Shop our Airplanes Products Drone Products Sales
Thread Tools
Feb 08, 2018, 03:51 PM
OlliW

UC4H ESC node with DSHOT and telemetry is working!


so, folks

as it seems I do have DSHOT and ESC telemetry working !!!

I just did a test flight with my testcopter, equipped with KISS 32A, and a DIY UC4H ESC node, and the telemetry cables from the ESCs connected to the UC4H ESC node. For telemetry to work this of course requires DSHOT. Also BetaCopter needed to be adapted, which was however surprisingly simple, didn't expected this when I started coding.

Attached below some pics of the flight.
pic1: rpm vs time
pic2: current vs time
pic3: rc input signals vs time
Please note the ESC1,ESC2,ESC3,ESC4 data fields in the bottom left list.
The rc input I've added so you see that I did some roll sweeps as well as some yaw turns.

How cool is that.

Not everything is yet perfect, some little things can be improved, but the basic ground work seems to have been done.

I'm running out of ideas what other UC4H nodes to do LOL

Have fun,
Olli
Sign up now
to remove ads between posts
Feb 08, 2018, 03:59 PM
Registered User
Those are some promising results! Well done, and keep up the good work
Feb 08, 2018, 04:09 PM
OlliW
thx, I will

the plots are actually quite interesting, the current of motor 4 is clearly much higher in hover than for the other motors. Not sure yet, but this motor has the by far most damaged prop. I guess I should swap the prop with motor 3, or install a new one to see if that is indeed the effect. But obviously, I could work on motor 4. Or is that simply a bit of imbalance? I guess will be interesting to see why that is.

is it maybe that already the first flight with ESC telemetry is going to pay out for me ??? LOL
Feb 08, 2018, 09:39 PM
Wisconsin
Quote:
Originally Posted by OlliW
thx, I will

the plots are actually quite interesting, the current of motor 4 is clearly much higher in hover than for the other motors. Not sure yet, but this motor has the by far most damaged prop. I guess I should swap the prop with motor 3, or install a new one to see if that is indeed the effect. But obviously, I could work on motor 4. Or is that simply a bit of imbalance? I guess will be interesting to see why that is.

is it maybe that already the first flight with ESC telemetry is going to pay out for me ??? LOL
Great progress! Such a fun project. So now you could have time to work on the UC4H fc with built in SLcan

How about a OLed status display, buzzer, switch module?
Feb 09, 2018, 01:32 AM
OlliW
yeah, I must admit, I'm really very happy with the telemetry (as much as I am with my power brick). I mean, this doesn't just replace the wires, like the gps-mag node essentially does, but really gives some new benefit which wasn't available to me before.

oh, yes, integrating a SLCAN adapter into ArduPilot would be great, as said, I had that idea too. But I assume that this is a bit more work. But it's definitely not out of my mind!

There is actually a whole lot one could (or has to?) do to make using UAVCAN much more user friendly, e.g., having the parameters accessible in the GCS. The basics has been established, QGC can do that, and as much as I know MissionPlanner is ready too. It's just that ArduPilot hasn't respective code. I was once thinking trying this, but the larger problem is that this would require quite some changes in the code, which bears the risk that all the work becomes obsolete because of a change in ArduPilot coming out over night.

I actually also thought about OLeds, etc. pp., I actually though about that before going with the ESCs. But here again, the issue is that there is no defined UAVCAN Notifier messages, commands. There is actually an issue/pr on ardupilots git. Sadly, it seems, they again go the short rout to just do LEDs, but not the general route even though it's visible already now.
You may understand that I am never ever going again to discuss improvements to the uavcan dsdl messages.

What I want to do, at some point, is to add LED(s) to the gps-mag node. I really would want to see what's up by looking at them. I long ago tried that when I still was Uart/I2C, but stopped because I found that to much work.With the node now, this would not take much effort hardware-wise.

