Ant+ FE-C broadcasting? #71
Replies: 11 comments 98 replies
-
|
Hi Carlito, I don't know if this is possible (I am not familiar with the ANT+ implementation as that is Lars' cup of tea. A workaround might be to upload your data to Garmin directly from ORM. As a workaround, the tcx file can be uploaded by hand, or automatically by using tools like https://github.com/La0/garmin-uploader I use RowsAndAll as an intermediate stage (you can e-mail workouts with much more detailed info in the csv), which can upload to Garmin as well (my current route). |
Beta Was this translation helpful? Give feedback.
-
|
Garmin supports any rowing machine with ANT+ and can capture
You can't even imagine how many of us are looking for such connectivity with our current not-so-smart rowing machines. |
Beta Was this translation helpful? Give feedback.
-
|
I’ve had a look at doing this, and it’s currently beyond my understanding, and I don’t have time to invest in learning both how to get the Ant+ package to broadcast Ant+ FE-C information, and also to learn how modify the openrower code. I’m also super keen to understand how to improve the concept 2 code so the rower can work with the ErgZone app. So even if I find time I’m still not sure which of these changes I’d want to focus on. Also, it looks like laberning hasn’t worked on the code for some months now. |
Beta Was this translation helpful? Give feedback.
-
|
Ant+ is a bad option regardless. I've played with it (I own a Concept2 and Garmin Venue), and it is a huge dissapointment. See http://runnermate.eu/garmin/concept2/rowing/2020/07/13/Concept2_GarminConnect.html for other people who tried and had the same experience. Stroke distance isn't recorderd by Garmin, regardless how you upload through ANT+. Garmin itself seems to abanon the idea of ANT+ as well: some (more expensive) Garmins also support BLE now, wich works much better (see http://runnermate.eu/garmin/concept2/rowing/2022/07/20/Concept2_reloaded.html). I personally hope they will bring this to more affordable models. The best approach is to upload a tcx to Garmin directly. I did that for over a year, and it can be automated using the script I mentioned above. There is tooling known to create fit files: https://github.com/inonoob/Coxswain2Fit. However, there are some subtle differences in the tcx-file this tool expects and we produce, so I never got a stable output. As it didn't provide a benefit for me, I abandoned that path. When you upload through this, the only thing that goes missing is the total stroke count, which is a known issue on Garmin's side (and we are not the only ones struggling with this). |
Beta Was this translation helpful? Give feedback.
-
|
I have a forerunner 255, it supports ant fe-c protocol. Every smart bicycle trainer supports this protocol. I don't think Garmin is abandoning this at all. Actually in my opinion Garmin determines the supported sensors based on the watch type (e.g. multisport, triatlon and cycling watches support the fe-c in general while smart watches and dedicated running watches do not). A tcx file does not scratch the surface what a Garmin fit file holds in terms of data (not to mention the health metrics recorded by a Garmin watch). |
Beta Was this translation helpful? Give feedback.
-
In theory, Yes. But in all honesty, I am extremely disappointed by Garmin in that sense. When I look at http://runnermate.eu/garmin/concept2/rowing/2022/07/20/Concept2_reloaded.html, I see several data elements that are exchanged and reported between a PM5 and a Garmin Fenix 6:
The last element is new, but can easily be calculated from substracting the distance between the individual strokes (the TCX already contains this information as datapoints are per stroke). So, it isn't that special to be honest. If you really want detailed data from your rowing, you need to use RowsAndAll/RowingData format, which can be read in Excel as well (I have added this as an export option in the Sandbox). That is much richer and surpasses the PM5, even with CoxSwain or PainSled attached. In collaboration with the owner of RowsAndAll/RowingData, the dataformat also provides in-stroke PowerCurve, ForceCurve and SpeedCurve which can be dynamically analysed (currently in Beta both on my end and their end). |
Beta Was this translation helpful? Give feedback.
