MEGABOBRA: Bosch Smart System Derestriction for Rim Magnet - DIY Project with Support

[AlumiPro]

Like this View attachment 127864

 

[AlumiPro]

Magura rotors also have the magnet mounting hole, which I wanted to use but my swing arm didn’t allow room for a reed switch to get close enough to the rotor magnet

 

[AlumiPro]

I believe the Bosch Rim/valve magnet is made up of TWO magnets, one on each side of the valve hole. I’m finding two S poles and two N poles. See my picture showing the yellow as each magnet inside the plastic shell and the location of the S and N poles.
I also attached pictures of the aftermarket spoke magnets offered.

 

[AlumiPro]

This picture shows the yellow circled reed switch position and how the spoke magnet S & N pole passes by it. The reed switch is in between the poles.
I have now spun the rear wheel by hand(not pedals) many times while watching the speed display, I’ve done this with my other style spoke magnets as well. Moving their placement and even rotating the magnet on the spoke to aim the S & N poles in different directions. Every scenario works, creating a smooth speed. I’m able to keep a consistent speed shown on the display and increase or decreasing the speed smoothly. The only time the speed has a drastic jump is when I spin the wheel really fast; it’ll jump up to 14-16mph then drop drastically to 8-10mph(I’m told this is the Megabobra engaging, reducing the visible speed). It will also do this when the speed is gradually slowing down to around 4mph, it will then drop drastically to around 2mph then instantly spike back up to around 5mph then gradually come down to zero as the wheel slowly comes to a stop. (I’m told this is also normal as the Megabobra is “shutting off” to allow the visible speed to catch up to the actual speed)
Keep in mind this current testing is being done with the bike held off the ground in a stand with me rotating the wheel by hand(not pedals)

 

[manu.w]

Here the “signature” of a perpendicular vs // passing by magnet.
So the ring magnet is just a strong long N-S magnet, consisting of 2 separate n-s
N-S||•||N-S

New video by Manu Weyers

photos.app.goo.gl

 

[manu.w]

... The only time the speed has a drastic jump is when I spin the wheel really fast; it’ll jump up to 14-16mph then drop drastically to 8-10mph(I’m told this is the Megabobra engaging, reducing the visible speed). It will also do this when the speed is gradually slowing down to around 4mph, it will then drop drastically to around 2mph then instantly spike back up to around 5mph then gradually come down to zero as the wheel slowly comes to a stop. (I’m told this is also normal as the Megabobra is “shutting off” to allow the visible speed to catch up to the actual speed)

View attachment 127890

I had that phenomena too at high wheel revs. Repositioning the electromagnet solved that for me.
see for exact location here ...

 

[manu.w]

If they want, they can have full control over our ebike motor…, have a good relation with an authorized bosch dealer..., "dangerous" ...

eBike tuning – illegal and dangerous

Technical manipulations to the pedelec are a safety risk. If an accident occurs, it may result in high liability costs and criminal prosecution. Bosch asks all eBike riders and providers to refrain from tuning.

www.bosch-ebike.com

"The software of the Bosch eBike systems can detect whether the pedelec has been tuned by means of a sensor while riding. It then switches to emergency mode. Consequence: an error code in the display indicates manipulation and support is reduced. The eBiker can restore the original riding settings by riding the pedelec for around 90 minutes. As soon as the system returns to normal mode, however, the manipulation check will be repeated. After the third time, limp home mode can only be deactivated by a specialist dealer using the Bosch DiagnosticTool."

 

[AlumiPro]

If they want, they can have full control over our ebike motor…, have a good relation with an authorized bosch dealer..., "dangerous" ...

eBike tuning – illegal and dangerous

Technical manipulations to the pedelec are a safety risk. If an accident occurs, it may result in high liability costs and criminal prosecution. Bosch asks all eBike riders and providers to refrain from tuning.

www.bosch-ebike.com

"The software of the Bosch eBike systems can detect whether the pedelec has been tuned by means of a sensor while riding. It then switches to emergency mode. Consequence: an error code in the display indicates manipulation and support is reduced. The eBiker can restore the original riding settings by riding the pedelec for around 90 minutes. As soon as the system returns to normal mode, however, the manipulation check will be repeated. After the third time, limp home mode can only be deactivated by a specialist dealer using the Bosch DiagnosticTool."

Bosch must be referencing a different error code than the two error codes I’ve been getting because I just power off and back on to regain full assisted power out of the motor, as well as eliminating the error code on the Kiox. There hasn’t been a 90min ride time reset or emergency/limp mode.

 

[megabobra]

Bosch must be referencing a different error code than the two error codes I’ve been getting because I just power off and back on to regain full assisted power out of the motor, as well as eliminating the error code on the Kiox. There hasn’t been a 90min ride time reset or emergency/limp mode.

