TSDZ2 OSF for all displays, VLCD5-VLCD6-XH18, LCD3, 860C-850C-SW102.
- Thread starter mbrusa
- Start date
mbrusa said:
Thanks mbrusa,
I am very pleased to know that you will be able to try to solve this "Power Delay Issue". Thanks.
I think my video can provide important information.
For me, Casainho's explanation about the current behavior in the motor seems to make sense.
But I don't know how the firmware was developed. I also don't know how the motor works.
I will fix, very quickly, the mechanical part of my engine and then install your firmware. So I can test the firmware and give some feedback.
If you need any tests that I can do, just ask.
I can see the pedaling power on my Garmin, because I have MTB pedals, with a power sensor on both pedals.
Thanks
Azur
Is this delay problem something you notice also in the stock firmware? Because it was troubling me as well with the OSF, and thus I've been using the stock fw and mostly I've been quite content with it. 52V battery and 36V motor gives assist with decent cadence, a bit more would be optimal but v.1.1.0 with field weakening did not make much difference. Efficiency is lower too, range is reduced 10-20% compared to OSF. But for me too the delay after stopping pedaling is a game changer on single track riding.
ilu said:
Is this delay problem something you notice also in the stock firmware?
I don't know.
But for me too the delay after stopping pedaling is a game changer on single track riding.
This is exactly what I think.
Thank you for being clear in the explanation
Thanks
Azur
Mr.Flibble
10 W
- Joined
- May 5, 2020
- Messages
- 90
mbrusa said:
Sorry to be a bit slow, but does that mean it is reversible for the display as well as the motor?
Also, the configurator flashes both the motor and the display at once ?
Thx.
HughF
100 W
devboy-greg said:
This is my experience too - this firmware is very very responsive when AWOP is enabled
jeff.page.rides
100 W
- Joined
- Aug 13, 2019
- Messages
- 295
Hi All,
I want to share my personal saga of working with key people on this forum to get the TSDZ2 firmware to where version 20.1C is now. I am so happy and satisfied with version 20.1C that I have decided to share my Hand-Cycle story to express how much I appreciate what everyone has contributed to getting us where we are today.
First, I will tell you the abridged version of my biking life and where I am now. I started mountain biking in 1986, and fell in love instantly, rode my mountain bikes and road bikes year-round three times each week, and enjoyed every minute of it.
Then while on a mountain biking vacation in Steamboat Springs, Colorado, in 2015, I rode off a 6-foot drop that I didn't see and was going way too slow to make the landing. I broke my neck, and I am a C6-7 quadriplegic. I'm paralyzed from the middle of my arms down and from my chest down. Nothing works. No muscles and no feeling. The very worst part is my hands don't work at all.
So, as you can imagine, I was devastated, and while I was in the hospital for three months, I searched for what type of outdoor activities I could do even as a quadriplegic. As I searched, I found that the two activities that I wanted to do were skiing and biking, and every post or video I saw was all about paraplegics and amputees. They both have working hands and some core strength, depending on where the break-in their back is. To make it simple, if someone breaks their neck and doesn't have working hands, they have quadriplegia, and if someone breaks their back and has working hands, they have paraplegia.
Before my injury, I was in great shape and more robust than an average person. But that all ended quickly, and I had to retrain what muscles did still work, how to work again. Before my injury, I worked with my body all day long every day, plus I rode bikes, skied, and snowboarded. After my injury, I have about 60% of the bicep’s strength of an average out-of-shape person and 30% of an average out-of-shape person's triceps strength. So, after nine months of no biking or skiing, I found https://www.utrails.us/ They had hand-cycles that I could try at a park 15 miles from my home, and they offered instructors to ride along with you if something went wrong during the ride, which it often did. So, I went there and gave it a try. On that first ride, it took me two hours to make two loops around that park. But it was something I could do outside. I kept at it, and a very kind Lady, I wish I could remember her name, gave me a used Top End XLT hand-cycle. After some time, I also found https://wasatchadaptivesports.org/; they also offered hand-cycles to ride if you didn't have one. The most significant difference was they were very close to my home. I didn't have to drive Uptown and ride at the same park time after time. I could ride even from home and meet them along the trail. Then shortly after this, I discovered https://discovernac.org/. They offered the same thing and the ability to ride on dirt and not just pavement. Riding off-road was exciting because it was more like mountain biking which I enjoyed so much before my injury. But here, I found out that I needed a power assist that had more low-end torque and was a lot safer than required on the pavement.