-
|
Open Rowing Monitor currently supports broadcasting the metrics via different BLE protocols (two FTMS protocols and parts of the C2 PM protocol). These can be broadcasted with Raspis native BLE chip. It might be possible to broadcast via Ant+ FE-C as well (with additional Ant+ Hardware attached to the Raspi). But to be honest that's probably nothing I personally would invest time in, as I don't use devices that require Ant+ FE-C. Open Rowing Monitor does not generate Improving the Concept2 PM implementation is definitely something that would be great! This proprietary protocol is kind of hard to nail correctly as the spec leaves some room for interpretation. Also I haven't implemented the whole spec, as there are some really fancy metrics inside it that are difficult (if at all possible) to extract from the physics engine. Different apps use different parts of these metrics so it is very app specific if the integration works or not. However there is definitely room for improvement, feel free to experiment with it and report or fix stuff that you find. |
Beta Was this translation helpful? Give feedback.
-
|
You are absolutely correct with the ant fe-c, my comment was referring to the fact that concept2 has its own ble protocol that is partially implemented by ORM. But that is not native to Garmin. A good workaround that I use to over come this is CPS where you set the wheel circumference to a value that you calculate from the drag factor. |
Beta Was this translation helpful? Give feedback.
-
I have "some" improvements in the works (see https://github.com/JaapvanEkris/openrowingmonitor/tree/Sandbox and especially https://github.com/JaapvanEkris/openrowingmonitor/blob/Sandbox/docs/Architecture.md), and I can add some additional metrics to the PM5 interface if you want. Some are already added as I could easily calculate these metrics using the new RowingEngine. Some reported metrics are dependent on interval settings (like projectedWorkDistance or projectedWorkTime) and thus depend on the session object that has to be filled before they become active. My current implementation doesn't require the session object to be filled, and it will fall back to current behaviour if it is empty. The issue I have is with AdditionalStatus2, where the Split is also reported, which isn't defined in my session object yet (I could do that, but then I actually have to add it to the tcx as a lap as well, which is not my cup of tea). Setting the session object is a different ballgame and way out of my comfort zone: it requires either input from the PM5/BLE interface (this is what ErgZone does on the PM5) or user input on the GUI directly. The last one is a big requirement for me. My current implementation isn't in the way for normal users, but it isn't production ready either because to set a target you have to modify server.js to modify the defined object manually. So if anyone can help out there, we could make this a new feature (and remove it from Lars' backlog). |
Beta Was this translation helpful? Give feedback.
-
|
The Fenix 6 should also support BLE, and setting ORM to PM5 or FTMS should deliver you a set of sensors. |
Beta Was this translation helpful? Give feedback.
-
|
That's right, on Epix 2 I can search for ANT+ FE-C smart trainer and it connects to it too, also the ANT+ chest HR meter can be connected. I've tried the same with all of the BLE profiles the ORM offers, but without luck. According to Garmin, this watch can also connect to a BLE sensor but somehow the FTMS profile is not supported. Tried also with ArduRowerIQ app from CIQ but without luck too. I've tried Wahoo Fitness App, MyHomeFit and even TrainerDay and all of them connect to my Rpi thingy. But the Garmin watch does not. Funny thing, but the Concept2 app finds the ORM too with the serial number 12345678 and when connecting to it, it crashes to the Android main screen. So a native ANT+ support would be a nice thing. |
Beta Was this translation helpful? Give feedback.
-
|
I am not aware of any Garmin watch that supports BLE FTMS natively (i.e. without a CIQ app). I believe this is a concise design choice from their side as they are pushing their own protocol (ant+) that enables 1 to many connections (for which Garmin thinks ant+ is superior...). I spent considerable time looking around for any kind of example of an ant+ fe-c server project with no real luck (I was able to find clients but never a server). So this probably would need to be implemented from scratch. I believe longhorn's project provides the necessary API, but the fe-c profile is not the simplest unfortunately... :) |
Beta Was this translation helpful? Give feedback.
