Forums > Windsurfing   Gps and Speed talk

Another DIY GPS logger approach

Reply
Created by rp6conrad > 9 months ago, 2 May 2021
boardsurfr
WA, 2202 posts
12 Apr 2022 7:58AM
Thumbs Up

Select to expand quote
rp6conrad said..
Two people have now access to the source code : Freezer and myself. The reason that the sw is not completely open source is because it is now released by GP3S. There is a certain trust that the generated GPS-data is not manipulated by the sw.


That is somewhat funny, considering the GP3S team worked closely with Coros to manipulate the watch data to their liking (that is, speeds for 2 second runs lower by ~ 1/2 knot are considered ok). Coros won't give much details about what is happening in the software without permission from GP3S, which the GP3S team is not willing to give.

Even ignoring that GP3S/Coros use GPS data that are manipulated by software which they keep as a trade secret, the argument that "the generated GPS-data is not manipulated by the sw" is a rather weak one. Anyone who would want to manipulate GPS data using software can do so, and just upload the resulting file for analysis. For GPSTC, that's not even necessary - you can still post results directly (albeit GPSTC might ask for the file if the postings look "funny"). Anyone who'd really want to manipulate data in software would need to have be rather masochistic to do it in Arduinio code!

Nor does the "trust" argument really hold up to closer inspection. With closed-source firmware, you rely on blind trust into the vendor or developer. Not that it ever happened that a vendor published firmware versions that gave bad results . With open source, others can check the software to see what's happening. That has at least the promise that problems and bugs are found and fixed faster, and that anyone can know exactly what's going on.

There are some good reasons why one may not want to go open source, like plans to commercialize a device, or that the software was written in a way that never was intended to be shared, or just unwillingness to deal with the potential overhead that could result of going open source. But fear of "data being manipulated by the sw" is not a good reason.

Windxtasy
WA, 4013 posts
12 Apr 2022 9:30AM
Thumbs Up

Select to expand quote
boardsurfr said..
Great reports and ideas here, keep it coming! The "problem reports" are very useful, too. I'm waiting to get parts from China, and got a 3D printer to make a boom mount like Flex did, but after Windxtasy's post, I might stick to the GoPro case route for a while. Maybe I'll follow Elmo's suggestion and print a frame for the components with switches that line up with the GoPro buttons.





boardsurfr, Flex has designed a 3D printed surround, that will hide the wiring and solve the ghosting problem, and redesigned the 3D printed box to be a bit deeper to accommodate it, so as long as you use the dichloromethane glue he recommends you should have no problem. I am sure Flex has put the updated design on thingyverse.
I would recommend that you seal the outside of the box with something before filling with potting epoxy, to stop messy leakage, but do not seal the inside.

Meanwhile I have come up with an elegant solution for my problem.
1. Attach scratch resistant polycarbonate cover to the screen/surround assembly with potting epoxy to remove internal spaces and the second surface for relection. My previous effort showed that potting epoxy (flexible) adhered much better than hard epoxy
2. If you can't glue dissimilar plastics, use mechanical means ie a frame with a lip, to hold the polycarbonate in place.
Make the frame from the same material as the box, ie 3D printed, so you can glue similar plastics with a substance known to work ie botecote epoxy.
3. seal the surround with epoxy for added strength.

Thanks to Flex' custom 3D prints for the surround with lip. I think it looks better than the original design anyway!
I will attcha a photo of the assembly before glueing as curing is currently in progress.




elmo
WA, 8659 posts
12 Apr 2022 1:01PM
Thumbs Up

Anita,

I had no problems with my clear plastic bonding as I used Mr Hot glue gun, a bit daggy on outside edge but stuck it down well and no seepage of resin.

Probably look at the Gopro case option next (I went and 'bricked" one of my units through my own stupidity).