I needed Power-Assist to work for a hand-cycle with a coaster brake and someone with very week arms. I also needed a way to shift gears because my hands were strapped to the pedals. I started searching for something that would work for me. I called the major companies that sell power assists, and they all told me that it did not exist. I was on the Internet continually trying to find things, and there was nothing out there. Then after a year of searching, I found Mark at https://www.electrifybike.com; he had just started his business and was working in a small garage at his home only a few miles from my house. He told me he had the power assist I needed a TSDZ2 that would work with a coaster brake. So, we installed it on my bike, and I was off and riding, NOT! It did not give me enough boost because my arms were too weak to get the torque sensor to apply power. After months of trying, I figured out that if I pulled on the cranks against the brake when I booted up the system, it would assist me with the power I needed. That was a game-changer.
I was also was able to find XShifter and work with the owner and inventor Paul Gallagher. With his help, we were able to use his current electric bike shifting the XShifter so that I could either shift with my voice and then eventually with my chin so that I could be in the right gear at the right time. Before XShifter I had to remove my hand that was strapped to a pedal from the cranks reached down make a shift put my hand back into the cranks and pedal again then repeat for each shift. It was very difficult and frustrating because I would lose all my momentum in the process of shifting. But with the ability to shift with my voice or chin I could shift very quickly without removing my hands from the cranks what a big difference that made. I am now the distributor for XShifter it's an awesome product that can help anybody shift simpler, easier and faster. https://www.electrifybike.com/store/p155/Xshifter.html#/
Shortly after that, Mark let me know about a guy called Casainho, the name he goes by on Endless-Sphere.com, an ebike technical forum for DIYers. He lives in Portugal, and he was working on Flexible Open-Source Firmware for the TSDZ2 with an LCD3 Display. We contacted him and let him know what we were trying to do. He informed us that there are many riders with weaker legs like his young son that could also benefit. He made a lot of improvement over time, but I still had to pull back on the cranks when booting up the system to get the power I needed. And if you pulled too much against the brake when the system boots up, it was difficult to stop rotation when you needed to brake.
Then Buba, an EndlessSphere.com engineer in Sweden, started adding to the TSDZ2 LCD3 OSF project. He reached out to us, and we started working together to make the safety and usability changes we needed. He made some changes that made it much safer to pull backward for the coaster brake. And give it much more power at a low rpm and from a stop. But you still needed to pull back as it boots up to get the amount of power required. At that point, we needed to stay with V20 because V19 was not safe for braking with a coaster brake.
We started converting hand-cycles for those programs listed above and individuals that had their own hand-cycle. And I could see that I was making a difference for others in the same situation as I was. It also made hand-cycling much better for paraplegics and amputees.
I stayed active on the forum, continually asking for help, then Covid 19 hit. Mbrusa, an Endless-Sphere engineer in Italy, offered to help as he was stuck at home and had some time to try and help make V20 for the LCD3 better, and he did. For the first time with his changes, you no longer needed to pull back against the brake as the system boots to get the required power. At about that same time, Casainho released V1.0.0 for the 850C. My new hand-cycle didn't need a coaster brake because I designed a new elbow brake. I wanted to try V1.0.0, but I was not too fond of the 850C display, So I kept asking Casainho to get version 1.0.0 to work on the 860C. Mark sent him an 860C display, and I think he acquired some 860C’s on his own. After he got it working with V1.0.0, I put an 860C with V1.0.0 on the new hand-cycle and left the older one with mbrusa’s version of 20.1 on the LCD3 he did for me on the older hand-cycle. Long story short, the 860C display was in every way the best display that I had tried. But for me, V1.0.0 had no low end and did not perform near as good as the older 20.1 on the LCD3. Even after a year and countless hours of trying different settings.
Then mbrusa stepped up and brought the best of both worlds together. He used the 860c display code and the incredible 20.1 motor firmware. I thought it was going to take another year to get this working. But he got it done very quickly, and it is so awesome to be able to use the 20.1 motor firmware on the 860C and blend it with the display firmware that Casainho had done so well on the 860C.
