TSDZ2 wireless mid drive motor (like Specialized Turbo Levo) -- Flexible OpenSource firmware

[casainho]

See the project page for the most updated information: https://github.com/OpenSource-EBike-firmware/TSDZ2_wireless

---

The inspiration for this project is the famous Specialized Turbo Levo Ebike.

There is a standard for a wireless ebike: ANT+ Light Electric Vehicle (LEV). As a comparison, you can connect any USB mouse to your computer and it will just work. If you have an ebike that implements the wireless ANT+ LEV, you can connect the many available wireless GPS cycling computers displays and they will just work. That is what happens with the Specialized Turbo Levo Ebike. And there are many wireless sensors for cycling, from pedal power meters, heart rate bands up to tire pressure sensors -- see this good example on Turbo Levo:



Advantages of wireless Ebike ANT+ LEV

1. Cheap and less attractive to thieves: no display needed, or alternatively you can use your smart watch or easily removables GPS cycling computer and remote buttons.

2. More robust ebike, less risk to have a no working ebike after a fall: wireless means less risk to break the display cables (really important for when you are far from home).

3. Freedom to choose different displays and sensors: many available on the market, starting at 62€ and up to 600€. Also is easily to DIY with Arduino and a 3D printer.

4. Advanced features - map navigation, cycling power (and other advanced fitness metrics), heart rate, tire pressure sensors, etc: these advanced features are provided by the GPS cycling computer, like Garmin Edge models.


What is needed to have TSDZ2 mid drive ebike to be fully wireless like Tubo Levo

(Note that this are planning ideas as no one did develop yet such system as OpenSource)

1. Connect a small electronic board on the TSDZ2 display connector. This board will implement the wireless ANT+ LEV + Bluetooth and there will be no more a display connected by wires to TSDZ2.
This board may cost only 20€ (and this price for DIY, it will be very simple to source and DIY).

Example: this boards (but without the red board) from @mspider65 on the TSDZ2-ESP32 project:


2. You can not use any display or use a display GPS cycling computer:

2.1 No display - you will have to use, at least once, your mobile phone for initial configurations. You will be able to use it also as a display to see and control TSDZ2 motor.

2.2 Attach to the handle bars a wireless GPS cycling computer (note that you can use as alternative a smart watch like the Garmin) - many available on the market, starting at 62€ and increase up to 150, 250, 400 or 600€.

Ebike app running on Edge GPS cycling computer:


3. Attach to the handle bars a wireless buttons remote control - this are easily removable and cost 55€ (38€ the remote + 16€ the support). Alternatively, will be possible to wire a buttons and a throttle to the wireless board, just like the brake sensors.



Note that will be possible to do DIY displays, remote buttons, sensors, etc, using cheap Arduino board that supports ANT+ LEV + Bluetooth. The plastic enclosures can be 3d printed.

See more notes about this project:

So do you use your GPS cycling computer also on your eMtb with TSDZ2? If so, had you wish to have TSDZ2 sending data as a field and control it from your GPS cycling computer and avoid the TSDZ2 display?
And which TSDZ2 display are you using?

Yes i use my edge 1000 also with eMtb with TSDZ2. Yes i wish that comunicate TSDZ2 with Garmin Edge 1000. It would be great for me if i get Data from TSDZ2 and Control it from Edge 1000. I use 860C Display. I have Powermeter on Race Bycicle. Its would be great if i get human power from TSDZ2 on my Garmin cycling computer.

I see. So on your handle bar you have 860C display and Edge 1000, which both have similar sizes and are big ones!!

I see you would have advantage in not have to buy 860C display and reuse the same display for both road and MTB bicycles, with the most important that is to have the same fitness metrics on both bicycles because you would be sharing the same Garmin device that measures them - yes, we can not expect the TSDZ2 torque sensor to be so good at measuring pedal power as a commercial pedals but I think it will anyway output some good approximation.

The way I see this happening for fully wireless TSDZ2 ebike (needs the implementation of a small electronic board and firmware, as also mobile app) :
1. Bluetooth mobile app for configurations
(firmware and mobile app development): use at least once Bluetooth mobile app to configure the TSDZ2 ebike configurations: the same as actual configurations on the display.

