Differences are small, and certainly in the normal "range". I always sail with 4 esp-gps, 2 on the head and 2 on the boom. Head is always a little faster (avg diff 0.02 knots or less).
You can compare each datapoint with gps-speedreader, and check for the nr of sats and sacc values : ecwindfest.org/GPS/GPSSpeedreader.html
The case is milled polycarbonat/plexiglas and then glued ? Which kind of glue ?

Greetings, Jan

Thanks for that test Actimel87.
Very interesting and closer than I expected. But the +/- values, are worse on one of them, I suspect the one with antenna behind display.

But the actual results are very close. I'd always want to post from the more accurate one. But antenna behind display is obviously more than adequate for on the water feedback, and if you can get it small enough for wrist mounting, ideal for that purpose.

I think the diffrence hapen during the start. The cold start is described with 27sec I guess kind of optimistic.
Also the ESP was during startup some initializing maybe the configuration was du this not equal.

The housing is milled out from 4mm and 2mm Polycarbonate(Hobbyglas).
ESP1 is glued with Tangit PVC-U. Works the last year but now it is getting loose.
ESP2 is glued with ACRIFIX 1R0192. Not tested.
Both are not waterproof!
I read something about double sided duct tape. I will give it a go next time.

I can provide a PDF template to cut a housing with a fret saw or similar?

Is there something besides the small antenna what speaks against the BE-122?

I read something about double sided duct tape. I will give it a go next time.

Simon used 3m BHV foam tape of 1 mm thickness.I will use it too to adhere the 0.7 mm PVC transparant top plate to the 3D printed part and cover it up with a 3D printed part secured by the 2 mm screws to keep the foam tape under 50% load.

Paco

One test is not enough to see any difference or draw valid conclusions. Good or bad.

But as Decrepit said, wearing on the wrist has been proven very clearly, to be throwing away a LOT of accuracy and consistency.
The best compromise is on the upper arm where it can still be easily viewed, (unless you are very longsighted. ) but also the boom mounting on both sides and averaging of the results has been shown to be very useful for a lot of sailors.

Select to expand quote

sailquik said..
One test is not enough to see any difference or draw valid conclusions. Good or bad.

But as Decrepit said, wearing on the wrist has been proven very clearly, to be throwing away a LOT of accuracy and consistency.
The best compromise is on the upper arm where it can still be easily viewed, (unless you are very longsighted. ) but also the boom mounting on both sides and averaging of the results has been shown to be very useful for a lot of sailors.

I wonder if somebody try to mount on the top of the mast already? It must be compete mess !!!

On the latest 213BN version should there be a red LED for when the wireless charging is active?

Paco

No, only with the older BN versions.
If you need indication for Charging you can wire a LED.

Select to expand quote

PacoRaapNL said..
On the latest 213BN version should there be a red LED for when the wireless charging is active?

Paco

I guess you got them at www.tinytronics.nl/shop/nl/development-boards/microcontroller-boards/met-wi-fi/lilygo-ttgo-t5-v2.3.2-esp32-met-2.13-inch-e-paper-e-ink-display
I also have the ones with the USBC versions without the serial chip and are version v2.3.2 and have LEDs again. I also notices it.
The olders ones you find on AliExpress at the Lilygo Official Store are v2.3.1 and had no LEDs:
nl.aliexpress.com/item/32869729970.html

Not sure if you can simply unsolder them. They do consume a bit of power, and not visble through the housing.

Select to expand quote

Freezer said..

PacoRaapNL said..
On the latest 213BN version should there be a red LED for when the wireless charging is active?

Paco

I guess you got them at www.tinytronics.nl/shop/nl/development-boards/microcontroller-boards/met-wi-fi/lilygo-ttgo-t5-v2.3.2-esp32-met-2.13-inch-e-paper-e-ink-display
I also have the ones with the USBC versions without the serial chip and are version v2.3.2 and have LEDs again. I also notices it.
The olders ones you find on AliExpress at the Lilygo Official Store are v2.3.1 and had no LEDs:
nl.aliexpress.com/item/32869729970.html

Not sure if you can simply unsolder them. They do consume a bit of power, and not visble through the housing.