Will have to figure out how to mount it to the boom which shouldn't be to tricky.

boardsurfr
WA, 2202 posts
12 Apr 2022 10:24PM
Thumbs Up

Select to expand quote
Windxtasy said..
I would recommend that you seal the outside of the box with something before filling with potting epoxy, to stop messy leakage, but do not seal the inside.


Thanks, Anita. So I guess the "messy leakage" means that the 3-D printed boxes are not watertight, and therefore have to be filled with epoxy? I think I'd rather keep the ability to replace parts, access the USB port, and so on, so probably will start out with the GoPro case.

But if the 3-d printed cases are indead leaky, this makes me wonder if some of the LCD problems that the original Motions had were due to small amounts of water getting into the case. Even the GT-31, which usually lasted many years, would show some signs of corrosion when opening it up. And don't get me started on the GW-60...

boardsurfr
WA, 2202 posts
13 Apr 2022 6:54AM
Thumbs Up

Checking on whether 3D-printed boxes are watertight, I quickly learned that they are not (which matches what Flex had seen in his early tries). One article that describes various ways to make 3D prints watertight (3dprinterly.com/how-to-make-your-3d-prints-waterproof-airtight/) suggested to use MG Chemicals Silicone Conformal Coating. That's what I had used in the past to "paint" the insides of bluetooth headphones that were sold as "IPX7", but had visible holes in to let the water in. Those headphones still work, and have so far survived longer than the ones I had used previously (and not waterproofed).

Thought I'd mention it as an alternative to the things Anita had suggested. But even when using a GoPro case, it may be a good idea to paint some of the electronic components with this stuff. While GoPro cases usually don't let any water in, and the one I use for my u-blox Openlog GPS has survived a few hundred sessions, I have also had a few GoPro cases that over time started to let water in. The GPS I currently use if on top of my helmet, so it often does not get wet during a session; a GPS mounted on the boom will probably get wet a lot more, so a bit of extra waterproofing may well be worth it.

decrepit
WA, 11829 posts
13 Apr 2022 1:35PM
Thumbs Up

Select to expand quote
boardsurfr said..>>>>a GPS mounted on the boom will probably get wet a lot more, so a bit of extra waterproofing may well be worth it.

It's not just "getting wet" it's hitting the water at high speed, so there can be a lot of pressure involved.

hardpole
WA, 565 posts
13 Apr 2022 3:37PM
Thumbs Up

Have built my first one of these (purchased parts for 2 but so far haven't fried anything).
Just has a couple of questions, it might just be me not understanding but thought the people here would know the answers.

What value should be put for the cal_bat field in the config.txt. I have mine set to 1.7 and even when the battery (labelled 3.7v 2000mAh) is showing 3.8 or 4 on the display when charging when logging the battery graphic is never full. Should I adjust the cal_bat value?

When connected to wifi at home and I shut the unit down it seems to come back on a minute or so later. If I disconnect from the wifi (turn of that SSID on my wifi) then shutdown it stays off.

When it is off the satellite LEDs seem to flash continually, see the picture below of its "state" when the LED is flashing.

None of this stops it working and it is logging perfectly but I just wondered if the variation in board / my assembly might have caused it to be not working properly. I havent sealed the case yet as I thought I would make sure its all working before I do that.

My plan is to wrap it like a birthday present in one layer of woven fibreglass and then use clear coat resin to seal it all up. It still should be transparent and charge but be waterproof unless I crack it. I dont know if it matters if the resin actually sticks to the plastic box but if it does that would be a bonus.

I did think I had an issue as I was finding that after charging it would sometimes lock up and I would have to open it up and use the switch on the side to turn it off/on - even the reset button on the display didnt help. But I then realised it was just overheating and if I left it to cool down (put it in the fridge the first time) it would start working again. I was also using a very strong magnet and had put the reed switch on the top side near the screen connector. I moved that to the bottom of the case and used a "normal" magnet and that seemed to help. But now Im thinking that issue was just the overheating.

The epaper display cards I got were slightly different to the ones in the build documents, slightly different layout and more importantly pins instead of holes for the connections. But the software and everything else seems to work perfectly well and I just had to identify the power chip to connect the charging pad to by its shape.






