Next Steps

[eglin73]

It has been a bit of a struggle as someone new to this, but I believe I may now be getting some metrics that compare reasonably with both my (limited) rowing experience, and with Concept2's pace-to-watt calculator, which is pretty exciting and rewarding. Thanks to those who are behind the project and working on improving it.

https://www.concept2.com/indoor-rowers/training/calculators/watts-calculator
(Average Watts 103, Average pace per 500 meters 2:30.3)

While my ORM reported a steady 500m row in 2:30, SPM ~20, Power ~105watts

However there are some errors in the log file (attached) which may mean I'm not there yet... Looking at other posts, and the guidance, it seems some errors are to be expected and just the noise reduction mechanisms doing their job, however I'm not sure if I should try reduce the noise a little further or not, or further tune the config.

The common errors seem to relate to minimum recovery time, which is 1.2. (I see some have set this to 0.9)

I'm a little confused by the two variations of this error which refer to 'flank suggests'.

The first one makes sense, as my layman's interpretation reads that the errors suggests ORM is reluctant to consider it the start of a power stroke as the minimum recovery time has not been reached.

flank suggests power (19.5 rad/s2), but waiting for recoveryPhaseLength (1.1523 sec) to exceed minimumRecoveryTime (1.2 sec)

However the second I don't understand as it suggests both the flank and the minimum recovery time are in agreement that the power stroke has not started, so I'm not sure what the concern is.

flank suggests no power (-41.0 rad/s2), but waiting for for recoveryPhaseLength (1.2447 sec) to exceed minimumRecoveryTime (1.2 sec)

This is what my config file looks like at the moment, and I will attach the logs. Hopefully someone can advise whether these errors a concern or just ORM noise reduction working as it should.

export default{
rowerSettings: {
numOfImpulsesPerRevolution: 3,
flywheelInertia: 0.16,
dragFactor: 550
}
}

I also have some questions regarding the dragFactor setting and autoAdjustDragFactor: true,

Q1. I read it is needed for the first stoke, which makes sense. I'm just not sure how much difference to the calculated metrics it makes beyond that first stroke?

Q2. I don't have a damper setting on my rower, but my calculated dragFactor does seem to fluctuate between about 650 and 750 is this normal? (it is outdoors my rower so I suppose wind could make a difference?)

Q3. Should I set the config file dragFactor to match the calculated? (The guidance suggests to look for "Goodness of Fit" when using calculated dragFactor in the config, but my logs don't show that?)

Many thanks

500m calibration row .txt

[JaapvanEkris]

First of all, please switch to V1Beta, see #118 how to do this.

Then, you can follow this manual to set up the machine: https://github.com/JaapvanEkris/openrowingmonitor/blob/v1beta_updates/docs/rower_settings.md

[eglin73]

Thanks Jaap, that manual (and other discussions) are what I have been working through. There are unfortunately aspects I can't reconcile - the lack of 'goodness of fit' in my logs - I'll keep trying.

I will also try get a second card to try V1Beta.

Thanks
Ian

[Abasz]

You need to move to v1beta. The docs Jaap posted covers the V1Beta settings. The "old" ORM implementation does not use regression, hence you dont see GoF.

[eglin73]

Ok, thanks for that context. I'll see what I can do about upgrading.

I had seen #118 and it looks straightforward enough, but I've found that my inexperience means things that should be/are simple if you know what you are doing, are not for me and so I didn't think I'm a good candidate for a beta version.

[JaapvanEkris]

I had seen #118 and it looks straightforward enough, but I've found that my inexperience means things that should be/are simple if you know what you are doing, are not for me and so I didn't think I'm a good candidate for a beta version.

V1Beta is still in Beta because we need to find an easy upgrade path for existing users. The new V1 engine is much more robust against noise and configuration errors and the upcoming update will make it much forgiving for beginning users as well. From a code maturity perspective, V1 also has made much more meters than 0.8.2 has, under better monitored conditions, across more machines. So in all honesty, I think that V1Beta is much more suited for new users than the current 0.8.2 (and that is aside the much cooler features @Abasz added in the latest GUI and ANT+ update).

[eglin73]

That sounds great, looking forward to the new features - I see exciting things in the logs about V02 too. As someone who has only just dipped a toe in the water, I don't know how you guys do it but it is impressive! And much appreciated as is the help and patience with people like me!

[eglin73]

Hi, I have upgraded to v1Beta, and I have a working config.js which is producing plausible results on the monitor.

I have followed the guide, enable raw data files and am looking at them in excel (well Google docs spreadsheets). I have several raw data files, and I'm at the stage in the guide of visualizing the data, but my raw data files don't contain impulse delta times. I have read in one of the threads that the excel has been discontinued, maybe the delta times are no longer collected, but I'm just trying to work through the guide step by step.

I've attached two raw data files, one with the dragfactor static and one with it auto adjusting.

Looking at the log, I now see goodness of fit, but I do still see errors which are similar to before the upgrade (ie minimum recovery time not reached)

log_18_04_2023_1723.txt
2023-04-18_06-46-35_rowingData.csv
2023-04-18_07-01-02_rowingData.csv

[eglin73]

Thanks for your replay, yes of course. It is a Johnson JAR 5100. Quite an old air rower and not very popular as far as I know. Happy to share the settings, but not sure how many people might find it useful.

I believe I have enabled the generation of raw data files in the config.js. with the line createRawDataFiles: true,

Once I did that files started appearing in the data/recordings/2023/04 directory (only the .tcx files were there before I added that line).

My config.js is below.

export default{
loglevel: {
default: 'info',
RowingEngine: 'trace'
},
rowerSettings: {
numOfImpulsesPerRevolution: 3,
flywheelInertia: 0.15,
dragFactor: 550,
createRawDataFiles: true,
autoAdjustDragFactor: true,
autoAdjustRecoverySlope: true
}
}

I guess my question really is whether these errors (information) are a concern and I should continue to work through the manual Jaap has provided or whether they are as you say just informational and not something to be too concerned about.