Hello Simon, yes those are the tinytronics versions I have.
Saves me for reading all responses again.
Removing SMD LED is question of heating with extra solder and use small tweezer to move it from its place.I have new spool with transparant PETG in my basket at Tiny tronics. Will print a transparant one so internal LED (red) are visible. :-)

Paco

Select to expand quote

PacoRaapNL said..

Freezer said..

PacoRaapNL said..
On the latest 213BN version should there be a red LED for when the wireless charging is active?

Paco

I guess you got them at www.tinytronics.nl/shop/nl/development-boards/microcontroller-boards/met-wi-fi/lilygo-ttgo-t5-v2.3.2-esp32-met-2.13-inch-e-paper-e-ink-display
I also have the ones with the USBC versions without the serial chip and are version v2.3.2 and have LEDs again. I also notices it.
The olders ones you find on AliExpress at the Lilygo Official Store are v2.3.1 and had no LEDs:
nl.aliexpress.com/item/32869729970.html

Not sure if you can simply unsolder them. They do consume a bit of power, and not visble through the housing.

Hello Simon, yes those are the tinytronics versions I have.
Saves me for reading all responses again.
Removing SMD LED is question of heating with extra solder and use small tweezer to move it from its place.I have new spool with transparant PETG in my basket at Tiny tronics. Will print a transparant one so internal LED (red) are visible. :-)

Paco

Mind that blue PETG filters RED light and flashing is not visible. I know as I have a transparent blue PETG housing and the led on the Beitian flashing is not visible (althoug you can see the wires. So te color of your transparent PETG will determine how visible it is.

The dry testing for the wireless charging.

The no high tech waterproof testing before electronics go in. It keeps floating so I put it upside down so the 3M foam tape seal is under water constantly. The 0.45 mm PVC sheet is clamped in-between the bottom and top PETG part. The large hole is for the IO39 switch. The small hole for the duo color led for charging and switch press. Both led colors are off in normal use. The PVC sheet in-between the parts can compensate for pressure differences as it is 1 mm above the ink screen.


The upside down test. :-)

First simple water proof test failed. :-(
From what I can see the foam tape has flat pressure bonding to the PVC sheet.

Things I think I need to take care of is that the top surface of the PETG needs to be sanded at smooth as possible.

As the case is 80x80 mm I looked for this size 3M foam tape/patches in a single square form at Ali but I cant find it yet in this size.

Currently the 4 seams are pressed together.

I also threaded the M2 screws through the tape which distorted the structure.

The next test version I will use a hot needle to poke/make the holes through the tape so the M2 does not chew the tape.I need to change the order of assembly.

Hey Paco,

did you consider gluing the window and using an O-Ring or a round Cord in a Labyrinth seal?
Also I think I would reduce the size of the "window part"to an absolut minimum. A smaller window is more robust and has less area where you can get a leaking.

Our beloved ESP32 in the T5-epaper configuration has a total of 16MB flash memory. When I started the project, I had no clue and I choose the Default "partition" scheme with only 4MB in the Arduino IDE. The partition can only changed over usb, so not with OTA. This was the root cause for troubles when using the ESP32 2.0x library instead of the 1.0x library !
However, a interesting feature is use of the flash memory for data-storage. So, without the SD card, we could even use minimal 12 MB for our log files ! With the compact .gpy format, this stands for 15 hours data @ 10 Hz ! (or 30 hours data @ 5 Hz).
I did some tests with a lilygo T7-S3, and it seems to work just fine. No lost points @ 10Hz gpy.
My idea is now to adapt the SW for the T5-e-paper, if no SD card is found, the internal flash will be activated for data-storage.
Another idea is to use other (smaller) ESP32 boards. There is a very small Lilygo board with a small LCD-display : www.lilygo.cc/products/lilygo%C2%AE-ttgo-t-display-1-14-inch-lcd-esp32-control-board
The 16MB version goes for only 9$ !
The biggest advantage is the smaller formfactor, and off course, dead cheap hardware...


