TSDZ2 mid drive with 860C, 850C or SW102 displays only -- Flexible OpenSource firmware (Casainho code only)

[NexusG]

Great work casainho !

But this 25% loss of efficiency is when field weakening is active, isn't it ? Since we aren't riding at 120rpm all the time, that's not a big problem

At which cadence the field weakening begin to work ? Also, what is the influence of battery voltage on it ?

 

[casainho]

1. But this 25% loss of efficiency is when field weakening is active, isn't it ? Since we aren't riding at 120rpm all the time, that's not a big problem

2. At which cadence the field weakening begin to work ? Also, what is the influence of battery voltage on it ?

1. yes, only it is automatically activated. If we disable field weakening on the configurations, it will never get active.

2. It get's activated only when PWM duty-cycle is max - 100% - (meaning that full battery voltage is now applied to motor coils, there is no way to increase the voltage) AND system is asking for more motor assistance power. So, the activation happen earlier when battery is more discharged and when motor is pulling higher current because the battery voltage get's lower and this voltage drop depends on the battery internal resistance, the higher, the higher the voltage drop.

Here is the current code:

    if ((ui8_g_duty_cycle >= PWM_DUTY_CYCLE_MAX) && // max voltage already applied to motor windings, enter or keep in field weakening state
        ui8_g_field_weakening_enable_state)
    {
      if (ui16_g_adc_motor_current < ui16_controller_adc_max_current)
      {
        if (ui16_counter_field_weakening_ramp_up++ >= FIELD_WEAKENING_RAMP_UP_INVERSE_STEP)
        {
          ui16_counter_field_weakening_ramp_up = 0;

          if (ui8_g_field_weakening_angle < FIELD_WEAKENING_ANGLE_MAX)
            ++ui8_g_field_weakening_angle;
        }
      }
      else if (ui16_g_adc_motor_current > ui16_controller_adc_max_current)
      {
        if (ui16_counter_field_weakening_ramp_down++ >= FIELD_WEAKENING_RAMP_DOWN_INVERSE_STEP)
        {
          ui16_counter_field_weakening_ramp_down = 0;

          if (ui8_g_field_weakening_angle)
          {
            --ui8_g_field_weakening_angle;
          }
          else
          {
            --ui8_g_duty_cycle; // exit from field weakening state
          }
        }
      }
    }

 

[NexusG]

I see, so a 15S battery will be still the best in terms of efficiency and max RPM, even with the field weakening active

But there is also a problem with PWM frequency, leading to bad efficiency at very high RPM, right ?

 

[casainho]

But there is also a problem with PWM frequency, leading to bad efficiency at very high RPM, right ?

PWM frequency is fixed and was increased by 1.21. I don't know how to calculate the efficiency loss but I think is very low. Anyway, previous, the max cadence was 92 even if you had higher voltage. 92 is kind of a low value, I think that at least 110 and PWM frequency is set to 110RPM cadence and after that a bit loss of resolution and efficiency but I think will be low.

As for cadence, after testing, maybe the best limit would be even 130 or 135 and not 120. Well, let's first test the next version...

 

[AZUR]

1. But this 25% loss of efficiency is when field weakening is active, isn't it ? Since we aren't riding at 120rpm all the time, that's not a big problem

2. At which cadence the field weakening begin to work ? Also, what is the influence of battery voltage on it ?

1. yes, only it is automatically activated. If we disable field weakening on the configurations, it will never get active.

2. It get's activated only when PWM duty-cycle is max - 100% - (meaning that full battery voltage is now applied to motor coils, there is no way to increase the voltage) AND system is asking for more motor assistance power. So, the activation happen earlier when battery is more discharged and when motor is pulling higher current because the battery voltage get's lower and this voltage drop depends on the battery internal resistance, the higher, the higher the voltage drop.

