Select to expand quoteAlhop said..PacoRaapNL said..
Drawn different...
same outcome.
Should pad 38 be 19 like the other schematic ?
You are correct, should be GIOP19.
Paco
Select to expand quoteAlhop said..Freezer said..Alhop said..
I used easyeda. It's free and pretty easy to use. There is then a link to JLCPCB who makes the PCB and adds all the parts. Plenty of good tutorials on YouTube. Takes about 5 days to get it.
the other chip is a voltage regulator for charging.
happy to share but not sure it works yet so will try it out at some point.
Thanks for the great tip Alhop. I have been playing around with the website from JLCPCB and must admit a lot of good things are available there. The Gerber-output is easily read from Fusion360. The Bill-Of-Materials and the Placement file required for assembly requirea bit of tinkering as the standard format was not recognized.
Keeping everything as much as possible standard keeps the cost low. I just ordered 30 pieces of 12.5x8.9mm PCB only without the components. Payment with PayPall would cost US$7,52, inlcuding shipment and tax to Europe. Not bad at all! The components however are pretty small (SOT23-6, SOD323 and R0805), but still doable with tiny soldering iron.
I found that the minimum size for assembly is 10x10mm. So I adjusted the footprint to match it exactly and redid all the placement and wiring. Fixing the packages in the library was a bit of a pain in the *s, but fixed after all (including the 3D views of the packages). I was able to find all part numbers at JLCPCB-componentsearch. At placement I only needed to fix the pin1 of the SOT23-6 by rotating it 180deg and then it looked OK. Obviously having all the components assembled on the PCB is more expensive but 30 assembled PCBs US$36,91 without shipment is not too bad. Thinking about how long it would take me to solder these tiny components it will be worthwhile. I will wait with this order until I have the 1st batch without assembled components (I have the parts already in house), so might take 1-2wks ( I choose the cheapest shipment option).
Yet another tickmark on the DYI journey with the ESP-GPS. It was ~30 years ago when I designed a PCB on Ultiboard on a 486DX PC :)
This PCB should disconnect the ESP from the battery after shutting down preserving battery life. According the spec-sheet it should be 0,005uA (www.ti.com/lit/ds/symlink/tps27081a.pdf). It is using just a single pole switch (could use the current reed switch everybody is using), connection to battery and using an extra GPIO pin on the ESP as it needs to get a second trigger to shut-down (disconnecting the batter would simply skip saving the file), Jan made and extra addition in the firmware for this.
So do I just connect SW_IN to pin 39 and SW_OUT to ground and the existing Reed switch in that location will make it all work ?
Also where can I get the ESPGPS logo from. Want to add it to my board too.
first board with GPS SD REED and charging seems to work ok. Onto next iteration now. Going to try SD NAND FLASH instead of the card. No idea what I'm doing but Google is my friend and it's all working so far.
cheers
Al
Select to expand quotePacoRaapNL said..
Drawn different... same outcome.
Ok - think I understand how this works now. When using this I assume you no longer need to connect the Reed to pin 39 ?
Select to expand quoteAlhop said..PacoRaapNL said..
Drawn different... same outcome.
Ok - think I understand how this works now. When using this I assume you no longer need to connect the Reed to pin 39 ?
Correct, depending on tthe firmware version.Jan made some BETA firmware for this high side switch setup which is not available yet in the webpage.
When GPS unit is OFF.When reed or button is pressed GPIO21 puts the ON/OFF high on pin 5.
When reed or button is pressed a puls goes to the GPS to switch screens as usual.
When reed or button is pressed for longer as 3 seconds the GPS saves the log file and after a while the GPIO21 pin goes low.
Now GPS unit is OFF and the battery is electrically disconnected but only a few micro amps are still flowing (instead of miliamps) which is equal to the self discharge of the lipo.
Makes this sense?
Paco
Select to expand quotePacoRaapNL said..Alhop said..PacoRaapNL said..
Drawn different... same outcome.
Ok - think I understand how this works now. When using this I assume you no longer need to connect the Reed to pin 39 ?
Correct, depending on tthe firmware version.Jan made some BETA firmware for this high side switch setup which is not available yet in the webpage.
When GPS unit is OFF.When reed or button is pressed GPIO21 puts the ON/OFF high on pin 5.
When reed or button is pressed a puls goes to the GPS to switch screens as usual.
When reed or button is pressed for longer as 3 seconds the GPS saves the log file and after a while the GPIO21 pin goes low.
Now GPS unit is OFF and the battery is electrically disconnected but only a few micro amps are still flowing (instead of miliamps) which is equal to the self discharge of the lipo.
