New arduino Q

[Francis Roussel]

Hi,

https://www.qualcomm.com/news/releases/2025/10/qualcomm-to-acquire-arduino-accelerating-developers--access-to-i
https://docs.arduino.cc/hardware/uno-q/
For our sailboats, there's the microcontroller solution: ESP32, STM32 for example, but also the single-board microprocessor Raspberry Pi under Linux.
Qualcomm has just acquired Arduino and is releasing a mix of the two: the Uno Q.
The promises are AI with visual recognition, but also an integrated microcontroller for our sensors and actuators, all on the same board and plenty of connectors for tailored solutions.
Towards a translation of the environment https://ardupilot.org/?

fr

[Oliver epsom]

Hi Francis.

I agree with you, there are a huge range of low cost processors available.

However for small boats (say 2m in length) power consumption is likely the limiting factor. Solar cells are not very efficient and every water counts. The computer power needed to navigate is tiny - pocket calculator stuff - so I would argue that the huge capabilities of these embedded systems is largely wasted on microtransat.  Even analysis of a route based on updated weather only needs to be done once a day or so because the boats move so slowly. So personally I'd recommend the lowest power machine available and use several for redundancy.

--
You received this message because you are subscribed to the Google Groups "Microtransat" group.
To unsubscribe from this group and stop receiving emails from it, send an email to microtransat...@googlegroups.com.
To view this discussion, visit https://groups.google.com/d/msgid/microtransat/5b626583-882b-496b-ad4d-f1f7bc741d02n%40googlegroups.com.

[Dermot Tynan]

That’s the architecture I’ve used in the past. An AVR chip like the ATmega328p to steer and sail trim. Also to monitor solar power and battery SoC. It can then power up a single-board computer a few times a day, which has a GPS and an FPU to do the route computations as well as reporting back to base via an Iridium modem. It mimics the way racing sailboats divide the tasks.

But STM32F4xx chips have floating point units and low power consumption so nowadays it’s possible to combine both. The Qualcomm purchase is interesting. Time will tell. They’ve certainly been fond of their chip monopolies in the 3/4/5G space. Who knows?

        - Der

—

Dermot Tynan

Kalopa Robotics

https://kalopa.com

On 12 Oct 2025, at 09:14, Oliver epsom <olive...@gmail.com> wrote:



To view this discussion, visit https://groups.google.com/d/msgid/microtransat/CAJ0SNALyJhrywvZzj8wMHq50m8aAJQ9_azFaCH%3D5QP_N6STCbw%40mail.gmail.com.

[Philip Smith]

I use the STM32L412 on my boat: it is one of the low power range from ST Microelectronics. I actually use the NUCLEO development board as it has very good debugging capability built in, power supplies an LED and a push button. As Dermot says, it has floating point and more capability than you could want. (Although perhaps not AI image recognition.) I run it at only 2MHz to keep the current down and it is plenty fast enough (you can disable debugging and USB once the program is OK to get low power). Current is about 1mA. BUT it is MUCH MUCH more complicated to program than an Arduino (which I used on my first boat). For example, PWM output on the Arduino is very easy to use the ready-made functions. On STM I had to investigate timers down to a very basic level, read the manual a lot, and keep referring to the internet for help. So unless you really know about software, or want to, be careful of jumping in to STM. However, on the plus side, the debugging capability of the STM is vastly superior to the Arduino I used: real time monitoring of variables, pause the program to investigate variables, step through a line at a time etc.  I haven't looked into the (I think it is called) Blue Pill, which is an STM processor pretending to be an Arduino, so you hopefully get the advantage of easy programming and low power. Worth exploring I suggest. 

So because the micro is such low current I just leave it running all the time, even though it is mostly just doing rudder control. There are various sleep modes available but I didn't think they are necessary. I also keep the compass on all the time as it is also very low current, and required to keep to heading. Everything else is only one for a short time, eg GPS every hour for a fix once out at sea. The main issue with current is how to reduce the rudder servo power whilst still going in the right direction. I'm not sure I've really got a good solution to this. I think the best answer is to get a boat that naturally keeps going in a straight line, but sailing boats don't seem to be stable like this (at least mine isn't). 

One tip to reduce the current: remove the power indicator LED from the CPU board: this uses about 3mA on its own, so once you have sorted the PSU, remove it.

Phil Smith

[andy robinson]

Re: Rudder servo the answer may be to have a servo that will keep position when power is removed (I seem to remember another post on this forum suggesting the use of linear actuators, and I have been using a servo to control a worm drive motor).    You then don't need the boat to maintain course at all points of sail with no rudder, just to be able to converge on a rudder deflection and remove the power.   At least in theory.   My experience of sailing this year seems to suggest that it would work sometimes, and at other times you need to constantly adjust.