I think my plan for the weekend though is to get the STorM32's horizon drift and yaw drift compensations working with ArduCopter.
Last edited by OlliW; Feb 09, 2018 at 01:51 AM.
Feb 09, 2018, 02:36 AM
OlliW
as regards git

I thought that this time I could do a proper fork of ardupilot, and work on that fork, so I did and besides a little hazzle here and there it went ok and I was happy ... just to realize now that I can't switch the branch because of a nastiness with a submodule which I just can't figure out ... I've spend now more time on trying to solve that issue with fail then what it took me to implement the escStatus message and DataFlash storing and doing the test flights ... what a waste of time ... I'm done with git, and went back to my hobby-grade techniques, which however just work and let you allow to progress

summary, I finally understood it:

It is no coincidence that git rhymes with s.it!!
Feb 09, 2018, 11:52 AM
Wisconsin
I think that is the major point. I am sure there are compelling reasons to use GIT as a day in day out professional coder. But for most hobbyist the time it takes to master GIT could be larger than your whole project. So it creates an unfortunate gap between valuable volunteers and the dev team. Randy recently posted a note talking about the need for more part time participants but GIT gets in the way.
Feb 09, 2018, 12:52 PM
Registered User
americanpelican's Avatar
OlliW, first of all many thanks for your UC4H initiative. It nicely fills the gap of UAVCAN-ready hobby devices.

As background, I'm replacing the PWM ESCs of my hobby PixHawk drone with BLHeli 32/DShot ones, and I just have built UC4H CAN adapter and ESC Actuator.

Now I'm configuring the actuator's parameters via UAVCAN GUI... my primary objective is to eliminate the need for ESC calibration (DShot) and take advantage of ESC telemetry in the future.

With that, could you point me to documentation about UC4H ESC actuator parameters? Specifically, I want to ensure they are set with the highest DShot option available. (I'm guessing the parameters OutA/OutB/OutC Mode would allow choice across PWM, OneShot, DShot)

Any guidance would be much appreciated!

Luis
Feb 09, 2018, 02:13 PM
OlliW
Hey
welcome

from your text I infer that you successfully can see the uc4h esc node in the UavcanGuiTool, i.e. that both the node and the slcan adapter are working. If, so, great job sir.

I can't point you to a documentation, since there isn't any; the project parts are still in development flow (and the user base is better approximated by 0 than by 10 ).

At first, I would recommend that you flash the latest v008 firmware, it's in the git repro.

You should then see parameters like the ones shown in the attachment.

"OutX Mode" means
0: the two associated pins are both disabled
1: Actuator PWM
2: ESC PWM
3: ESC DSHOT

The actuator function is not yet implemented.

"OutXY Index" means
-1: this specific output is disabled
0...20: this is the index in the esc.RawCommand to which this output listens.

Thus, to enable a pin both Mode and Index must be valid (this can be quite handy in testing).

Thus, more than one output could have the same index, but that's probably not useful. For ArduPilot the indices should start from 0 to number of motors-1.

"OutX PwmPeriod" means
sets the time in us between two PWM pulses, only relevant for OutX Mode = 1 or 2

"EscStatus Rate" means
0: disabled
>0: sets the time in us in which the escStatus messages are emitted. Since each escStatus messages covers just one motor, they are send with gaps of ca 5 ms.

"Motor Pole Pairs" means
0: disabled
>0: the DSHOT telemetry provides the electrical rpm, in order to convert to mechanical rpm the number of motor poles need to be set

As regards the highest DSHOT, the node doesn't offer any options. This is because the F1 processor which it (currently) uses is limited in that respect. The rules are this:
if only one output is enabled, then DSHOT600 is used
if two or more outputs are enabled, then DSHOT300 is used

the limitation of the F1 makes that the more outputs are enabled the less precise the DSHOT low/high times are. With e.g. 4 outputs enabled, there is already quite some variation, and one is relying on the ESC. I have not tested what happens with all 6 outputs enabled. I guess I wouldn't recommend this.

You may find this disappointing, but it's really a non-issue. With DSHOT300 an additional latency of < 60 us is introduced which you should compare with the latency due to the CAN bus, of 150-250us. With DSHOT600 it's just a ca. 10% increase of latency. I think this can be considered irrelevant (if you think you need significantly lower latencies, you shouldn't use CAN). With only one output and the right pin one could go to DSHOT1200, but as said I think that's pointless, and hence I played it safe and chose DSHOT600. To improve that, someone would have to identify a small and cheap STM32 which is (known to be) much better than the F1.

As regards building and testing the setup, I want to direct you to the ESC Panel in the UavcanGuiTool (I think Mike has discussed it in his ArduPilot.Discuss blog). I now generally first test and set the motors with this tool, before I go in MissionPlanner to proceed.

Also, my probably best "advice": I would first ensure that with the ESCs driven as usual per PWM from the flight controller the copter is perfectly flying !!! Otherwise it can be difficult to sort out issues, in case there are some.