2. ANT+ Light electric vehicles (EBike) (firmware development): implements ANT+ Light electric vehicles (LEVs) profile that sends typical ebike data as also permits typical control of the ebike. Note that you need to have an ANT+ remote control on your handle bar and paired with your Garmin Edge to be able to control the ebike, like changing the assist level.



3. ANT+ Bicycle Power sensor (firmware development): implements ANT+ Bicycle Power profile that transmits Power-only data, Torque, Cadence and Speed data.



COSTS

Current costs for TSDZ2 with 860C display: 80€.

Option for MTB medium cost
1. Wireless TSDZ2 board -- 20€ and very easy to DIY: Bluetooth and ANT+, small black box and to be connected on the TSDZ2 display connector.
2. Wireless GPS cycling computer Garmin Edge 130 -- 150€.
3. Wireless Remote control -- 55€ (38€ the remote + 16€ the support).

TOTAL cost: +145€ compared to use 860C display. Only use this wireless option instead of regular 860C if user values the robustness of wireless (no cables to break on a fall) and the extra metrics given by the GPS cycling computer.
Note that if user may already have 2. and 3. and then the cost would be half then installing 860C display.

Option for most cheap city ebike, MTB, etc
1. Wireless TSDZ2 board -- 20€ and very easy to DIY: Bluetooth and ANT+, small black box and to be connected on the TSDZ2 display connector.
2. Wireless Remote control -- 55€ (38€ the remote + 16€ the support).

TOTAL cost: -5€ compared to use 860C display. Only use this wireless option instead of regular 860C if user values the robustness of wireless (no cables to break on a fall) and probably a clean installation (good to avoid attraction to thieves, even because the remote can be easily removed so not way to start the motor).

The most small and cheap GPS display would be the Edge 130 model, that seems to be relatively small and have a very good reading at sunlight. It also have good specific mounts for road bikes and MTB:



We can take much more value from Edge 130 than 860C display although the price is not much more:


There are on the market more options for ANT+ LEV GPS cycling computers (all of them will work!!) and there is one that costs 62€ - Coachsmart LEV:

[Brlowe]

Thanks for the link to the new thread. I will be following this closely and will build one of these once you have a working model and an app for Garmin or iphone or both.
Do you plan on this having the motor temp option for display?

[mallesepp]

Thanks for the link to the new thread. I will be following this closely and will build one of these once you have a working model and an app for Garmin or iphone or both.
Do you plan on this having the motor temp option for display?

I am also interested on this build. I will do my best to bring forward this. I think it is a great benifit by comparsion to existing professional produced ebikes.

[casainho]

I found that the most cheap Garmin Edge 130 does not support ANT+ LEV as also does not support CIQ apps, unfortunately, because it is important to have cheap display.

The next version of Edge to support ANT+ LEV is the Edge 530 that costs 250€.

I found that remote buttons are very expensive, the one from Garmin specific for Ebike costs 80€.

Developing the ANT+ and Bluetooth board firmware

And for developing ANT+, I don´t find almost nothing on web. Seems there are no one DIY their own sensors...

At least there are small and cheap NRF52840 boards on ebay, as also Arduino with NRF52840.

I was looking at NRF52 SDK and I fond there and example of firmware that runs both Bluetooth and ANT+, this is what what we need!!
Here is the example on SDK: nRF5_SDK_12.3.0_d7731ad/examples/multiprotocol/ble_ant_app_hrm and the notes on main.c file:

Code: Select all

 * The application uses the BLE Heart Rate Service (and also the Device Information
 * services), and the ANT HRM RX profile.
 *
 * It will open a receive channel which will connect to an ANT HRM TX profile device when the
 * application starts. The received data will be propagated to a BLE central through the
 * BLE Heart Rate Service.
 *
 * The ANT HRM TX profile device simulator SDK application
 * (Board\pca10003\ant\ant_hrm\hrm_tx_buttons) can be used as a peer ANT device. By changing
 * ANT_HRMRX_NETWORK_KEY to the ANT+ Network Key, the application will instead be able to connect to
 * an ANT heart rate belt.

