Derestricting bosch smart system

[Koekie]

I'm using an ESP32 Arduino (for now an M5Stack Core2) using Bluetooth Low Energy it updates a speed meter on my phone.
This is a nice gimmick but it helped to develop the system. On the phone I'm using the free app RemoteXY and that finds my Arduino and shows the following interface. The connection is not always stable...

 

[Koekie]

Maybe interesting is that the code I'm using was written by ChatGPT4. I explained that I want a speedometer and an output pulse that was slower. Code worked first time.
Later I added variable pulse width, the RemoteXY interface and some logging on the screen of my Arduino.
Please find the code attached.

 

[Gaxax]

Na nějakém německém fóru jsem četl, že poslední aktualizace leden/únor (5.xx) způsobuje chybu 524001

Nějaké zprávy o tom?

Po 1000 km mi přišla zpráva 524001. Motor běžel 90 minut v omezeném režimu.

 

[Koekie]

Po 1000 km mi přišla zpráva 524001. Motor běžel 90 minut v omezeném režimu.

What system are you using? What coil and which Arduino code?

 

[Gaxax]

Jaký systém používáte? Jaká cívka a jaký Arduino kód?

SpeedBox 1.0b

 

[stratosa]

Maybe interesting is that the code I'm using was written by ChatGPT4. I explained that I want a speedometer and an output pulse that was slower. Code worked first time.
Later I added variable pulse width, the RemoteXY interface and some logging on the screen of my Arduino.
Please find the code attached.

Because like is not enough, i have to say BRAVO to you for the entire project.

 

[-Matt]

So i have taken a punt and bought another speedbox, just gone for the standard 1.0 - less to go wrong.

Im told speedbox updated the firmware in Feb this year that helps stop the detections.
Its a shame once you have bought one, its not possible to update the firmware on the speedbox.

Will be ditching the Bosch App, and no more bike updates!

Will install it tomorrow and go for a test shred, hopefully goes better than my experiance with a 1.0B last year that lasted less than 10 miles before going crazy!

Will update on how it goes.

 

[Koekie]

I reworked my code to work on a Wemos D1 mini, that is a ESP8266 microcontroller. (see attached)
Together with a DC/DC convertor and a Bosch power cable it is now nicely integrated in my bike.

Bosch (HPP) cable: Bosch Additional Power Cable for Accessories - Smart System
DC/DC convertor: 0.75￡ 5% OFF|MH KC24 QC3.0 QC2.0 Usb DC DC Buck Converter Opladen Step Down Module 6 32V 9V 12V 24V Naar Fast Quick Charger Printplaat 5V|Geintegreerde Schakelingen| - AliExpress
Wemos D1: 1.32￡ |Wemos D1 Mini Pro V3.0 Nodemcu 4Mb/16Mb Bytes Lua Wifi Internet Van Dingen Development Board Gebaseerd ESP8266 CH340G Nodemcu V2|nodemcu 4m|d1 minimini nodemcu - AliExpress

 

[megabobra]

I reworked my code to work on a Wemos D1 mini, that is a ESP8266 microcontroller. (see attached)
Together with a DC/DC convertor and a Bosch power cable it is now nicely integrated in my bike.

Bosch (HPP) cable: Bosch Additional Power Cable for Accessories - Smart System
DC/DC convertor: 0.75￡ 5% OFF|MH KC24 QC3.0 QC2.0 Usb DC DC Buck Converter Opladen Step Down Module 6 32V 9V 12V 24V Naar Fast Quick Charger Printplaat 5V|Geintegreerde Schakelingen| - AliExpress
Wemos D1: 1.32￡ |Wemos D1 Mini Pro V3.0 Nodemcu 4Mb/16Mb Bytes Lua Wifi Internet Van Dingen Development Board Gebaseerd ESP8266 CH340G Nodemcu V2|nodemcu 4m|d1 minimini nodemcu - AliExpress

Thanks very much Koekie. PS such a great idea to debounce the hall sensor input - I didn't think of that!

I wonder how using the static delay (160ms in your case) compares to the multiplier method? Below I plotted true input speed against both output methods to see how it looks. I used a 96ms delay to give a max speed of 35kph, aligning with a 1.4x multiplier.

Interestingly the static delay gives a small curve in its output while the multiplier is linear. I'd argue the non-linear relationship between input and output is more likely to be detected, BUT when you consider the wide range of available gearing between chainring and cassette, there's no way for the motor to know that you're not simply working through the gears. So it should be ok?

How is this setup going for you?

 

[Fuzz]

Is it just me, or do others feel like these guys are electrical wizards out to haunt us?
Please, just a dummy's version, so we can make our bikes go faster...