Hope this helps a bit.
Olli
Last edited by OlliW; Feb 09, 2018 at 02:22 PM.
Feb 10, 2018, 03:50 PM
OlliW
as a continuation to my UC4H-Solo mod to overcome the motorpod issue, see post #464, I today did a "speed" test with my Solo.

As expected, I could not observe any issues.

However, I put "speed" in brackets since I'm not yet satisfied with how it flies, I think it should do better. It seems that this speed test has not really stressed the Solo. Please see the attached pic for some data of the flight. >50 km/h speed, not sure if that's good.

Most strangely, the current did not went above hover-level !!!! It's a bit like the earlier finding where I couldn't get large currents. I doubt this can be right, even at full stops after full speed the current stays low. Is there something wrong with the Solo battery and AC3.6?

Anyway, the mod has passed also this test without issues. Certainly increases confidence.

have fun, Olli
Feb 11, 2018, 03:58 AM
OlliW
ok, this was now a much better stress test of the Solo
increased ANGLE_MAX from 30 to 45, and the effect was very noticeable

max speed 82 km/h
max current 41 A

no noticeable issues with the UC4H-modded Solo
inspect the the data in the pic for yourself

Feb 11, 2018, 09:21 AM
Wisconsin
Quote:
Originally Posted by OlliW
ok, this was now a much better stress test of the Solo
increased ANGLE_MAX from 30 to 45, and the effect was very noticeable

max speed 82 km/h
max current 41 A

no noticeable issues with the UC4H-modded Solo
inspect the the data in the pic for yourself

Lookin good!
Feb 18, 2018, 05:04 AM
OlliW

My rehauled, fully UC4H-ized copter


Hey Folks

so, the UC4H ESC Kiss carrier boards arrived, and I of course immediately built them, tested them, installed them, and test-flew them - they are just GREAT!

I also got annoyed by the fact that the pixhawk doesn't give visual information on the info I really want to have, so I just quickly did also a UC4H Notify node. Sometimes it's actually just funny how things fall into place, so I have to tell this story: I had WS2812 LEDs in my box since very long, but never got around to use them. Looking now at them again I realized that the protocol is essentially exactly what is used for DSHOT!!! Well, it's the over way around, DSHOT uses the same protocol as the WS2812, but the bottom line is: I just had to re-purpose what I already did before, namely the UC4H ESC nodes . Anyway. just great, this UC4H Notify node. It's not yet displaying exactly what I want, but that I obviously can improve.

So, with all these new pieces, I decided to give my test copter a complete rehaul, and to throw out everything which I can replace by UC4H nodes. This you can read as an indication of my level of confidence in these components. This doesn't mean that everything is absolutely perfect. E.g., the cold weather right now had required improvements to the magnetometer code. But it's way perfect enough to fly safely and comfortably - at least, since I'm using these UC4H components, I myself had never an issue, where e.g. the copter would come down because of them. So, whatever anyone out there in this big world thinks has to say about the UC4H stuff, and I'm sure there are sufficiently many naynay sayers in ArduPilot world, but ... well, that's reason why I do have a flying, fully UAVCAN-ized copter . Anyway, I'm flying confidently with them . Hence this rehaul.

Just for reference, that's what it features are:
* UC4H ESC Kiss32A, running DSHOT600, providing full live telemetry (and betacopter stores that data for later analysis)
* UC4H GPS-Magnetometer node, dual-GPS setup
* UC4H Power Brick, with precision current and charge measurement
* UC4H Notify, just for the sake of fun

Pictures of all that below.



Have fun,
Olli
Feb 18, 2018, 12:54 PM
KD2PBU - Fly No Evil
davidbitton's Avatar
How much of a code change to ArduCopter is necessary for this to work right now?
Feb 18, 2018, 01:09 PM
OlliW
fair question
I've put together some info here http://www.olliw.eu/2017/uavcan-for-hobbyists/#chapter0
should answer this I hope
it's though not yet updated to include also the UC4H Notify node, will happen soon
from that list you can see that the gps-mag node and the esc node work with current master
the uc4h power brick, uc4h esc telemetry and uc4h notify need few code additions, you can cherry pick them from my fork if you want, or just use that fork
Feb 18, 2018, 02:40 PM
Registered User
americanpelican's Avatar

ArduCopter 3.6 master not generating esc.rawcommand frames


Olli, Mike et al.,

I would appreciate any insight on the setup of UC4H ESC Actuator and Pixhawk/ArduCopter 3.6 master.

