TSDZ2 OSF for all displays, VLCD5-VLCD6-XH18, LCD3, 860C-850C-SW102.
- Thread starter mbrusa
- Start date
Danger_Toast
1 µW
- Joined
- May 10, 2020
- Messages
- 3
mbrusa said:
Oh boy am I embarrased :/ thankyou lol.
Fixed the power, is the field weaking/high cadence thing about the same as 1.0, except for the fact it can't be disabled?
Mr.Flibble
10 W
- Joined
- May 5, 2020
- Messages
- 90
mbrusa said:To better understand what it depends on, you should take a test in Power assist mode with:Mr.Flibble said:I have stripped the motor, dried everything out, re-greased, physically calibrated the torque sensor and reassembled. But it is still ghost peddling after a while
As It's fine when I test it on the flat, I suspect its heat acting on a dry controller joint.
Anyone got any other ideas?
Min current ADC step = 0
Assist w / o pedal rotation = disabled
Startup boost = disabled
If it still fails, immediately check the Pedal torque delta and Pedal cadence values in the Technical menu.
I have done this (but in EMTB mode).
When it's working fine the torque sensor registers 0.
When it's ghost peddling is registers 4-6.
I did a test of 40km on LCD3.
I used simple power mode with no boost just basic. I tried to put the motor to stress by keeping it for a while between 40-50 km/h with a top speed of 53, on flat with standard 42 gear.
Motor feels definitely more powerful at higher cadence since it was not possible to keep this speed before for me. At 40-45km/h for a few km the temperatures settled around 49C at outside temperature of 15C. This looks ok to me, similar to older versions with thermal pads installed.
However after 30km the battery died completely even my watthour counter showed only 300wh consumed out of 630wh. Also the battery indicator never works fine even after a full charge is not full bar. This is a serious problem if I can not count on my battery range.
I think that the voltages are messed up in LCD3 and thus the weird behavior like Elinx said. I will do an other test to make sure.
@Mbrusa Do you have an LCD3 to check if the problem is reproduceable for you too?
I used simple power mode with no boost just basic. I tried to put the motor to stress by keeping it for a while between 40-50 km/h with a top speed of 53, on flat with standard 42 gear.
Motor feels definitely more powerful at higher cadence since it was not possible to keep this speed before for me. At 40-45km/h for a few km the temperatures settled around 49C at outside temperature of 15C. This looks ok to me, similar to older versions with thermal pads installed.
However after 30km the battery died completely even my watthour counter showed only 300wh consumed out of 630wh. Also the battery indicator never works fine even after a full charge is not full bar. This is a serious problem if I can not count on my battery range.
I think that the voltages are messed up in LCD3 and thus the weird behavior like Elinx said. I will do an other test to make sure.
@Mbrusa Do you have an LCD3 to check if the problem is reproduceable for you too?
No unfortunately I don't have an LCD3 to test.
In any case, it is a problem to be investigated.
The consumption in Wh e and the battery bars are controlled differently in the three versions.
- VLCD5/6, XH18
The calculation of the Wh used is done in the controller. I found a difference between the calculated Wh and those measured with an external instrument, this difference is compensated for in the code.
The battery bars are calculated with reference to the voltage, so they may not correspond to the SOC%, but it is possible to correct them by changing the values of Cell x / 4 or Cell x / 6 in the configurator. It is also possible to manually reset the SOC% in case you put a not fully charged battery.
- LCD3
The calculation of Wh and bars is in the LCD3 code and has remained that of the original 20beta1. I have not made any checks on the correspondence of the calculated Wh.
Also in this version, the battery bars are calculated with reference to the voltage, but if there is no correspondence with the SOC%, a correction is not possible.
However, it is possible to enter the effective capacity and have a precise SOC%. But this in all versions.
If you put a not fully charged battery, a manual SOC% reset is not possible.
- 860C
The calculation of Wh and bars is in the 860C code and the same as in v1.1.0 remained, here too I found a difference between Wh calculated and measured, but I did not foresee any compensation.
The bars of the battery are calculated with reference to SOC%, they are therefore very precise, as long as the real value of the capacity is entered and the battery always fully charged.
If a not fully charged battery is inserted, a manual reset is required, otherwise SOC% and bars are not updated. The logic of the manual reset is different from that of the stock displays.
Perhaps it would be better to standardize the versions with regard to SOC%, Wh consumed and battery bars.
Now I ask those who use LCD3 or 860C and have an external wattmeter if they find differences by comparing the consumption.
What are the preferences for calculating bars, referring to voltage or SOC%? There are pros and cons.