Here is the current code:

    if ((ui8_g_duty_cycle >= PWM_DUTY_CYCLE_MAX) && // max voltage already applied to motor windings, enter or keep in field weakening state
        ui8_g_field_weakening_enable_state)
    {
      if (ui16_g_adc_motor_current < ui16_controller_adc_max_current)
      {
        if (ui16_counter_field_weakening_ramp_up++ >= FIELD_WEAKENING_RAMP_UP_INVERSE_STEP)
        {
          ui16_counter_field_weakening_ramp_up = 0;

          if (ui8_g_field_weakening_angle < FIELD_WEAKENING_ANGLE_MAX)
            ++ui8_g_field_weakening_angle;
        }
      }
      else if (ui16_g_adc_motor_current > ui16_controller_adc_max_current)
      {
        if (ui16_counter_field_weakening_ramp_down++ >= FIELD_WEAKENING_RAMP_DOWN_INVERSE_STEP)
        {
          ui16_counter_field_weakening_ramp_down = 0;

          if (ui8_g_field_weakening_angle)
          {
            --ui8_g_field_weakening_angle;
          }
          else
          {
            --ui8_g_duty_cycle; // exit from field weakening state
          }
        }
      }
    }

Hi Casainho,

Good work on Field Weakening.

For a 48V motor and a 14S battery when the battery is at maximum charge (58.8V), Field W is "activated" (the effect of the FW begins to be felt), for example, at 100 rpm, but when the voltage drops to 50.4V, Field W is eventually "activated" at a lower cadence, for example 80 rpm. Does it work like that?

So in this case, for a voltage of 50.8v, we are only operating less efficiently on the engine when we exceed the cadence of 80 rpm rate. Right?

regards

 

[AZUR]

I see, so a 15S battery will be still the best in terms of efficiency and max RPM, even with the field weakening active

But there is also a problem with PWM frequency, leading to bad efficiency at very high RPM, right ?

Hi,

The 15s battery is still in testing. Maximum 63V. 15x4.2 = 63

If we apply a very high voltage to the controller, we can burn it.

After COVIDE 19 passed, I think I can have the final tests completed.

Azur

 

[casainho]

For a 48V motor and a 14S battery when the battery is at maximum charge (58.8V), Field W is "activated" (the effect of the FW begins to be felt), for example, at 100 rpm, but when the voltage drops to 50.4V, Field W is eventually "activated" at a lower cadence, for example 80 rpm. Does it work like that?

So in this case, for a voltage of 50.8v, we are only operating less efficiently on the engine when we exceed the cadence of 80 rpm rate. Right?

Well, it depends for a few variables but the easiest way is for you understand is to look on your display main screen PWM duty-cycle variable, cadence, battery current and motor power and try see how they change!!

The more current you ask from the battery / the faster the field weakening will be enabled. In practice, everytime you see the PWM duty-cycle at 100 means your motor is now limited and so on new firmware, the field weakening will start at the time.

 

[Veko]

I want to understand the efficiency trade-offs between 36V and 48V motor. I was trying to wrap my head around this subject and researched a bit so below are all my calculations, please let me know if something is incorrect!

Is it so that both motors have 4000 rpm unloaded motor speed or is it slighly different to that? If it'd be the 4000 rpm, the motors would have Kv factors of:

36V: 4000 rpm / 36 V = 111.11...
48V: 4000 rpm / 48 V = 83.333...

Does anyone have info of the actual Kv factors of the motors? Are the calculations even done with the nominal voltage or should max charge pack voltage be used instead?

