Select to expand quote

LISAWindsurfing said..
Hi all,

I only provived Brendan with a water proof case for Luderitz ;) . This is indeed a ESP GPS here, so not a LISA dev (not yet )

Now I understand the rush to get the parts........Luderitz.

Select to expand quote

Anybody tried using magnetic pogo-pins to charge the ESP? I'm just experimenting. Although I could connect all 4 pins to USB to make it connect fully to USB, the connections are too tiny to connect on the PCB. For now I only use 2 for 5V and GND. The magnets on the side are reversed (N/S) such it connects only 1 way. Curious to see how these pins hold against corosion. I'm not sailing in salty water so that helps. Reason to try this is to prevent heat during wireless charging when the airbag gets too big.
I also added a watertight button instead of the reed contact for simplicity and tactile feedback.

Freezer, the contacts corrode pretty badly. I used them to charge the battery for my light fin. To be fair, I treated them with complete neglect...probably should have rinsed it after each session. I know Slowboat uses the connector to charge his arm unit and swears its great and lasted a year so far. So if rinse and dry after each session I'm sure they'll be fine

The last version before closing with the acrylic tape and PVC plate.
The white tube contains the NC (normally closed) reed contact to switch off the voltage to the ESP32.
The battery can still be charged.
Carlos Sainz almost ready.

I was about to pot my first unit, but Im considering to get rid of the wireless coil charger as It may cause heat problems or fail, and go try the pogo connectors, its a great idea, some more experience about It? The install on the fin is quite radical ;) but as many sports watches use this solution It sounds good to me. Probably those are better quality than we can afford at Aliespress, but Ill give It a go. Also found a model with the magnets under the plastic, so the only worry will be the very pins which supposedly are covered with gold and should resist decently.
Paco how do you deal with the push of the button while keeping all watertight?

Select to expand quote

veton said..
I was about to pot my first unit, but Im considering to get rid of the wireless coil charger as It may cause heat problems or fail, and go try the pogo connectors, its a great idea, some more experience about It? The install on the fin is quite radical ;) but as many sports watches use this solution It sounds good to me. Probably those are better quality than we can afford at Aliespress, but Ill give It a go. Also found a model with the magnets under the plastic, so the only worry will be the very pins which supposedly are covered with gold and should resist decently.
Paco how do you deal with the push of the button while keeping all watertight?

Veton, pouring would make it waterproof but that has your mentioned limitations when things go wrong.

The technique I use differs and I do not talk about waterproof but making the unit waterresistant for a certain amount of time.
The PVC plate is 0.3 mm thick and is flexible enough to press the button.
The coverplate protects the PVC plate and prevents the PVC plate and the acrylic double sided tape to come off.
The coverplate is fixed to the main case with two 2/3mm hollow carbon tube which also enhances the strength that is put on the unit by the flexible velcro straps.

The unit is now ready to be shipped this week.


With PVC plate and the double sided acrylic tape pressed for a few hours.


The top cover locked by the 2/3 mm carbon tube.

Very nice with the button, and most likely more practical than the magnet. Very Cool dev.
Have tested both the one from rp6conrad and from BigBoss in Luderitz.
They both worked Perfectly... so did the Motion.
Very very small difference in between the 3. like 0.1 0.2 max
same with the one from Manfred (reserved to wssrc competition)

p.s: wanted to test that the 2sec could display 3 digits in kmh ahah but I couldnt ... 2025 objective

Select to expand quote

PacoRaapNL said..

LISAWindsurfing said..
Hi all,

I only provived Brendan with a water proof case for Luderitz ;) . This is indeed a ESP GPS here, so not a LISA dev (not yet )

Now I understand the rush to get the parts........Luderitz.

ahah yes everyone did super good and I got them all before my departure. Thank you so much everyone involved.

Select to expand quote

brendanF113 said..
Very nice with the button, and most likely more practical than the magnet. Very Cool dev.
Have tested both the one from rp6conrad and from BigBoss in Luderitz.
They both worked Perfectly... so did the Motion.
Very very small difference in between the 3. like 0.1 0.2 max
same with the one from Manfred (reserved to wssrc competition)

p.s: wanted to test that the 2sec could display 3 digits in kmh ahah but I couldnt ... 2025 objective

PacoRaapNL said..