-
|
Seeing that ANT+ broadcasting is competely uncharted territory here, I don't consider it a viable option. I would be the first to provide a data structure to test, but in all honesty I'm not looking forward to reverse engineer yet another propriatary protocol (PM5-BLE has proven to be tough enough as it is from a data elements perspectve, where Lars found/adopted a great base product to begin with and that isn't even working flawlessly yet either), especially if nobody has documented anything about it. What is the goal here? Get as much data into Garmin connect as possible? I get the feeling people get stuck in a specific way of getting data to Garmin, instead of the higher goal to just get data into Garmin. The ANT+ IMHO is a disappointing strategy to begin with as the datafield extra isn't that interesting to begin with, and my experience wth ANT+ connections between ORM and Garmin is dissapointing to say the least. Because it might be doable to create a FIT file as some people have reverse engineered the rowing elements of FIT already and perhaps even desirable as several sites also accept FIT and some additional fields might be interesting. When I manually upload FIT file from C2's logbook to Garmin, all recorded fields are nicely filled, suggesting that these fields fro the FIT are indeed used. What's the opinion about such a strategy here? |
Beta Was this translation helpful? Give feedback.
-
I also would like to set session parameters, either through the GUI or a BLE-connected app like ErgZone. My experience is that it sounds cool, but the only app that really works flawlessly and quick with a PM5 is the new ErgData app. ErgZone sometimes can't find the PM5, or it takes a long time to load a straightforward timebased training into the PM5. EXR can't push the training schedule to the PM5 (yet?), so there isn't much to learn there. But somehow being able to set a target time or distance would be an improvement for users.
I agree that Garmin is shrouding itself in obscurity here.Please note that several are reverse engineering the FIT files, so the issue with unknown fields isn't just our problem. For example, at https://github.com/scrapper/fit4ruby/blob/master/lib/fit4ruby/GlobalFitDictionaries.rb they already identified where VO2Max is residing, as well as HR Recovery (I think Garmin typically uses HRR1). And even for rowing specific fields, once we have an example file, we could decode it with their own SDK toolkit and add these fields as these files are self-contained. As long as you know what to look for (a value for VO2Max for example), it should be doable. Could be that Garmin checks things somewhere, but it would be worth a try.
There are some other projects, most notably https://github.com/inonoob/Coxswain2Fit which produces indoor rowing specific FIT files, but they are very specifically tuned towards CoxSwain's TCX-output, which isn't compatible to ours (ORM has more precission in several fields). It is written in Python (never could get the hang of that one) without the SDK present, and it produces valid FIT files, so it can be done. https://github.com/mayfield/jsfit is another tool, which is in Javascript and seems to set the basics. https://github.com/trustao/gpx2fit is a TypeScript tool (close to JavaScript, but Lars can probably tell how close they are in practice), that can actually add "record" elements (i.e. Trackpoints in TCX terms). I haven't tried any of them, but they are bits of pieces that could provide a way to something working.
Could be an option, but mapping the fields isn't trivial, as you have shown. There we probably need some guidance either from you (how to map our internal fields like Drag to Rotation Distance). I would be willing to help with the data mapping, but here Lars would have to decide what library to use/extend, as we are already using a BLE library. |
Beta Was this translation helpful? Give feedback.
-
IF we know the distance between the magnets, why can't be just the amount of impulses calculated as distance? When the flywheel is sipinning down, it rotates too and those ticks can be counted as well. |
Beta Was this translation helpful? Give feedback.
-
The problem isn't counting ticks, it is the issue that in rowing the distance per rotation/tick varies as we make a transition between angular (rotational) to linear distance where the dragfactor is involved. In cycling, the ratio circumfence to linear distance travelled is fixed (and even 1:1), so consuming devices probably won't expect a varying measurement of that activity (although Abasz could shine some light on that). |
Beta Was this translation helpful? Give feedback.