 

[megabobra]

Is it just me, or do others feel like these guys are electrical wizards out to haunt us?
Please, just a dummy's version, so we can make our bikes go faster...

Did you check out my build tutorial earlier in this thread? Anything I can do to make it clearer/easier?

 

[Koekie]

Thanks very much Koekie. PS such a great idea to debounce the hall sensor input - I didn't think of that!

I wonder how using the static delay (160ms in your case) compares to the multiplier method? Below I plotted true input speed against both output methods to see how it looks. I used a 96ms delay to give a max speed of 35kph, aligning with a 1.4x multiplier.

View attachment 114507

Interestingly the static delay gives a small curve in its output while the multiplier is linear. I'd argue the non-linear relationship between input and output is more likely to be detected, BUT when you consider the wide range of available gearing between chainring and cassette, there's no way for the motor to know that you're not simply working through the gears. So it should be ok?

How is this setup going for you?

The 160 ms works well for me. I think the nice thing is at low speeds it is kind of x1. Also I'm not limiting myself at 35 km/h since sometimes I want to go faster that that.

I'm just starting to use it.
I hope it will not be detected by the motor, will only know that for sure after 1000 km.
Since I got that delay idea and the 160 number from an existing device (that is stated to be compatible with Bosch Smart systems) I hope that helps in that.

 

[BeBiker]

Is it just me, or do others feel like these guys are electrical wizards out to haunt us?
Please, just a dummy's version, so we can make our bikes go faster...

Don't worry, It will not take long until there's a dummy web page with only a "Pay Now" button

 

[Fuzz]

Don't worry, It will not take long until there's a dummy web page with only a "Pay Now" button

I'm looking forward to it, my brain just glitches trying to understand the DIY version. I'd hazard a guess, I'm not the only one.

 

[Sander]

I placed the resistor not to overload the transistor base. Probably better in the long run to do that. Maybe smaller resistor values will also do the job. Transistor Base Resistor Calculator

Hey Koekie, great device! I'm riding with the megabobra with no issue. I build one for a friend mine. He seems to have issues. We're playing with the position of the selfmade coil, but we also noticed that the transistor is becoming that hot that is melts the case a little bit. So soldering a 1k resistor at the base of transistor will prevent overheating? My coil is 10m 0.3 mm enameled copper with a coil diameter of 25mm.
Again, my device has no issues at all but better to dial in the resistors? In case if meltdown the same device of the friend of mine? Thx! Sander

 

[megabobra]

Hey Koekie, great device! I'm riding with the megabobra with no issue. I build one for a friend mine. He seems to have issues. We're playing with the position of the selfmade coil, but we also noticed that the transistor is becoming that hot that is melts the case a little bit. So soldering a 1k resistor at the base of transistor will prevent overheating? My coil is 10m 0.3 mm enameled copper with a coil diameter of 25mm.
Again, my device has no issues at all but better to dial in the resistors? In case if meltdown the same device of the friend of mine? Thx! Sander

Hey Sander, the thing to check here is the resistance of your coil - probably it's too low so it's pulling more current than the transistor can handle.

Ps I know what you mean with the placement, it's a pain that it's so sensitive. It's hard not knowing exactly where the sensor in the motor is, nor knowing the correct magnetic field strength that it's looking for. I tried measuring the standard magnet's field, but I only had my smartphone and some free app! Then built the coil to give a similar field strength given its location within the frame/against the motor. I bet if we could do this more accurately there would be a greater tolerance on the placement of the coil.

 

[Koekie]

Hey Koekie, great device! I'm riding with the megabobra with no issue. I build one for a friend mine. He seems to have issues. We're playing with the position of the selfmade coil, but we also noticed that the transistor is becoming that hot that is melts the case a little bit. So soldering a 1k resistor at the base of transistor will prevent overheating? My coil is 10m 0.3 mm enameled copper with a coil diameter of 25mm.
Again, my device has no issues at all but better to dial in the resistors? In case if meltdown the same device of the friend of mine? Thx! Sander

To avoid overheating the resistor should be in series with your coil. That is on the emiter of the transistor. I do not know if the resistor in on the base would cause overheating.

 

[Sander]

Hey Sander, the thing to check here is the resistance of your coil - probably it's too low so it's pulling more current than the transistor can handle.

Ps I know what you mean with the placement, it's a pain that it's so sensitive. It's hard not knowing exactly where the sensor in the motor is, nor knowing the correct magnetic field strength that it's looking for. I tried measuring the standard magnet's field, but I only had my smartphone and some free app! Then built the coil to give a similar field strength given its location within the frame/against the motor. I bet if we could do this more accurately there would be a greater tolerance on the placement of the coil.