I can't believe the stuff they sell on aliexpress - but why after purchasing this they think (know) that I want a scale model steam engine I cant work out. I ended up with a couple of extra 1.5 inch epaper displays I purchased by mistake - making a temperature logger for my sun room out of them now that I have the arduino bug.

Windxtasy
WA, 4013 posts
13 Apr 2022 8:56PM
Thumbs Up

Select to expand quote
hardpole said..

What value should be put for the cal_bat field in the config.txt. I have mine set to 1.7 and even when the battery (labelled 3.7v 2000mAh) is showing 3.8 or 4 on the display when charging when logging the battery graphic is never full. Should I adjust the cal_bat value?
my battery graphic is never completely full either even though it gets to about 4.2. config file is set to 1.75. I think 1.75 is the level the unit cuts out at to stop the battery from running dangerously low.
My config file settings:
{
"cal_bat": 1.75,
"cal_speed":1.944,
"sample_rate":10,
"logOAO":1,
"logUBX":1,
"speed_field":1,
"stat_field":6,
"dynamic_model":1,
"timezone":-16,
"bar_length":1852,
"UBXfile":"BoomRight",
"ssid": (home wifi)
"password": (home wifi password)
}
When connected to wifi at home and I shut the unit down it seems to come back on a minute or so later. If I disconnect from the wifi (turn of that SSID on my wifi) then shutdown it stays off.
That doesn't happen with mine. How are you turning it off?

When it is off the satellite LEDs seem to flash continually, see the picture below of its "state" when the LED is flashing.
When the unit is off there are no lights on

I did think I had an issue as I was finding that after charging it would sometimes lock up and I would have to open it up and use the switch on the side to turn it off/on - even the reset button on the display didnt help. But I then realised it was just overheating and if I left it to cool down (put it in the fridge the first time) it would start working again. I was also using a very strong magnet and had put the reed switch on the top side near the screen connector. I moved that to the bottom of the case and used a "normal" magnet and that seemed to help. But now Im thinking that issue was just the overheating.
I had one unit that was sometimes slow and occasionally impossible to turn off. Especially seemed to happen after glueing the polycarbonate top on. Maybe that was heat? After some trouble shooting, changing the reed switch and chats with Jan I have removed the on off switch and that seems to have solved the problem.

The epaper display cards I got were slightly different to the ones in the build documents, slightly different layout and more importantly pins instead of holes for the connections. But the software and everything else seems to work perfectly well and I just had to identify the power chip to connect the charging pad to by its shape.
Your induction coil looks much smaller than the one I have which is about 4cm X 3cm

hardpole
WA, 565 posts
13 Apr 2022 9:22PM
Thumbs Up

Sorry should have said that's a speaker that it came with. I disconnected it as I didn't see much point in keeping it. Coil about 4x3.

Turning it off by long hold of the magnet. It's not a big deal but might be todo with the slightly different board. Even though lights are flashing battery doesn't seem to go flat. Left it logging last night by accident and it had turned off this morning.

rp6conrad
288 posts
13 Apr 2022 10:29PM
Thumbs Up

I have no experience with this version (T5 V2.4 ?), but I assume that the blue LED on the board is directly connected to the 3.3V. So as long there is power from the battery, the LED will stay on. Here, the standby (deepsleep) current is around 1.5 mA, where the version without LED only has0.5 mA.
If you switch off the board, the LEDs on the GPS (Beitian BN220 / 280) should immediately switch off. After 10s, you get then the "deep sleep" screen. Wifi connection active / not active should make no difference.
When switching ON, the blue LED on the GPS will flash at the sample rate that you have in the config. The red LED will flash @ 1Hz if there is a gps fix.
A full LIPO battery has a voltage around 4.2 V, completely empty is around 3.2 V. The symbol should match with these voltages. The 1.7 factor in the config.txt can be adjusted until you have the correct reading (measure the exact voltage with a multimeter on the Vbat /GND pins).
Which version of the sw you are using now ?
Before potting, make sure the reed switch works fine, and the unit works like it should be !
Greetings, Jan.

hardpole
WA, 565 posts
14 Apr 2022 7:41PM
Thumbs Up

Thanks Jan, the blue led in the photo was the one on the GPS module. There is only one on the board which lights up when charging, I notice now that when connected with USB cable it goes out when fully charged. It didnt seem to ever do that with wireless charging.Is that photo I posted above the deep sleep screen?