-
|
What I meant here is that Garmin Watches only read data from a power meter sensor when you select a cycling activity (i.e. if you go with an indoor rowing, the watch will not read the data from the sensor). But the caveat of selecting cycling activity is that the Garmin Watch does not calculate the distance per stroke (i.e. does not include it in the FIT file generated) and hence it will not be visible in Garmin Connect even after changing the activity type there. Of course if you start an indoor cycling activity on the watch and you set the wheel circumference according to distance equivalent of one rotation, the value which should derive from the drag factor (with this formula: Math.pow(102 / 2.8, 1/3) * (2.0 * Math.PI) * 10; i.e. a drag factor of 102 is approx 208mm) speed and distance will be calculated by the watch. Here is an important distinction (or limitation) compared to an FE-C profile: distance is calculated by the watch based on the reported number of wheel revolution in the CPS Measurement Characteristic (i.e. distance is not advertised directly, only the number of rotations of the wheel and a time stamp of the last rotation), while in the FE-C distance and speed is actually calculated by the server and the actual values are advertised to the watch (i.e. these are not calculated by the watch it self merely registered) |
Beta Was this translation helpful? Give feedback.
-
That is correct. With CPS there is no way to change the "distance per rotation" from stroke to stroke (when the DF changes). This results in potential deviation on machines where DF fluctuates (e.g. my Copy Concept 2 has 2-4% higher DF on slower speed than on faster, that is caused by bearing drag, inaccuracies in the assembly, unbalanced flywheel etc). Note this does not seem to be an issue on a real Concept 2 that is very well assembled machine and the changes to the DF from stroke to stroke is very subtle (at least based on my limited tests I was able to conduct on the real C2). Again in my view this is an acceptable trade off, I am still testing with various data but the distance error with this approach seems to be less than 1%. |
Beta Was this translation helpful? Give feedback.
-
|
I figured out how to report correct distance (that is conscious about the changes to the drag factor) via BLE cycling speed and cadence service and cycling power service. This would help people that want to record activities on their smart watches which do not support FTMS (e.g. many Garmin watches), without having to implement ANT FEC. Its surprisingly straight forward, I will create a pull request in the upcoming weeks. I have one question: should i create it against Jaap's version or to the original ORM? The implementation is fully independent of the rowing engine. They work by broadcasting certain data that allows the client to calculate the necessary metrics:
The actual distance, speed, cadence etc. metrics are not exposed directly but rather converted to "wheel revolution count" in a way that when the "wheel circumference" is set to the appropriate length on the client (e.g. to 10mm) it can actually calculate the speed and distance by looking at the deltas of the rotation count that occurred over a known delta time. The only limitation (at least on the Garmin watches) is that those watches only connect to these sensors when cycling activity is selected, but the activity type can be changed to rowing in Garmin Connect. Alternatively the FIT file produced by the watch may be manipulated directly (for this I actually created a nodejs UI app that uses a C# console app implementing the FIT SDK). Of course this is a workaround but learning Ant+ FEC with zero experience turned out to be quite a pain due to the lack of good examples... (I have not given up on it just not quire there yet :) |
Beta Was this translation helpful? Give feedback.
-
|
@JaapvanEkris I have a few questions on the metrics exposed to the BLE module: As I said I would need to expose a few timing data so the client can calculate the metrics. Can you please confirm the following:
Your help would is very much appreciated |
Beta Was this translation helpful? Give feedback.
-
And I'm very glad to answer them. Your contributions have proven to be very valueable. Before I dive into your question, let me explain the architecture a bit (some additional descriptions and pictures can be found in https://github.com/JaapvanEkris/openrowingmonitor/blob/Backend_Redesign/docs/Architecture.md). All flywheel data and dependent metrics are updated whenever possible, so a gpio-alert indeed messages
Assuming this would be an addition to the already supported BLE profiles, you could add it to the BLE profiles (see
Please note the disconnect here between the flywheel impulses and the updates of the BLE interface (both time and state driven): the BLE interface only gets time and distance updates every x ms, and when the stroke state changes.
Not at the moment. If we start introducing sessions with fixed intervals that might change, but I think we then should maintain both a total and interval distance. For some things it is really helpful if there is a time and distance always moving in direction.
We could expose it, as |
Beta Was this translation helpful? Give feedback.
-
There is no need to differentiate between states (these profiles do not care about state, that is fully managed by the client i.e. the watches)
Exactly what I am doing. These will be new profiles selectable in settings like 'CPS', 'CSC' (instead of 'FTMS') Should be plug and play
That is exactly what is needed, no data change should be broadcast while stopped (unchanged data can be broadcast, the client always look for deltas, apart from power where 0 power should be broadcast so the value resets on the watch.