I have built the UC4H SLCAN Adapter, and using UAVCAN GUI, I'm able to monitor "status" frames from both ArduCopter and UC4H ESC Actuator, but after arming the FC it does not generate any esc.rawcommand (or similar) frames, reacting to me moving throttle up or down.

Tests performed with ArduCopter 3.5.5, 3.6 Master, and Olli's BetaCopter 3.6, all the same behavior.

My CAN set up is not complex, and likely missing something obvious to you... but now stuck on figuring out why ArduCopter CAN node is only generating status frames.

CAN_P1_DRIVER=1
CAN_D1_PROTOCOL=1
CAN_P2_DRIVER=1
CAN_D2_PROTOCOL=0 // using a single bus

CAN_D1_UC_ESC_BM=15 // 4 motors
CAN_D1_UC_NODE=1 // have tested with different IDs
CAN_D1_UC_SRV_BM=0 // no servos
CAN_P1_BITRATE=1000000
CAN_P1_DEBUG=1

Hopefully some of you have crossed this bridge already.

Thank you,

Luis
Feb 18, 2018, 05:50 PM
KD2PBU - Fly No Evil
davidbitton's Avatar
Quote:
Originally Posted by OlliW
fair question
I've put together some info here http://www.olliw.eu/2017/uavcan-for-hobbyists/#chapter0
should answer this I hope
it's though not yet updated to include also the UC4H Notify node, will happen soon
from that list you can see that the gps-mag node and the esc node work with current master
the uc4h power brick, uc4h esc telemetry and uc4h notify need few code additions, you can cherry pick them from my fork if you want, or just use that fork
thanks. BTW, is the .sch file for the magnetometer node available?
Feb 18, 2018, 06:30 PM
OlliW
Quote:
Originally Posted by americanpelican
CAN_P1_DRIVER=1
CAN_D1_PROTOCOL=1
CAN_P2_DRIVER=1
CAN_D2_PROTOCOL=0 // using a single bus

CAN_D1_UC_ESC_BM=15 // 4 motors
CAN_D1_UC_NODE=1 // have tested with different IDs
CAN_D1_UC_SRV_BM=0 // no servos
CAN_P1_BITRATE=1000000
CAN_P1_DEBUG=1
look below for my params
I'm not sure what CAN_P2_DRIVER=1, CAN_D2_PROTOCOL=0 means, especially if it really means that it just uses one bus, or what protocol = 0 is actually doing
pl choose the parameters as reported
I also would use NODE = 10, it's pointless changing it
you should NOT assume that AC's UAVCAN support is working flawlessly, e.g. choosing settings which one might think are perfectly reasonable and hence would expects them to work may very often in fact not work, it's a bit of a mess ... we spend a lot of time to figure out what works and what doesn't, so, pl stick to recommended params
it is meaningless to try 3.5.5, and both AC3.6 and BC are known to work, so choose one, that's not your problem
it would be helpful to be more precise, I only can assume that "Status" means the NodeStatus, and not the esc Status.
you must see the esc.RawCommand messages in the UAVCAN GUI Tool, my first guess was that you didn't arm, but you say you did, so not sure what's going on, maybe you have enabled some other checks?
I only can tell that it would be best to provide proper info, e.g. in the past we fooled around for long to find an issue a user had, and it kind of immediately was identified the moment we eventually saw pictures of the setup ... not sure why it always has to be repeated that pictures can be helpful ... here I guess I would have liked to see the UAVCAN GUI output
Feb 18, 2018, 09:36 PM
Wisconsin
I agree to turn off the second bus. In addition a lot of the changes in CANbus settings require reboots. FYI.

What is the status of your ESC node(s)? It is it "operational"?
Feb 19, 2018, 07:37 PM
Registered User
americanpelican's Avatar
Olli,

Appreciate the prompt reply! I did follow your suggestion and brought all pre-arm checks on except GPS (workbench is in the basement with poor signal), and FC using 3.6 Master started generating esc.RawCommand messages.

Conclusion is that besides arming FC, it is necessary that certain pre-checks take place. I will play with pre-check switches and figure the one(s) prevents esc.RawCommand to be generated.

One more thing, I have set my FC CAN node to 1, as lowest possible binary address, to give the FC-generated CAN messages priority over any other devices on CAN bus. Do you think it makes sense?

Anyway, thanks for the help!