LISAWindsurfing said..
Hi all,

I only provived Brendan with a water proof case for Luderitz ;) . This is indeed a ESP GPS here, so not a LISA dev (not yet )

Now I understand the rush to get the parts........Luderitz.

ahah yes everyone did super good and I got them all before my departure. Thank you so much everyone involved.

Should not bee too difficult displaying 2sec with 3 digits, or do you mean decimals (behind the comma). 2 seconds contains 40 data points at 20Hz. At 50 knots 0.001kn might sound great when you beat AA but I fear that the accuracy is not that high on the ESP-GPS with M10 (although they are good). Just compare 2 identical units on the same arm, they differ 0.05kn pretty easy on a 10sec run. Looking at the 3rd decimals would be noise. 0.001kn/s = 0.72cm difference in 2sec after just crossing half a soccerfield.
Same reason people should forget max speed. Typically there is a sawtooth on the datapoints, some noticable from the chop, but when testing in a car at steady speed it is still there.

Most M10 units can log 10Hz, but I usually set them to 5Hz to save memory without loosing accuracy (or tiny). Not sure where I read that longer integration times on GNSS antennas would increase accuracy. In that case there must be a sweetspot between highest polling speeds and shorter integration time and lower logging speed with higher integration times.

Not sure what the framerate of the video timing is at Luderitz. Are they using high framerate cameras? At 50Hz at 100km/h ~ 27.7/50 => 0.56m in 0.02sec. But interpolation of 2 frames at the finishline might improve accuracy. Would be interesting to hear how they do this. Perhaps Michael already explained it somewhere and I missed it.

Select to expand quote

Freezer said..

brendanF113 said..
Very nice with the button, and most likely more practical than the magnet. Very Cool dev.
Have tested both the one from rp6conrad and from BigBoss in Luderitz.
They both worked Perfectly... so did the Motion.
Very very small difference in between the 3. like 0.1 0.2 max
same with the one from Manfred (reserved to wssrc competition)

p.s: wanted to test that the 2sec could display 3 digits in kmh ahah but I couldnt ... 2025 objective

PacoRaapNL said..

LISAWindsurfing said..
Hi all,

I only provived Brendan with a water proof case for Luderitz ;) . This is indeed a ESP GPS here, so not a LISA dev (not yet )

Now I understand the rush to get the parts........Luderitz.

ahah yes everyone did super good and I got them all before my departure. Thank you so much everyone involved.

Should not bee too difficult displaying 2sec with 3 digits, or do you mean decimals (behind the comma). 2 seconds contains 40 data points at 20Hz. At 50 knots 0.001kn might sound great when you beat AA but I fear that the accuracy is not that high on the ESP-GPS with M10 (although they are good). Just compare 2 identical units on the same arm, they differ 0.05kn pretty easy on a 10sec run. Looking at the 3rd decimals would be noise. 0.001kn/s = 0.72cm difference in 2sec after just crossing half a soccerfield.
Same reason people should forget max speed. Typically there is a sawtooth on the datapoints, some noticable from the chop, but when testing in a car at steady speed it is still there.

Most M10 units can log 10Hz, but I usually set them to 5Hz to save memory without loosing accuracy (or tiny). Not sure where I read that longer integration times on GNSS antennas would increase accuracy. In that case there must be a sweetspot between highest polling speeds and shorter integration time and lower logging speed with higher integration times.

Not sure what the framerate of the video timing is at Luderitz. Are they using high framerate cameras? At 50Hz at 100km/h ~ 27.7/50 => 0.56m in 0.02sec. But interpolation of 2 frames at the finishline might improve accuracy. Would be interesting to hear how they do this. Perhaps Michael already explained it somewhere and I missed it.

Now I read it again, you mean 3 digits => 100km/h for 2 sec... sorry for the confusion

Select to expand quote

Freezer said..


Now I read it again, you mean 3 digits => 100km/h dor 2 sec... sorry for the confusion

yes

Select to expand quote

PacoRaapNL said..
The last version before closing with the acrylic tape and PVC plate.
The white tube contains the NC (normally closed) reed contact to switch off the voltage to the ESP32.
The battery can still be charged.
Carlos Sainz almost ready.

Can't wait :) Nice to see it comes pre-configured....