Developing our custom Garmin app

On the Garmin side, if we want to develop a custom app, there is a sample on Garmin SDK that communicates with Moxy muscle oxigen sensor: connectiq-sdk-lin-3.1.8/samples/MO2Display/source



There is also EBike field on Garmin app store that seems communicate with ebikes but seems to be reading only information and not control ebike assist level. Also this app is not OpenSource so we can not use it to learn for our needs.

[AZUR]

Casainho,

This is an outstanding solution.

If Casainho manages to do this development he will take the TSDZ2 FW to the Next Level.

It is amazing that he has the motivation to do this.

Watch this 2016 video from TurboLevo.

Attention in the video in the part between 1m12s and 1m21s where you can see the Garmin Remote and Garmin Gps.



Aspects that I think are very important:

1 - Possibility of having the TSDZ2 information on the Garmin GPS (Human Power, Candence, Assit level, Battery level) through Ant+.

2 - Possibility to connect to the smartphone via bluetooth

3 - Possibility of configuring the TSDZ2 via smartphone

4 - Possibility of having a wireless remote to control the assistance levels.

I make a proposal that for the first stage to make the connection through ANT+ with Garmin, keeping the Display 860 with the buttons.
Eventually, in this first phase, you can, simultaneously / parallel, change the assistance level with the Garmin Remote.

In a second phase, remove the 860C and leave everything in wireless mode including the TSDZ2 configuration, via the smartphone, and the Garmin Remote.

I am afraid that the Garmin Wireless Remote does not act very quickly in changing the level. On the other hand, you cannot have Walk Assist because the button must always be pressed. WA is important in MTB and Backpacking.

We neeed to test and see how remote works.

The Turbolevo video is already from 2016 and not many people are currently using the Garmin Remote. And I know who uses Turbolevo with a Garmin 1000 and never used the Garmin remote.

However, having everything wireless is a big goal and I support 150%.

Casainho, Let's take the TSDZ2 to the Next Level!

Thank you

[AZUR]

Hi,

And, also as in the TurboLevo, if you need you can use a 10€ 850C Switch.

switch para 850C.jpg (53.16 KiB) Viewed 2639 times

https://www.aliexpress.com/item/4000849 ... WUezC&mp=1

Also the TurboLevo Mission Control Smartphone App, does the log of the Bike Ride.

The Smatphone uses the internal GPS to calculate the KM/h, Avg Speed, Ascent Elevation, etc, and gets the Heart Rate directly from the cardiac sensor.
----------------------------------------------------------------------------------------------
BLEvo v3.5.6 iOS

Statistical data for my ride Sun, 22 Mar 2020 12:28:31 WET

Short statistics:

- Elapsed time: 1:29:23
- KM total: 28.46 km
- Speed average: 19.1 km/h
- Ascent total: +535 m
- Battery Consumed: 24% (106 Wh)
- Consumption average: 3.72 Wh/km
- Wh ride: 272Wh
- Wh Biker: 61.0% (166Wh)
- Wh Battery: 39.0% (106Wh)

------------------
Full statistics:

Firmware: 4.22.3
User Settings:
"Advanced user": 20/50/100 CM 80

Assistance average: 20.0%
- ECO: 20.0%

Ride Time:
- Start time: Sun, 22 Mar 2020 12:28:31 WET
- Stop time: Sun, 22 Mar 2020 14:07:31 WET
- Elapsed time:1:29:23
- ECO: 1:28:49 (99.4%)
- TURBO: 0:00:32 (0.6%)

Battery:
- Start: 99% (429 Wh)
- End: 75% (323 Wh)
- Consumed: 24% (106 Wh)
- ECO: 24.0% (106 Wh)
- TURBO: 0.0% (0 Wh)

Consumption average: 3.72 Wh/km
- ECO: 3.73 Wh/km
- TURBO: 0.00 Wh/km

Battery Temperature:
- Min: 16°C
- Max: 22°C
- Average: 19°C

Motor Temperature:
- Min: 16°C
- Max: 36°C
- Average: 31°C

KM Total: 28.46 km
- ECO: 28.44 (99.9%)
- TURBO: 0.02 (0.1%)