I am so thankful for everyone that has contributed to this OSF project. I mention the main three friends that I've worked with and the countless private messages that have gone back and forth to get to this point. Casainho, buba and for me most of all mbrusa. I can call the current state of this project close to perfection. It's taken four years to get here and at the point where it's very safe, very powerful at no or slow cadence, very usable, very friendly, very smooth, and on and on.
My main goal for all of this was if someone looks into hand-cycling, they will know that it is possible to ride a hand-cycle safely, and all they need is biceps. Knowing this would have made a massive difference in my life for that first year after I was injured. When I started, I had biceps and triceps, but my average speed was five mph, with my longest ride of 25 miles, which took 6.5 hours. At that slow speed, it is all work and not much fun. Also, without a headwind, it was easy to overheat. Now my average speed is 15mph, and I can ride 30 miles in less than 2.5 hours. My longest ride so far is 110 miles. In one day, I use three 17.5 Ah batteries to ride that far, and I have done that ride once in each of the last two years. Now that riding is much safer and much more fun for me and everyone else, they can take advantage of all the contributions that everyone has made to this OSF Project!
Some of you may know that last year I started to try to liquid cool the TSDZ2. The primary purpose of this was that because I'm 190 pounds, my bike, pack, and three batteries are 100 pounds. The TSDZ2 has to power 300 pounds up and down the trails. With watts turned up to 900 watts, it was super fun, but anything over 450 watts would kill the TSDZ2. After several coolant leaks and other problems, it was working great. I am still going to share this "how-to" with everyone. I chose to put it off because getting to this point with the firmware was more critical than keeping cool at high watts.
I have been riding my old hand-cycle with the 860C 20.1C for testing. And I have been riding my newer hand-cycle with a BBS02 because Mark wanted me to compare them. I have to say that a torque sensor is the only way to go now that we have 20.1C on the 860C! To compare the TSDZ2 at 450 Watts to the BBS02 at 1100 watts is not fair. But with the TSDZ2 turned up to 900 watts with liquid cooling, the only two things that the BBS02 does better is quiet sound and runs cooler without liquid cooling.
Many more people have worked coding, testing, and writing the TSDZ2 open-source firmware project. Thanks to you all. I am grateful to see this come together after three years of effort.
Thanks, Jeff
I want to share my personal saga of working with key people on this forum to get the TSDZ2 firmware to where version 20.1C is now. I am so happy and satisfied with version 20.1C that I have decided to share my Hand-Cycle story to express how much I appreciate what everyone has contributed to getting us where we are today.
First, I will tell you the abridged version of my biking life and where I am now. I started mountain biking in 1986, and fell in love instantly, rode my mountain bikes and road bikes year-round three times each week, and enjoyed every minute of it.
Then while on a mountain biking vacation in Steamboat Springs, Colorado, in 2015, I rode off a 6-foot drop that I didn't see and was going way too slow to make the landing. I broke my neck, and I am a C6-7 quadriplegic. I'm paralyzed from the middle of my arms down and from my chest down. Nothing works. No muscles and no feeling. The very worst part is my hands don't work at all.
So, as you can imagine, I was devastated, and while I was in the hospital for three months, I searched for what type of outdoor activities I could do even as a quadriplegic. As I searched, I found that the two activities that I wanted to do were skiing and biking, and every post or video I saw was all about paraplegics and amputees. They both have working hands and some core strength, depending on where the break-in their back is. To make it simple, if someone breaks their neck and doesn't have working hands, they have quadriplegia, and if someone breaks their back and has working hands, they have paraplegia.
Before my injury, I was in great shape and more robust than an average person. But that all ended quickly, and I had to retrain what muscles did still work, how to work again. Before my injury, I worked with my body all day long every day, plus I rode bikes, skied, and snowboarded. After my injury, I have about 60% of the bicep’s strength of an average out-of-shape person and 30% of an average out-of-shape person's triceps strength. So, after nine months of no biking or skiing, I found https://www.utrails.us/ They had hand-cycles that I could try at a park 15 miles from my home, and they offered instructors to ride along with you if something went wrong during the ride, which it often did. So, I went there and gave it a try. On that first ride, it took me two hours to make two loops around that park. But it was something I could do outside. I kept at it, and a very kind Lady, I wish I could remember her name, gave me a used Top End XLT hand-cycle. After some time, I also found https://wasatchadaptivesports.org/; they also offered hand-cycles to ride if you didn't have one. The most significant difference was they were very close to my home. I didn't have to drive Uptown and ride at the same park time after time. I could ride even from home and meet them along the trail. Then shortly after this, I discovered https://discovernac.org/. They offered the same thing and the ability to ride on dirt and not just pavement. Riding off-road was exciting because it was more like mountain biking which I enjoyed so much before my injury. But here, I found out that I needed a power assist that had more low-end torque and was a lot safer than required on the pavement.