Weirdly since posting I havent had the turn itself back on issue so maybe that was just a temporory thing (or something I was doing wrong).

Am i right that there are really only two "controls" a short battery / switch close to change screens and a long one to turn on /off ?

I am running version 5.4 of the software. I will experiment with the cal_bat value to get that graphic correct. Now that its charging via cable I can be confident its fully charged. I have used shrink wrap on all the leads so will have to unplug the battery to measure the voltage, might just use trial and error instead.

Am planning on making a little cosmetic cover to leave only the display and lights showing through the transparent cover. Just have to make sure it all lines up neatly before closing it up.

rp6conrad
288 posts
15 Apr 2022 3:33AM
Thumbs Up

Select to expand quote
hardpole said..
Am i right that there are really only two "controls" a short battery / switch close to change screens and a long one to turn on /off ?


That is almost correct, the GPIO39 will turn the device on, and after startup a long push will send it back to sleep. But if you have a fix, a short push will toggle the upper field (speed_field in the config.txt, see attached screendump).
If the speed_field is set to 1, the upper field will change depending on the actual run distance or jibe.
In the newer version of the the sw, there is a second GPIO12 active. With this one, you can direct choose a stat screen, and toggle through stats-screens. I will send you the latest version (SW 5.46)
Greetings, Jan.






Windxtasy
WA, 4013 posts
15 Apr 2022 7:41AM
Thumbs Up

Select to expand quote
rp6conrad said..

hardpole said..
Am i right that there are really only two "controls" a short battery / switch close to change screens and a long one to turn on /off ?



That is almost correct, the GPIO39 will turn the device on, and after startup a long push will send it back to sleep. But if you have a fix, a short push will toggle the upper field (speed_field in the config.txt, see attached screendump).
If the speed_field is set to 1, the upper field will change depending on the actual run distance or jibe.
In the newer version of the the sw, there is a second GPIO12 active. With this one, you can direct choose a stat screen, and toggle through stats-screens. I will send you the latest version (SW 5.46)
Greetings, Jan.







I discovered the new alpha screens when I took my units for a test drive yesterday. (not ready for the water yet)
I like the improvements - the way it switches to an alpha screen after detecting a turn - you don't need to be on an alpha screen to start with.
The window (proximity) and distance to exit numbers are very useful, as is the miss message when your alpha did not reach the required proximity.

hardpole
WA, 565 posts
16 Apr 2022 11:33AM
Thumbs Up

My problem fixed with Jan's help.

anyone else who purchases the 2.4.1 version of the board, read the instructions properly. This board doesnt have pin 25 so I connected the 3.3v output instead. Of course the software doesnt turn that one off when in sleep mode. Removing that connection now means the GPS satellite device powers off. It does mean it only has 2 sources of power but that seems to be enough to drive it.

So assembler error !!

many thanks again to everyone for their help and most especially Jan who probably had better things to do that research my stuffup.

I set the cal_bat to 1.81 and that seems to now show a "full" battery for me. Displaying "Bat: 4.2" when fully charged.

Windxtasy
WA, 4013 posts
19 Apr 2022 2:43PM
Thumbs Up

Select to expand quote
hardpole said..
My problem fixed with Jan's help.

anyone else who purchases the 2.4.1 version of the board, read the instructions properly. This board doesnt have pin 25 so I connected the 3.3v output instead. Of course the software doesnt turn that one off when in sleep mode. Removing that connection now means the GPS satellite device powers off. It does mean it only has 2 sources of power but that seems to be enough to drive it.