[Damon McMillan]

I think it would depend a great deal on the particular servo.  Some are much harder to back-drive than others.

It also depends on how your rudder is hinged.  A large rudder hinged at the leading edge will produce a lot of force on the servo linkage and therefore easily back-drive the servo if the servo is turned off (and draw a lot of current when the servo is turned on!).  A rudder that is hinged at the center of pressure will (in theory) not produce any force on the servo linkage.  In theory.

Damon

Damon McMillan

Blue Trail Engineering LLC

1567 Skyway Dr. Unit B

Longmont, CO 80504

720-526-6388 (business)

801-520-4953 (cell)

www.bluetrailengineering.com

To view this discussion, visit https://groups.google.com/d/msgid/microtransat/272b8a8f-b765-47f7-91e4-9ebb81e65c83n%40googlegroups.com.

[Francis ROUSSEL]

Hi,

Le 17/10/2025 à 18:48, Damon McMillan a écrit :
> I also keep the compass on all the time as it is also very low current,

I don't plan on using the Arduino Q; I just wanted to let the group know
about this new feature in case you missed it.

However, I'm interested in the name of the chip used for the compass.
Is it capable of functioning even with a more or less severe heel, and
can it restart without necessarily being on a horizontal plane?

In the meantime, since it's important to conduct tests and prove to a
potential sponsor that they can contribute to the purchase of a
satellite transmitter, I'm testing the IoT (Internet of Things) with a
SIM7080G card, showing that the boat is capable of autonomous navigation
(and recovery in case of a problem). All you have to do is stay within
range of a phone operator; the card accepts all operators.

fr

[Philip Smith]

Notes on Compass.

The compass I'm using is the STM LMS303D type. Mine is now out of date, but if you search for LSM303 you get lots of options. They are tiny, so best to get one already mounted on a breakout board. But to keep to a low current, get one with as little extra components as possible. It is I2C communication, using 2 addresses, one for the magnetometer and one for accelerometer. You need both set of data (3 axes on each) to calculate the heading. There are various software examples to help you do this. It can give heading whatever the heel of you boat, and can restart from any orientation. The reason lots of compasses need to be set horizontal is to re-calibrate the magnetometer, which can drift, but also different parts of the earth have different magnetic field strength and direction (slope into the ground). It's not easy to keep your boat flat whilst recalibrating, so I don't bother, and hope its accuracy is OK. One thing you do need to do though is to find the calibration of the 3 magnetometer axes and put these in your program (hard coded so available after power-up) so that you can allow for the specific chip you have (zero offset and gain for each axis). The accelerometer seems to be OK without individual calibration.

Rudder servo:

Having a rudder that doesn't back-drive is probably useful, but even if the rudder is fixed into a position, my boat still doesn't keep to the desired heading. It basically isn't stable (in a control sense) so keeps needing adjustment. I tried adding a second mast and sail to provide stability, but the boat just went into wind and then I had to wade into the pond to get it out of the reeds! So I gave up that idea. I've been trying to work out a low-current strategy for steering, but it seems like the lower the current the worse the heading. So I'm just having to live with the fact that the battery will probably go flat overnight and make no useful progress for several hours until sunrise. 

SIM card

Mobile phone coverage isn't much good out at sea. When retrieving my boat from just off the Isle of Wight, phone signal was pretty much non-existent, and that was with a proper phone about 3m above sea level. So a sim in a small boat is going to struggle, even with an aerial on top of the mast. For useful information you need a satellite transmitting module (eg Iridium), or satellite internet (if you can fit the dish on your boat).

Phil

[Philip Smith]

Update to Compass information

I've just looked up 9 degree of freedom (9DoF) sensors and have found that they are now very low current. Previously all the gyros I'd seen were several mA so I've made do with just 6DoF (magnetometer and accelerometer). I've just ordered an Adafruit ISM330DHCX + LIS3MDL which according to the data sheets should run at under 2mA for all sensors (magnetometer, accelerometer and gyro). When it arrives I'll check this and post results. 

To explain why I'm looking into this. The problem with a 6DoF sensor eg the LSM303D is that it gives inaccurate readings if it is being accelerated, as the signals and maths involved think the sensor is a different orientation (relative to the earth) than it actually is. So in effect it doesn't know where "down" is.This means the heading result from the maths is incorrect. On a small boat there is a significant amount of random acceleration due to waves which therefore causes an error. (If the sensor is not accelerating, it can correctly calculate heading regardless of its orientation). To reduce this error I filter the heading readings over time, but this delays any real change of direction for a while. So the heading sent into the steering software is always a bit behind the times. If the boat went is a straight line this wouldn't be too much of a problem. However, when trying to navigate a small loop course in a pond with waves, the filtering delay makes a noticeable difference when tacking or gybing. Perhaps not a real problem when at sea, but it should be interesting to see if I can get better instantaneous readings from the 9DoF sensor. In theory, with the extra information from the gyro it should be able to cope with acceleration caused by the waves.