KM Total with assistance: 22.69/28.46 km (79.7 %)
- ECO: 22.69/28.44 km (79.8%)
- TURBO: 0.00/0.02 km (0.0%)

Speed average: 19.1 km/h
- ECO: 19.2 km/h
- TURBO: 2.0 km/h

Speed Max: 49.5 km/h
- ECO: 49.5 km/h (13:08:58 - km 11.11)
- TURBO: 2.1 km/h (12:28:47 - km 0.01)

Cadence average: 65 rpm
- ECO: 65 rpm

Cadence Max: 95 rpm
- ECO: 95 rpm (12:46:41 - km 3.67)

Heart rate average: 123 bpm
- ECO: 123 bpm
- TURBO: 102 bpm

Heart rate Max: 142 bpm
- ECO: 142 bpm (13:07:23 - km 10.37)
- TURBO: 102 bpm

Kcal consumed: 652 Kcal
- ECO: 652 Kcal

Biker power Max: 519 W
- ECO: 519 W (13:11:11 - km 11.66)

Total Wh Biker: 166 Wh
- ECO: 166 Wh (100.0 %)
- TURBO: 0 Wh (0.0 %)

Motor power Max: 111 W
- ECO: 111 W (13:38:14 - km 20.21)

Total Wh motor: 104 Wh
- ECO: 104 Wh (100.0 %)
- TURBO: 0 Wh (0.0 %)

Max Altitude: 68 m
Min Altitude: 2

Ascent total: +535 m
- ECO: 535 m (100.0 %)
- TURBO: 0 m (0.0 %)

Descent total: -527 m
- ECO: 527 m (100.0 %)
- TURBO: 0 m (0.0 %)
----------------------------------------------------------------------

[raw]

I am afraid that the Garmin Wireless Remote does not act very quickly in changing the level. On the other hand, you cannot have Walk Assist because the button must always be pressed. WA is important in MTB and Backpacking.

We neeed to test and see how remote works.

why not use a cheap bluetooth bike button?

it could be paired directly to the wireless extension.

@Casainho the nRF52840 board arrived today. It looks like we need the S340 softdevice for concurrent BLE+ANT support which needs registration and manual activation for the download. I hope our usecase counts as "Evaluation Use".
i dont think we are allowed to redistribute the softdevice, so every user may have to register and wait for activation for their own
if the project reaches a working state, we may ask for an open-source exception, as we are not evaluating for commercial purposes. The non-ANT softdevices (like S140) allow redistributon. maybe it is easy to have two builds, one BLE only with S140 and one with BLE+ANT for ANT users.

[casainho]

@Casainho the nRF52840 board arrived today.

That was fast!!

It looks like we need the S340 softdevice for concurrent BLE+ANT support which needs registration and manual activation for the download. I hope our usecase counts as "Evaluation Use".

So they say:

Evaluation Use:
• Not to be used for any commercial or revenue-generating purpose.

I will not be selling anything and not having any revenue-generation purpose, so, I have not problem with it.

I hope we can distribute our firmware final HEX file and put a note on download, explaining that users should not use for any commercial or revenue-generating purpose, otherwise they should license with ANT - this will be the case of current shops selling TSDZ2 with the OpenSource firmware installed.

[casainho]

Aspects that I think are very important:

1 - Possibility of having the TSDZ2 information on the Garmin GPS (Human Power, Candence, Assit level, Battery level) through Ant+.

2 - Possibility to connect to the smartphone via bluetooth

3 - Possibility of configuring the TSDZ2 via smartphone

4 - Possibility of having a wireless remote to control the assistance levels.

1. ok, and should not be the most complex part of the project.
2. well, I think the need for smartphone is only for configurations. The issue is that some users may only have an iphone and that is a big problem as there are few iphone developers, even because there are high costs to develop for iphone while for Android is free and there are a lot of free resources for Android development. So, maybe the best thing is to develop an app for Garmin that will specifically make the configurations - this would mean the display devices must be: Edge 530, 830 and 1030.