So assembler error !!

many thanks again to everyone for their help and most especially Jan who probably had better things to do that research my stuffup.

I set the cal_bat to 1.81 and that seems to now show a "full" battery for me. Displaying "Bat: 4.2" when fully charged.


I have been experimenting with different cal bat values but I can't get any closer to full than one thin white line at the top off the battery symbol. 1.69 gives the "fullest" looking symbol with 4.15V

Windxtasy
WA, 4013 posts
19 Apr 2022 6:53PM
Thumbs Up

Today I had my first sail with my ESP32 GPS. It looked great with my orange KA sails.
Building it has been an adventure that has taken me out of my comfort zone many times over the last couple of months, from microsoldering components, to navigating Arduino, learning to upload and download from the device using wifi and its own IP address, learning about 3D printing, various epoxies and adhesives, and what works with polycarbonate and what doesn't. In trying to make the unit look tidy I made a discovery along the way too, that the surround of the E paper needs to be shielded so it doesn't ghost in bright sunlight.
At times I was almost in tears with frustration (don't try to do this with windows 7) but thanks to a lot of encouragement from Flex and help form Jan I now have two well functioning units.
I am very happy with the end result, and I have enjoyed the challenge despite the frustrations, and I might be a teensy bit sad the adventure is over.
Jim was right when he said "I could flash the board for you but you will feel a lot better if you do it yourself".








decrepit
WA, 11829 posts
19 Apr 2022 9:17PM
Thumbs Up

Fantastic Anita!

Windxtasy
WA, 4013 posts
19 Apr 2022 9:21PM
Thumbs Up

Select to expand quote
decrepit said..
Fantastic Anita!


Thanks Mike. I look forward to sending you some tracks when the epoxy on the other unit is cured

Stretchy
WA, 914 posts
19 Apr 2022 10:03PM
Thumbs Up

Great effort Anita. I'm not game, I'd stuff it up for sure. Can you or Flex make me one?

Windxtasy
WA, 4013 posts
20 Apr 2022 7:05PM
Thumbs Up

Select to expand quote
Stretchy said..
Great effort Anita. I'm not game, I'd stuff it up for sure. Can you or Flex make me one?


Stretch, I was afraid of stuffing it up also, many times! especially when I was having a problem with turning off one unit and Jan said he had solved that in one unit by removing the on off switch. Then he said "be careful you don't damage the adjacent resistor as you do it"... I had practically resigned myself to building another unit but removing the switch solved the problem.
It was because the build etc was such a challenge that I feel a real sense of achievement having built it. Equally as good as my first 35 kn.
As Flex said "I could flash the board for you but you will feel a lot better if you do it yourself."
I wouldn't want to deny you that feeling of achievement.
Advice is available if you need help.

Freezer
89 posts
27 Apr 2022 6:01AM
Thumbs Up

I absolutely love the 3D housing that we developed. Similar to the design of Flex on the boom it is split in 2 zones. 1 with the battery and GPS, and the 2nd with the wireless receiver and the T5 board with the reed contacts.
We used the same construction as the standard boxes from China with lid and seal to make the top/bottom waterproof construction. In the lid we mount plexiglass with double-sided tape, similar to the GT31. The inserts in the lid and mounting screws from the bottom it can be screwed firmly together. The PLA is not waterproof in itself so I need to apply a coat of DD 2k to solve that. The velcro strap need to be connected with a 2mm pen (will try a bicycle spoke ??) epoxied into place.
the result looks stunning and it its nicely and snugg. I'm still using a BN280, so with a BN220 or BN180 the hump can be made smaller. First trails I have done last Saturday, but still in a Paqua housing since I haven't put the coating and straps on.
One problem that I have is the heat from the wireless charger. The round shape of the charger is preventing the coils to line-up and causing losses and therefore heat. Perhaps the thickness of the housing is also a bit too thick at the receiver, so might need to thin things for better connection and less heat.
I opened the charger and found that the 2 power fets are generating a lot of heat. And they are not actively or passively cooled, nor do I notice it reduces power based on temperature. So this a hazard in the making. I will probably create a custom housing that will align the coils and use some thermal paste and alu strip to cool the power fets.
I thought that my phone did not heat up the charger, but that's not the case. It so gets hot as well, coming from the charger/fets.
When I inceased the distance of the coil, while it was still charging the temperature went up even faster. So the more losses, the higher the temperature I guess.