I am not sure I fully understand this. Are you saying that these are not continuous at the moment? By continuous I mean that after a stop when rowing restarts these are not starting from zero but rather from the last vale (i.e. if we stop at 100sec and start after 20 sec the time is counted from 100 and not 120)
Clear, the above is great help. I did a fork and created a separate branch to host the changes better isolation of the code changes |
Beta Was this translation helpful? Give feedback.
-
So you probably map the state transitions to nothing (i.e. not sending any message), and only respond to the timed triggers then (i.e.
After a pause (i.e. somebody stopped rowing for more than The only exception is a reset of OpenRowingMonitor, which resets all metrics.
If you need some help, don't hessitate to ask. |
Beta Was this translation helpful? Give feedback.
-
|
Hi @Abasz , did some testing on your improvements. Great news: CSC works on my Garmin Venue, which doesn't officially support BLE. However, I have an issue with the distance recording: I simulated a 2K row on a Concept2, and according to Garmin, I did a 417,65 km cycle activity. I actually like all the badges I just collected, but a bit unexpected. Somewhere, we need some conversion :). PS: I had some modifications in the pipeline which needed to be cleared before I could test your improvements, so if you could merge them into your repro, that will save us some issues later on. PPS: Shouldn't we rename the profiles to "BLE Bike Speed + Cadence" and "BLE Bike Power"? |
Beta Was this translation helpful? Give feedback.
-
|
So I'm probably jumping the gun on this, but Abasz's work got me thinking again, and I'm probably going to jinx myself here, but I just successfully recorded a session with Open Rowing Monitor broadcasting metrics over Ant+ to my Garmin Watch I need to sense check a few other things and record a proper rowing session before I share the code, but here's the example activity: Anyways, exciting developments ahead! |
Beta Was this translation helpful? Give feedback.
-
|
Okay, I am ready to show the work I've done to have the Open Rowing Monitor broadcast rowing metrics over ANT+ This allows a user with a Garmin watch to connect to the rowing machine, and have the following metrics recorded as part of their indoor rowing activity:
All of these fields are recorded as "native" Garmin fields, which means you get all the usual comparison metrics as well (Time in Power Zones, distance per stroke etc) So I guess the main question is, Why do this? So why use ANT? Why not just use the FTMS BLE functionality? Okay, enough jibber jabber. You can find my code below. I'm not a developer by any means, so apologies in advance for the mess I've created. Also, as you may note from the URL below, this branch is using the new engine as developed by Jaap, so keep that in mind if you're wanting to try it out, because if you don't understand some of the changes introduced by the new engine your metrics will go crazy! Oh and also, full credit to the only example I could find of someone building an ANT Server goes to PTX2, who built a Bicycle Power Meter broadcaster, which formed the basis of what I did to build my Fitness Equipment server. You can find their work linked here If you want to try it out then when your rowing machine is up and running you'll need to first search for the Ant Server on your watch. For the Fenix 6 you'd do this by going to [Sensors & Accessories], then [Add New], then choose [Smart Trainer]. So, a few things to note:
So what's my ultimate end goal with this? @Abasz @Scx8R5SvSlYB3xcM0psn |
Beta Was this translation helpful? Give feedback.
-
|
ok, I played around with the UI, the code is not too complicated, I will implement an MVP that can be polished if need be :) |
Beta Was this translation helpful? Give feedback.
-
|
I'm looking at extending the session object and |
Beta Was this translation helpful? Give feedback.
-
Sorry, I only just saw this now @JaapvanEkris |
Beta Was this translation helpful? Give feedback.
-
Thanks, this is extremely helpful. What I see in the tcx is that the distance in the laps always continues, despite labs being inserted. This already gives me some hints how to create decent trainings in the tcx interface. It will take some time to crack the .fit nut, but I see a nice challenge for 2023 :) |
Beta Was this translation helpful? Give feedback.