With the above values, motor rpm at (nominal voltages) would be

36V motor max rpm:
36V battery = 4000 rpm
48V battery = 4000 rpm + (48V - 36V) * 111.11... = ~5300 rpm
52V battery = 4000 rpm + (52V - 36V) * 111.11... = ~5700 rpm

48V motor max rpm:
48V battery = 4000 rpm
52V battery = 4000 rpm + (52V - 48V) * 83.333... = 4300 rpm

I read from earlier in this thread that the gear reduction ratio is 41.4, so with those values cadence would be

36V motor max cadence:
36V battery = 4000 rpm / 41.8 = ~96
48V battery = 5300 rpm / 41.8 = ~128
52V battery = 5700 rpm / 41.8 = ~138

48V motor max cadence:
48V battery = 4000 rpm / 41.8 = ~96
52V battery = 4300 rpm / 41.8 = ~104


The efficiency vs rpm curve for DC motors in general should look something like this:
https://imgur.com/a/QzmV9V0

I'm interested in what cadence would you get for those >60% efficiency range limits. It seems that around 60% rotational speed you get around 60% efficiency and after ~90% of max motor rotational speed the efficiency plummets fast and around 95% motor speed you get to around 60% efficiency. These values of course would be slightly different on TSDZ2, but I guess those could be good estimates.

Based on earlier calculations, the cadence range (60% to 95% motor speed) where you achieve >60% efficiency would be following:

36V motor >60% efficiency cadence range:
36V battery (max cadence ~96) = 96*0.6 to 96 * 0.95 = ~57 to ~91
48V battery (max cadence ~128) = 128*0.6 to 128 * 0.95 = ~77 to ~121
52V battery (max cadence ~138) = 138*0.6 to 138 * 0.95 = ~83 to ~131

48V motor >60% efficiency cadence range:
48V battery (max cadence ~97) =97*0.6 to 97 * 0.95 = ~57 to ~91
52V battery (max cadence ~104) = 104*0.6 to 104 * 0.95 = ~62 to ~98

It would be fun to know the actual efficiency curve of TSDZ2 so you could optimize the cadence to match the peak efficiency. For example it does seem like the 48V motor with 52V battery reaches the peak efficiency at cadence close to 90, while with 48V battery it could be closer to 80 (depending on the SoC of the pack of course).

Do these assumptions and calculations make sense and does the motor behavior match these numbers in practice? If the Kv factor should be calculated with max charge pack voltage, the 36V motor would start to look better, but with these calculations it would seem that 36V motor is quite inefficient at lower cadences with higher voltage batteries, but does have high efficiency even in the very high cadence range. I guess the cadence range is pushed towards even more inefficiency when 52V battery pack has max charge of 58.8V. On the other hand, at the lower half of SoC, the 36V motor with 52V battery does seem to move towards a very good range of cadence for high efficiency.

Now with the field weakening it seems that 48V motor with either battery offers the benefit of higher inherent torque as well as a really good efficient range of cadence, while also getting high enough power output in the bursts of very high cadences. I'm starting to second guess my choice of motors now Either way, it seems that 36V doesn't need to utilize field weakening, but the 48V motor version gains a huge benefit for ease of use. I'm getting excited to test these things in practice when I get all the installation supplies from the postal office!

Edit: Changed gear ratio to 41.8, which affected some cadence calculations by 1-2 rpm

 

[casainho]

Now with the field weakening it seems that 48V motor with either battery offers the benefit of higher inherent torque as well as a really good efficient range of cadence, while also getting high enough power output in the bursts of very high cadences. I'm starting to second guess my choice of motors now Either way, it seems that 36V doesn't need to utilize field weakening, but the 48V motor version gains a huge benefit for ease of use. I'm getting excited to test these things in practice when I get all the installation supplies from the postal office!

Do you know a way for the system calculate the efficiency?

And 41.8 reduction ratio: https://www.electricbike.com/tsdz2-750w-mid-drive-torque-sensing/

Seems that users are saying that 36V heats faster than the 48V motor...

 

[barrettrussell]

Thanks for your tireless advancements with this software to Casainho et al.

I am wondering about the calibration of the torq. sen. I see there is an initial calibration to be done, but the link on the github is dead. And then, in addition, another way to calibrate thereafter? Confused as to whether one calibration is done via st link on computer and another on the display? why the need for two formal calibration processes?

I am running sw102 and mostly interested in this feature as the current torque appears to wane at <30% of my force - hoping to make more natural feeling with better curve and endpoints.