I am afraid that the Garmin Wireless Remote does not act very quickly in changing the level. On the other hand, you cannot have Walk Assist because the button must always be pressed. WA is important in MTB and Backpacking.

We neeed to test and see how remote works.

I wish to develop my own button and 3D print it, using the same wireless board, and have 7 buttons:
- 4 for control TSDZ2
- 3 for control the Edge
[/quote]

[casainho]

@Casainho the nRF52840 board arrived today.

My suggestion is for you to reuse the project of the SW102, because it uses an NRF51 microcontroller and has a good Makefile to deal with softdevice, etc. The current version implements UART over BLE with pairing - the code for the cycling profiles are not active but I think they should work simultaneous if you enable the code again.

This board should be like a SW102 display but without the display itself.

And for ANT+ and Bluetooth, then follow the Nordic example of SDK that has both Bluetooth and ANT+ working simultaneous.

Maybe the easiest way to start, would be to read an ANT+ HR band/watch sensor (that are cheap and maybe you even have one) and just bridge the motor communications to your current display AND overwrite the motor temperature value byte with the HR data - then you can see your heart rate on your 860 display, even plot on the graph -- I know this seems out of the topic of this project BUT this would be a good way to expand 860C and SW102 displays, to add support to some ANT+ sensors like heart rate -- I would love to have heart rate, I want to train for heart rate zones and I have heart rate always on my pulse watch that is always broadcasting the value over ANT+.

[AZUR]

I am afraid that the Garmin Wireless Remote does not act very quickly in changing the level. On the other hand, you cannot have Walk Assist because the button must always be pressed. WA is important in MTB and Backpacking.

We neeed to test and see how remote works.

why not use a cheap bluetooth bike button?

it could be paired directly to the wireless extension.

Hi,

It is always good to have more than one remote bike button to test.
If we connect the Remote directly to the Wireless extension, it may work well. If we connect via Garmin edge there may be some delay.


But the turbolevo has three systems
1 - Buttons connected with cable
2 - Ant + wireless connection for Garmin
3 - Bluetooth connection for the smartphone

levo remote.jpg (30.65 KiB) Viewed 2553 times

levo 2 remote.jpg (49.19 KiB) Viewed 2553 times

[AZUR]

Aspects that I think are very important:

1 - Possibility of having the TSDZ2 information on the Garmin GPS (Human Power, Candence, Assit level, Battery level) through Ant+.

2 - Possibility to connect to the smartphone via bluetooth

3 - Possibility of configuring the TSDZ2 via smartphone

4 - Possibility of having a wireless remote to control the assistance levels.

1. ok, and should not be the most complex part of the project.
2. well, I think the need for smartphone is only for configurations. The issue is that some users may only have an iphone and that is a big problem as there are few iphone developers, even because there are high costs to develop for iphone while for Android is free and there are a lot of free resources for Android development. So, maybe the best thing is to develop an app for Garmin that will specifically make the configurations - this would mean the display devices must be: Edge 530, 830 and 1030.

I am afraid that the Garmin Wireless Remote does not act very quickly in changing the level. On the other hand, you cannot have Walk Assist because the button must always be pressed. WA is important in MTB and Backpacking.

We neeed to test and see how remote works.

I wish to develop my own button and 3D print it, using the same wireless board, and have 7 buttons:
- 4 for control TSDZ2
- 3 for control the Edge

[/quote]

android.jpg (88.67 KiB) Viewed 2550 times

Yes the smartphone will be used mainly to configure the TSDZ2.

But with a market share in 2020 of 71%, it is enough to develop for Android. All families have access to an android.

We are not going to do it like another project that failed because someone intended to launch a product for bikes on the market and initially launched it on IOS.
Then the money ran out, there was no money to make the App for Android and the project died.

[casainho]

It is always good to have more than one remote bike button to test.
If we connect the Remote directly to the Wireless extension, it may work well. If we connect via Garmin edge there may be some delay.

But the turbolevo has three systems
1 - Buttons connected with cable
2 - Ant + wireless connection for Garmin
3 - Bluetooth connection for the smartphone

In one side I would like to go fully wireless for robustness but on the other side there is the issue of battery power failing on the remote and then the ebike will not work at all -- well, maybe this is not the case if we diversify the possibilities as you say. I guess we will always carry at least our mobile phone, it should be able to turn on the system and change assist level.