Luis
Feb 19, 2018, 07:44 PM
OlliW
FANTASTIC that you got it working

looking forward to see your first flight test

in UAVCAN, the node id has some effect on the priority, but only a very very weak one. So: No, choosing 1 instead of 10 doesn't really make sense. Especially since the other nodes would usually be at higher ids anyhow.
see e.g. http://uavcan.org/Specification/4._C...n-frame-format
Feb 19, 2018, 07:54 PM
Registered User
americanpelican's Avatar

ESC carrier board for BLHeli 32?


Olli,

Is the carrier board design limited to KISS ESCs? I got a set of new BLHeli 32 and they are 27mm x 14mm and they do have Telemetry pad in the same position as KISS does.

Unfortunately, not all BLHeli 32 have the same size and the ones with onboard LEDs are a bit larger.

Wishful thinking... it would certainly make the wiring much cleaner.

Luis
Feb 19, 2018, 08:16 PM
OlliW
I have designed them for the KISS. It may happen that there are BLHeli32 ESCs which also fit, but this would be then by pure luck.
As you mentioned yourself, there are just too many form factors for the BLHeli ESCs. Doing a carrier for them would only make sense, IMHO, if we, the community could agree on one particular kind, but I can't imagine this to happen.
The KISS have 27x19mm or so, so size-wise the ESC you show maybe (only maybe !!!) could fit, but the pins are neither in the exact same location, nor have the same sequence, i.e, one would have to do some wiring anyhow. One would have to try and see what one gets.
I think for the BLHeli32 ESCs the mini pcb makes more sense as a general item
Feb 21, 2018, 05:33 AM
OlliW
I've looked again a bit at the latest BLHeli_32 ESCs

the Wraith32 35A V2 from Airbot looks quite suitable to me for installing in a carrier

it should in fact be easy to adapt the KISS carrier to this board, this ESC is also only 17$, and it also has LEDs which might allow nice signaling, so, this might be a very viable option (since I just recently invested quite a bit into ESCs, I though don't want to buy some)
maybe we should suggest this to Airbot ???
Last edited by OlliW; Feb 21, 2018 at 05:51 AM.
Feb 21, 2018, 05:49 AM
OlliW
for your information:

I just released the Eagle layout files for the UC4H ESC KISS carrier board, and the UC4H ESC Mini board.

Let's hope that this will help getting them available. I will try to get a quote from the ShenZhen2U pcb assembler, and if the prices are reasonable we could organize a group buy, but I have a feeling that the prices won't be reasonable.
Feb 23, 2018, 12:13 AM
OlliW

Group order assempled UC4H Kiss Carrier Boards


Hey Folks

I got the quotes from ShenZhen2U for the production and assembly of UC4H ESC KISS carrier boards - and it doesn't look that bad .

I asked for quotes for 12, 20, 32, and 50 boards. Below you find a chart which shows the cost per board, resolved into cost for the electronic parts, PCB boards, assembly, and the final total including shipping to me. I find the costs quite reasonable. For instance, the materials cost is ca 7$ per board (the STM32F103 e.g. is just 2.80$, which is good), which is significantly less than what we would pay at digikey or mouser. Assembly drops to just 3.xx$ for 50 boards. On top of that would be, of course, the shipping cost from me to you, of few Eur.

So, I think that if we can get together about 30 or more boards, the cost will be less than ca 13$ per board. This is still a whopping 52$ for 4 boards ... but hey, we can't expect Asian mass production prices for such a limited quantity.

As regards the involved risks:
I have never done such a PCB assembly myself, nor know anyone having done that with ShenZhen2U. So I can't know nor predict with certainty what we would get. However, this is a quite simple board. Only few parts, none of which critical. The layout is simple. And so on. As compared to the boards I build myself, see above report, I have changed the layout just marginally, i.e., the layout is proven. So, I would think that this is a perfect project for doing a PCB assembly for the first time, since there isn't much to go wrong. But, as said, I can't know. And of course, if we get crap, for whatever reasons, their fault, my fault, shipping fault, etc, we all would lose our money. So, there is some risk, but I think it hardly can be lower for such an adventure.

Given that, I herewith start the attempt of such a group buy.

The procedure will be:
The offer includes fully assembled UC4H ESC Kiss carrier boards, but nothing else, i.e., no wires, no plugs, no KISS ESCs, and so on. The procedure will be this:
i) I will let this run for ca two weeks to collect the participants
ii) I will place the Eagle files and the BOM into the github repository, so that anyone who wants can look at them and double- and triple-check. This is to wed out all mistakes I might have done.
iii) After closing (i) I will ask each participant to confirm that she/he understands the nature, and risks, of this group buy, and will collect 67% of the money, via paypal, from the participants. That is, in case of a loss, each participant will lose their 67%, and I will lose 33% of the total. This will run for few days.
iv) Once (iii) is completed, I will place the order at ShenZhen2U. I will order few spare boards in addition, so that we can cope with few failure boards. The costs for them will be added to each participant.
v) Once the boards have arrived me, which could take few weeks, I will flash and test them.
vi) I will then collect the remaining 33% from you, and ship the boards out to you. Completing steps (v) and (vi) may take a week, at most two.