Thanks

 

[ezrider1199]

I'm in absolutely the same boat. I use to flash the latest firmwares but after a couple of incidents i decided to stick to one that's more tested with time. I can confirm 0.7 works ok (https://github.com/OpenSource-EBike-firmware/Color_LCD/releases/tag/860C_850C_SW102_v0.7.0). This one is the latest firmware where the throttle works OK also. I dont think the issues you described regarding initialization misfires have been totally fixed because sometimes after gear changes the motor provides barely any assist and it takes like 2-3 seconds for it to start going again. So some problems you described are still present but it may be better than the version you're on.

I'm running v0.6.5 on my 850C but it has become unpractical because it gets too many initialization misfires (where it either reads the torque sensor wrong, provides no assist, provides way too much assist on barely touching the pedals, or makes a horrible noise as it assists.) I understand this has been fixed in later releases.

So I'll have to flash a new version on the motor and the display. The thing is that I need this bicycle to be 100% functional and reliable: it's my only means of transportation and I depend on it.

With this requirement of dependability, which version should I go for? Which version is the most stable for daily use?

Thanks!

 

[Veko]

Do you know a way for the system calculate the efficiency?

And 41.8 reduction ratio: https://www.electricbike.com/tsdz2-750w-mid-drive-torque-sensing/

Seems that users are saying that 36V heats faster than the 48V motor...

I'm not expert in this area so I don't really know whether this would work, but you mentioned earlier that you have a bike trainer and you can put a constant load on the motor, so I think if you can extract the torque, battery voltage and current draw at different motor speeds, you could get a decent estimate of the efficiency curve (similar to my linked image) with for example 1 data point per 5% speed, so 20 points in total from 0% to 100% motor speed. We wouldn't know the absolute maximum efficiency, but we'd know relative efficiency at different points of the motor rpm range.

 

[Waynemarlow]

Didn't Buba do just that way back somewhere in the original thread, he had lots of figures on the FOC and had started it in V20.

 

[NexusG]

I see, so a 15S battery will be still the best in terms of efficiency and max RPM, even with the field weakening active

But there is also a problem with PWM frequency, leading to bad efficiency at very high RPM, right ?

If we apply a very high voltage to the controller, we can burn it.

Yes, the capacitors are rated at 63V, there is no margin
Fortunately, as you said before, the battery voltage decreases rapidly at his nominal value : 55.5V

Did you burn a controller during your tests ?
Which tests remain to be done ?

I want to build a 15S3P battery with 21700 Samsung 50E cells, it will be a better riding experience

 

[AZUR]

For a 48V motor and a 14S battery when the battery is at maximum charge (58.8V), Field W is "activated" (the effect of the FW begins to be felt), for example, at 100 rpm, but when the voltage drops to 50.4V, Field W is eventually "activated" at a lower cadence, for example 80 rpm. Does it work like that?

So in this case, for a voltage of 50.8v, we are only operating less efficiently on the engine when we exceed the cadence of 80 rpm rate. Right?

Well, it depends for a few variables but the easiest way is for you understand is to look on your display main screen PWM duty-cycle variable, cadence, battery current and motor power and try see how they change!!

The more current you ask from the battery / the faster the field weakening will be enabled. In practice, everytime you see the PWM duty-cycle at 100 means your motor is now limited and so on new firmware, the field weakening will start at the time.

Thanks. I will try and give some feedback.

 

[skestans]

I'm in absolutely the same boat. I use to flash the latest firmwares but after a couple of incidents i decided to stick to one that's more tested with time. I can confirm 0.7 works ok (https://github.com/OpenSource-EBike-firmware/Color_LCD/releases/tag/860C_850C_SW102_v0.7.0). This one is the latest firmware where the throttle works OK also. I dont think the issues you described regarding initialization misfires have been totally fixed because sometimes after gear changes the motor provides barely any assist and it takes like 2-3 seconds for it to start going again. So some problems you described are still present but it may be better than the version you're on.