I'm trying to confirm the data is good enough to focus on whether the moment of inertia, drag factor and magic constant are correct, hence why I was trying to visualise the data as suggested in the manual.

Thanks
Ian

[eglin73]

Thanks for clarifying the excel Jaap, I will move on with the manual!

Ok, so I have turned off the autoadjust features as you suggest.

export default{
loglevel: {
default: 'info',
RowingEngine: 'trace'
},
rowerSettings: {
numOfImpulsesPerRevolution: 3,
flywheelInertia: 0.15,
dragFactor: 550,
createRawDataFiles: true,
autoAdjustDragFactor: false,
autoAdjustRecoverySlope: false
}
}

I have 3 magnets and have not adjusted the flank length (or anything else apart from the flywheel inertia, drag factor and number of impulses).

As to why it is noisy, I'm not sure. My mounting of the sensor could be a little better once I get it working I will make it more permanent but visually it looks good - an LED lights up every time one of the magnets pass. At slow enough speed this seems to indicate it picking up each impulse correctly.

Thanks

[JaapvanEkris]

Thanks for your replay, yes of course. It is a Johnson JAR 5100. Quite an old air rower and not very popular as far as I know. Happy to share the settings, but not sure how many people might find it useful.

It is quite a nice looking machine. You never know if people have a similar machine lying around, giving it a very usefull second life.

export default{ loglevel: { default: 'info', RowingEngine: 'trace' }, rowerSettings: { numOfImpulsesPerRevolution: 3, flywheelInertia: 0.15, dragFactor: 550, createRawDataFiles: true, autoAdjustDragFactor: true, autoAdjustRecoverySlope: true } }
I believe I have enabled the generation of raw data files in the config.js. with the line createRawDataFiles: true,
Ah, there are some things going on here. You must place the createRawDataFiles outside the rower settings (thank you for helping us improve our manuals :)):

export default{
loglevel: {
 default: 'info',
 RowingEngine: 'trace'
 },

createRawDataFiles: true,

rowerSettings: {
 numOfImpulsesPerRevolution: 3,
 flywheelInertia: 0.15,
 dragFactor: 550,
 autoAdjustDragFactor: true,
 autoAdjustRecoverySlope: true
 }
 }

May I suggest to add flankLength: 3, to the rowersettings as well, making it clear what the flanklength is you use?

I'm trying to confirm the data is good enough to focus on whether the moment of inertia, drag factor and magic constant are correct, hence why I was trying to visualise the data as suggested in the manual.

The output should be generated once you move the createRawDataFiles flag.

[eglin73]

Great, thanks. I will give that a try.

Yes, it is quite a nice machine to row on - and more affordable than a Concept2 :)

[Abasz]

Since you have 3 magnets I would start with a flnakLength of something between 6-8 (approx. 2.5 rotation).

[eglin73]

I tried first with a flank length of 3, and the log looked pretty much the same, and then with a flank length of 8 (thanks Abasz) and it seems like that particular error is not present in the logs following that change to 8.

2023-04-18_08-39-06_rowingData.csv
2023-04-18_09-24-55_rowingData.csv

2023-04-18_08-39-06_rowing_log.txt
2023-04-18_09-24_rowing_log.txt

Interestingly the csv files don't look any different since I moved the createRawDataFiles: true, out of the rowersettings. My current config.js is below

import rowerProfiles from './rowerProfiles.js'

export default{
loglevel: {
default: 'info',
RowingEngine: 'trace'
},

createRawDataFile: true,

rowerSettings: {
numOfImpulsesPerRevolution: 3,
flywheelInertia: 0.15,
dragFactor: 550,
flankLength:8,
autoAdjustDragFactor: false,
autoAdjustRecoverySlope: false
}
}

Thanks
Ian

[Abasz]

There should be a file named {DATE}_raw.csv but based on the log it was not created, but that is probably because you are not using the correct property name. You are using createRawDataFile instead of createRawDataFiles. It is plural and you are using it as singular.

[eglin73]

Thanks Abasz, I did not spot that error. I have changed it to the plural and have a raw data file now.

2023-04-19_10-14_rowing_log.txt
2023-04-18_23-10-14_raw.csv.gz
2023-04-18_23-10-14_rowingData.csv

[JaapvanEkris]

Thanks Abasz, I did not spot that error. I have changed it to the plural and have a raw data file now.

When I look at the log, it looks pretty clean!

[eglin73]

Yes, I thought so too :) I'm very happy with the outcome, thanks very much for the help!

I have documented (attached) some of the things I have learnt along the way as someone who is not a developer, not familiar with github, json, or with Rasberry Pi (or linux) but was interested in getting an updated monitor for an old machine.

Maybe they will help someone else in a similar situation, or save Jaap and Abasz some time dealing with others trying to do the same.

Or maybe the upgrade of main to V1.

Thanks again to those who have helped me on the journey. Much appreciated.

Working config.js

export default{
loglevel: {
default: 'info',
RowingEngine: 'debug'
},

createRawDataFiles: true,
// Strava
stravaClientId: "xxxxxx",
stravaClientSecret: "yyyyyyyyy",

rowerSettings: {
numOfImpulsesPerRevolution: 3,
flywheelInertia: 0.15,
dragFactor: 630,
flankLength:8,
autoAdjustDragFactor: false,
autoAdjustRecoverySlope: false
},
}

Open Rower Learnings (1).pdf

[JaapvanEkris]

Thanks for the feedback! I'll add some points to the manuals where needed. Slowly we will get there.

Jaap