EDIT: github repository with the materials for double checking: https://github.com/olliw42/uavcan4ho...ouporder-stuff

So, let's start a list of participants. If you want to join in, please copy the text in the next post and add in your order!.

Let's see if we get 30 or more boards together . Cheers,
Olli
Last edited by OlliW; Feb 24, 2018 at 07:02 AM.
Feb 23, 2018, 12:14 AM
OlliW
UC4H ESC Kiss board v1.00 group buy, participants
------------------------------------------------
4 x OlliW
add yourself here
Last edited by OlliW; Feb 23, 2018 at 12:57 AM.
Feb 23, 2018, 09:55 PM
Wisconsin
Quote:
Originally Posted by OlliW
UC4H ESC Kiss board v1.00 group buy, participants
------------------------------------------------
4 x OlliW
add yourself here
4 x Mike_Kelly
Feb 24, 2018, 01:35 AM
OlliW
UC4H ESC Kiss board v1.00 group buy, participants
------------------------------------------------
4 x OlliW
4 x Mike_Kelly
add yourself here
Feb 24, 2018, 04:12 AM
Registered User

Group buy


4 x for James Pattison
Feb 24, 2018, 03:24 PM
OlliW
UC4H ESC Kiss board v1.00 group buy, participants
------------------------------------------------
4 x OlliW
4 x Mike_Kelly
4 x for James Pattison (auturgy)
add yourself here
Feb 25, 2018, 04:46 PM
OlliW
I have to share this, just because I liked so much seeing this on my desktop

I've released some new firmwares, among them a new version of BetaCopter, 3.6-dev v006, and a new version of the STorM32 firmware, v2.36e. All installed on my UC4H-ized flamewheel, gave the attached screen shot.

I'm obviously satisfied with that development, STorM32-UC4H-BetaCopter have become kind of a coherent ecosystem now. Also the set of new features and possibilities coming with it I find just great. In some sense, the only significant piece missing now would be smart shots, right
Feb 25, 2018, 08:04 PM
Wisconsin
Quote:
Originally Posted by OlliW
I have to share this, just because I liked so much seeing this on my desktop

I've released some new firmwares, among them a new version of BetaCopter, 3.6-dev v006, and a new version of the STorM32 firmware, v2.36e. All installed on my UC4H-ized flamewheel, gave the attached screen shot.