I needed Power-Assist to work for a hand-cycle with a coaster brake and someone with very week arms. I also needed a way to shift gears because my hands were strapped to the pedals. I started searching for something that would work for me. I called the major companies that sell power assists, and they all told me that it did not exist. I was on the Internet continually trying to find things, and there was nothing out there. Then after a year of searching, I found Mark at https://www.electrifybike.com; he had just started his business and was working in a small garage at his home only a few miles from my house. He told me he had the power assist I needed a TSDZ2 that would work with a coaster brake. So, we installed it on my bike, and I was off and riding, NOT! It did not give me enough boost because my arms were too weak to get the torque sensor to apply power. After months of trying, I figured out that if I pulled on the cranks against the brake when I booted up the system, it would assist me with the power I needed. That was a game-changer.
I was also was able to find XShifter and work with the owner and inventor Paul Gallagher. With his help, we were able to use his current electric bike shifting the XShifter so that I could either shift with my voice and then eventually with my chin so that I could be in the right gear at the right time. Before XShifter I had to remove my hand that was strapped to a pedal from the cranks reached down make a shift put my hand back into the cranks and pedal again then repeat for each shift. It was very difficult and frustrating because I would lose all my momentum in the process of shifting. But with the ability to shift with my voice or chin I could shift very quickly without removing my hands from the cranks what a big difference that made. I am now the distributor for XShifter it's an awesome product that can help anybody shift simpler, easier and faster. https://www.electrifybike.com/store/p155/Xshifter.html#/
Shortly after that, Mark let me know about a guy called Casainho, the name he goes by on Endless-Sphere.com, an ebike technical forum for DIYers. He lives in Portugal, and he was working on Flexible Open-Source Firmware for the TSDZ2 with an LCD3 Display. We contacted him and let him know what we were trying to do. He informed us that there are many riders with weaker legs like his young son that could also benefit. He made a lot of improvement over time, but I still had to pull back on the cranks when booting up the system to get the power I needed. And if you pulled too much against the brake when the system boots up, it was difficult to stop rotation when you needed to brake.
Then Buba, an EndlessSphere.com engineer in Sweden, started adding to the TSDZ2 LCD3 OSF project. He reached out to us, and we started working together to make the safety and usability changes we needed. He made some changes that made it much safer to pull backward for the coaster brake. And give it much more power at a low rpm and from a stop. But you still needed to pull back as it boots up to get the amount of power required. At that point, we needed to stay with V20 because V19 was not safe for braking with a coaster brake.
We started converting hand-cycles for those programs listed above and individuals that had their own hand-cycle. And I could see that I was making a difference for others in the same situation as I was. It also made hand-cycling much better for paraplegics and amputees.
I stayed active on the forum, continually asking for help, then Covid 19 hit. Mbrusa, an Endless-Sphere engineer in Italy, offered to help as he was stuck at home and had some time to try and help make V20 for the LCD3 better, and he did. For the first time with his changes, you no longer needed to pull back against the brake as the system boots to get the required power. At about that same time, Casainho released V1.0.0 for the 850C. My new hand-cycle didn't need a coaster brake because I designed a new elbow brake. I wanted to try V1.0.0, but I was not too fond of the 850C display, So I kept asking Casainho to get version 1.0.0 to work on the 860C. Mark sent him an 860C display, and I think he acquired some 860C’s on his own. After he got it working with V1.0.0, I put an 860C with V1.0.0 on the new hand-cycle and left the older one with mbrusa’s version of 20.1 on the LCD3 he did for me on the older hand-cycle. Long story short, the 860C display was in every way the best display that I had tried. But for me, V1.0.0 had no low end and did not perform near as good as the older 20.1 on the LCD3. Even after a year and countless hours of trying different settings.