Makes this sense?
Paco
I have been shutting off power to my units for a while now with similar circuit using mosfets.
I modified/compiled a version of code to use GPIO22 as the power control pin.
Will the standard code moving forward have GPIO21 configured as power control or just BETA builds?
This GPIO function would be great as I could change pins and hopefully be able to use stock builds.
Does the BETA version of code with GPIO21 control still have a brief sleep before displaying session stats after shutdown initiated?
I had to use an RC circuit to sustain power through this brief sleep/wake as GPIO pins cycle during the sleep/wake up.
Glad to see others going down the same path as me and would be great if the stock code got this function.
Thanks to those who shared their power control development, happy to provide info on my setup if anyone interested but others have already shared pretty similar designs.
Colin
Select to expand quotecckiteau said..
Will the standard code moving forward have GPIO21 configured as power control or just BETA builds?
This GPIO function would be great as I could change pins and hopefully be able to use stock builds.
Colin
Yes, power control on GPIO21 will be included in next builds. Power will be cut down after the sleep-screen is ready.
Greetings, Jan.
Select to expand quoterp6conrad said..cckiteau said..
Will the standard code moving forward have GPIO21 configured as power control or just BETA builds?
This GPIO function would be great as I could change pins and hopefully be able to use stock builds.
Colin
Yes, power control on GPIO21 will be included in next builds. Power will be cut down after the sleep-screen is ready.
Greetings, Jan.
Thanks,
Looking forward to the new build.
Colin
Select to expand quotecckiteau said..PacoRaapNL said..Alhop said..PacoRaapNL said..
Drawn different... same outcome.
Ok - think I understand how this works now. When using this I assume you no longer need to connect the Reed to pin 39 ?
Correct, depending on tthe firmware version.Jan made some BETA firmware for this high side switch setup which is not available yet in the webpage.
When GPS unit is OFF.When reed or button is pressed GPIO21 puts the ON/OFF high on pin 5.
When reed or button is pressed a puls goes to the GPS to switch screens as usual.
When reed or button is pressed for longer as 3 seconds the GPS saves the log file and after a while the GPIO21 pin goes low.
Now GPS unit is OFF and the battery is electrically disconnected but only a few micro amps are still flowing (instead of miliamps) which is equal to the self discharge of the lipo.
Makes this sense?
Paco
I have been shutting off power to my units for a while now with similar circuit using mosfets.
I modified/compiled a version of code to use GPIO22 as the power control pin.
Will the standard code moving forward have GPIO21 configured as power control or just BETA builds?
This GPIO function would be great as I could change pins and hopefully be able to use stock builds.
Does the BETA version of code with GPIO21 control still have a brief sleep before displaying session stats after shutdown initiated?
I had to use an RC circuit to sustain power through this brief sleep/wake as GPIO pins cycle during the sleep/wake up.
Glad to see others going down the same path as me and would be great if the stock code got this function.
Thanks to those who shared their power control development, happy to provide info on my setup if anyone interested but others have already shared pretty similar designs.
Colin
Hello Colin, sharing knowledge in what kind of form is always nice.
Sometimes one thing leads to another.
Paco
Ther is a SW5.91 update on Github with next changes :
Added 115200 Baud support for Beitian Bx122 (connector has other pinout !!!)
Bugfix for M8@38400 bd
Added estimated logtime left (webserver+display). Calculates logtime with actual settings sample-rate, file selection and free space.
GPIO 19 Go to sleep pin PULLDOWN, alternative for GPIO39
GPIO 21 hold HIGH SIDE SWITCH, for if you want to use a high side switch to cut power
GPIO2 pulldown in deepsleep, was still floating
Added statscreen B , last run, best run, slowest run and average
Stat screen 500m / runs bugfix time reading : time 18:6 -> 18:06
Adapt FIR filter bat reading 0.1 -> 0.02, more consistent measurement
Auto calibration, full lipo 4.2V -> 4.3 V
Before Wifi start, min 3.4V
Override lipo calibration possible, only auto-calibrate with more then 5% too high reading
Deep sleep current -> 80 ?A without SD, with SD 550 ?A, this by deleting "hold_gpio_enable()" !!!
Here a screenshot from a T5 with 16MB flash :
Greetings, Jan.
Progress...Progress
Flash > slide on PCB > plug in GPS and it works..nearly
Trying a SD NAND chip instead of the SD card but that bits not working yet.
super happy with the PCB on top to make the window and have some printing on it.
next version hopefully have the SD working and High side switch also.
then it will be Flash > PCB on (solder) > plug in GPS > plug in battery > plug in charger > POT > sail !