For having a Garmin Edge always on the handle bar, for daily riding, it will not work well because it runs on his own battery and there is not in the market a base connector to which we can power from the battery -- this is a big disadvantage of this project!! Unless we let users connect the display anyway.

I think I will want an ANT+ remote and my Garmin watch, I think I will not get without battery simultaneous on both, as daily drivers. For MTB events I then can use the expensive Gairmin Edge GPS cycling computer. And yes, mobile phone will always be present but can be the first one failing, to be without battery.

So, much more redundancy and robustness than today

[orangecrumble]

Looks like a great project. I currently have sw102 build that is frustrating as it freezes if i try to calibrate torque sensor (Still think it's great). I would personally love the minimalist build as i never know when I will crash into bushes.

Also have 860c, vlcd5 +6 and XH18 builds available. would be happy to help test, i'm reasonably technically minded but biased toward mechanical rather than electronics, so please excuse me if my questions seem dumb.

the lev coachsmart o-synce looks like a great way of introducing a lot of control for little money. If a device like this could couple ANT+ would scaling back project to just a BLE module like https://hackaday.io/project/21160-hackable, allow workflow be simplified in the first instance with ability to expand to ANT+ coming later if needed?

Could standard displays work as a set of connected buttons? Not quite clean installation of wireless but everyone has at least one variation as it comes with motor kit and it could work around remote control battery life issues,

A lot of people also have outdated phones lying around after upgrades ( or know someone who has one lying around) that using them as displays would appeal from a no cost perspective. There is a lot of computing power lying about in drawers. A lot of older phones didn't have BLE so would backward compatibility of bluetooth be viable?

[raw]

Yes the smartphone will be used mainly to configure the TSDZ2.

But with a market share in 2020 of 71%, it is enough to develop for Android. All families have access to an android.

We are not going to do it like another project that failed because someone intended to launch a product for bikes on the market and initially launched it on IOS.
Then the money ran out, there was no money to make the App for Android and the project died.

First idea that comes in my mind is to write the app using Flutter. My company developed two apps in the last months without much struggle. I strongly advice against writing a native app, even android only, because maintaining compatibility with older android versions can be very time consuming and frustrating (you may have to implement one thing multiple times to cover most common android versions). Futter has its own API to interact with hardware and files and takes care of the differences between android versions and android and ios. it is new so there are some bugs, especially when using plugins, but it is still better than having to mess with older android SDK versions bugs.

@casainho: i had no time to start playing around with the dongle yet and i dont know when i will find the time.
Using the SW display source tree sounds like a good idea - it would be nice to add the wireless dongle as 4th display and keep it in the main Color_LCD repository.

[casainho]

the lev coachsmart o-synce looks like a great way of introducing a lot of control for little money. If a device like this could couple ANT+ would scaling back project to just a BLE module like https://hackaday.io/project/21160-hackable, allow workflow be simplified in the first instance with ability to expand to ANT+ coming later if needed?

There are so many possibilities...
I think the 860C has great display quality but it is expensive for me. And it is locked, can't be used on this project.

I was looking for how to build a cheap wireless display and DIY, and with small size if possible, 3D print enclosure and with battery or power from the ebike battery.

Costs
1. wireless microcontroller board: 12€
2. display 2'' IPS screen, 240×320: 10€
3. 60V to 5V converter module: 2.5€
4. 3D print enclosure and other bits: 5€
TOTAL: 30€

I like to idea to have a relative small, cheap and very flexible display - one that If I burn for some reason, I can easily repair and exchange only the damaged parts.

Wireless microcontroller board

This is always the same board based on NRF52840 microcontroller with BLE and ANT+, 1MBytes flash memory (8x more memory than the SW102). This board can be used to connect to the TSDZ2 motor, for a remote button or display.