Greetings, Jan.

Nice one Jan!!!!

Nice find and thinkering Jan.

Reducing the fomfactor by standard by using internal memory would be a welcome new feature for new to build versions.
New firmware for the T5 ink would be nice.

This way it is easier to build wrist versions like Actimel tries to achieve and also less wires too.
But smaller display means less readability?LCD consumes more current?

Paco

Select to expand quote

Actimel87 said..
Hey Paco,

did you consider gluing the window and using an O-Ring or a round Cord in a Labyrinth seal?
Also I think I would reduce the size of the "window part"to an absolut minimum. A smaller window is more robust and has less area where you can get a leaking.

Actimel, yes I thinkered about both things.
I am still in a low tinkering stage.
I first like to have the project that I can open and close again with using less parts as possible.
Replacing the tape and transparant window are not the cost if opening the box is needed due to an electronic problem.

The larger size of the transparant window lets the pressure in the box compensate like a membrane (that is my thinking) depending on depht and temperature changes.

If you glue the transparant window the glue joint might break or leak more easily, therefor Simon his 3M VHB foam tape would solve that as it is a flexible foam and so with temperature changes the whole can expand and retract giving less stress on the whole structure.
Or you use a flexible adhering butene kit which might be a bit more messy.

The top plate over the large transparant window has two functions.
To devide the pressure over the foam tape as regular as possible.
Estathic and only viewable parts are exposed reducing the possibility to reduce impact from the outside.
My thinking is it would survive as it is no brick object anymore.

For the sealing with an O-ring or 2 mm diameter slicone foam tube I already looked at Ali for it but not yet ordered.
nl.aliexpress.com/item/1005005325164575.html?spm=a2g0o.cart.0.0.506c61d7kFiHhr&mp=1&gatewayAdapt=glo2nld

However the flat 3M tape has more sealing surface as a compressed tube.I currently use 10 mm width but might opt for other test to 5mm width.

So much tinkering and decisions,

Paco

Select to expand quote

PacoRaapNL said..

Actimel87 said..
Hey Paco,

did you consider gluing the window and using an O-Ring or a round Cord in a Labyrinth seal?
Also I think I would reduce the size of the "window part"to an absolut minimum. A smaller window is more robust and has less area where you can get a leaking.

Actimel, yes I thinkered about both things.
I am still in a low tinkering stage.
I first like to have the project that I can open and close again with using less parts as possible.
Replacing the tape and transparant window are not the cost if opening the box is needed due to an electronic problem.

The larger size of the transparant window lets the pressure in the box compensate like a membrane (that is my thinking) depending on depht and temperature changes.

If you glue the transparant window the glue joint might break or leak more easily, therefor Simon his 3M VHB foam tape would solve that as it is a flexible foam and so with temperature changes the whole can expand and retract giving less stress on the whole structure.
Or you use a flexible adhering butene kit which might be a bit more messy.

The top plate over the large transparant window has two functions.
To devide the pressure over the foam tape as regular as possible.
Estathic and only viewable parts are exposed reducing the possibility to reduce impact from the outside.
My thinking is it would survive as it is no brick object anymore.

For the sealing with an O-ring or 2 mm diameter slicone foam tube I already looked at Ali for it but not yet ordered.
nl.aliexpress.com/item/1005005325164575.html?spm=a2g0o.cart.0.0.506c61d7kFiHhr&mp=1&gatewayAdapt=glo2nld

However the flat 3M tape has more sealing surface as a compressed tube.I currently use 10 mm width but might opt for other test to 5mm width.

So much tinkering and decisions,

Paco

Do you think the smaller version would fit in a GoPro waterproof housing ?