Phil Smith

[James]

Phillip:

I have been very successful using a BNO085 9-DOF sensor from Adafruit. It may draw more electrons than you wish… but the libraries that Adafruit have put together do all of the “maths” you refer to below.

James Hull

Norfolk, Virginia 

Sent from my Glade air freshener.

On Oct 21, 2025, at 9:10 AM, Philip Smith <philip...@gmail.com> wrote:

Update to Compass information

--

You received this message because you are subscribed to the Google Groups "Microtransat" group.
To unsubscribe from this group and stop receiving emails from it, send an email to microtransat...@googlegroups.com.

To view this discussion, visit https://groups.google.com/d/msgid/microtransat/bea8240f-dd79-4e15-af53-8ced1f9c0059n%40googlegroups.com.

[Francis Roussel]

Hello Phil,

Thanks for all these details.
The use of an IoT connection is only for testing, initially on a lake and then along the coast, so the range is theoretically sufficient.

The issue of compass heading stability is important because the variation can be significant when the boat heels.

Regarding course stability, I think we need to look at the wind vanes. We trim the sails according to the course, we cut the signal to the rudder servo, and it's the wind vane that will adapt the boat's course according to the wind direction. The wind at sea is steady, so for several minutes it will stay in the same direction. Mechanical self-steering gears were widely used in the early days of ocean racing before computers became widespread. https://seatizens.org/wp-content/uploads/2023/02/Menu-1.jpeg
This model is very efficient; it just needs to be adjusted.

Older modelers who practiced "free sailing" before radio control used "vane gear."
https://www.sailboat-cruising.com/self-steering-gear.html

Attached is an achievement found I don't know where but which shows how to remove the gears with a simple U-shaped piono rope which operates the small additional rudder.

[Philip Smith]

Regarding wind vanes: I did consider this, and even 3D printed some gears to try and work out a suitable design. I was trying to use the sail as the vane, and gear the rudder to that with a differential gearbox, the third input being the servo. I didn't ever do any trials as it all seemed a bit complicated. My main concern with a vane was how to make a vane steering system robust enough to survive the elements. Even real vane auto-steer systems for single handed ocean sailing often seem to be destroyed in a storm. So I thought the risks outweighed the benefit. Plus it would be a lot of testing, and I can't really test on the sea. Pond sailing is a completely different matter, and vanes would make sense there.

One of the microtransat entries from several years ago (Craig Gorton) had a boat where the mast was rotated by a servo and this angled the sail, so the boat would sail at a constant angle relative to the wind. I never found out how this worked exactly but an interesting solution. It suffered from leaks around the shaft seal and back-driving the servo by wind and waves hitting the sail I think. Old posts probably have some useful information.

Regarding the suggestion of the gyro board: thanks for the suggestion. I now have my new 9DoF Adafruit board (see below), and first measurements of the current gave about 0.7mA: very low and good enough to leave on permanently. Its called a featherwing, and has newer chips than the one you suggest which are lower current (I think). Adafruit do an Arduino library for it, and even better for me is that ST do a C library for use with their IDE and microcomputers so I don't have to translate from Arduino into STmicro. I've "just" got to program it up, learn how to "fuse" the accelerometer and gyro, do some trigonometry using quaternions and make sure it actually gives the right answer. Should be done by christmas!

Phil Smith

[Patrick Coole]

I have been working on a smallscale example of this system which uses a sort of cam setup to control a windvane. I struggle to understand how it would work in downwind scenarios but will soon find out once I get my little vessel in the water for testing. https://www.youtube.com/watch?v=LNTstks9U_s. If it works it would make for an automatic mechanical system. The servo would then be the only moving electronics. 

[Francis Roussel]

Je travaille depuis quelques mois sur le concept de Peter Worsley (en pointillés). Il est excellent car avec un indicateur de gîte (facile avec un MPU9250 ou un autre qui fait aussi compas)  on peut automatiquement programmer une réduction de la "puissance" transmise aux voiles.

La navigation en vent arrière est impossible , il faut naviguer comme contre le vent, la programmation est donc la même. (quand on l'a réussie ce qui n'est pas mon cas je me réserve le travail une fois que j'aurais un compas qui fonctionne comme je veux), je travaille à d'autres projets).

Le principal est de baisser la voile au plus près du pont car le système de Peter est trop encombrant., de même il faut envoyer l'aileron de queue le plus haut possible afin de ne pas être embêté par les vagues..

J'ai réfléchi à une coque et en utilisant FreeShip 3.43 j'ai dessiné MicroFlamem dont les images sont jointes, j'ai commencé à le construire avec la méthode traditionnelle l'hiver dernier  mais poncer quelle fatigue...