-
I am able to create complex schedules now, all you need to do is fill the intervalSettings[] object in You can add several elements to the array, and thus add several intervals. I'll upload the changes after a week of testing comming thursday. |
Beta Was this translation helpful? Give feedback.
-
|
WOW! Chrismas must have come a couple of weeks early! WOW! I'll start testing today (the play phase) and see if I can incorporate @Abasz pull-request in a week or two (a have a bit of work in the coming weeks). I'll certanly will be playing with this. @carlito1979 : I'll start playing with this as well, but we'll have to figure out how to place it in the architecture. As BLE and ANT can coexist (by your design :)), and you indicate the ANT+ is powerhungry (probably it will ramp up the CPU cycles), it makes sense to isolate it as a seperate process. So spawning it into a dedicated server like you done now would make much sense. In working on the PREEMPT version, I stumbled on an approach of CPU isolation: basically you give specific processes their own dedicated CPU (as the RPI 4B has 4), isolating it from other disturbing factors. Giving the GPIO-thread (which is high frequency AND time critical) its own CPU makes sense, and putting a high CPU consuming server in a stand-alone service on a dedicated CPU also makes sense. But I think @laberning should have the final say in this approach. I can inmagine that he wants to do some code refactoring and review the message subscription model here. Currently, RowingStatistics.js is aware of the (two) types of clients (web en pheripiral) and their implementation specifics (i.e. their frequency). There is a case to be made to add a third, making it web, BLEpheripiral and ANTpheripiral and get on with our lives. There is also a strong case to be made to simplify this and fix it at the other end: RowingStatistics already has different types of messages it broadcasts across a messagebus, and the subscribers should decide for themselves what messages they accept and in what frequency. The later is a bit cleaner as it rids RowingStatistics from some implementation knowledge, but it potentially is much more CPU intensive for the pheripirals. |
Beta Was this translation helpful? Give feedback.
-
|
So the power value is native just the data field displaying it is not. The value is coming directly from the Activity.Info data that is available for Data Fiels in the SDK (I'm an not sure how familiar you are with connect iq ask). So I have native power in connect iq not the one flagged with IQ (https://apps.garmin.com/en-GB/apps/90ee803c-80d8-4248-8a6b-ccd8c95582e6 this is the link). So is not like when you install those apps you refer to. I use native indoor rowing activity and different data fields on the data pages |
Beta Was this translation helpful? Give feedback.
-
|
Did a quick check and your version works just fine, although the new buttons are causing the web GUI to go wonky on my 7” screen |
Beta Was this translation helpful? Give feedback.
-
|
Thanks, can you please reattach the picture, as it seems the link does not work. I am aware that the GUI needs more love... For now all it got from me is hatred:):)):). I will fix that in the upcoming days. Thanks for the feedback |
Beta Was this translation helpful? Give feedback.
-
Hahahahaha :) A small update of my status. I have updated the https://github.com/JaapvanEkris/openrowingmonitor/tree/Backend_Redesign branch, to include more complex workouts, allowing @carlito1979 to change these parameters. Most changes currently are limited to server.js and RowingStatistics.js. I also added a change to the PM5 code in GeneralStatus.js, where I replaced with I'll run a test this evening and then I'll put everything in a PR for @laberning to include in V1Beta. There are some ToDo's in this setup which aren't critical:
For now, all these outputs will ignore the interval-information in their reporting, but the GUI will respond accordingly. This will allow @carlito1979 to build the PM5 service interface and store the interval data (and use it!), and when we can take the rest from there. Looking a bit at the architecture and the Interval object, I realize that we omit one possibility that the PM5 really has: a PauseBefore option for each interval. I have some ideas, but I will not get around to implement them before mid january. We need the RowingStatistics to manage this information (allowing a countdown timer on the display and setting the correct session and interval state), but I need to find out how this interacts with the PM5 code, and integrate that in @Abasz his code. |
Beta Was this translation helpful? Give feedback.
-
|
I think I fixed the GUI (note I had to force push to the branch). It should look ok on every screen. |
Beta Was this translation helpful? Give feedback.