The placement of the coil of my friends bike is now good with my megabobra. And I have no issues at all. So I'll make another arduino for my friend and I'll put a resistor in serie with the coil. The coil now has a resistance of 2.4 Ohm (R = rho copper * 10m / A) so adding 10 Ohm should do the job I guess.

To avoid overheating the resistor should be in series with your coil. That is on the emiter of the transistor. I do not know if the resistor in on the base would cause overheating.

so overloading of the transitor does not mean overheating? And adding the 1k resistor is to lengthen the lifespan of the transitor?

 

[megabobra]

The placement of the coil of my friends bike is now good with my megabobra. And I have no issues at all. So I'll make another arduino for my friend and I'll put a resistor in serie with the coil. The coil now has a resistance of 2.4 Ohm (R = rho copper * 10m / A) so adding 10 Ohm should do the job I guess.

so overloading of the transitor does not mean overheating? And adding the 1k resistor is to lengthen the lifespan of the transitor?

You have the magic touch for the placement - still takes me ages to get it settled in the right place!

Yes at 2.4Ohms you will be pulling way too much current: V=IR, @ 5v you're at 5v/2.4ohm = 2.08Amps! From memory the 2222 transistor's current limit is ~0.8A.

 

[Sander]

You have the magic touch for the placement - still takes me ages to get it settled in the right place!

Yes at 2.4Ohms you will be pulling way too much current: V=IR, @ 5v you're at 5v/2.4ohm = 2.08Amps! From memory the 2222 transistor's current limit is ~0.8A.

I guess the magnetic force will be less when placing a resistor in the coil and thus the position of the coil would be more precise?

 

[megabobra]

I guess the magnetic force will be less when placing a resistor in the coil and thus the position of the coil would be more precise?

Yes that's correct.

 

[Sander]

Yes that's correct.

Well, I'll keep my bobra as it is, without resitors. If it stops working I know what to do

 

[Sander]

Yes that's correct.

And what about a transistor that can handle more than 0.8 Amps? Less affecting the position of the coil.

 

[megabobra]

And what about a transistor that can handle more than 0.8 Amps? Less affecting the position of the coil.

With the coil, more current doesn't mean a greater magnetic field. A better way to achieve that is with more turns of the copper. So where maybe you have 10m, use 20m - you should see an increase in your magnetic field strength, increase in your resistance and decrease in current. However, you can reach a point where too many turns will create so much resistance that at 5v you won't be able to push enough current through it. In that case, you would need to increase the voltage (rather than the current).

But just note that since the motor has a hall sensor internally, you can definitely have a field which is too strong - just try putting the standard rim magnet up against the motor; you'll immediately see an error in the Flow up about magnet field strength.

So increasing the strength of the coil might not help with placement, as you'll just need to position it further away in order to still give the correct reading to the motor's sensor.

But please try and post your findings!

 

[CliffP]

Is it just me, or do others feel like these guys are electrical wizards out to haunt us?
Please, just a dummy's version, so we can make our bikes go faster...

I agree

 

[Koekie]

Indeed, you do not need much magnetic field to give a pulse to the motor.
But it might be difficult to find the right spot.

I used a simple program that generates a fixed pulse train to find the spot (see attached). Use this program and look at your speed display until it reads a fixed speed. (the wheel does not have to turn)

 

[megabobra]

Indeed, you do not need much magnetic field to give a pulse to the motor.
But it might be difficult to find the right spot.

I used a simple program that generates a fixed pulse train to find the spot (see attached). Use this program and look at your speed display until it reads a fixed speed. (the wheel does not have to turn)

Ha that's such a great idea. I've lost count of how many times I've stood holding the wheel off the ground so I can spit it an watch the display!

 

[RobbieMcManus]

So, what is possible in regard to having a US/Canadian dealer change the settings through a Teamviewer application to get to the 20MPH mark as the base setting, then not using the app anymore so the system thinks you are just running it in the US? 20MPH is all that most need and the bike is build for it. Legally they can not decline any warranty claim based on that setting.

 

[Sander23]

shop for E Bike tuning - bikespeed

Unser Shop für E Bike Tuning für die Marken Bosch, Brose, Yamaha, Panasonic und TQ. Tunen Sie Ihr E Bike auf 50km/h und mehr. Online bestellen.

www.shop.bikespeed.de

This also seems like a good solution.
I've been using their products for 7 years on 3 different bikes and has been working very well. Support is also fast and good if you have an issue.
( not yet have a smart motor)

 

[stratosa]

( not yet have a smart motor)

I agree with you, i also have been using their products for 7 years on 5 different bikes and has been working without any problem. Support is also Excellent. People do not know them well because they sell chips themselves. THEY DO HAVE FOR BOSCH SMART MOTOR , not with rim magnet.