Then mbrusa stepped up and brought the best of both worlds together. He used the 860c display code and the incredible 20.1 motor firmware. I thought it was going to take another year to get this working. But he got it done very quickly, and it is so awesome to be able to use the 20.1 motor firmware on the 860C and blend it with the display firmware that Casainho had done so well on the 860C.
I am so thankful for everyone that has contributed to this OSF project. I mention the main three friends that I've worked with and the countless private messages that have gone back and forth to get to this point. Casainho, buba and for me most of all mbrusa. I can call the current state of this project close to perfection. It's taken four years to get here and at the point where it's very safe, very powerful at no or slow cadence, very usable, very friendly, very smooth, and on and on.
My main goal for all of this was if someone looks into hand-cycling, they will know that it is possible to ride a hand-cycle safely, and all they need is biceps. Knowing this would have made a massive difference in my life for that first year after I was injured. When I started, I had biceps and triceps, but my average speed was five mph, with my longest ride of 25 miles, which took 6.5 hours. At that slow speed, it is all work and not much fun. Also, without a headwind, it was easy to overheat. Now my average speed is 15mph, and I can ride 30 miles in less than 2.5 hours. My longest ride so far is 110 miles. In one day, I use three 17.5 Ah batteries to ride that far, and I have done that ride once in each of the last two years. Now that riding is much safer and much more fun for me and everyone else, they can take advantage of all the contributions that everyone has made to this OSF Project!
Some of you may know that last year I started to try to liquid cool the TSDZ2. The primary purpose of this was that because I'm 190 pounds, my bike, pack, and three batteries are 100 pounds. The TSDZ2 has to power 300 pounds up and down the trails. With watts turned up to 900 watts, it was super fun, but anything over 450 watts would kill the TSDZ2. After several coolant leaks and other problems, it was working great. I am still going to share this "how-to" with everyone. I chose to put it off because getting to this point with the firmware was more critical than keeping cool at high watts.
I have been riding my old hand-cycle with the 860C 20.1C for testing. And I have been riding my newer hand-cycle with a BBS02 because Mark wanted me to compare them. I have to say that a torque sensor is the only way to go now that we have 20.1C on the 860C! To compare the TSDZ2 at 450 Watts to the BBS02 at 1100 watts is not fair. But with the TSDZ2 turned up to 900 watts with liquid cooling, the only two things that the BBS02 does better is quiet sound and runs cooler without liquid cooling.
Many more people have worked coding, testing, and writing the TSDZ2 open-source firmware project. Thanks to you all. I am grateful to see this come together after three years of effort.
Thanks, Jeff
HughF said:
Are you using Power mode or Torque mode?
I am talking about Power mode.
But as michih says,
"from the stand, yes. but when rolling, starting and responding is delayed."
Are you saying "this firmware is very responsive when AWOP is enabled" from standing or when rolling?
What is AWOP ?
Azur
Peacepirate
100 mW
- Joined
- Sep 22, 2020
- Messages
- 41
mctubster said:Could you explain a scenario when you need this behaviour?Lii said:I've tried different modes. But none of the modes can provide a smooth start when you press the pedal lightly and at the same time high power when you press the pedal firmly. To achieve a smooth start, you have to limit the overall level of support. As a result, the maximum power is limited. I'm trying to find parameters with which I can change the non-linearity curve. Yesterday I experimented a lot with this but there is no result. An excessive change in the calibration of the torque sensor resulted in an "E02" error. Apparently I will have to study the code. I have some experience with C ++ programming with Arduino.
I believe to understand very well what Lii desires and he explained it quite straightforward.
Imagine not having those two buttons for changing these - typically - 5 assist levels. When i am on assist level 1 and i push the pedals slightly the motor assists with light support(maybe 25w). Logical. Pedaling hard on assist level 1 results in more motor power output (150w for example) but the motor output could be (much) higher if i would change the assist level to 2 or 3 etc.
On the other side when pedaling slightly on assist level 5 the minimal motor output will be always higher than the 25w mentioned before (75w f. e.) on assist level 1.
I understand him in that way that he wishes a mode where there is only 1 assist level. He wants the whole band width of control with the low motor power support equal to assist mode 1 and high motor power support equal to assist lvl 5 - controlled by his legs only without buttons.