The wireless board costs 12€ with the full flexibility for us to implement what we want:


It can be powered from 3.3V up to 5.5V, meaning directly from a lithium battery or 5V. It has also an RTC, meaning while it is powered it will keep the clock time. And this board should be very low power, I think it can be powered from a coin cell, for when building an ANT+ remote button:



Display

There is a 10€, 2'' IPS screen, 240×320 resolution that seems good, not to big and not to small: https://www.waveshare.com/product/displ ... module.htm





60V to 5V converter

The 60V to 5V converter module is this one, found on ebay:

[casainho]

Ideas for a DIY switch

For the switch, other than try to go with the BLE one, I would like to explore to use an ebike button. The 850C switch is fragile, easy to damage and the best switch yet I used is the one for 860C, very robust and with the advantage that has 4 buttons - but I can not find this switch on selling.

The other switch available is the VLCD5, that costs 4€ on PSWPower and can be part of the TSDZ2 kit (free). This has 4 buttons, which is good:



To make a VLCD5 button wireless:

Costs
1. wireless microcontroller board: 12€
2. VLCD5 button: 4€
3. coin cell CR2032: 1€
TOTAL: 17€

The wireless board is ready to have the VLCD5 button wires soldered. Then, the coin cell powers the board and the firmware will have the be optimized for the button to work at least 1 year with the same coin cell.

Optimimizations
- the wireless board can be cut to remove the USB connector (will be 30% smaller)
- I think a 3D print case can be built to replace the VLCD5 base part original one (the one that attach to handle bar), but keeping the buttons board and the soft keypad that also provides water proofing. This way the wireless board and the coin cell could be inside the full button case, this way making a fully wireless water prof button for about 20€ and not 80€ (seems the Garmin ANT+ specific ebike button costs like 80€ or more).

[casainho]

This display can be fully water prof if there is no connection to outside.

If we put a battery inside and a board for wireless charging, then, there will be no need for connection to outside.





https://www.epanorama.net/blog/2016/04/ ... -teardown/

I was looking for how to build a cheap wireless display and DIY, and with small size if possible, 3D print enclosure and with battery or power from the ebike battery.

Costs
1. wireless microcontroller board: 12€
2. display 2'' IPS screen, 240×320: 10€
3. 60V to 5V converter module: 2.5€
4. 3D print enclosure and other bits: 5€
TOTAL: 30€

I like to idea to have a relative small, cheap and very flexible display - one that If I burn for some reason, I can easily repair and exchange only the damaged parts.

Wireless microcontroller board

This is always the same board based on NRF52840 microcontroller with BLE and ANT+, 1MBytes flash memory (8x more memory than the SW102). This board can be used to connect to the TSDZ2 motor, for a remote button or display.

The wireless board costs 12€ with the full flexibility for us to implement what we want:


It can be powered from 3.3V up to 5.5V, meaning directly from a lithium battery or 5V. It has also an RTC, meaning while it is powered it will keep the clock time. And this board should be very low power, I think it can be powered from a coin cell, for when building an ANT+ remote button:



Display

There is a 10€, 2'' IPS screen, 240×320 resolution that seems good, not to big and not to small: https://www.waveshare.com/product/displ ... module.htm





60V to 5V converter

The 60V to 5V converter module is this one, found on ebay:

[orangecrumble]

Ideas for a DIY switch

If you want more buttons and robust how about one like this, other 2 buttons are underneath, not very discrete but it will be easy to use with gloved hands and might have space inside for the board and battery.
https://www.ebay.co.uk/itm/7-Buttons-AT ... 3770684526

[casainho]

If you want more buttons and robust how about one like this, other 2 buttons are underneath, not very discrete but it will be easy to use with gloved hands and might have space inside for the board and battery.

Thanks, I was not aware of that kind of product on ebayy. But they are clearly to big and colorful for an ebike, they may be ok for the motorcycle for which they seem to be designed.

[casainho]

Here is an good example of an cycling computer done with the exactly same microcontroller NRF52840:



That board is nice however has a small screen and to large board, so, BIG bezels and not good looking.



That is why I decided to buy this display because the microcontroller board is smaller than it, so, hopefully the bezels will be much smaller.

Another interesting thing is that they are using metal contacts for detecting touch, this is the capacitive touch detection with the microcontroller libraries.