I'm running v0.6.5 on my 850C but it has become unpractical because it gets too many initialization misfires (where it either reads the torque sensor wrong, provides no assist, provides way too much assist on barely touching the pedals, or makes a horrible noise as it assists.) I understand this has been fixed in later releases.

So I'll have to flash a new version on the motor and the display. The thing is that I need this bicycle to be 100% functional and reliable: it's my only means of transportation and I depend on it.

With this requirement of dependability, which version should I go for? Which version is the most stable for daily use?

Thanks!

Yours sound different from mine. I have the issue when turning the display on, not while riding. 0.7.0 is labeled pre-release, but it’s working well for you?

 

[ezrider1199]

I'm in absolutely the same boat. I use to flash the latest firmwares but after a couple of incidents i decided to stick to one that's more tested with time. I can confirm 0.7 works ok (https://github.com/OpenSource-EBike-firmware/Color_LCD/releases/tag/860C_850C_SW102_v0.7.0). This one is the latest firmware where the throttle works OK also. I dont think the issues you described regarding initialization misfires have been totally fixed because sometimes after gear changes the motor provides barely any assist and it takes like 2-3 seconds for it to start going again. So some problems you described are still present but it may be better than the version you're on.

I'm running v0.6.5 on my 850C but it has become unpractical because it gets too many initialization misfires (where it either reads the torque sensor wrong, provides no assist, provides way too much assist on barely touching the pedals, or makes a horrible noise as it assists.) I understand this has been fixed in later releases.

So I'll have to flash a new version on the motor and the display. The thing is that I need this bicycle to be 100% functional and reliable: it's my only means of transportation and I depend on it.

With this requirement of dependability, which version should I go for? Which version is the most stable for daily use?

Thanks!

Yours sound different from mine. I have the issue when turning the display on, not while riding. 0.7.0 is labeled pre-release, but it’s working well for you?

I think 0.7 is like 0.6.5 but with more fixes so it might be worth trying. The versions after 0.7 introduce new features and changes.

edit: plus this is the version with responsive buttons (no more lag in UI and assist changes) - a very important update.

 

[AZUR]

Hi,

On the screens below, taken from a Garmin 1000 edge, you can see the human power and cadence, sent from TurboLevo to Garmin. The data refer to an activity carried out on 15 March.

Thanks for sharing. I think the only metric missing on our firmware is the consumption in Wh/km, which I plan to implement soon with the Field Weakening feature.

I see you use the Garmin 1000 edge and that costs around 550€. I never looked at this type of devices and for what I see from Garmin, I guess it has a big degree of customization of data fields and custom apps (although I guess is totally locked to install other apps that does not follow the Garmin SDK). I know there commercial bikes that share data over ANT+ and that data can be seen on that Garmin edge devices as also on Garmin watches, and, bike can even be controlled on this devices, like changing the assist level.

I would say a big points on Edge are (at least) small size, light, and long battery range. Also the navigation and connection to the external sensors by ANT+.

If the Garmin Edge costs +500€, I wounder if a good investment would be to make a DIY display in 150€ parts costs, for ebikes that can supply power to the display from the battery. Would be a Raspberry PI + display + GPS + USB ANT+ stick + 3D print enclosure.
For software, the Android would make a very fast development platform and probably use the popular navigation OpenSource Android app OsmAnd -- this would be a BIG project!!!

Hi Casainho,

Some ideas:

The fields I would like to have are:

Short statistics:

- Elapsed time: 1:29:23
- Total KM: 28.46 km
- Battery Consumed: 24% (106 Wh)
- Wh ride: 272Wh
- Wh Biker: 61.0% (166Wh)
- Wh Battery: 39.0% (106Wh)
- Total Wh engine: (104 Wh)
- Biker W: 111W
- Biker Normalized power: 147W

In addition to these fields I would also like to have the normalized power of the Biker.
Normalized Biker Power = Average Biker Power without the time the Biker is not pedaling.