Waynemarlow
10 kW
AZUR said:
Sorry Azur I suspect you are talking about a mechanical problem that we have seen at times rather than a electrical response. The large sprag bearing on the drive side on some motors can have that 1/2 to even a full rotation before engagement when the grease inside it becomes contaminated with water. It’s a really odd feeling when it happens as you have to peddle maybe a a half turn or what feels more sometimes and only then the sprags on that bearing engage. Remember the torque sensor only then can sense a change in torque once the sprags have engaged and the motor can respond once there is strain on the peddles.
It is only a momentarily thing as you can at standstill load up the peddles normally and its for some reason only whilst underway ( sorry I have no reasoning other than vibration for this ). On both motors we have had this symptom we replaced the sprag bearing with a good quality bearing and the problem was solved. One of the motors was an almost new motor and right from new the sprag bearing was always slow to engage, but I do suspect from a pure engineering point of view that these large grease filled one way bearings are always going to be slower to respond than say the ungreased rear hub where it’s an almost instant engagement due to the actual toothed design.
for me it's not mechanical. you can see human power on the 850c display. it goes high but the engine reacts with a delay.
MFG Michael
OK, that's moaning at a high level. the software from mbrusa is really very good! but I have the comparison to the Rocky Mountain Powerplay. it just works better. also better than Bosch
MFG Michael
OK, that's moaning at a high level. the software from mbrusa is really very good! but I have the comparison to the Rocky Mountain Powerplay. it just works better. also better than Bosch
I lowered the help levels from 20 to 13 in MTB mode and lowered the upper calibration value of the torque sensor from 280 (this is the real maximum value) to 270 because the bike will be used by a woman who will not be able to pedal with such force. The reaction to pressing and accelerating the bike is now working properly. But the problem that other users have already written about has not disappeared: The reaction to the beginning and stopping of the pedal movement comes with a big slowdown. This gives an unpleasant and dangerous feeling when you have already stopped pedaling but the bike continues to accelerate. I am trying to find the place in the code where this delay is being generated. This is similar to a "delay" function that pauses program execution or a very high filtering factor for cadence or torque sensors. Can you tell me where this delay can be in the code?Lii said:mbrusa said:
I've tried different modes. But none of the modes can provide a smooth start when you press the pedal lightly and at the same time high power when you press the pedal firmly. To achieve a smooth start, you have to limit the overall level of support. As a result, the maximum power is limited. I'm trying to find parameters with which I can change the non-linearity curve. Yesterday I experimented a lot with this but there is no result. An excessive change in the calibration of the torque sensor resulted in an "E02" error. Apparently I will have to study the code. I have some experience with C ++ programming with Arduino.
In general, this is a very big difference between the original program code. Thank you!
You have described exactly what I want to get! The bike will be controlled by a woman who does not need extra buttons, modes, etc toys). Just sit down and go!)Peacepirate said:mctubster said:Could you explain a scenario when you need this behaviour?Lii said:I've tried different modes. But none of the modes can provide a smooth start when you press the pedal lightly and at the same time high power when you press the pedal firmly. To achieve a smooth start, you have to limit the overall level of support. As a result, the maximum power is limited. I'm trying to find parameters with which I can change the non-linearity curve. Yesterday I experimented a lot with this but there is no result. An excessive change in the calibration of the torque sensor resulted in an "E02" error. Apparently I will have to study the code. I have some experience with C ++ programming with Arduino.
I believe to understand very well what Lii desires and he explained it quite straightforward.
Imagine not having those two buttons for changing these - typically - 5 assist levels. When i am on assist level 1 and i push the pedals slightly the motor assists with light support(maybe 25w). Logical. Pedaling hard on assist level 1 results in more motor power output (150w for example) but the motor output could be (much) higher if i would change the assist level to 2 or 3 etc.
On the other side when pedaling slightly on assist level 5 the minimal motor output will be always higher than the 25w mentioned before (75w f. e.) on assist level 1.
I understand him in that way that he wishes a mode where there is only 1 assist level. He wants the whole band width of control with the low motor power support equal to assist mode 1 and high motor power support equal to assist lvl 5 - controlled by his legs only without buttons.