Select to expand quote

rp6conrad said..
Our beloved ESP32 in the T5-epaper configuration has a total of 16MB flash memory. When I started the project, I had no clue and I choose the Default "partition" scheme with only 4MB in the Arduino IDE. The partition can only changed over usb, so not with OTA. This was the root cause for troubles when using the ESP32 2.0x library instead of the 1.0x library !
However, a interesting feature is use of the flash memory for data-storage. So, without the SD card, we could even use minimal 12 MB for our log files ! With the compact .gpy format, this stands for 15 hours data @ 10 Hz ! (or 30 hours data @ 5 Hz).
I did some tests with a lilygo T7-S3, and it seems to work just fine. No lost points @ 10Hz gpy.
My idea is now to adapt the SW for the T5-e-paper, if no SD card is found, the internal flash will be activated for data-storage.
Another idea is to use other (smaller) ESP32 boards. There is a very small Lilygo board with a small LCD-display : www.lilygo.cc/products/lilygo%C2%AE-ttgo-t-display-1-14-inch-lcd-esp32-control-board
The 16MB version goes for only 9$ !
The biggest advantage is the smaller formfactor, and off course, dead cheap hardware...


Greetings, Jan.

Nice sollution for a display-less sollution Jan. The lack of needing an sdcard in the new firmware will lower the cost if you're willing to upload the tracks more often and don't keep history on your device. I'm not sure yet if I'm willing to drop it this.

I am in favor of an easy to ready display and I love the epaper for clarity. I'm targeting wearing at my (upper/lower)arm (not wrist) the formfactor does not need to shrink. Actually I'm looking for larger batteries for longer opertation time and with the larger space required also accomodate an increased screen (2.66"). I'm working on the firmware to change the resolution based on the increased amount of pixels available. It looks quite impressive as the 2.13" looks tiny.

The Lilygo TTGO T8 that Jan is mentioning is LCD based. Although the small screen size (1.14") the resolution is about the same 135x240 (vs. 122x250 of the 2.13"epaper). It might be small to read while planing. Not sure what the power-draw would look like, I expect a lot more and might reduce operations with 25-50%.

For all users that come from wrist sollutions and would like watches, it should be possible to use the ESP32 wrist sollutions with a 1.54"screen right? nl.aliexpress.com/item/1005003473613797.html
This would require some specific screen size / resolution tinkering in the software and organize the screen layout as the formfactor is square (with the 2.66" I only change the resolution of the font).
How to make the wrist sollution waterproof is similar to the other display formfactors. I still expect it will be quite bulky. The reduced GPS signal reception from the low grip on the boom might not be OK for everybody. But not everybody is looking for highest accuracy.

This platform can serve many people.

By removing the sd card I could reduce the Size by maybe 5mm.





to use the sd card as an optional storage extension is a very nice idea!

Update van onze (Freezer(Simon) Bart en ik) versie. We noemden het de "compact". Het gebruikt een kleine batterij (1500mAh) en het m10 (1025tf) display is b74 of 213 miljard. Ik heb gelezen dat de Ali-versie (liligo-winkel) geen leds heeft, niet waar, dat doen ze wel. Het maakt niet uit of je bij Tiny Electronics of Ali bestelt. De BN zonder chip heeft leds. De BN met chip heeft geen LED's.
wat we veranderden, omdat eerdere pogingen mislukten toen we het apparaat met zeer dunne en heldere epoxy porden. Nu alles is aangesloten (soldeer) is de achterkant van het scherm (de print) volledig bedekt met flexibele kit. Monteer vervolgens de componenten in de compacte behuizing. Vervolgens wordt haar voor 95% opgevuld met normale epoxy tot net onder het scherm. Scheur het plastic van het scherm. Plaats de displayafdekking. Wacht 4-8 uur. Breng de zeer heldere en dunne epoxy aan over het scherm en de behuizing. Deze epoxy heeft een uv-blocker (normale epoxy niet). Nadat de epoxy is uitgehard, sluit u de band aan.
Het is een beetje (duh) tijdrovend om te bouwen, maar het resultaat is prachtig. Zie de foto's. Het scherm is zeer helder en ligt hoog op het oppervlak. Sluit mooi aan op de arm. Gps-chip wijst naar de lucht. Scherm een ??beetje naar je toe.