The latest 3D files I use.
The one with ESP GPS is the battery breaker coverplate.It contains the magnet that is switching OFF the Normally Closed reed contact.
The one in the white tube from previous picture.
The magnet is placed in the round small tube that is glued to the cover top right hand side hole.
This way the battery will not be drained by the 2 mA in sleep mode anymore.
Take cover OFF a day before and charge the unit.
If not used for a few days put the cover on when the voltage is 3.9 volt.

Paco

Select to expand quote

Freezer said..

PacoRaapNL said..
Who has built that left one?

That is lisawindsurfing.com as indicated on the picture.

Anybody tried using magnetic pogo-pins to charge the ESP? I'm just experimenting. Although I could connect all 4 pins to USB to make it connect fully to USB, the connections are too tiny to connect on the PCB. For now I only use 2 for 5V and GND. The magnets on the side are reversed (N/S) such it connects only 1 way. Curious to see how these pins hold against corosion. I'm not sailing in salty water so that helps. Reason to try this is to prevent heat during wireless charging when the airbag gets too big.
I also added a watertight button instead of the reed contact for simplicity and tactile feedback.

I have built several with the plug and cable shown but have now swapped over to a round 2 pin plug as I insert a silicon plug to keep it dry.
I now also use waterproof switch instead of reed switch and the switching circuit totally disconnects the battery when off to prevent drain when not in use

Colin

We've been evaluating the FR255 watch for accuracy.
it seems the AG3335M chipset with Multi-Band technology really improves it's accuracy.
What's the chances we could also improve the accuracy of the esp's, instead of using the UBOX chip???

Select to expand quote

decrepit said..
We've been evaluating the FR255 watch for accuracy.
it seems the AG3335M chipset with Multi-Band technology really improves it's accuracy.
What's the chances we could also improve the accuracy of the esp's, instead of using the UBOX chip???

Recently, Ublox released a L1/L5 gnss chip (www.u-blox.com/en/product/max-f10s-module). As soon as these gps-modules become available at a reasonable price, it is worth to try them out. I did not found a AG3335M gpsmodule, nor did I found a datasheet how to integrate these chips. Do they support 10 Hz ? The chips used in watches are optimized for low power, and the setting is 1 Hz (as far as I know). I see 5 Hz / 10 Hz as the main advantage compared to Coros / Garmin.

Select to expand quote

rp6conrad said..

I did not found a AG3335M gpsmodule, nor did I found a datasheet how to integrate these chips. Do they support 10 Hz ? The chips used in watches are optimized for low power, and the setting is 1 Hz (as far as I know). I see 5 Hz / 10 Hz as the main advantage compared to Coros / Garmin.

Modules using the AG3335 variants are available from a number of manufacturers, including SIMCom, Quectel and Locosys. I've got product sheets and reference material, if required.

The Airoha supports fix rates up to 10 Hz and numerous power modes, all of which can be switched on / off. 1-sigma accuracy estimates for position and speed are also produced and available via $PAIR commands (proprietary Airoha). The AG3335 platform also implements RAIM.

You'll find a lot of information, plus links to a couple of useful protocol references on these pages.

logiqx.github.io/gps-details/devices/garmin/review/airoha.html

logiqx.github.io/gps-details/devices/coros/smoothing/update.html

The second of these pages was written after reviewing the technical docs for possible causes of the issue that affected Garmin watches in 2022, and COROS watches since May 2024.

This Max10s with the addition of Multi Band would be very interesting to compare for our windsurfing.

The current M8, M9 and M10 Ublox chipsets give us extreme Doppler speed precision now. I am seeing correlation of under 0.02 knots in controlled side by side testing. It quite rare that we see differences of more than 0.05 knots in real life windsurfing category comparisons between ublox 5Hz and 10Hz devices worn side by side in good conditions.

The most common times we see variation between devices is during Alphas (probably due to sudden orientation change which leads to switching, or loss of some satellites) and when less than ideal situations are encountered, like crashes and submergence with sudden loss of signal.

It is entirely feasible that it in these kinds of situations where multi path technology may help. It will be very interesting to find out.

One big advantage of the Ublox chipsets are that they are readily set up for our needs with Binary UBX messages through the readily available U-center tool. Another is that there are modules and designs readily available, and any new modules utilising these chipsets should be easily integrated into existing designs.

What software do I need to run this on my Android guys?

Select to expand quote

zoloft said..
What software do I need to run this on my Android guys?