It would be interesting to have a screen with only the statistics.
The statistics could be reset all manually, at once.

At the end of the activity, you could take a picture of the statistics screen.



Is it to much?

At least insert the fields that don't exist yet.

Thamks

 

[AZUR]

I see, so a 15S battery will be still the best in terms of efficiency and max RPM, even with the field weakening active

But there is also a problem with PWM frequency, leading to bad efficiency at very high RPM, right ?

If we apply a very high voltage to the controller, we can burn it.

Yes, the capacitors are rated at 63V, there is no margin
Fortunately, as you said before, the battery voltage decreases rapidly at his nominal value : 55.5V

Did you burn a controller during your tests ?
Which tests remain to be done ?

I want to build a 15S3P battery with 21700 Samsung 50E cells, it will be a better riding experience

Hi NexusG,

I've done 500km with a 15S1P battery. But in early March the controller broke down. I don't know what caused the malfunction. I don't know if it was the controller that failed or if it was the battery that caused the controller to fail.
I already bought a new controller but I haven't installed it yet because I can't leave my home because of the COVID-19 virus.
Whenever I have to charge the battery, I disconnect it from the engine. When reconnecting the battery, a large spark occurs because the battery is fully charged.
A hypothesis for the source of the fault may be the high voltage of the battery, 63V causing a spike in the controller and over time the sparks will wear the controller.
For this reason I will start to disconnect the battery before connecting it to the engine. This way, the spark does not occur during connection.

In the figures below the first figure shows the locked battery and the second figure shows the battery ready to work.





What I'm going to do is continue to use the engine, taking these precautions and checking that there are no more problems.

The problem happened in March, when a new 15S2P battery reached full discharge.

Anyway the TSDZ2 works very well with the 63V, 15S battery.

I hope this Help.

 

[AZUR]

Hi,

I have a friend who has a Specialized Turbo Levo (TL).

It uses a Gramin Edge 1030 GPS. The 1030 has a specific application to connect to the Turbo Levo.

TL connects to Garmin via ANT + and sends the instantaneous engine power and human instantaneous power data to the 1030.

Garmin records data for each activity.

At the end, or during the activity, it is possible to know the average engine power and the average human power.

I also use a Garmin and I have no experience with similar applications for Android smartphone.

I have the following questions to ask the forum.

1 - Which Android applications allow you to show and record activity data. Data such as speed, power, distance, cadence, etc? And also record GPS locations?

2 - What is the format of the activity files? Garmin .FIT file format? GPX format?

3 - Which android applications will allow to develop applications to connect other peripherals, such as cadence sensors, power sensors? etc. Or connect to TSDZ2 firmware? using bluetooth.

Thanks

I moved from Fitbit watch to Garmin watch recently. Garmin watch broads cast user HR over ANT+ and I would like to have my HR on the 860C display while riding. I think TSDZ2 ESP32 / NRF52 BLE and ANT+ project is a good idea, it can be seen like a gateway in the middle of an ebike motor and display, and that connect to external sensors as also a mobile app. Also the TSDZ2 ESP32 Android app seems to have active development.

Note that:
- seems that there are many bicycle sensors that are only ANT+ and not Bluetooth
- SW102 display does not support ANT, only Bluetooth
- ESP32 board does not support ANT
- Arduino Nano 33 BLE supports ANT+ and Bluetooth and also includes an accelerometer, a gyroscope, and a magnetometer

I think the Arduino Nano 33 BLE is a better choice because it would support the ANT+ sensors and the accelerometer, a gyroscope that could be used to detect hills and increase or decrease motor assist level.

• nRF52840 Flash 1MB and RAM 256KB, 3.3V
• BLE and ANT+
• accelerometer, a gyroscope, and a magnetometer with 3-axis resolution each
• 45mm x 18mm

Casainho,
Yes I think also the Arduino Nano 33 BLE is a good solution.

As I already told, TurboLevo has now a solution to detect hills and descends and increase or decrease motor assist level. Also its maibe used in the eMTB mode.