Update of our (Freezer(Simon) Bart and I) version. We called it the "compact". It uses a small battery (1500mAh) and the m10 (1025tf) display is b74 or 213 BN. I read that the Ali version (Lilygo store) doesn't have LEDs, not true, they do. It doesn't matter whether you order from Tiny Electronics or Ali. The BN without chip has LEDs. The BN with chip has no LEDs. Wat did we changed, because previous attempts failed when we poked the device with very thin and clear epoxy? Now after everything is connected (solder) the back of the screen (the PCB) is completely covered with flexible sealant. Then mount the components in the compact housing. It is then filled 95% with normal epoxy to just below the screen. Tear the plastic off the screen. Install the display cover. Wait 4-8 hours. Apply the very clear and thin epoxy over the screen and case. This epoxy has a UV blocker (normal epoxy does not). After the epoxy has hardened, connect the strap. It's a bit (duh) time consuming to build, but the result is beautiful. See the pictures. The screen is very bright and sits high on the surface. Fits nicely on the arm. GPS chip points to the sky. Screen slightly towards you.


strangly enough is de last post translated bij Google translate in Chrome ??

Bloody hell, I like the idea of two stage pouring of the epoxy, such a simple solution for the worst part of the job.

Very nice design on the arm holder.

Select to expand quote

BigBoss said..
Update of our (Freezer(Simon) Bart and I) version. We called it the "compact". It uses a small battery (1500mAh) and the m10 (1025tf) display is b74 or 213 BN. I read that the Ali version (Lilygo store) doesn't have LEDs, not true, they do. It doesn't matter whether you order from Tiny Electronics or Ali. The BN without chip has LEDs. The BN with chip has no LEDs. Wat did we changed, because previous attempts failed when we poked the device with very thin and clear epoxy? Now after everything is connected (solder) the back of the screen (the PCB) is completely covered with flexible sealant. Then mount the components in the compact housing. It is then filled 95% with normal epoxy to just below the screen. Tear the plastic off the screen. Install the display cover. Wait 4-8 hours. Apply the very clear and thin epoxy over the screen and case. This epoxy has a UV blocker (normal epoxy does not). After the epoxy has hardened, connect the strap. It's a bit (duh) time consuming to build, but the result is beautiful. See the pictures. The screen is very bright and sits high on the surface. Fits nicely on the arm. GPS chip points to the sky. Screen slightly towards you.


strangly enough is de last post translated bij Google translate in Chrome ??

Hello Bart and SImon, thanks for the info.

Would you be so kind to share the type of epoxy's used and the brand names or the location off-line or online you bought the stuff.I see there are also seperate UV blockers available to go through epoxy.
itsokay.nl/products/uv-blocker-voor-epoxyhttps://itsokay.nl/products/uv-blocker-voor-epoxy

Would a single pour then take one step less?Do you guys also bristle the expoy over the outside with a brush or is just to overfill the last 5% and let it drip over the outside of the 3D part?
Sorry for all the questions.

Paco

yes, we first pour the GPS in one go. But after a few days or weeks the GPS stopped working. We deliberately used slow epoxy because it creates less heat during curing. We think that the components are damaged, either due to elongation and shrinkage due to temperature differences or because the epoxy is too thin. We used this epoxy:

www.mrboat.nl/epoxy-giethars/epoxy-giethars-ultra-clear-75/

The flexible sealant is intended to absorb the stretch and shrinkage. The sealant also ensures that the epoxy does not reach (all) components. This is also what universal epoxy is intended for. It's a bit thicker. Yes, you can add UV blocker to epoxy, but the amounts are so small that getting a good ratio is difficult. And the ultra Clear is Ultra Clear :-).


I used this universal epoxy
www.mrboat.nl/mrboat-epoxy/universele-epoxy-langzaam/
is has a workingtime of 45 min.

