How stable are the dragfactor numbers for you? #40
Replies: 8 comments 8 replies
-
|
It tends to depend on the rower. For a hybrid rower, the balance between magnetic and air drag changes when it is pulled harder, so the dragfactor really changes. For a pure air-rower, it should be fairly stable although there are some physical effects that sublty change the dragfactor. However, It is a known issue that pauses create havoc among dynamically calclated dragfactors (see Shane Farmer "complain" about ths issue on a Concept2 rower: https://www.youtube.com/watch?v=-aDELD29Zxo). OpenRowingMonitor tries to ignore wildly varying dragfactors, but I am quickly to admit that we need more data here: I havan;t got a clue what dragfactors normally hover around and how much they change. So I statically coded some boundaries. We could make them a setting you can tune to your machine if you'd like. |
Beta Was this translation helpful? Give feedback.
-
|
Hello Japp, thank you for the quick answer. In my case I am using an old Tunturi R740, a purely magnetic rower with a rather heavy flywheel of 19kg. Drag factor hovers around 22000 - 25000 when rowing seriously (splits of about 02:05 - 02;10). During rest periods (splits of about 2:25 and less SPMs) the drag factor rises to 30000 - 35000. When I then start rowing seriously again, log reports and I see split times of 01:45 and correspondingly high power numbers (300 - 320w instead of the expected 180 - 200w). After 60 - 90 seconds things get back to normal and splits are back at 02:05 - 02:10. My gut feeling is that the drag factor (or some parameter used to calculate splits and power) is buffered over time or averaged, that this average is too high after a rest period and that this leads to incorrect calculations. Cheers, |
Beta Was this translation helpful? Give feedback.
-
|
Hi Gordon, I see two things going wrong here:
|
Beta Was this translation helpful? Give feedback.
-
|
😄 sound like a plan. Michael, I think you are the first person to use OpenRowingMonitor on a purely magnetic rower with a pretty heavy flywheel. So it's also pretty interesting for us to see how our physics model behaves there. Would be cool if you could create some pieces of recorded raw rower data and send it to us (set recordRawData: true). We'll add the 10 stroke recording to our automated test suite to prevent regression issue with future version and use the other recording to take the drag factor calculation to a test-run... |
Beta Was this translation helpful? Give feedback.
-
|
Hi there, I woud be more that happy to help! Unfortunately I won't have acces to the rower over the weekend but I had already compiled a short description of the issue I encountered - some info within 'Dragfactor_issue.txt' will be a rehash of what I wrote within this thread. 'dragfactor.zip' contains
I can supply more tests/logs/trainings after the weekend. Thanks for looking into this, Michael |
Beta Was this translation helpful? Give feedback.
-
|
Hi Michael, Lars has pushed the first part of the bugfix to the main release. After updating OpenRowingMonitor you get two new settings in the rowers settings:
Please let us know if this helps. I'm still looking at the magnetic resistance part to fix the physics model. Aside from a footnote that it the dragfactor is linear instead of quadratic, I didn't learn much. So I guess I'll have to e-mail some people and do some math.... |
Beta Was this translation helpful? Give feedback.
-
|
Hello Jaap, thank you very much for the fixes - I will try them in the next days. Cheers, |
Beta Was this translation helpful? Give feedback.
-
|
Hello, I'd like to add my (future) two cents here as I am building ORM in the next days for my magnetic rower. I hope, I can provide you some data as my rower has internal power meter and the drag factor could be somehow correlated with the readings on the Display. |
Beta Was this translation helpful? Give feedback.