Maybe we can take advantage of that, by using small metal screens on the 3D printed enclosure and use the metal contacts on outside as button, to make buttons but water prof.

Other idea for charging from outside, using the same metallic screws as pins:

[orangecrumble]

[/quote]
they are clearly to big and colorful for an ebike
[/quote]

How about
https://www.aliexpress.com/i/3296133760 ... 2372iPWIq9

Uses an A23 battery. It pairs by pressing either + and - together or A and B together when headset is set to pairing. There are at least two different brands out there and same brand but different model says it uses a csr1010 ic. which is widely available as modules.

I am hoping that I am helping by trying to find inexpensive parts that could be integrated to your project if you think they are suitable. It seems that your focus at the moment is to design the screen/controller.

The metal pins sound great for being waterproof but a similar layout is used in light up bath toys for children, could that lead to random self programming in the rain?(just think how often people posted about random glitches from your code, and code is easier than hardware changes if your design has flaws)

I also thought that designing the motor end of things to fit underneath the plastic cover could be both elegant and protected.

[casainho]

I am being thinking why using Garmin is so relevant (I just bought the Edge 830 that I will use for development and test/use):

- I own a Garmin Forerunner watch and it has advanced metrics for running. But I can't run everyday and for recovery, cycling and swimming are recommended.

- With a Garmin Edge cycling computer (with TSDZ2 integration), that has advanced metrics metrics for cycling, like MTB flow, grit and climb pro, then I can join cycling and running together and have the best metrics of each sport so I can assess my fitness level and health (on the same platform). The watch also have advanced metrics for swimming.

Climb pro:


MTB jumps, grit and flow:

How about

I am hoping that I am helping by trying to find inexpensive parts that could be integrated to your project if you think they are suitable. It seems that your focus at the moment is to design the screen/controller.

That one even seems to have a near profile of SW102 - seems good, I don't think we will get something better than that on the market.

I would like to do and so learn most possible about ANT+ possibilities, there is the only way to have something that can really work for us. For instance, Garmin sells remote buttons, ones specific for ebikes (I guess for ANT+ LEV), other for Edge control and other for lights control. I don't understand why I can't use the cheaper remote and I have to buy the most expensive for the ebike control - I guess it is just an ID and I would like to have full control and for that I need to have knowledge.

Also I saw that people is 3D printing the pads in flexible filament, that means it can be a fully 3D printed enclosure, with like 3 up to 6, 8, etc buttons and fully water prof. Buttons can be regular ones with wires soldered inside, very easy.

[orangecrumble]

I am being thinking why using Garmin is so relevant (I just bought the Edge 830 that I will use for development and test/use):

MTB jumps, grit and flow:

I had to look up what grit and flow were. My competitive MTB days were on fully rigid bikes in the early nineties. Metrics were obtained by the amount of abuse or praise you got from team mates. Iv'e only recently started to get interested in MTB suspension. My tsdz2 is guilty for the suspension interest as you have to to ride an ebike differently.
I am motivated toward the minimalist approach so I brake things less. I still don't record trips on strava etc. as I mostly like to have fun but still crash. My minimalist approach was applied to my first ebike build in that motor control was only by throttle. (no weak blue gear in that one, but i did brake motor mounts by doing 20kph on frozen snow)
The tsdz2 feels much more natural to cycle, especially with your firmware. I would be happy to help with development and testing on the no display or minimalist display approach (although if I had to invest in a display to assist develop I would). I assume that the NRF52840 module would have to be wired to the motor but can also contain the firmware, if so would an LED or two to help confirm successful initialisation of the motor also be sensible redundancy for more advanced options that might fail.

[raw]

i finally had 2 hours time to play with my dongle. sadly i did not get very far as i had problems flashing the firmware.
basically the same as here but nothing helped to get the usb flashing to work. dongle was recognized by OS, slowly red pulsing - nothing worked. i tested it on linux and windows. on linux i also applied the udev rules and added my user to the correct group.

my "fix" now was to add some layers of electric tape to the back of the usb plug so it fits tightly into the USB port. After a reboot(!) it started to work then on linux. strange that is was recognized and identified so i could select it on the connect application but was unable to communicate.