I just let the epoxy drip over it. Yes, that makes a mess. If I were to order again I would take this one. hardens a lot faster.

www.mrboat.nl/epoxy-giethars/epoxy-giethars-ultra-clear-03/

Select to expand quote

BigBoss said..
yes, we first pour the GPS in one go. But after a few days or weeks the GPS stopped working. We deliberately used slow epoxy because it creates less heat during curing. We think that the components are damaged, either due to elongation and shrinkage due to temperature differences or because the epoxy is too thin. We used this epoxy:

www.mrboat.nl/epoxy-giethars/epoxy-giethars-ultra-clear-75/

The flexible sealant is intended to absorb the stretch and shrinkage. The sealant also ensures that the epoxy does not reach (all) components. This is also what universal epoxy is intended for. It's a bit thicker. Yes, you can add UV blocker to epoxy, but the amounts are so small that getting a good ratio is difficult. And the ultra Clear is Ultra Clear :-).


I used this universal epoxy
www.mrboat.nl/mrboat-epoxy/universele-epoxy-langzaam/
is has a workingtime of 45 min.

I just let the epoxy drip over it. Yes, that makes a mess. If I were to order again I would take this one. hardens a lot faster.

www.mrboat.nl/epoxy-giethars/epoxy-giethars-ultra-clear-03/

Bas thanks, still I am puzzled which epoxy in which stage you used for the last double stage pouring of the last produced ESP32-GPS.Dont talk about the past...

For all the DIY guys here who want to play copycat, am I correct?

Stage 1 solder all electronic components and dry test.

Stage 2 conformal coat the rear of the T5 E-ink display with for example nl.rs-online.com/web/p/conformal-coatings/0714462

Stage 3 place all the electronics in the 3D printed box and fix them on their correct place.

Stage 4 Make sure the top opening of the 3D printed box is spirit level and fill for 95% with www.mrboat.nl/mrboat-epoxy/universele-epoxy-langzaam/ (slow cure epoxy) so the epoxy is just below the top of the E-ink display and let it cure.

Stage 5 remove the protective film from the T5 E-ink display.

Stage 6 fill the top just above the T5 E-ink display with XXXXXX???????? epoxy and press the plastic cover on the top of the display and press gently on the plastic top cover.

Stage 7 overfill on top of the plastic top cover with the rest of the XXXXXX??????? epoxy.

Paco

Select to expand quote

PacoRaapNL said..

BigBoss said..
yes, we first pour the GPS in one go. But after a few days or weeks the GPS stopped working. We deliberately used slow epoxy because it creates less heat during curing. We think that the components are damaged, either due to elongation and shrinkage due to temperature differences or because the epoxy is too thin. We used this epoxy:

www.mrboat.nl/epoxy-giethars/epoxy-giethars-ultra-clear-75/

The flexible sealant is intended to absorb the stretch and shrinkage. The sealant also ensures that the epoxy does not reach (all) components. This is also what universal epoxy is intended for. It's a bit thicker. Yes, you can add UV blocker to epoxy, but the amounts are so small that getting a good ratio is difficult. And the ultra Clear is Ultra Clear :-).


I used this universal epoxy
www.mrboat.nl/mrboat-epoxy/universele-epoxy-langzaam/
is has a workingtime of 45 min.

I just let the epoxy drip over it. Yes, that makes a mess. If I were to order again I would take this one. hardens a lot faster.

www.mrboat.nl/epoxy-giethars/epoxy-giethars-ultra-clear-03/

Bas thanks, still I am puzzled which epoxy in which stage you used for the last double stage pouring of the last produced ESP32-GPS.Dont talk about the past...

For all the DIY guys here who want to play copycat, am I correct?

Stage 1 solder all electronic components and dry test.

Stage 2 conformal coat the rear of the T5 E-ink display with for example nl.rs-online.com/web/p/conformal-coatings/0714462

Stage 3 place all the electronics in the 3D printed box and fix them on their correct place.

Stage 4 Make sure the top opening of the 3D printed box is spirit level and fill for 95% with www.mrboat.nl/mrboat-epoxy/universele-epoxy-langzaam/ (slow cure epoxy) so the epoxy is just below the top of the E-ink display and let it cure.

Stage 5 remove the protective film from the T5 E-ink display.

Stage 6 fill the top just above the T5 E-ink display with XXXXXX???????? epoxy and press the plastic cover on the top of the display and press gently on the plastic top cover.