I saw that Julien also had issues with the chargers/heat of the motions. I don't think there is a handshake between the devices to start charging only with certified (Qi) like with the fast-wired chargers. I guess the charger just starts based on an impedance sense.

I had a wireless charger from Mi that has active cooling, but it stopped working half a year ago. I only have these cheap Chinese ones now, but they get too hot and require modifications to stay safe.

Looking forward to your experience on wireless chargers and suggestions.

Flex2
WA, 269 posts
27 Apr 2022 6:29PM
Thumbs Up

I've noticed with my units I can put them on the charge mat and the indicator light on mat changes to confirm charging...however, despite getting quite hot sometimes (actually quite often) the units don't charge even if left for 24hrs. The work around/solution is turn the GPS on, wait the 10 secs to get past the AP option then you get realtime battery voltage (up until you get GPS lock). You then move/angle your GPS on the charge mat....I find there is a tiny sweet spot where the voltage jumps 0.1 to 0.2V above any other position...this is the position to charge in...and the unit can be fully charged from flat in around 2hrs. The unit 'seems' cooler when it is charging in this position but I haven't made measurements to confirm this...I will try to remember to stick the infrared gun on in both cases. The sweet spot position is definitely not dead centre.

Flex2
WA, 269 posts
27 Apr 2022 9:34PM
Thumbs Up

Freezer, I did a quick test...may not be representative as my unit was basically fully charged for both cases but there was significant temperature difference to worth mention. First I placed my unit dead centre (photo 1) for an hour where you would instinctively put it (I know this doesn't charge in this position but the charge mat light comes on regardless). On the display side my infrared thermo read a max of 55degC...underneath on the charge mat side I got 65degC...this is hot. Then I left to cool then repeated with battery voltage optimise trick..you can see its off centre (photo 2)....left another hr and max temp I could get on display side was 36degC and underneath was 45degC. This is nearly 20 degC difference. When measuring the display temp the hottest part seemed to be 1/3 up the display and underneath was directly underneath those pesky FETs.

After playing with my Tesla and junk Tesla battery for quite some time I know Lipo batteries have their optimal charge temp for both longevity and speed of charge at 55degC. The Tesla engineers I figure are far smarter than me so they are doing it for good reason. Whenever you charge a Tesla it makes a beeline straight to 55degC bat temp if charging over about 11KW. 65degC though is way too high and certainly damaging. Not sure if you can extrapolate a 75KWHr Battery backwards to an 8Whr battery but if so a charge rate anywhere near 1W for the battery Jan is using is the same as 'fast charging a Tesla'. i.e. 250mA charge rate on a 2000mAhr battery is equivalent to 11KW on a 75KHr battery (please correct me if I got it wrong). Since I can charge from near flat to full in 2-3 hrs it must be around 750mA Hr peak current. Out of interest, Tesla define fast charging as any charge rate requiring and increase in battery temp to maximise longevity. If your battery is at 0degC, you need to heat it even to get 1W in.

Worth mentioning my charge coil is more on left of unit due to the electronics and wires on right so its more centred on mat in optimal position than appears in photo. (Photo 3 for rough placement of coil and charge electronics in unit) The point is though, the charge mat can register its delivering power but the receiver position needs to be optimised else much/all of the energy seems to go to heat. Not exactly sure why the heat seems to be dumped out the FET's on the Rx coil side with slight misalignment and not delivered to battery but that is the observation.













rp6conrad
288 posts
28 Apr 2022 2:04AM
Thumbs Up

I decided to make the project completely open source : Code is on github in a public repository : github.com/RP6conrad/ESP-GPS-Logger
I use the Arduino IDE 1.8.1, code is mainly in C, with some C++ objects. Do not except a well written readable code, as I am not a professional in C / C++.
Greetings, Jan.