Watch or phone?

Select to expand quote

zoloft said..
What software do I need to run this on my Android guys?

Run what exactly?

Remery may be a mind reader, but I'm not.

After many changes and prints this is my first unit. Poliurethane covered electronics, flex resin and the pogo connector, which is worth giving a try. Also added a led. it is very compact, only 101 grams with a 1800 Lipo and clear field for the GPS antenna.
Nobody talks about print specs, this is made out of PCTG, similar to PETG, which I also used in a lot of prototypes. In both cases I didn't find the need of spraying them with laquer, they don't loose a drop, printed full solid in 3mm walls. Time will tell;)

Would it be possible to use a hall effect switch to replace the reed switch ? one of these aliexpress(dot)com/item/1005003574024310.html
Would it be more reliable ?

@veton have you posted the files for the enclosure somewhere ? also what is "flex" resin ?

Select to expand quote

yekokata said..
Would it be possible to use a hall effect switch to replace the reed switch ? one of these aliexpress(dot)com/item/1005003574024310.html
Would it be more reliable ?

@veton have you posted the files for the enclosure somewhere ? also what is "flex" resin ?

@ yekokata, if the hall switch is designed for 5 volt power it should work.
Problem is, these switches drain some miliamps when in use which you do not want.
Reed switches do not consume power.
Also the size is not so small.
Some users are working in the background for another electronic switch solution based on hi side switching with a latch function using a special IC.
You want the battery to be disconnected when not in use to prevent the battery draining.

Select to expand quote

veton said..
After many changes and prints this is my first unit. Poliurethane covered electronics, flex resin and the pogo connector, which is worth giving a try. Also added a led. it is very compact, only 101 grams with a 1800 Lipo and clear field for the GPS antenna.
Nobody talks about print specs, this is made out of PCTG, similar to PETG, which I also used in a lot of prototypes. In both cases I didn't find the need of spraying them with laquer, they don't loose a drop, printed full solid in 3mm walls. Time will tell;)

Very nice! I like the color usage and the logos.
The thickness looks a bit high compared to the width. The width of the strap looks also a bit narrow. On a neo it probably doesn't matter but on skin one might feel it. How long can you log with this setup?

Select to expand quote

PacoRaapNL said..

yekokata said..
Would it be possible to use a hall effect switch to replace the reed switch ? one of these aliexpress(dot)com/item/1005003574024310.html
Would it be more reliable ?

@veton have you posted the files for the enclosure somewhere ? also what is "flex" resin ?

@ yekokata, if the hall switch is designed for 5 volt power it should work.
Problem is, these switches drain some miliamps when in use which you do not want.
Reed switches do not consume power.
Also the size is not so small.
Some users are working in the background for another electronic switch solution based on hi side switching with a latch function using a special IC.
You want the battery to be disconnected when not in use to prevent the battery draining.

There are these
www.melexis.com/en/product/MLX92216-MLX92217/Micropower-microwatt-hall-effect-latch-switch
file:///C:/Users/61412/Downloads/MLX92216-MLX92217-Datasheet-Melexis.pdf

I don't know how this compares but here is the power consumption


Only thing is they a seiously small.

Select to expand quote

Freezer said..

veton said..
After many changes and prints this is my first unit. Poliurethane covered electronics, flex resin and the pogo connector, which is worth giving a try. Also added a led. it is very compact, only 101 grams with a 1800 Lipo and clear field for the GPS antenna.
Nobody talks about print specs, this is made out of PCTG, similar to PETG, which I also used in a lot of prototypes. In both cases I didn't find the need of spraying them with laquer, they don't loose a drop, printed full solid in 3mm walls. Time will tell;)

Very nice! I like the color usage and the logos.
The thickness looks a bit high compared to the width. The width of the strap looks also a bit narrow. On a neo it probably doesn't matter but on skin one might feel it. How long can you log with this setup?

I didnt try It yet but I think It will be fine on skin, It is a bit lighter than the motion, and if so we can try a neoprene strap cover as in motions.
Im out but soon will upload the model. It is a bit crazy with tight fits, but allows a clear vision for the GPS in a very compact size.
It is 1800ma so I guess enough for a couple or three sessions.