Lii said:I lowered the help levels from 20 to 13 in MTB mode and lowered the upper calibration value of the torque sensor from 280 (this is the real maximum value) to 270 because the bike will be used by a woman who will not be able to pedal with such force. The reaction to pressing and accelerating the bike is now working properly. But the problem that other users have already written about has not disappeared: The reaction to the beginning and stopping of the pedal movement comes with a big slowdown. This gives an unpleasant and dangerous feeling when you have already stopped pedaling but the bike continues to accelerate. I am trying to find the place in the code where this delay is being generated. This is similar to a "delay" function that pauses program execution or a very high filtering factor for cadence or torque sensors. Can you tell me where this delay can be in the code?Lii said:mbrusa said:
I've tried different modes. But none of the modes can provide a smooth start when you press the pedal lightly and at the same time high power when you press the pedal firmly. To achieve a smooth start, you have to limit the overall level of support. As a result, the maximum power is limited. I'm trying to find parameters with which I can change the non-linearity curve. Yesterday I experimented a lot with this but there is no result. An excessive change in the calibration of the torque sensor resulted in an "E02" error. Apparently I will have to study the code. I have some experience with C ++ programming with Arduino.
In general, this is a very big difference between the original program code. Thank you!
I also think that it was interpreted that way.
When I made the first rides on my Rocky Mountain, you clearly noticed a non-circular rotation.
if the response is deliberately delayed, the engine runs "smoother". but has the effect that everything is implemented with a delay.
MFG Michael
Lii said:I lowered the help levels from 20 to 13 in MTB mode and lowered the upper calibration value of the torque sensor from 280 (this is the real maximum value) to 270 because the bike will be used by a woman who will not be able to pedal with such force. The reaction to pressing and accelerating the bike is now working properly. But the problem that other users have already written about has not disappeared: The reaction to the beginning and stopping of the pedal movement comes with a big slowdown. This gives an unpleasant and dangerous feeling when you have already stopped pedaling but the bike continues to accelerate. I am trying to find the place in the code where this delay is being generated. This is similar to a "delay" function that pauses program execution or a very high filtering factor for cadence or torque sensors. Can you tell me where this delay can be in the code?Lii said:mbrusa said:
I've tried different modes. But none of the modes can provide a smooth start when you press the pedal lightly and at the same time high power when you press the pedal firmly. To achieve a smooth start, you have to limit the overall level of support. As a result, the maximum power is limited. I'm trying to find parameters with which I can change the non-linearity curve. Yesterday I experimented a lot with this but there is no result. An excessive change in the calibration of the torque sensor resulted in an "E02" error. Apparently I will have to study the code. I have some experience with C ++ programming with Arduino.
In general, this is a very big difference between the original program code. Thank you!
Casainho thinks that the problem happens as described below.
Maybe he can give some additional information, as he has already solved the problem with the lights, as he refers, below in the image.
https://endless-sphere.com/forums/viewtopic.php?f=30&t=93818&start=6275#p1585446
.
POWER DELAY ISSUE
The images were taken sequentially, during the presentation of the video. The values for PWM, Human Power, Motor Speed and Cadence are in the table below.
POWER DELAY ISSUE
We can see in these images some screens taken from the videos I made in September 2019.
Casainho V 1.0.0-beta.1
48V motor, 15S battery.
Display 850
The assistance mode is "power mode".
As can be seen only after 1m482ms of starting the pedaling, the PWM is greater than zero.
PWM = 0 means that the motor is not delivering power.
PWM = 100 maximum power
Only after pedaling for 1m482ms does the motor start to assist.
It is also visible that the motor measures human power from the beginning.
Video
https://www.youtube.com/watch?v=GAUzipgIwoc&t=10s
The video can be downloaded here:
https://meocloud.pt/link/31a9b10c-6a12-488c-a192-4fc5831cc481/TSDZ2 Power delay Issue/
I use a video player, free for windows.
I downloaded it here: https://www.nchsoftware.com/videopad/index.html
Azur
POWER DELAY ISSUE
The images were taken sequentially, during the presentation of the video. The values for PWM, Human Power, Motor Speed and Cadence are in the table below.
POWER DELAY ISSUE
We can see in these images some screens taken from the videos I made in September 2019.