-
|
I completed the rework of the peripherals (and addition of the ANT profile) as well as tested the code and the different modes. There are two things:
What have been done:
Code is available here: https://github.com/Abasz/openrowingmonitor/tree/extend-peripherals A side note: at some point there needs to be a decision whether ES6 classes or original functions declarations should be used (my preference would be to refactor the whole code to ES6 classes as that is, in my view, cleaner to maintain) because it is a mix now (it was already a mix in the original version, so I kept the logic as is). Also the server.js is becoming a mess so that needs a refactor as well at some point, separating out features. |
Beta Was this translation helpful? Give feedback.
-
Sorry this was a reference to the messy PR you had in place that has been resolved so far :) I have not checked branch conflict with the new v1beta_updates branch. I will able to do it during the weekend. but you are right in that the package-lock.json gets messed up, but for this reason I dont include that in my commits (generally). I havent checked and I hope that this does not get contradicted :D |
Beta Was this translation helpful? Give feedback.
-
Yeah, it was easier than expected. I think I'll postpone the changes to the session object, as our other discussion would imply that some stuff gets moved outside RowingStatistics (HRR measurement) and session management gets moved in. So I might pick that up when you implement the new architecture pattern, and not before. But I have to look if it is a complex move. I do have one improvement planned for the V1Beta update: checking the config file and reporting errors when data is wrong. I have a working prototype, all I have to do is apply it to all settings.
I got some conflicts when installing on a cleaned directory. |
Beta Was this translation helpful? Give feedback.
-
|
Just a quick clarification or confirmation, when you say:
Do you imply that the cleaning up and setting up the server.js in line with the new architecture should be assign to me? :) not that I mind its just it may take a little longer as I need to familiarizes with the engine part of the code that I have not done yet :) |
Beta Was this translation helpful? Give feedback.
-
I think we should do this together. I like the concept, but I have no clue how to get the messaging structure working. I'll take care of the engine, PM5 and workoutrecorder parts as soon as I understand how the messaging structure works :) I guess we have to decide which events to broadcast, where most would originate from the RowingEngine. I could add them in another branch, as a copy from the "V1Beta updates". I would use messages similar to the current datamessages, just as eventmessages where the content is the event (and not containing any data). I could also expose both the simple and complete metrics through two functions. But how these functions should be wired to respond to a message request, is something I haven't done in Javascript before. I guess I could provide a single function (processDataRequest) with a case statement (distinguishing between the basic and complete dataset) that would broadcast the requested dataset, but I don't see how you only respond to the specific feature. |
Beta Was this translation helpful? Give feedback.
-
Completed the code change after a rebase to V1Beta, see #114 The pull request is sent, I hope @laberning will integrate this swiftly. As agreed with @Abasz , the interval management is now completely managed by 'RowingStatistics' internally, in anticipation of a cleanup of server.js. |
Beta Was this translation helpful? Give feedback.
-
|
@JaapvanEkris I did a test on the C2 with ANT+ Trainer on my Garmin FR255 with the indoor rowing profile, and total distance was spot on for me (there is 5m difference that is due to the fact that I stopped the activity on the watch a sec after the interval training finished).:
But splits did not transfer to the watch from the workout for me (but the C2 I used had old firmware actually so it is possible that there has been changes). On the watch the whole workout is one lap. |
Beta Was this translation helpful? Give feedback.
-
|
I manually installed your expansion on top of V1Beta-Updates. As my install is an install of unnown pedigree anyways (too much testing has been done on this one), why not play around more :). I'll focus on using that, as that is what we try to perfect. Approach was quite simple: just throw away ANT and BLE, and copy peripherals, update the webclient, server.js and config manually, and its working. I did the first tests, worked quite well, both on the Garmin and on EXR (tested a 6K both in FTMS and PM5 mode). I'll check if there aren't any missed changes (there is one: the PM5 serial number should start with a 43....., to make it consistent with the HW and SW revision, so I changed it to 431234567). |
Beta Was this translation helpful? Give feedback.