That's right. The messages you're being shown are just issues with the magnet pickup, not a detection error

 

[aegidius]

The error code for detected tampering used to be 504. It's now split out into several different 6-digit codes, and there are also sensor errors as well (which are not regarded as tampering and do not affect the motor long term)
A Bosch manual is at

Vebuka.com - BoschAE_HB2020_UK

BoschAE_HB2020_UK. Pages count: 124

vebuka.com

 

[AlumiPro]

Went on another 22mile ride. I tried the small round spoke magnet with no luck, it triggered the same error code within a few minutes of the ride. I moved its position a bit and it triggered the same error code again right away. I reinstalled the black plastic covered cylindrical magnet that I used on my first ride with the Megabobra installed and it seems to work the best, but I still eventually get the same 523003 error code, sometimes followed by a 523001 error code. At times I am able ride 10, 15, even close to 20miles with this magnet, error free.
I’m curious if anyone has had success with the Megabobra with the programmed 1.6 / 32mph limit?? At this point I’m wondering if the Megabobra engagement and disengagement with the programmed 32mph limit setting is to much speed fluctuation and causing these sporadic error codes when using the cylindrical black plastic covered magnet?
On this ride my max speed was 32.6mph but if lowering the Megabobra max speed to 30 or even 28mph would eleminate the error codes I’d be ok with it!

 

[aegidius]

Just a thought, when the device transitions from passing magnet pulses straight through to generating its own pulses (or back the other way) is there a delay or a repeat pulse at that point? That might be enough to trigger the 523 error.

 

[megabobra]

Just a thought, when the device transitions from passing magnet pulses straight through to generating its own pulses (or back the other way) is there a delay or a repeat pulse at that point? That might be enough to trigger the 523 error.

When it begins introducing the delayed signal, it does so gradually over five wheel revolutions. There's no possibility of repeats in this case. There is however a chance of a repeat when slowing to less than ~5kph as the delayed signal is substituted immediately for the real signal. If the delayed signal has just fired, then deferred to the real signal and the magnet just happened to pass, there's opportunity for a second pulse in close succession with the first. To be honest I'd been meaning to code a fix in for it, but while I've seen it take place (thanks to my external LEDS), I've never had an error or loss of power - I guess because the speed is already so low - so never got around to it...

On a side note, I finally passed 1000km on my megabobra over the weekend. FYI my multiplier is set to 1.4. I'm the first to admit that it's tricky to get it setup and working reliably; it's certainly not as straight forward as those without the rim magnet setup.

 

[manu.w]

Went on another 22mile ride. I tried the small round spoke magnet with no luck, it triggered the same error code within a few minutes of the ride. I moved its position a bit and it triggered the same error code again right away. I reinstalled the black plastic covered cylindrical magnet that I used on my first ride with the Megabobra installed and it seems to work the best, but I still eventually get the same 523003 error code, sometimes followed by a 523001 error code. At times I am able ride 10, 15, even close to 20miles with this magnet, error free.
I’m curious if anyone has had success with the Megabobra with the programmed 1.6 / 32mph limit?? At this point I’m wondering if the Megabobra engagement and disengagement with the programmed 32mph limit setting is to much speed fluctuation and causing these sporadic error codes when using the cylindrical black plastic covered magnet?
On this ride my max speed was 32.6mph but if lowering the Megabobra max speed to 30 or even 28mph would eleminate the error codes I’d be ok with it!

View attachment 128042 View attachment 128043

If that is the case, the magnet could not be strong enough when passing very fast against the reedswitch, not able to close…
Try to stack another coin magnet on top of one other…

 

[manu.w]

Here the “signature” of a perpendicular vs // passing by magnet.
So the ring magnet is just a strong long N-S magnet, consisting of 2 separate n-s
N-S||•||N-S

New video by Manu Weyers

photos.app.goo.gl

I am quite sure that the hall sensor in the Bosch motor is only picking up the total (xyz) magnetic field, just because of not knowing if the rim magnet n-s or s-n placing!
In green the different signature by passing the magnet n-s or s-n (see the symmetry)
So, generating just the white (positive) pulse will be enough.

 

[AlumiPro]

The magnet that works the best so far, I’ve moved in different positions while on a ride. The only thing I haven’t done is rotating the reed switch to a different position facing the magnet.

 

[megabobra]

The magnet that works the best so far, I’ve moved in different positions while on a ride. The only thing I haven’t done is rotating the reed switch to a different position facing the magnet.

Definitely give this a try. If the contacts in the reed switch are not facing the magnet they'll be pulled in the wrong direction as the magnet passes, ie not towards each other.

