Concept 2, RowErg

[JaapvanEkris]

I'm struggling to get a clear solution to connect a new Concept2 Rowerg to OpenRowingMonitor. As I understood from a post on the Concept2 forum (see https://www.c2forum.com/viewtopic.php?f=7&t=194719) the signal output is around 18V (!). As I understood from the Raspberry (see https://raspberrypi.stackexchange.com/questions/3209/what-are-the-min-max-voltage-current-values-the-gpio-pins-can-handle) the Raspberry Pi can only handle 3.3V and 0.5mA as an input signal., and to be seen as a raising flank it needs at least 1.8V.

I've looked at the solution for an Arduino (see https://www.instructables.com/ARDUINO-MONITOR-FOR-CONCEPT2-MODEL-B-C-D/), but they can handle 5 volts and use an analog input that they sample for reaching zero from a non-zero state (i.e. the measure the falling flank and they don't care what the exact value of 1 is).

Does ayone have an idea how to connect this all?

[laberning]

Hi Jaap,

on thing that might work is to build a small voltage divider (i.e. https://pimylifeup.com/voltage-dividers) and choose the resistors so you end up in the voltage range that the Raspberry Pi can consume.

If you want to play it safe for the internal GPIO-pins you can decouple them electronically i.e. by using an optocoupler (an introduction on optocouplers: https://www.youtube.com/watch?v=pYENAGK8qH4).

[JaapvanEkris]

Hi Lars,

Thanks for your good advice. Given the unfamiliarity of the Concept2's construction, I'll go for optocouplers. I'll chase down that avenue and see what it turns up.

Jaap

[JaapvanEkris]

I found this board: https://mschoeffler.com/2021/05/13/optocoupler-isolation-board-dst-1r4pn/

It is an interesting board as it seems to contain all necessary resistors and diodes/leds to transfer 16V to 3.3V. I choose to buy the 24V version (as there are no dedicated 16V versions), as a video (see https://www.youtube.com/watch?v=JvtKvXNsGQ0) showed that it will switch around 12V. I went for the DST-1R4P-P version (bought it here: https://www.plexishop.it/en/robotics-and-automation/converters-for-arduino/24v-to-3-3v-4-channel-plc-optocoupler-based-isolator.html) as this version seems to support 80KHz (which is 400 times faster than a Concept2 will produce). I'll report when all components are in and make a Wiki page when it works.

[carlito1979]

Hi there. Have you had any luck getting a concept 2 rower to work? I have a no name rower that uses the same sensor coil type sensor as the C2, but has a terrible monitor. So hoping you’ve had some success that you can share?

[JaapvanEkris]

Hi Carlito, know the feeling. OpenRowingMonitor works very well with the Concept2 coil, and the new rowingengine even is very close to Concept2's results (i.e. within 0,3%). It is very reliable and I've rowed around 300KM on it without issues.

Please note that Concept2 silently changed coil a couple of years ago, where the new coil delivers around 15V on the middle pin (essentially charging the monitor) and 15V signal on the point. For that, you need to use this board: https://mschoeffler.com/2021/05/13/optocoupler-isolation-board-dst-1r4pn/. The wiring schematics can be found here: https://github.com/JaapvanEkris/openrowingmonitor/blob/Sandbox/docs/img/Concept2_Optocoupler.jpg

I'm not familiar enough with the electronics of the previous coil they used, so some measurement might be needed. When you read this post, you see that the old coil gives around 0.6V, which is too low for a RPi to pick up. So you need some low-voltage circuit to amplify this signal to the 3.3V the RPi expects. If you have this kind of monitor and have electronics that works, I would be very interested in this to include it in the manual.

[carlito1979]

Thanks for the response JaapvanEkris. I've ordered a raspberry pi, which could take some time to arrive given the supply chain crunches at the moment. Just to clarify, you have a concept 2 model D hooked up? What rower settings are you using? Are you using the default settings from rowerProfiles.js or have you set up your own C2 profile that isn't in the source code? I couldn't see one in your branched version of the code either?

So curiously when I first tested the voltage output from my rower (this model https://australianfitnesssupplies.com.au/products/force-usa-r3-air-rower) I was getting values close to 15 volts, but since I opened the rower up and detached the sensor coil it now reports no more than 8 volts. In any case it does still work with an ESP32 based rowing monitor (https://github.com/zpukr/ArduRower), but that is seriously limited compared to what this project could do.

As an aside, one of the reasons I bought my rower was because it claimed bluetooth compatibility with Kinomap (the other reason was it was about half the price of a C2). There is a clear bug with the implementation of the BT interface in the monitor which means it doesn't report speed or distance properly (off by a factor of 5) which makes Kinomap completely useless. The supplier was initially very helpful (even sending me a second monitor) but they've stopped responding to me recently, as I think they've realised it's a firmware/hardware issue that can't be easily resolved.

This is what has led me down the path of looking at alternatives and this one on paper looks pretty promising, even if I don't quite understand the disagreements yourself and Laberning have had over some of the calculations.

[JaapvanEkris]

Thanks for the response JaapvanEkris. I've ordered a raspberry pi, which could take some time to arrive given the supply chain crunches at the moment. Just to clarify, you have a concept 2 model D hooked up? What rower settings are you using? Are you using the default settings from rowerProfiles.js or have you set up your own C2 profile that isn't in the source code? I couldn't see one in your branched version of the code either?

I test it on a Model D (now RowErg) on almost a daily basis. There is 0,3% deviation I don't like, so I'm tweaking a bit of the code. As it is a rewrite of the physics engine, some settings are gone, some have changed. I could upload it to the as-is-branch you if you like, but retrofitting is tough. But most important settings are the flywheel inertia (0,1001 or 0,1015), the number of flanks (10, maximum errors 2), and the number of magnets (6). Minimum time between impulses is around 0.005, maximum around 0.022.

My new rowingengine is still based on the 0.7.2 version of ORM, and I have to retrofit a lot of the Engine to the 0.8.2 functionality, which is pretty tricky. It is on my to-do-list, but it is a complex project. Using these settings, is a bit less accurate and robust than the PM5, but it will still beat most monitors out there.

As an aside, one of the reasons I bought my rower was because it claimed bluetooth compatibility with Kinomap (the other reason was it was about half the price of a C2). There is a clear bug with the implementation of the BT interface in the monitor which means it doesn't report speed or distance properly (off by a factor of 5) which makes Kinomap completely useless. The supplier was initially very helpful (even sending me a second monitor) but they've stopped responding to me recently, as I think they've realised it's a firmware/hardware issue that can't be easily resolved.

I don't use KinoMap, but as I understood it, Lars does. I use EXR which is pretty nice and works quite well both on FTMS and the PM5 protocols (tested them both).