective it's a bargain.
(and despite it's huge price point you did had issues with it!).

To be fair, I want to make that clear: My above results with the FuriBee's must not be taken as representative for BLHeli_32 ESCs! This probably was just bad luck.
Feb 06, 2018, 05:52 PM
Registered User
dk7xe.g's Avatar
@OlliW will you sell this adapter boards? Felix just notified me about this since i use his 32A ESC together with our PX4 reference design. Would really love to go for UAVCAN on the long run.
Feb 06, 2018, 06:07 PM
OlliW
no, I don't sell anything, I do that just for my personal fun
Felix is closer to this selling thing than me
if they are working as I hope, we might find some way to get you a set
funny alias btw
Feb 06, 2018, 08:48 PM
Registered User
Olli -- I do not want to compile Betacopter (since I am not making any changes) do you have a Pixhawk binary. I did not see a binary here: https://github.com/olliw42/storm32bg...005/ArduCopter
Feb 06, 2018, 09:15 PM
OlliW
choose one of the ArduPilot-v?.px4 files, v2 for pixhawk, v4 for pixracer, v3 for cube (I think LOL)
Feb 07, 2018, 08:30 AM
Wisconsin
Quote:
Originally Posted by Marc Dornan
Olli -- I do not want to compile Betacopter (since I am not making any changes) do you have a Pixhawk binary. I did not see a binary here: https://github.com/olliw42/storm32bg...005/ArduCopter
Marc, the binaries are mixed in with the source under the Arducopter folder.

OlliW I don't think you ever posted the V3 version.
Feb 07, 2018, 08:52 AM
OlliW
before I responded to Marc I looked up the git link, and it has a v3
I admit I was surprised since I don't recall having done this, but it was a positive surprise LOL
Feb 07, 2018, 09:51 AM
Registered User
Quote:
Originally Posted by mike_kelly
Marc, the binaries are mixed in with the source under the Arducopter folder.

OlliW I don't think you ever posted the V3 version.
Thanks -- I see them now.
Feb 07, 2018, 03:14 PM
Wisconsin
Quote:
Originally Posted by OlliW
before I responded to Marc I looked up the git link, and it has a v3
I admit I was surprised since I don't recall having done this, but it was a positive surprise LOL
Ah, it is the zip package that does not have the V3.
Feb 07, 2018, 03:18 PM
OlliW
yes, it's not in the .zip, but it is in this link https://github.com/olliw42/storm32bg...005/ArduCopter which Marc was giving
ArduCopter-v2.px4
ArduCopter-v3.px4
ArduCopter-v4.px4
in that folder you should not look for 'betacopter', that's because of how the build system works
I guess I should remove the zip

EDIT: should be clearer now
Last edited by OlliW; Feb 07, 2018 at 03:31 PM.
Feb 07, 2018, 04:40 PM
Registered User
Quote:
Originally Posted by OlliW
yes, it's not in the .zip, but it is in this link https://github.com/olliw42/storm32bg...005/ArduCopter which Marc was giving
ArduCopter-v2.px4
ArduCopter-v3.px4
ArduCopter-v4.px4
in that folder you should not look for 'betacopter', that's because of how the build system works
I guess I should remove the zip

EDIT: should be clearer now
I forgot that you must use the RAW button and Save as. Git is not intuitive when it comes to getting binaries.
Feb 07, 2018, 08:03 PM
Wisconsin
Quote:
Originally Posted by Marc Dornan
I forgot that you must use the RAW button and Save as. Git is not intuitive when it comes to getting binaries.
GIT is NOT intuitive at anything.


Quick Reply
Message:

Thread Tools

Similar Threads
Category Thread Thread Starter Forum Replies Last Post
Question Tools and specifically hobby knifes for arthritic hobbiests maxadventure The Builders Workshop 2 Jun 13, 2012 08:12 PM
Discussion You might be a Hardcore R/C hobbiest if... graphixman Dock Talk 55 Jun 18, 2008 08:58 AM
Competitions for Hobbiest andyg UAV - Unmanned Aerial Vehicles 21 Jul 04, 2005 11:37 AM
Need some advice from Ducted fan Hobbiest.. XiiiXiii Fuel Jet Talk 6 Mar 31, 2004 02:03 PM