Casainho V 1.0.0-beta.1
48V motor, 15S battery.
Display 850
The assistance mode is "power mode".
As can be seen only after 1m482ms of starting the pedaling, the PWM is greater than zero.
PWM = 0 means that the motor is not delivering power.
PWM = 100 maximum power
Only after pedaling for 1m482ms does the motor start to assist.
It is also visible that the motor measures human power from the beginning.
Video
https://www.youtube.com/watch?v=GAUzipgIwoc&t=10s
The video can be downloaded here:
https://meocloud.pt/link/31a9b10c-6a12-488c-a192-4fc5831cc481/TSDZ2 Power delay Issue/
I use a video player, free for windows.
I downloaded it here: https://www.nchsoftware.com/videopad/index.html
Azur
Waynemarlow said:
Thanks Waynemarlow,
At first I also thought it could be a hardware problem and so I bought a new engine.
As you can see, in the previous post, the controller/firmware is able to measure the force applied to the torque sensor. It is proof, as you say, that the hardware is working.
But I'm not the only one to experience the Power Delay Issue.
Michel, Lii, ilu and also mbrusa have also identified the problem.
Altogether there are already 5 users of the firmware that are concerned with this matter.
I was the only one, who managed to show in what situation and how the problem happens.
I was able to measure the problem and show evidence.
I'm just trying to help.
Azur
.
POWER DELAY ISSUE - Motor Current Analysis
The images were taken sequentially, during the presentation of the video. The values for PWM, Motor Current, Motor Speed and Cadence are in the table below.
In this test you can see the representation of the motor current as a function of pedaling.
In the previous test that I showed earlier, you can see the "Human power"
POWER DELAY ISSUE - Motor Current Analysis
We can see in these images some screens taken from the videos I made in September 2020.
Casainho V 1.0.0-beta.1
48V motor, 15S battery.
Display 850
The assistance mode is "power mode".
As can be seen only after 1s914ms (almost 2 seconds) of starting the pedaling, PWM is greater than zero.
PWM = 0 means that the motor is not delivering power.
PWM = 100 maximum power
Only after pedaling for 1s914ms does the motor start to assist.
Video
https://www.youtube.com/watch?v=xosa4rvcS7c
The video can be downloaded here:
https://meocloud.pt/link/31a9b10c-6a12- ... y Issue/
I use a video player, free for windows.
I downloaded it here: https://www.nchsoftware.com/videopad/index.html
Azur
POWER DELAY ISSUE - Motor Current Analysis
The images were taken sequentially, during the presentation of the video. The values for PWM, Motor Current, Motor Speed and Cadence are in the table below.
In this test you can see the representation of the motor current as a function of pedaling.
In the previous test that I showed earlier, you can see the "Human power"
POWER DELAY ISSUE - Motor Current Analysis
We can see in these images some screens taken from the videos I made in September 2020.
Casainho V 1.0.0-beta.1
48V motor, 15S battery.
Display 850
The assistance mode is "power mode".
As can be seen only after 1s914ms (almost 2 seconds) of starting the pedaling, PWM is greater than zero.
PWM = 0 means that the motor is not delivering power.
PWM = 100 maximum power
Only after pedaling for 1s914ms does the motor start to assist.
Video
https://www.youtube.com/watch?v=xosa4rvcS7c
The video can be downloaded here:
https://meocloud.pt/link/31a9b10c-6a12- ... y Issue/
I use a video player, free for windows.
I downloaded it here: https://www.nchsoftware.com/videopad/index.html
Azur
Peacepirate
100 mW
- Joined
- Sep 22, 2020
- Messages
- 41
Rydon said:
Please read carefully what others write and mean instead of posting those senseless one-liners. Actually i am not really interested in any e-mtb mode or in installing this firmware in particular because i won’t deal with any tsdz2 firmware that has lag/response issues.
But to be honest i don‘t really understand all the whining speech of the other users here and their problems in terms of late/lagged motor response. R0nkos TSDZ2 fork works with all those modern displays like my sw102, it is tork only based and there is no delayed motor response/support no matter the bike starts from 0km/h or within the ride when one stops and continues pedaling. Motor starts always immediately and r0nkos fork is based on casainhos 0.80 firmware, means it is in a more or less advanced state and development status. Just my 2 cents...