In any case, it is a problem to be investigated.
The consumption in Wh e and the battery bars are controlled differently in the three versions.
- VLCD5/6, XH18
The calculation of the Wh used is done in the controller. I found a difference between the calculated Wh and those measured with an external instrument, this difference is compensated for in the code.
The battery bars are calculated with reference to the voltage, so they may not correspond to the SOC%, but it is possible to correct them by changing the values of Cell x / 4 or Cell x / 6 in the configurator. It is also possible to manually reset the SOC% in case you put a not fully charged battery.
- LCD3
The calculation of Wh and bars is in the LCD3 code and has remained that of the original 20beta1. I have not made any checks on the correspondence of the calculated Wh.
Also in this version, the battery bars are calculated with reference to the voltage, but if there is no correspondence with the SOC%, a correction is not possible.
However, it is possible to enter the effective capacity and have a precise SOC%. But this in all versions.
If you put a not fully charged battery, a manual SOC% reset is not possible.
- 860C
The calculation of Wh and bars is in the 860C code and the same as in v1.1.0 remained, here too I found a difference between Wh calculated and measured, but I did not foresee any compensation.
The bars of the battery are calculated with reference to SOC%, they are therefore very precise, as long as the real value of the capacity is entered and the battery always fully charged.
If a not fully charged battery is inserted, a manual reset is required, otherwise SOC% and bars are not updated. The logic of the manual reset is different from that of the stock displays.
Perhaps it would be better to standardize the versions with regard to SOC%, Wh consumed and battery bars.
Now I ask those who use LCD3 or 860C and have an external wattmeter if they find differences by comparing the consumption.
What are the preferences for calculating bars, referring to voltage or SOC%? There are pros and cons.
Waynemarlow
10 kW
Keep it simple, relying on entering precise values for total W/h’s, reliance on fully charging a battery and using the same battery ( I have 3 all using different cells and total W/h’s ) each time makes me vote for simple voltage level only on the battery bars and an alternative SOC reading customisable to your battery for those who want that extra bit of info.
casainho
10 GW
- Joined
- Feb 14, 2011
- Messages
- 6,045
The battery SOC I implemented on 860C and SW102 is even more complex than this, it includes the battery resistance value that is configured on the display.mbrusa said:
To have the correct SOC estimation, user will need:
- use the automatic measure of battery resistance done by the 860C and SW102
- make the battery low voltage configuration very low to use the max real capacity of the cells
- fully discharge the battery and see the counted value of capacity in Wh
Also be aware that battery resistance can change a lot with ambient temperature and so will SOC also change. Like SOC at 0 degrees or SOC at 35 degrees ambient temperature, may change like -+ 20%.
maximusdm said:
I have also noticed a discrepancy for SOC on my 860c running 0.20c ... I also run an external coulomb counter ... with the external coulomb counter at around 20% capacity left, the 860c is showing 33%.
I just double checked my coulomb counter and it seems I de-rated the battery in that by more than I remember ... the 860C has 650Wh configured, the coulomb counter 16.6Ah or 614Wh (@ 3.7 nominal) so the error is slightly higher again, about 15% out.
I will check settings and test again, but likely a week to use a full battery.
mbrusa said:
I use the validated real battery capacity in WH = 630, inserted in battery menu. It worked fine and the precision is to the procent.
Indeed the battery state in the left corner seems to be Voltage rated, but has also internal resistance set as well (130).
I think that the Voltage difference that Elinx was talking about is the root of all evil.
1. The cut out voltage is wrong and cuts the motor power.
2. The power outputted on display is smaller than in reality and thus the higher temperature on a similar watt usage on display.
I really do not understand how Buba volts are smaller than Mspiders volts
)). These formulas are 200 year old, and do not see how same things can have different values, no matter the implementation.If you check the real battery capacity and enter the real value, the SOC% will be accurate in all versions.
The Wh consumed instead, for different reasons, may be lower than the real ones.
The Wh consumed instead, for different reasons, may be lower than the real ones.
The volts of Buba are the same as those of Mspiders, it is the Amps that are smallermaximusdm said:...
I really do not understand how Buba volts are smaller than Mspiders volts
Your torque sensor changes value as the temperature rises like everyone, yours is probably more sensitive.Mr.Flibble said:mbrusa said:To better understand what it depends on, you should take a test in Power assist mode with:Mr.Flibble said:I have stripped the motor, dried everything out, re-greased, physically calibrated the torque sensor and reassembled. But it is still ghost peddling after a while
As It's fine when I test it on the flat, I suspect its heat acting on a dry controller joint.