I'm obviously satisfied with that development, STorM32-UC4H-BetaCopter have become kind of a coherent ecosystem now. Also the set of new features and possibilities coming with it I find just great. In some sense, the only significant piece missing now would be smart shots, right
You definitely have earned the ALL-UC4H badge! Very cool.
Feb 26, 2018, 01:30 AM
OlliW
I can't win my badges - only users can
I mean, how simple would that be to win one's own Olympic Gold medals LOL
btw, you're still a bit away, but clearly in the lead
Feb 26, 2018, 06:05 AM
Registered User
Quote:
Originally Posted by OlliW
I can't win my badges - only users can
I mean, how simple would that be to win one's own Olympic Gold medals LOL
btw, you're still a bit away, but clearly in the lead
Excellent work Olliw! Thank you. I definitely want to use it in my extremely slow solo rework project. It's very cool that there is now one convenient way to connect everything. What's next? An obstacle detection module with something like this https://www.sparkfun.com/datasheets/...y0a02yk_e.pdf? Or something very unusual?
Feb 26, 2018, 06:20 AM
OlliW
probably none of both
I think everything beyond the basic I should (and want to) leave to the professionals.
the current nodes all have potential for improvement, I might add a buzzer and OLed to the Notify, LEDs to the gps-magnetometer, the esc node could support motor reversal, the power brick could offer a scheme to power also the CAN bus, etc. pp.
if a plane user is found I definitely would want to do an actuator/servo with current telemetry, maybe a barometer, or airspeed
but from my current perspective I wouldn't want to do anything, beyond the gps, which involves a difficult driver. The major thing missing is now actually a bootloader, with that many nodes on a system upgrading the firmware has become a bit of a pain LOL.
but you know, any plan is only good for as long as the next plan comes up LOL
suggestions are welcome of course too, like you did, even if I may not bite
Feb 26, 2018, 09:18 AM
Registered User
americanpelican's Avatar
UC4H ESC Kiss board v1.00 group buy, participants
------------------------------------------------
4 x OlliW
4 x Mike_Kelly
4 x for James Pattison (auturgy)
4 x Americanpelican (Luis)
add yourself here
Feb 26, 2018, 09:36 AM
Registered User
Quote:
Originally Posted by OlliW
but you know, any plan is only good for as long as the next plan comes up LOL
:-) You never know...
Ok, thanks Olliw. By the way, I remember you wrote about four escs on one node and it was kind of unstable. I can not find your post. I plan to use 4in1 esc with your node. It works? Or should I use single regulators for each motor?
Last edited by P43YM; Feb 26, 2018 at 09:53 AM.
Feb 26, 2018, 10:07 AM
OlliW
it worked fine for me, I flew it for some weeks
I just "tuned" it down to use "only" DSHOT300, instead of DSHOT600, since with DSHOT600 the timing I considered to imprecise
also, I would be careful with going to more than 4 outputs, it might work well, but I have not tested it, and as a rule, the more outputs the more imprecise the DSHOT signal
one probably could optimize the timing, with some effort, but since it appeared to work for me, I didn't
one node instead of 4 has an advantage in that the current draw on CAN 5V is lower ... with my 8 nodes, and one of them driving 3 WS LEDs and another one a gps, I actually had to "boost" the CAN 5V
not sure how your 4in1 would look like, if it has a BEC itself, one could consider powering the node from this, has power advantages, but the node would be live only when powered by battery

this might be the post you were searching: https://www.rcgroups.com/forums/show...&postcount=509

btw,
it seems that the support for the UAVCAN esc.Status message may go into ArduPilot master, at least there is such a PR ... thx to auturgy/James !! ... that is, the UC4H ESC nodes may be fully used without BetaCopter in some near future
Feb 26, 2018, 10:22 AM
Registered User
Quote:
Originally Posted by OlliW
btw,
it seems that the support for the UAVCAN esc.Status message may go into ArduPilot master, at least there is such a PR ... thx to auturgy/James !! ... that is, the UC4H ESC nodes may be fully used without BetaCopter in some near future
Great news!

As for the escs,I'm definitely going to try it. I am primarily concerned not with the current draw but with the size of everything. I have one crazy idea about remaking solo into a drone the size of a Mavic and I'm implementing it. I think this is a more fun and interesting task than assembling a huge frame. And your work with the canbus is a very beautiful and convenient solution for a problem with 3.3 volts in solo, and even with additional features.
Feb 26, 2018, 10:41 AM
OlliW
well, if you want to go for size, and solo with stock cube, the 4in1 uavcan thing might not the best approach
the uavcan converter adds to size, no way around it
also, please note, in the UC4H solo motorpod mod I placed the UC4H node close to the motor pod, and powered it from the motorpod's 5V. If the node is far away from the motor pod, it might work too, but I wouldn't assume that from scratch! This might change the situation considerably, and it should be carefully re-evaluated, and it might turn out to not work.
on the other hand, if you're going to throw out the motor pods and use your "own" ESCs anyhow, which your 4in1 comment suggests, and have small distances, when you just have to pay better attention to potentials and you shouldn't need the UC4H nodes at all to overcome the stock solo's issue. I mean, nearly everyone is flying on 3.3V without issues, right. It's not rocket science to do it.
Feb 27, 2018, 06:16 PM
OlliW
hey folks,

I'd like to share with you my today's evening entertainment .

Exploring the new possibilities offered by the UC4H ESC KISS node, i.e., the advantage of allowing us to measure the rpm and current (and some other quantities), I thought that this should allow us to also estimate the thrust characteristics of the ESC-motor-prop combination. This should be highly useful, since the thrust characteristics is among the most important properties determining the behavior of a copter, but is not trivially measured. Usually relatively sophisticated thrust measurement stands are needed.

In fact, since that characteristic can be so important, ArduPilot has some parameters specifically for adjusting the motor output to the specific thrust characteristic. It's called MOT_THST_EXPO. You can find some very insightful explanations and comments by Leonard here: https://discuss.ardupilot.org/t/usin...e-expect/26172.