Select to expand quoteAlhop said..
Trying a SD NAND chip instead of the SD card but that bits not working yet.
Can you give me a link from these SD NAND chips ?
Greetings, Jan.
Select to expand quoterp6conrad said..Alhop said..
Trying a SD NAND chip instead of the SD card but that bits not working yet.
Can you give me a link from these SD NAND chips ?
Greetings, Jan.
Hi Jan
says it works 'similar' to SDcard so may need a code tweak ??
www.lcsc.com/datasheet/lcsc_datasheet_2411181354_XTX-XTSDG01GWSIGA_C7429710.pdf
www.adafruit.com/product/4899
Select to expand quoteAlhop said..rp6conrad said..Alhop said..
Trying a SD NAND chip instead of the SD card but that bits not working yet.
Can you give me a link from these SD NAND chips ?
Greetings, Jan.
Hi Jan
says it works 'similar' to SDcard so may need a code tweak ??
www.lcsc.com/datasheet/lcsc_datasheet_2411181354_XTX-XTSDG01GWSIGA_C7429710.pdf
www.adafruit.com/product/4899
Jan
Connected as per pics.
I should try to connect 3V3 to Vin of the SD NAND : see learn.adafruit.com/assets/99167
If you look to the schematics, the SS has only a pullup to Vin : learn.adafruit.com/adafruit-spi-flash-sd-card/downloads
Select to expand quoterp6conrad said..
I should try to connect 3V3 to Vin of the SD NAND : see learn.adafruit.com/assets/99167
If you look to the schematics, the SS has only a pullup to Vin : learn.adafruit.com/adafruit-spi-flash-sd-card/downloads
I already have 3V going to VDD. Given the T5 boards don't like all SD cards I'm a bit worried it may just not work. Will try a few things and let you know.
Select to expand quoteAlhop said..
Progress...Progress
Flash > slide on PCB > plug in GPS and it works..nearly
Trying a SD NAND chip instead of the SD card but that bits not working yet.
super happy with the PCB on top to make the window and have some printing on it.
next version hopefully have the SD working and High side switch also.
then it will be Flash > PCB on (solder) > plug in GPS > plug in battery > plug in charger > POT > sail !
Whoa! That's the idea. You're using the backside of the pcb as the front of the unit covering all but the epaper. That's jus great. If only the pins would not go through the PCB, it would be a great finish. How much clearance between the epaper and your PCB?
Select to expand quoteFreezer said..Alhop said..
Progress...Progress
Flash > slide on PCB > plug in GPS and it works..nearly
Trying a SD NAND chip instead of the SD card but that bits not working yet.
super happy with the PCB on top to make the window and have some printing on it.
next version hopefully have the SD working and High side switch also.
then it will be Flash > PCB on (solder) > plug in GPS > plug in battery > plug in charger > POT > sail !
Whoa! That's the idea. You're using the backside of the pcb as the front of the unit covering all but the epaper. That's jus great. If only the pins would not go through the PCB, it would be a great finish. How much clearance between the epaper and your PCB?
Clearance is 0mm. I took off the plastic spacers so its sits flush. There is an option for surface mount pins so could make it 'clean' on top. Need to move the window slightly to cover up the left better too.
Rather than that I'm thinking of having a few components on top also then pot it so its all under the resin but still visible. Bit of an industiral look :) LED for charging and on / off etc.
Cant get the XTSD NAND to work though (500MB) which is pissing me right off so might have to go back to SD Card unless someone smarter than me (Jan.......) can help. Not sure if its code or hardware issue yet ??
For those who want to go the High Side Switch route.
With the IC on a small board (take care you know what you are doing) soldered and two 1/8 watt resistors you are almost there.
On the Lilygo I soldered the other 1/8 Watt resistor, the 3 x 1N4148 diodes and the wires.
With one switch (push button or Normally Open reed contact) you can control the Lilygo in a regular way.
With no High Side Switch the Lilygo uses around 80 micro amps when OFF.
This is with no SD card but 16 Mb memory IC fitted.
When switched OFF with a High Side Switch the Lilygo uses around 30 micro amps.
It doesnt matter if you use a SD card or only onboard memory.
So this is almost three times less, making the time before you need to recharge 3 times longer.
Now if you find this not enough there is an other full electronic switch option.
This option needs an extra switch button or Normally Open reed contact but gives an OFF state current use of 0,2 micro ampere.
Yes, this means that the self discharge of the lipo is likely larger then the power drain from the switch.
All used and tested with firmware version 6.00
I heard the NAND memory from Alhop also works now. So more options.....