Anyone got any other ideas?
Min current ADC step = 0
Assist w / o pedal rotation = disabled
Startup boost = disabled
If it still fails, immediately check the Pedal torque delta and Pedal cadence values in the Technical menu.
I have done this (but in EMTB mode).
When it's working fine the torque sensor registers 0.
When it's ghost peddling is registers 4-6.
The advice to try with Power assist, I did not want you to change the mode you use, but only to have a confirmation that only depends on this.
You should solve by calibrating the torque sensor, increasing the "Pedal torque offset" value obtained from +5 to +10.
However, this affects all modes.
eMTB has a dedicated offset with value -10, the problem could occur with sensitivity eMTB > 12.
I can also just change eMTB offset and send you a modified version to try.
Waynemarlow
10 kW
Went out on a good ride with a couple of friends on a Levo and another TSDZ2
Interesting comparison between the Levo and the TSDZ2 in that we seemed to be drawing much the same W's for all the climbing with the Levo max at 744W and an average of nearly 200W's, On all the ride he used 370Wh's over the 990m of climb
On my 48 volt motor 52 volt battery, unmodified for heat dissipation TSDZ2 on V20.1C LCD3, unfortunately I don't have a max reading, but mine is set at 700W's max but on the display I saw regulalry 450W's and sometimes pushing into the 600W's and used 420Wh's. Now the Levo rider is fitter and lighter and the 50Wh's could easily be explained by the 15kg difference between riders.
The most interesting for me was the other TSDZ2, a 36 volt motor on 48 volt battery, modified with a sleeve around the motor filling the gap between case and motor and heat transfer grease to fill any voids between motor - sleeve - case. Again on V20.1C LCD3 but nowhere as physically fit rider but again 15kgs lighter, using only 340Wh's.
Those little 36 volt motors are great motors and I'm thinking of dropping back once I have an 8 pin controller with a temperature sensor on board.
Interesting comparison between the Levo and the TSDZ2 in that we seemed to be drawing much the same W's for all the climbing with the Levo max at 744W and an average of nearly 200W's, On all the ride he used 370Wh's over the 990m of climb
On my 48 volt motor 52 volt battery, unmodified for heat dissipation TSDZ2 on V20.1C LCD3, unfortunately I don't have a max reading, but mine is set at 700W's max but on the display I saw regulalry 450W's and sometimes pushing into the 600W's and used 420Wh's. Now the Levo rider is fitter and lighter and the 50Wh's could easily be explained by the 15kg difference between riders.
The most interesting for me was the other TSDZ2, a 36 volt motor on 48 volt battery, modified with a sleeve around the motor filling the gap between case and motor and heat transfer grease to fill any voids between motor - sleeve - case. Again on V20.1C LCD3 but nowhere as physically fit rider but again 15kgs lighter, using only 340Wh's.
Those little 36 volt motors are great motors and I'm thinking of dropping back once I have an 8 pin controller with a temperature sensor on board.
Waynemarlow
10 kW
Yes with Boost but no to rotational.
mbrusa said:If you check the real battery capacity and enter the real value, the SOC% will be accurate in all versions.
The Wh consumed instead, for different reasons, may be lower than the real ones.
The volts of Buba are the same as those of Mspiders, it is the Amps that are smaller
I am not sure what you mean by SOC %. Is it odo field or battery segments?
If I set the Wh in menu 1, the battery percentage from odo field is obtained by subtracting the consumed wh for the total wh. And so the percent is wrong because the consumed wh are wrong. For the battery segments i know is calculated based on voltage and internal battery resistance set in menu 1, but still this does not work fine since it is missing one segment with fully charged battery.
Any idea how to fix the amp difference? From what I saw in code, the display should only show the values sent by motor, not recalculate them. I can try to test some modifications, just don't have a clue where to start.
Waynemarlow
10 kW
We have 1 controller that reads the voltage incorrectly, always under reading and hence the top bar is never shown. Wasn’t the version for the VLCD5 able to put in an offset to correct the voltage level ?maximusdm said:
James Broadhurst
100 W
Well! with 52v battery & 48v motor and 850C was a revelation. Superb piece of work. Approx 6wh per km over 40km @17kph av over gentle road with one big hill which I thrashed up it with the temp staying below 30 (very cold here for us - just above freezing but wind chill at least -4) All on a 29er
Having read that it was worth trying the LCD3 I tried it with a 48v battery and motor on 28” Boardman hybrid. That was very special! Not done the power consumption but it’ll be slightly greater than the 29er but faster. Temperature the same.