Freezer
89 posts
28 Apr 2022 4:34AM
Thumbs Up

That's great news Jan! Now everyone can contribute. There are so many options and possibilities that can only be enabled with many contributers. It takes quite a bit of time, but it brings joy to all that like to tinker and tweak.

The commercial parties don't need to be scared. Their development might even benefit from the community that contributes in the open source.

Freezer
89 posts
28 Apr 2022 5:25AM
Thumbs Up

Select to expand quote
Flex2 said..
Freezer, I did a quick test...may not be representative as my unit was basically fully charged for both cases but there was significant temperature difference to worth mention. First I placed my unit dead centre (photo 1) for an hour where you would instinctively put it (I know this doesn't charge in this position but the charge mat light comes on regardless). On the display side my infrared thermo read a max of 55degC...underneath on the charge mat side I got 65degC...this is hot. Then I left to cool then repeated with battery voltage optimise trick..you can see its off centre (photo 2)....left another hr and max temp I could get on display side was 36degC and underneath was 45degC. This is nearly 20 degC difference. When measuring the display temp the hottest part seemed to be 1/3 up the display and underneath was directly underneath those pesky FETs.

After playing with my Tesla and junk Tesla battery for quite some time I know Lipo batteries have their optimal charge temp for both longevity and speed of charge at 55degC. The Tesla engineers I figure are far smarter than me so they are doing it for good reason. Whenever you charge a Tesla it makes a beeline straight to 55degC bat temp if charging over about 11KW. 65degC though is way too high and certainly damaging. Not sure if you can extrapolate a 75KWHr Battery backwards to an 8Whr battery but if so a charge rate anywhere near 1W for the battery Jan is using is the same as 'fast charging a Tesla'. i.e. 250mA charge rate on a 2000mAhr battery is equivalent to 11KW on a 75KHr battery (please correct me if I got it wrong). Since I can charge from near flat to full in 2-3 hrs it must be around 750mA Hr peak current. Out of interest, Tesla define fast charging as any charge rate requiring and increase in battery temp to maximise longevity. If your battery is at 0degC, you need to heat it even to get 1W in.

Worth mentioning my charge coil is more on left of unit due to the electronics and wires on right so its more centred on mat in optimal position than appears in photo. (Photo 3 for rough placement of coil and charge electronics in unit) The point is though, the charge mat can register its delivering power but the receiver position needs to be optimised else much/all of the energy seems to go to heat. Not exactly sure why the heat seems to be dumped out the FET's on the Rx coil side with slight misalignment and not delivered to battery but that is the observation.














Thanks for the experiment. I did exactly the same, switching on the esp32 to read out the battery voltage. Does it even charge without switching on? Do you charge it through the USB port or did you solder the wire to the pin on the esp32 board as described in Jan's documentation?

I would not compare the Tesla or even call this fast charging. They might be both full in 2 hrs, but that's all they have in common. I think I read that the esp32 can do 500mA. Even the lowest rated wireless chargers can deliver this.

You confirm basically my finding that non ideal placing of the coil causes losses an quite some heat. I also recognize the non center placement of the coil below the esp32. It is easy to forget or misplace it on a circular charger. I would recommend making some alignment markers on both GPS as well as the charger. Or even better print a 3d custom charger housing that will take care of the alignment and create some cooling for the electronics of the charger. We don't want to have more melted plastic or even worse, fires and accidents...