-
I did a rebase on the v1beta_updates (it was a clean rebase, so no conflicting commits). Finalized the feature, and I also updated the PM5 serial as per the above. I force pushed the changes to my git repo. It should be up-to-date and pretty final in terms of functionality. Of course there could be numerous things to clean up or make better but those are mostly architecture depended or I dont know whether other parts of the app uses it so I did not dare to touch to avoid possibly breaking something (these parts of the code will need to be cleaned before they legacy code slowly :D. I would do this following the finalization of the cleanup of server.js, etc). |
Beta Was this translation helpful? Give feedback.
-
You could create a pull-request to V1Betaupdates? This way we can easily track what is finished, and prevent a huge queue to merge into V1Beta. With the current acceptance rate, I guess a long queue won't help.
What parts are you concerned about whether it is used? Might it help if I looked at those parts? Do we need to look at package-lock.json to fix that one? |
Beta Was this translation helpful? Give feedback.
-
|
Did the PR to your forked branch I dont commit the package-lock.json in general, but an npm install will be necessary due to the new library. When that command is issued the lock.json is overridden automatically anyway. So there are parts like the status notification that is used by some of the profiles (but not by ANT or CSC or CPS) but implementation of empty handlers are still neccessary to avoid crash. This is due to the way the architecture was set up and the communication layers with different parts of the app. So for instance because the communication direction is from perihperalMenager to the peripheralFunctoin there are guards needed to avoid crash on an unimplemented function. function notifyStatus (status) {
if (bleMode !== 'OFF') { blePeripheral?.notifyStatus(status) }
if (antMode !== 'OFF') { antPeripheral?.notifyStatus(status) }
}Because the above would be better if peripheralManager subscirbes/unsibscribes to a message so when things are off there no need for the above guard. or this code in server.js, where basically there is an event deispatched on the peripheralManager and then a call back to the peripheralManager peripheralManager.on('control', (event) => {
switch (event?.req?.name) {
case 'requestControl':
event.res = true
break
case 'reset':
log.debug('reset requested')
resetWorkout()
event.res = true
break
// todo: we could use these controls once we implement a concept of a rowing session
case 'stop':
log.debug('stop requested')
stopWorkout()
peripheralManager.notifyStatus({ name: 'stoppedOrPausedByUser' })
event.res = true
break
[...]But anyway these can be cleaned up once there is a final decision and probably implementation of the new service.js structure. or there are thees logging things: bleno.on('platform', (event) => {
log.debug('platform', event)
})
bleno.on('addressChange', (event) => {
log.debug('addressChange', event)
})
bleno.on('mtuChange', (event) => {
log.debug('mtuChange', event)
})
bleno.on('advertisingStartError', (event) => {
log.debug('advertisingStartError', event)
})
bleno.on('servicesSetError', (event) => {
log.debug('servicesSetError', event)
})
bleno.on('rssiUpdate', (event) => {
log.debug('rssiUpdate', event)
}) |
Beta Was this translation helpful? Give feedback.
-
|
I have Openrowingmonitor connected to my Garmin watch via Ant+ FE. Is it possible to set FTP value for Power? When my heart rate gets close to max Garmin says that I'm in power zone 3 (152 to 181 watts Tempo) . |
Beta Was this translation helpful? Give feedback.
-
Please open discussions on https://github.com/JaapvanEkris/openrowingmonitor/discussions , as this repository is as dead as a dodo and hasn't accepted a PR in three years. |
Beta Was this translation helpful? Give feedback.
Uh oh!
There was an error while loading. Please reload this page.
-
Hi there. I understand if I get an ant+ usb stick I’ll be able to connect to an ant+ heart rate monitor, but I was wondering if anyone knew if it would be feasible to broadcast an ant+ signal with the fe-c profile so my Garmin Fenix 6 watch could pick up the data?
The PM5 monitor on the concept2 rower can do this
https://www.concept2.com/news/connecting-garmin-devices-to-pm5
I realise I can do this via the BLE FTMS implementation and using a ciq data field like ardurower to do this, but the ant+ fe-c signal is native and allows full integration of power, speed and pace into the indoor rower fit files.
Beta Was this translation helpful? Give feedback.
All reactions