Paco
Latest 2025 ESP-GPS test version. ready before the spring season start.
Equipped with :
Lilygo T5 213BN
16MB memory (replaced the standard 4MB)
Firmware 6.00
Low side electronic switch 0.2 microamp power drain when OFF.
Left and Right hand side push buttons.
USB-C charge port.
No reed contact needed to switch ON/OFF
The Low side switch with pushbutton / charger and two leds. Red is charging. Green is ON. Green LED uses 1.4 miliampere when the GPS is ON. If needed can be removed...
The USB-C port which should be waterproof according to the Chinese advertising on Aliexpress. We will see.
That port looks a bit like my Sena helmet headset. It has a rubber bung to keep the water out.
Select to expand quoteremery said..
That port looks a bit like my Sena helmet headset. It has a rubber bung to keep the water out.
Your words into a new design....
The aliexpress USB-C rubber covers did not fit.
I still had some TPU filament so I redesigned the outercover.
I will glue the small part of the TPU cover partly to the outer-cover so you cant loose it.
All fits snugly and with normal pressure.
Whats next in the ultimate design?
A POGO style connector unit is in preparation too.
Great design and mods
I'm concerned with the square corners on seal and cavity.
If able to would recommend re-doing with rounded corners. then making a mock up mold of the cavity a touch larger. You can then make a silicone plug with silastic to go in there.
Select to expand quotePa
The USB-C port which should be waterproof according to the Chinese advertising on Aliexpress. We will see.
With these USB C ports the issue is not the waterproofing, but the corrosion resistance toward salt water.
Select to expand quoteLISAWindsurfing said..Pa
The USB-C port which should be waterproof according to the Chinese advertising on Aliexpress. We will see.
With these USB C ports the issue is not the waterproofing, but the corrosion resistance toward salt water.
Correct. I do not sail saltwater.
For saltwater you can use wireless charging.
Or the pogo style as they use on Cronos watches.
To many options.
Like boards and rigs for specific usage.
Select to expand quotePacoRaapNL said..LISAWindsurfing said..Pa
The USB-C port which should be waterproof according to the Chinese advertising on Aliexpress. We will see.
With these USB C ports the issue is not the waterproofing, but the corrosion resistance toward salt water.
Correct. I do not sail saltwater.
For saltwater you can use wireless charging.
Or the pogo style as they use on Cronos watches.
To many options.
Like boards and rigs for specific usage.
Sure thing. By the way I already tested these USB C connectors on my LISA prototypes. It's great as it's easy to charge anywhere, but still even with clear water use only I could see some rust coming in the connector, and also be careful with sand and remaining water inside when connecting your cable ;) .
Actually the wireless charge is nice but I don't like the heat it generates, and it's not the easiest to use in the car for example. Pogopins are my choice for now.
Nico
Hello Nico,
Thanks for the letting us know the pogo style works.The wireless charger heat is depending on the charger you use.
The lower the power rating 5W or lower, the better.Takes longer to charge but generates less heat, so better in overall use.
People think that high rating charging 100W is better, but it is the opposite.
Easy wireless charging in the car is up to the design you use and the wireless charger.
I use the puck model with small radius and so it fits exactly under the main case in the center like shown in previous entries.
Paco
Select to expand quotePacoRaapNL said..
Hello Nico,
Thanks for the letting us know the pogo style works.The wireless charger heat is depending on the charger you use.
The lower the power rating 5W or lower, the better.Takes longer to charge but generates less heat, so better in overall use.
People think that high rating charging 100W is better, but it is the opposite.
Easy wireless charging in the car is up to the design you use and the wireless charger.
I use the puck model with small radius and so it fits exactly under the main case in the center like shown in previous entries.
Paco
Indeed, that why we only uses the 5 watt chargers (like the original iPhone charger, its 5v 1A). The wireless reciever can only output 800mA, so less than the charger, and the chargechip only 500mA. So faster (higer) then 500mA (=2,5 watt) is not posible. Nico has seen our design for Brendon.
But indeed we now are using pogo pins to.
Pogo style version ready too.
16 MB memory
High side switch
2000 mAh
push button inside maincase
So I made one with the S3 display (just a test box in the pic), this one is readable in sunlight much better than the T-Display, but still not as good as the E-paper display.
I've just received the T5 and I'm going to replace the flash to avoid using memory cards, can someone please tell me the right one to pick?
Is it this one reference: W25Q128JVSIQ ?
By the way the accuracy of these devices are impressive (M10 module), I did some tests and found the estimate error (that's what I use more to evaluate) in hundreds of a knot, and sat reception from 22 to 26.
thanks