@yekokata Flex resin is a flexible epoxi used mostly for doming, trying to give a bit more tolerance to thermal expansi?n and contractions as many think hard epoxi could cause some problems with electronics. I dont know really, just trying to avoid any posible cause of fail. Thats why i changed in the last minute to pogo connectors as the wireless coils could also be problematic.

I bought a wireless charging powerbank from itechworld. It has Magsafe which hold the phone in the correct.l position for charging. Unfortunately the magnets on the powerbank trigger the reed switches in the DIY-GPS and turn it on. It charges slowly and gets real hot. If gone back to the non Magsafe powerbank.

I made a video a week or so ago how to optimise wireless charging and minimise heating for these DIY GPS's. Video was meant for a few team members getting my design 3D printed units Boz and I are currently making. (Original batch we made were all cooked from overheating from poor position while charging and video trying to avoid a 3rd batch build). Since a few here are reporting units getting hot whilst charging (and slow charging), tips on how to optimise the position of the GPS charge receiver on the charge mat may be beneficial to all styles and users. Short summary if you don't want to watch video: 1) Wireless charging for 1hr if done correctly should give 40% battery capacity which is around 8hrs sailing
2) 3hrs charging should fully charge battery from 0 to 100%
3) Don't leave the units on charge mat overnight (quite dangerous for you or unit unless optimally positioned)
4) For all my 3D printed case designs (boom, mast, shoulder), the optimal position is not centred on charge mat, is 2-4cm off centre
5) To optimise charge position if unknown, turn unit on before placing on charge mat and note battery voltage and %, then place on charge mat and move unit around until battery voltage and % read the highest. Then turn unit off, set timer for 1hr no more than 2hrs unless you about to set a distance record in which case 3hrs. (mark optimal position for future reference)

The above times apply to generic 5W charge receivers..possibly you are using smaller 3W receivers so add roughly 1hr to above charge times

quick question, does anybody know the differences between the diferent usb male cables specs? im buynig these "type c FPC 2p EK"
I assume Fpc is something about the plastic used, 2p two pin, but what does "EK" means? I know is something about stabilizing charging but I wonder if it is dangerous or overcharges the PCB. I use them to connect t to the pogo pin connector and feed the board with it.

Is the B one, thanks!

Select to expand quote

Flex2 said..
I made a video a week or so ago how to optimise wireless charging and minimise heating for these DIY GPS's. Video was meant for a few team members getting my design 3D printed units Boz and I are currently making. (Original batch we made were all cooked from overheating from poor position while charging and video trying to avoid a 3rd batch build). Since a few here are reporting units getting hot whilst charging (and slow charging), tips on how to optimise the position of the GPS charge receiver on the charge mat may be beneficial to all styles and users. Short summary if you don't want to watch video: 1) Wireless charging for 1hr if done correctly should give 40% battery capacity which is around 8hrs sailing
2) 3hrs charging should fully charge battery from 0 to 100%
3) Don't leave the units on charge mat overnight (quite dangerous for you or unit unless optimally positioned)
4) For all my 3D printed case designs (boom, mast, shoulder), the optimal position is not centred on charge mat, is 2-4cm off centre
5) To optimise charge position if unknown, turn unit on before placing on charge mat and note battery voltage and %, then place on charge mat and move unit around until battery voltage and % read the highest. Then turn unit off, set timer for 1hr no more than 2hrs unless you about to set a distance record in which case 3hrs. (mark optimal position for future reference)

The above times apply to generic 5W charge receivers..possibly you are using smaller 3W receivers so add roughly 1hr to above charge times

Good video, nice to see we had the same experience. thats why we made the housing so, that the charger fit in one way.

We Advice to use only 5 watt chargers. Like the orginal iPone chargers. These are 5v 1 amp. The esp charge with a max of 500 mA so 2.5 Watt. The max current i have messeured was 400 mA (so max 2 watt) with the wireless reciever.

Select to expand quote

remery said..
I bought a wireless charging powerbank from itechworld. It has Magsafe which hold the phone in the correct.l position for charging. Unfortunately the magnets on the powerbank trigger the reed switches in the DIY-GPS and turn it on. It charges slowly and gets real hot. If gone back to the non Magsafe powerbank.

Yes, in our design a iPhone magnetic charge fits to. It works but indeed the magnet in the charger activate the reedswitch. I did not experience a lot of heat but it's not the intention that the reedswitch is closed .