And the lights! Watching the reaction from motorists behind as the lights blinked away with the brakes on was a hoot!
Both bikes on V20.1C running Hybrid mode.
There are a number of quite dangerous minor road junctions here where you’re joining from a lane onto a road with vehicles travelling at 70mph round a blind bend. Having a reliable and fast start is so much safer. Well done to all who made this software happen.
Having read that it was worth trying the LCD3 I tried it with a 48v battery and motor on 28” Boardman hybrid. That was very special! Not done the power consumption but it’ll be slightly greater than the 29er but faster. Temperature the same.
And the lights! Watching the reaction from motorists behind as the lights blinked away with the brakes on was a hoot!
Both bikes on V20.1C running Hybrid mode.
There are a number of quite dangerous minor road junctions here where you’re joining from a lane onto a road with vehicles travelling at 70mph round a blind bend. Having a reliable and fast start is so much safer. Well done to all who made this software happen.
SOC% is the battery capacity in percent.maximusdm said:mbrusa said:If you check the real battery capacity and enter the real value, the SOC% will be accurate in all versions.
The Wh consumed instead, for different reasons, may be lower than the real ones.
The volts of Buba are the same as those of Mspiders, it is the Amps that are smaller
I am not sure what you mean by SOC %. Is it odo field or battery segments?
If I set the Wh in menu 1, the battery percentage from odo field is obtained by subtracting the consumed wh for the total wh. And so the percent is wrong because the consumed wh are wrong. For the battery segments i know is calculated based on voltage and internal battery resistance set in menu 1, but still this does not work fine since it is missing one segment with fully charged battery.
Any idea how to fix the amp difference? From what I saw in code, the display should only show the values sent by motor, not recalculate them. I can try to test some modifications, just don't have a clue where to start.
Maybe I was misunderstood, the consumption in Wh displayed is not wrong, the 20% difference in the current / power calculation is only in the 20 beta 1 version.
To get the exact percentage, calibration is required.
I can't check it on the display, I read from the manual:
1.5 - Enabled (1) 100% -> 0%
1.7 - Total actual battery Wh
1.8 - Wh consumed
Fully charge the battery, the consumed Wh is zeroed.
At this point, use the bike until the battery is completely exhausted.
Check the Wh consumed 1.8
Set the actual total Wh of the battery 1.7 with the value of the consumed Wh.
The remaining battery capacity (SOC%) can be displayed in the odo field (option 2.1), or if you prefer in the temperature field, I believe it is at the bottom right of the screen, set menu item 8.10 to (2).
The battery bars with LCD3, are based on voltage, they cannot be precise, you should be able to set a voltage value for each bar as you do in the version with configurator.
mbrusa said:
Thank you for the clear explanation.
The procedure explained is exactly what I use for more than 2 years. I will redo de calibration and come with the results.
Just that my battery is 48v/17ah so aprox 816Wh. The real measured value by checking the consumed watts is 630Wh on 0.20.beta1.
I hope that the same value can be use also in current version, because in my latest test the consumed Wh was aprox 350wh much smaller than my battery.
Mr.Flibble
10 W
- Joined
- May 5, 2020
- Messages
- 90
mbrusa said:Your torque sensor changes value as the temperature rises like everyone, yours is probably more sensitive.Mr.Flibble said:mbrusa said:To better understand what it depends on, you should take a test in Power assist mode with:Mr.Flibble said:I have stripped the motor, dried everything out, re-greased, physically calibrated the torque sensor and reassembled. But it is still ghost peddling after a while
As It's fine when I test it on the flat, I suspect its heat acting on a dry controller joint.
Anyone got any other ideas?
Min current ADC step = 0
Assist w / o pedal rotation = disabled
Startup boost = disabled
If it still fails, immediately check the Pedal torque delta and Pedal cadence values in the Technical menu.
I have done this (but in EMTB mode).
When it's working fine the torque sensor registers 0.
When it's ghost peddling is registers 4-6.
The advice to try with Power assist, I did not want you to change the mode you use, but only to have a confirmation that only depends on this.
You should solve by calibrating the torque sensor, increasing the "Pedal torque offset" value obtained from +5 to +10.
However, this affects all modes.
eMTB has a dedicated offset with value -10, the problem could occur with sensitivity eMTB > 12.
I can also just change eMTB offset and send you a modified version to try.
That's very kind of you, but I have ordered a new controller as it has locked up, and gone off all by it's self.
If it was just a change in readings with temp, surely restarting would reset "zero" and fix it?