Stage 7 overfill on top of the plastic top cover with the rest of the XXXXXX??????? epoxy.

Paco

Hi Paco my name is Andres :-)

You asked why we didn't do it in a single pour. That's why I explained (and talked about the past).
Oke back to the present.
stage 1: yep

stage 2: yes. I first want to use "butylkit" (used on caravans). It's flexibel and removable, but had to wait to long for it to deliver so I used pattex powerfix transparant (from the Action). It's a polymer and it's acid-free. See photo (zoom in)

stage 3. Yes I don't use the inner frame. I fix the t5 With "seconden lijm /super glue" on the flatcable side (right side). It fits al very tight.

stage 4 and 5 yes no . eh almost.

stage 5 no the protective film is directly removed before curing (see stage 4) and the display cover is placed to. To pour the universal epoxy, I let it drip in at the flatcable (there is room) side (right side) and look at the left side to see what the level of the epoxy is. If the level is even with the pcb I stop, remove the protective film, install the display cover and wait 4-8 hour. On the picture I place some weight on the cover to be sure the cover is tight to the display.
stage 6 will be: remove the weight, and fill the display with the ultra clear uv resistant epoxy. because I was to impatient I accidentally used a bit to much on the display so it flows a bit on the rest of the housing hmm couldn't clean it . so I thought .. can't clean it, then the rest of the housing get epoxy to. cover it up so no dust wil destroy the end result

Select to expand quote

CarlosSainz said..

PacoRaapNL said..

Actimel87 said..
Hey Paco,

did you consider gluing the window and using an O-Ring or a round Cord in a Labyrinth seal?
Also I think I would reduce the size of the "window part"to an absolut minimum. A smaller window is more robust and has less area where you can get a leaking.

Actimel, yes I thinkered about both things.
I am still in a low tinkering stage.
I first like to have the project that I can open and close again with using less parts as possible.
Replacing the tape and transparant window are not the cost if opening the box is needed due to an electronic problem.

The larger size of the transparant window lets the pressure in the box compensate like a membrane (that is my thinking) depending on depht and temperature changes.

If you glue the transparant window the glue joint might break or leak more easily, therefor Simon his 3M VHB foam tape would solve that as it is a flexible foam and so with temperature changes the whole can expand and retract giving less stress on the whole structure.
Or you use a flexible adhering butene kit which might be a bit more messy.

The top plate over the large transparant window has two functions.
To devide the pressure over the foam tape as regular as possible.
Estathic and only viewable parts are exposed reducing the possibility to reduce impact from the outside.
My thinking is it would survive as it is no brick object anymore.

For the sealing with an O-ring or 2 mm diameter slicone foam tube I already looked at Ali for it but not yet ordered.
nl.aliexpress.com/item/1005005325164575.html?spm=a2g0o.cart.0.0.506c61d7kFiHhr&mp=1&gatewayAdapt=glo2nld

However the flat 3M tape has more sealing surface as a compressed tube.I currently use 10 mm width but might opt for other test to 5mm width.

So much tinkering and decisions,

Paco

Do you think the smaller version would fit in a GoPro waterproof housing ?

Actimel,

Here you GO-Pro... :-)


I designed an insert for this GO-Pro case.

Tight fit but it wors fine.

Has place for a round 1500 mah Lipo 14500

Equipped with a small switchbutton that can be operated with the left push button that is already in the case.


The gps is at the bottom face up.
I have no idea about sat reception this way but we will find out.
IO39 can be used by either small switch or reed switch.
For test I use both.


The wireless charge coil nicely fits in the lens cover.


Here the unit on the charge base.
Blue LED is charging.


Finally the 800 mah Lipo is also in.
5mm height.
So now I can proceed with the other versions.


@ Anders, thanks for the extra explanation with your epoxy route.

Paco

Yesterday I was doing a test run again.
The journey took about 8h30min.
The battery only dropped by 27% (LP103454).I don't know if the battery discharged is linear but to give you an hint about du duration.



I would wait with the epoxy pouring till the ESP runs without a SD card.