Hier j'ai appris que les fichiers obj pouvaient être lus comme les STL moi qui cherchais depuis des années à utiliser le résultat d'une coque dessinée et la convertir pour  pouvoir l'imprimer en 3D.

J'ai tenté aujourd'hui et ma prochaine activité sera d'imprimer un moule lequel sera bien meilleur que mon travail de ponçage moins fatigant et plus rapide. Méthode RamyRC  https://www.youtube.com/watch?v=Fi35p8bJMsA

J'ai dessiné la coque avec dans l'idée : autour de 1.2 m pour les essais , bien ventru pouvoir y mettre plein de choses et garder  de belles formes , des bouchains pour pouvoir finir vite en construction classique mais aussi car à la gîte un bouchain c' est utile enfin comme la partie haute est dévelopable on peut augmenté le franc bord facilement. le pont est plat (cellules solaires et dispositif Worsley).

Si voulez davantage de renseignements ou fichiers : message en privé.

I've been working on Peter Worsley's concept (dotted lines) for a few months. It's excellent because with a heel indicator (easily done with an MPU9250 or another that also acts as a compass), you can automatically program a reduction in the "power" transmitted to the sails.

Sail downwind is impossible; you have to sail as if against the wind, so the programming is the same. (Once you've mastered it, which isn't my case, I reserve the work for once I have a compass that works the way I want it to). I'm working on other projects.

The main thing is to lower the sail as close to the deck as possible because Peter's system is too bulky. Likewise, the tail fin must be raised as high as possible so as not to be bothered by the waves.

I've been thinking about a hull and using FreeShip 3.43, I designed MicroFlamem, the images of which are attached. I started building it using the traditional method last winter, but sanding was a real pain...
Yesterday, I learned that OBJ files could be read like STL files, and I'd been trying for years to use the result of a designed hull and convert it for 3D printing.

I tried today, and my next project will be to print a mold, which will be much better than my sanding work: less tiring and faster. RamyRC Method https://www.youtube.com/watch?v=Fi35p8bJMsA

I designed the hull with the idea: around 1.2 m for the tests, well rounded to be able to put lots of things in it and keep nice shapes, chines to be able to finish quickly in classic construction but also because when heeling a chine is useful finally as the upper part is developable we can increase the freeboard easily. the deck is flat (solar cells and Worsley device).
If you would like more information or files, please send a private message.

fr

[Francis Roussel]

J'ai oublié : Freeship est mauvais pour les modèles réduits donc remplacer maintenant m par dm et tonnes par kg.

I forgot: Freeship is bad for scale models so now replace m with dm and tonnes with kg.

fr

[Pierce Nichols]

There's at least a couple of STM32 based Arduino-compatible boards, so you should be able to use many Arduino libraries directly with your board, including the sensor fusion library that Adafruit has to use with their 9-DoF boards. Easier than coding a Madgewick, Mahoney, or Kalman filter from scratch.  

I think the only reasonable way to build a base-driven sail for a Microtransat competitor is to use a properly waterproof servo and a brake mechanism. A bicycle disk brake with an over-center mechanism to activate it seems like it would be a good mostly off the shelf choice for the braking mechanism. I don't like systems that use the gear train as the locking element, because that invites breakage; a properly designed and adjusted brake will slip before mechanical damage occurs as a result of a wave impact or similar event. The control system can then reset it on the next frame. 

Similarly, I think relying on the actuator gear train to lock your rudder between power-on periods is asking for breakage. The braking element should be separate and adjusted to slip before mechanical damage occurs in an overload situation. 

-p

To view this discussion, visit https://groups.google.com/d/msgid/microtransat/50b30459-43c1-41a4-a023-4ca40657afe0n%40googlegroups.com.

[Pierce Nichols]

The BNO085 has the sensor fusion built in -- it will output quaternions (among other things) directly. The datasheet has all the goodies: https://www.ceva-ip.com/wp-content/uploads/BNO080_085-Datasheet.pdf

-p

To view this discussion, visit https://groups.google.com/d/msgid/microtransat/90208C86-EC40-45BC-A596-A7904D09645A%40gmail.com.

[Francis Roussel]

https://fr.aliexpress.com/item/1005005691927793.html

A l'intérieur un ressort, l'écrou permet de comprimer plus ou moins le ressort.

Entre le deux pièces noires un vé .

En cas de choc les deux vés s'éloingnent en comprimant le ressort, puis le ressort permet de tout remettre en place.

Inside, there is a spring, and the nut allows you to compress the spring to a greater or lesser extent.
There is a V-shaped piece between the two black parts.

In case of impact, the two V-shaped pieces move apart, compressing the spring, and then the spring returns everything to its original position.

fr