-
|
Hi, I have a magnetic rower and have set the dragfactor to be calculated. On the first couple of stroked the calculation is too much but it hen gets better. config.js smoothing: 3, minimumTimeBetweenImpulses: 0.03, maximumTimeBetweenImpulses: 1.0, maximumDownwardChange: 0.05, maximumUpwardChange: 1.75, rowerSettings: { numOfImpulsesPerRevolution: 1, flywheelInertia: 0.10, magicConstant: 23, dragFactor: 60, dampingConstantSmoothing: 2, dampingConstantMaxChange: 0.30, autoAdjustDragFactor: true } Log root@rowingmonitor:/opt/openrowingmonitor/bin# ./openrowingmonitor.sh > [email protected] start > node app/server.js ==== Open Rowing Monitor 0.8.2 ==== bluetooth profile: FTMS Rower webserver running on port 80 websocket client connected workout is shorter than minimum workout time, skipping automatic creation of recordings... *** recovery phase completed Calculated drag factor: 5524.95, which is too far off the currently used dragfactor of 60 Time: 0.3136 sec, impulse 1: recoveryStartAngularVelocity = 448.80 rad/sec, recoveryEndAngularVelocity = 12.57 rad/sec, recoveryPhaseLength = 1.4000 sec *** Applied drag factor: 69 *** DRIVE phase started at time: 0.3136 sec, impulse number 1 stroke: 1, dur: 0.00s, power: 1657w, split: 02:00, ratio: 0.30, dist: 5.5m, cal: 0.0kcal, SPM: 30.0, speed: 14.98km/h, cal/hour: 0.0kcal, cal/minute: 0.0kcal *** drive phase completed *** RECOVERY phase started at time: 1.2426 sec, impuls number 15 Time: 2.2567 sec, impulse 34: flank suggests power (6.9 rad/s2), but waiting for recoveryPhaseLength (0.9836 sec) to exceed minimumRecoveryTime (1.2 sec) Time: 2.4350 sec, impulse 37: flank suggests power (17.0 rad/s2), but waiting for recoveryPhaseLength (1.1576 sec) to exceed minimumRecoveryTime (1.2 sec) *** recovery phase completed Calculated drag factor: 158.58, which is too far off the currently used dragfactor of 69 Time: 2.4907 sec, impulse 38: recoveryStartAngularVelocity = 132.04 rad/sec, recoveryEndAngularVelocity = 105.21 rad/sec, recoveryPhaseLength = 1.2180 sec *** Applied drag factor: 83.85000000000001 *** DRIVE phase started at time: 2.4907 sec, impulse number 38 stroke: 2, dur: 3.50s, power: 301w, split: 04:43, ratio: 0.58, dist: 9.0m, cal: 5.8kcal, SPM: 18.9, speed: 6.36km/h, cal/hour: 941.1kcal, cal/minute: 15.7kcal *** drive phase completed *** RECOVERY phase started at time: 3.0973 sec, impuls number 52 Time: 3.8571 sec, impulse 69: flank suggests power (28.0 rad/s2), but waiting for recoveryPhaseLength (0.7373 sec) to exceed minimumRecoveryTime (1.2 sec) Time: 4.2133 sec, impulse 76: flank suggests power (25.6 rad/s2), but waiting for recoveryPhaseLength (1.0812 sec) to exceed minimumRecoveryTime (1.2 sec) *** recovery phase completed Calculated drag factor: 171.09, which is too far off the currently used dragfactor of 83.85000000000001 Time: 4.3660 sec, impulse 79: recoveryStartAngularVelocity = 157.31 rad/sec, recoveryEndAngularVelocity = 118.01 rad/sec, recoveryPhaseLength = 1.2373 sec *** Applied drag factor: 99.3525 *** DRIVE phase started at time: 4.3660 sec, impulse number 79 stroke: 3, dur: 1.88s, power: 238w, split: 04:05, ratio: 0.40, dist: 13.1m, cal: 6.5kcal, SPM: 27.5, speed: 7.32km/h, cal/hour: 971.7kcal, cal/minute: 16.2kcal *** drive phase completed *** RECOVERY phase started at time: 4.8260 sec, impuls number 90 *** recovery phase completed Calculated drag factor: 190.60, which is too far off the currently used dragfactor of 99.3525 Time: 6.1581 sec, impulse 118: recoveryStartAngularVelocity = 159.14 rad/sec, recoveryEndAngularVelocity = 114.39 rad/sec, recoveryPhaseLength = 1.2899 sec *** Applied drag factor: 119.08162500000002 *** DRIVE phase started at time: 6.1581 sec, impulse number 118 stroke: 4, dur: 1.73s, power: 269w, split: 03:42, ratio: 0.31, dist: 17.3m, cal: 7.0kcal, SPM: 32.2, speed: 8.08km/h, cal/hour: 981.3kcal, cal/minute: 16.4kcal *** drive phase completed *** RECOVERY phase started at time: 6.5999 sec, impuls number 128 *** recovery phase completed Calculated drag factor: 212.87, which is too far off the currently used dragfactor of 119.08162500000002 Time: 7.8604 sec, impulse 154: recoveryStartAngularVelocity = 156.92 rad/sec, recoveryEndAngularVelocity = 111.59 rad/sec, recoveryPhaseLength = 1.2161 sec *** Applied drag factor: 141.98218125 *** DRIVE phase started at time: 7.8604 sec, impulse number 154 stroke: 5, dur: 1.77s, power: 312w, split: 03:30, ratio: 0.28, dist: 21.4m, cal: 7.6kcal, SPM: 34.5, speed: 8.57km/h, cal/hour: 996.6kcal, cal/minute: 16.6kcal *** drive phase completed *** RECOVERY phase started at time: 8.5938 sec, impuls number 170 Time: 9.7856 sec, impulse 193: flank suggests power (53.9 rad/s2), but waiting for recoveryPhaseLength (1.1512 sec) to exceed minimumRecoveryTime (1.2 sec) *** recovery phase completed Calculated drag factor: 202.95, which is too far off the currently used dragfactor of 141.98218125 Time: 9.8379 sec, impulse 194: recoveryStartAngularVelocity = 148.79 rad/sec, recoveryEndAngularVelocity = 108.96 rad/sec, recoveryPhaseLength = 1.2106 sec |
Beta Was this translation helpful? Give feedback.
-
There are some issues here: smoothing, minimumTimeBetweenImpulses, maximumTimeBetweenImpulses, maximumDownwardChange and maximumUpwardChange are part of the rowersettings as well (and thus should be inside the brackets). The value for magicConstant strikes me as odd: in essence this determines the ration between power and linear speed, and changing it from the default 2.8 to something else essentially makes you work harder or less hard for the same speed, where almost everybody uses the 2.8 factor. Is there a special reason for this?
When you look carefully, you see that stroke detection is still struggling a bit. As a side note: perhaps it is wise to upgrade to V1Beta (see #118) and look at your settings again. The V1Beta version is more robust. |
Beta Was this translation helpful? Give feedback.
-
|
thankyou very much I am just updating to V1Beta I set magicConstant using trial and error based on my rowing machines screen. My speed and power were a little bit great compared with the rowing monitor. The strokes did match. I am not sure about the calories calculated, on the machine's original monitor I see around 100 kCal for 20 minutes of rowing. I am not sure if this is right and be using it as a benchmark. I was interested in having the damping effect wattage. I would like to see when greater damping is selected, higher wattage is shown for a given stroke rate.
|
Beta Was this translation helpful? Give feedback.
-
|
Also is there a cheaper rowing virtual app. This is EXR, it is okay but a little expensive at $15.00 p month. I was thinking about trying to develop an open-source one oft these. |
Beta Was this translation helpful? Give feedback.
-
|
Just ran the v1beta, I see a small error
|
Beta Was this translation helpful? Give feedback.
-
That is the normal procedure, but you should modify the dragfactor and flywheelinertia and not the magicconstant, following this manual https://github.com/JaapvanEkris/openrowingmonitor/blob/v1beta_updates/docs/rower_settings.md
Most monitors, aside the Concept2, RowPerfect RP3 and the BioRower are total nonsense, so if you are going to compare, please not use anything aside these brands.
V1Beta's physics engine is validated against the PM5, which is tested by universities. We use the same formula's as they do, and get identical results.
It will in V1Beta, providing the settings are correct.
The yearly subscription is cheaper I think. But it is a great app with a nice community which organize rowalongs almost weekly. Used it since it left the beta stage, still loving it. An alternative is RowPro: https://www.digitalrowing.com/index.html which is cheaper but isn't tested against OpenRowingMonitor yet.
That is a tough project, but cool if you would succeed! |
Beta Was this translation helpful? Give feedback.
-
|
Hello, |
Beta Was this translation helpful? Give feedback.
-
|
got it thanks |
Beta Was this translation helpful? Give feedback.
-
Hi,
I'm seeing strange data for split times and power when doing intervals which seems to coincide with wildly fluctutating values for 'drag factor'. Before I flood the board with debug log files I'd like to know what min/max values for drag factor we can expect to see during a training session, especially when rowing hard vs rest periods. Or is it expected to remain (more or less) static?
push don't pull,
Michael
Beta Was this translation helpful? Give feedback.
All reactions