Problem is, as mentioned: Normal mortals like us have little chances to measure it. So we stick with the default value. This may change fundamentally, thanks to UAVCAN, and the substantially lowered entry price.

So, with a crucial hint by Pavel (thx, sir!!), I wrote a python script to ramp up and down the motor+prop, measure the rpm and current, estimate the thrust using some basic blade momentum and motor theory results, and fit the result to a curve specified by ArduPilot, to yield an estimate for MOT_THST_EXPO. I have to admit that the numerical value doesn't yet make sense to me, so, this needs to be further investigated. BUT: That's the key message, it's now all doable, and easily doable, for anyone of us

EDIT: The MOT_THST_EXPO value issue has been sorted out, I had a bug in my script . The correct value for my setup would be 0.80. See the new pic. The new script, with some other usability improvements can be found again in the git repo.

Some pics of the "measurement" setup, and of some results are attached below.

Have fun, cheers,
Olli
Last edited by OlliW; Feb 28, 2018 at 03:42 PM.
Mar 01, 2018, 09:31 AM
Wisconsin
UC4H ESC Kiss board v1.00 group buy, participants
------------------------------------------------
4 x OlliW
12 x Mike_Kelly
4 x for James Pattison (auturgy)
4 x Americanpelican (Luis)
add yourself here
Mar 01, 2018, 10:52 AM
OlliW


great

so, we are getting closer, we are short of only two more participants to reach the critical level for ordering ...
Mar 02, 2018, 05:03 AM
OlliW
REMINDER REMINDER REMINDER

I'd like to draw your attention to the attempt of a group buy of UC4H ESC KISS Carrier Boards, which can be used to build UAVCAN ESCs by installing KISS 32A or 24A ESCs in the carrier. The details of that effort are described in post #526

Currently 24 carrier boards have been "ordered", but I've (artificially) set a limit of a minimum 30 (or 32) to complete this effort. So, we are short of two participant.

The call is however still running for about a week, so, still time to sign in!

I will close this call by Thursday, 8. March., 24:00UTC.

(this is to get things synchronized with my day work )

Have fun, Olli
Mar 03, 2018, 04:46 PM
OlliW
with this many UAVCAN nodes, setting up all the things has become quite an effort. So, in an attempt to make that a bit easier, I wrote a Python script for configuring the esc indices (available in the UC4H git repo), and also quickly did a video, see below.

Clearly, the script is not perfect, and is only a first step, and is not what one expects UX wise in 2018. We all hope that eventually such things will be taken over by some ground control station, in some future. But for as long as this hasn't become reality, such scripts might be helpful. They might not only be useful, but I actually consider them a sort of "research vehicle" to explore how user-friendly setup procedures could look like.

For instance, standard UAVCAN suggests a method of enumerating the ESCs in a semi-automated way, which however requires that you put your fingers into the rotors while the ESCs are connected to the lipo (http://uavcan.org/Implementations/Py...C_enumeration/, see also the video in this link). Initially I found this method quite cool, but Felix has pointed out the risks to me and that this is certainly not what one should recommend as a default procedure in the hobby world. And I fully agree. No fingering with props attached to hot ESCs. Thus, in the script, I tried to explore a different approach.

Interactive Setup of UC4H ESCs (9 min 0 sec)
Mar 05, 2018, 01:04 PM
Registered User
Hi Olli! I've created 3d model from your github source for UC4H ESC. It is in the attachment.
Maybe you you'll find it helpful if you want to create housing or something else...
Anyway, I'd like to offer my help with UC4H ESC.
Best regards, Alex
Mar 06, 2018, 02:41 AM
OlliW
Quote:
Originally Posted by j3qq4hch
I've created 3d model from your github source for UC4H ESC.
GREAT. Thx a lot. I couldn't yet watch it, need to installer a proper viewer

Quote:
Originally Posted by j3qq4hch
Anyway, I'd like to offer my help with UC4H ESC.
you're highly welcome. Which kind of help you mean?
Mar 07, 2018, 03:45 AM
Registered User
Your UC4H KISS adapter needs to be modified a little bit for better compliance with UAVCAN hardware specification.
Also if you don't mind we could try change its construction and reduce its dimensions drastically.
By the way, there is browser version of sketchup. You can use it to check the model.
Here it is https://app.sketchup.com/app?hl=en
Mar 07, 2018, 03:47 AM
Registered User
Also I saw your fancy photos of these boards. Looks fantastic, but mounting boards with zip-ties(although it is pretty strong) may not be the best option. Couple mounting holes in the PCB may help with that


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