Roo
765 posts
28 Apr 2022 6:06AM
Thumbs Up

Select to expand quote
rp6conrad said..
I decided to make the project completely open source : Code is on github in a public repository : github.com/RP6conrad/ESP-GPS-Logger
I use the Arduino IDE 1.8.1, code is mainly in C, with some C++ objects. Do not except a well written readable code, as I am not a professional in C / C++.
Greetings, Jan.


Well done Jan,

fantastic effort putting this together. Couple of questions, with the latest firmware you can show the company logos on the final screen, how do you choose the correct logo? "Logo_choice":42, seems to give Board Brand 99 at the top right and the Neil Pryde logo below it.

Also there seems to be a few more variable available now in the config.txt.

{ "cal_bat": 1.75, "cal_speed":3.6, "sample_rate":5, "speed_field":8, "bar_lenght":1852, "stat_screens":321, "GPIO12_screens":421, "Logo_choice":42, "sleep_off_screen":11, "logOAO":1, "logUBX":1, "dynamic_model":0, "timezone":2, "UBXfile":"ESP_GPS", "Sleep_info":"Call +32 001 002 003", "ssid": "Your_SSID", "password": "Your_password"}

What different stat_screens and GPIO12_screens are available?

Loving the two units I have built up so far, thanks for all your assistance,

Roo

P.S. Now the source code's available I can write some arduino code to port the data to my heads up display glasses!

Flex2
WA, 269 posts
28 Apr 2022 6:43AM
Thumbs Up

Select to expand quote
Does it even charge without switching on? Do you charge it through the USB port or did you solder the wire to the pin on the esp32 board as described in Jan's documentation?



Yes, it will charge ok without switching on if you get the placement right. This particular unit has the soldered wire as per Jans documentation.

Nice one Jan, I'm sure this will evolve your software to its best potential.

rp6conrad
288 posts
28 Apr 2022 11:09PM
Thumbs Up

Select to expand quote

Roo said..



Also there seems to be a few more variable available now in the config.txt.

{ "cal_bat": 1.75, "cal_speed":3.6, "sample_rate":5, "speed_field":8, "bar_lenght":1852, "stat_screens":321, "GPIO12_screens":421, "Logo_choice":42, "sleep_off_screen":11, "logOAO":1, "logUBX":1, "dynamic_model":0, "timezone":2, "UBXfile":"ESP_GPS", "Sleep_info":"Call +32 001 002 003", "ssid": "Your_SSID", "password": "Your_password"}

What different stat_screens and GPIO12_screens are available?


The latest SW 5.49 has indeed some extensions in the config for the screens :
With the variable "Stat_screens": 321, you choose 3 different screens which will toggle when the speed drops below 1 m/s, in this case screen 1, screen 2 and screen 3. If you set 2, only screen 2 will show.
The same with the variable "GPIO12_screens":42, here screen 2 will show up when GPIO12 is pulled low (reed switch, or switch to GND), next push will show then screen 4. Long push will reset sthe stats on screen 4 and 5. (not the global stats !)
"slep_off_screen":11 : This will set the "Freezer off" and the "Freezer sleep" screens. If 0, then the old sleep screen will show.
"Sleep_info":"Your preferred txt in the sleep_screen" This line will show up in the "Freezer sleep_screen" !
"Logo_choice":23, will set Logo 3 for the board, and Logo 2 for the sail (only in "Freezer sleep_screen = 6"). Next logo's are now in the firmware :
// Board:
1 : Starboard_logoS_zwart
2 : Fanatic_logoS_zwart
3 : JP_logoS_zwart
4 :NoveNove_logoS_zwart
// Zeil:
1 : GAsails_logoS_zwart
2 : DuoTone_logoS_zwart
3 : Pryde_logoS_zwart
4 : NP_logoS_zwart
Numbers of different screens :

Greetings, Jan.


Roo
765 posts
28 Apr 2022 11:19PM
Thumbs Up

Thanks Jan,

explains it all, appreciate all the extra features.



Subscribe
Reply

Forums > Windsurfing   Gps and Speed talk


"Another DIY GPS logger approach" started by rp6conrad