If you're 90deg out of alignment (worst case) the switch can behave erratically.

 

[manu.w]

All you want to know about reedswitch

How reed switches work (magnetically operated switches)

An easy explanation of how we can use reed switches triggered by magnets.

www.explainthatstuff.com

 

[manu.w]

Just did an update of my drive unit from 4.21.0 to 5.17.0
, will see how it goes
…
All went well

 

[manu.w]

Just did an update of my drive unit from 4.21.0 to
, will see how it goes
…
All went well
View attachment 128084

Apparently I had to do a second update and now to 8.16.0 !
First quick short test run and all seems ok… (I am running at factor 2, easy to calculate )

 

[manu.w]

Good to know, now all four assistant modes can be tuned

 

[AlumiPro]

Apparently I had to do a second update and now to 8.16.0 !
First quick short test run and all seems ok… (I am running at factor 2, easy to calculate )

What reed switch and electric magnet coil are you using?

 

[manu.w]

This is my setup: (earlier posts)
https://www.emtbforums.com/threads/...et-diy-project-with-support.34779/post-507814

 

[manu.w]

Must e the location of the “programmable” , hall sensor…

 

[megabobra]

Must e the location of the “programmable” , hall sensor…

View attachment 128298 View attachment 128299 View attachment 128300

ahhh ha, great find!

 

[aegidius]

Has anyone does a teardown of the gen 4? I'd like to see if there's an accelerometer on the board.

 

[aegidius]

This one is interesting:
Method and device for checking the plausibility of speed data

Here are some more interesting patents I found.

Detection of sensor manipulation of an electric bicycle
Not Bosch, but a comparison on power used in air/rolling resistance, with hill climbing contributions stripped out, against expected power used for the sensor speed. Pretty much how I believed the Bosch system to work, until I saw that first patent (and the first patent gives enough wiggle room so they still could use that method)

Method and device for checking the plausibility of a sensor signal of a single-track vehicle
This one is by Bosch, it relates to detecting the gear ratios of the bike by looking at a histogram of wheel speed to cadence ratio. With everything working, there should be strong peaks in the histogram corresponding to the different gears.

It's more for detecting unreliable sensors than for detecting manipulation, but it could be of interest if a derestrictor were to switch between generated pulses and passed-through pulses. It could also catch out mode-switched derestrictors since they would look to the algorithm like they had two different sets of gear ratios, and the patent specifically mentions detecting the case where there were more histo peaks than physical gears. It is known too that the motor will throw an error if its programmed highest gear ratio is exceeded.

 

[megabobra]

Has anyone does a teardown of the gen 4? I'd like to see if there's an accelerometer on the board.

I believe there is, as one of Bosch's instructions to commence walk mode (after holding the button down) is to tip to the bike to the side.

 

[manu.w]

Here are some more interesting patents I found.

Detection of sensor manipulation of an electric bicycle
Not Bosch, but a comparison on power used in air/rolling resistance, with hill climbing contributions stripped out, against expected power used for the sensor speed. Pretty much how I believed the Bosch system to work, until I saw that first patent (and the first patent gives enough wiggle room so they still could use that method)

Method and device for checking the plausibility of a sensor signal of a single-track vehicle
This one is by Bosch, it relates to detecting the gear ratios of the bike by looking at a histogram of wheel speed to cadence ratio. With everything working, there should be strong peaks in the histogram corresponding to the different gears.

It's more for detecting unreliable sensors than for detecting manipulation, but it could be of interest if a derestrictor were to switch between generated pulses and passed-through pulses. It could also catch out mode-switched derestrictors since they would look to the algorithm like they had two different sets of gear ratios, and the patent specifically mentions detecting the case where there were more histo peaks than physical gears. It is known too that the motor will throw an error if its programmed highest gear ratio is exceeded.

I made my own setup (MEGABOBRA: Bosch Smart System Derestriction for Rim Magnet - DIY Project with Support)
On my original firmware i managed to have a quite easy rudimentary working setup.
When upgrading the motor system firmware from 4.21.0 to 8.16.0, troubles appears.
After a loooot up and downs i got a working setup.I
would like to share my experience, but if i see that others "steel" it an sell it over 300£ ..., i am a little bit reluctant. Sad, because i would love to share my findings.
What i will share is that the reedswich should work flawlessy, the way to generate the electromagnet pulse is also much more tricky on latest firmware! I only generated a "pusitive" pulse, no need to mimic the "passing by" permanent magnet signature...

 

[AlumiPro]

I preordered the new Speedbox 1.2B which is for the smart system + Rim Magnet and has Bluetooth connectivity. Scheduled to ship the end of January.
At which time I’ll be removing my Megabobra if anyone in the US wants to buy it.