Regards

 

[NexusG]

Whenever I have to charge the battery, I disconnect it from the engine. When reconnecting the battery, a large spark occurs because the battery is fully charged.
A hypothesis for the source of the fault may be the high voltage of the battery, 63V causing a spike in the controller and over time the sparks will wear the controller.
For this reason I will start to disconnect the battery before connecting it to the engine. This way, the spark does not occur during connection.

The spark is caused by current spike when the capacitors of the controller are charged up
From what I understood, you're using the BMS as a switch after physically connecting the battery to the motor ?
So there is no more apparent spark, but the current spike is still here and maybe damaging the controller

We could use an anti-spark connector, like the XT90S, which contains a precharge resistor
I don't know if the current spike was the real problem here but with a 63V battery, an anti-spark connector might be beneficial

Anyway thanks for your feedback, it is helpful
Regards

 

[cityboy19]

Hello everyone
I have a lighting problem.
I recently bought a front lamp. foto

(6V DC)
I connected to the cables from my TSDZ2. It started to light up. I think great ...
I got on the bike and started pedaling and it turned out that the engine was not working. After switching off the lamp, the engine works ... :-(
I measured the current on the lamp = 288mA

I also have a second lamp. Here, when connected, everything works (lights up and the engine works). Current = 277mA

Where is the problem?

 

[casainho]

Hello everyone
I have a lighting problem.
I recently bought a front lamp. foto

(6V DC)
I connected to the cables from my TSDZ2. It started to light up. I think great ...
I got on the bike and started pedaling and it turned out that the engine was not working. After switching off the lamp, the engine works ... :-(
I measured the current on the lamp = 288mA

I also have a second lamp. Here, when connected, everything works (lights up and the engine works). Current = 277mA

Where is the problem?

Please see how much battery current you see when the lights are on. Then I need to make an option on the firmware for that.

 

[cityboy19]

Hello everyone
I have a lighting problem.
I recently bought a front lamp. foto

(6V DC)
I connected to the cables from my TSDZ2. It started to light up. I think great ...
I got on the bike and started pedaling and it turned out that the engine was not working. After switching off the lamp, the engine works ... :-(
I measured the current on the lamp = 288mA

I also have a second lamp. Here, when connected, everything works (lights up and the engine works). Current = 277mA

Where is the problem?

Please see how much battery current you see when the lights are on. Then I need to make an option on the firmware for that.

Software v 0.8 TSDZ2 48V

Without the light on

With the light on

Now hit !!! During pedaling

20A motor current

 

[AZUR]

Whenever I have to charge the battery, I disconnect it from the engine. When reconnecting the battery, a large spark occurs because the battery is fully charged.
A hypothesis for the source of the fault may be the high voltage of the battery, 63V causing a spike in the controller and over time the sparks will wear the controller.
For this reason I will start to disconnect the battery before connecting it to the engine. This way, the spark does not occur during connection.

The spark is caused by current spike when the capacitors of the controller are charged up
From what I understood, you're using the BMS as a switch after physically connecting the battery to the motor ?
So there is no more apparent spark, but the current spike is still here and maybe damaging the controller

We could use an anti-spark connector, like the XT90S, which contains a precharge resistor
I don't know if the current spike was the real problem here but with a 63V battery, an anti-spark connector might be beneficial

Anyway thanks for your feedback, it is helpful
Regards

Yes. I am using BMS as a switch. You are correct, the spark can occur, but we don't see it.
I had previously thought about it.
The XT90S anti-spark is a good solution. Battery manufacturer Luna Cycle sells the XT90S for use with the batteries they manufacture.
The XT90S is perhaps a better solution.
Also, whenever possible, I will start charging the battery on the bike without disconnecting it from the engine. So I avoid the spark.

Thank you for your help.

These two videos that I present below show how the XT90S works.

[youtube]94FEjXYkgyY[/youtube]

[youtube]KX89FhDJuoU[/youtube]