Bench Testing a PM5

[cakeman]

Hello all..

I just inherited what looks to be a well maintained, but well used, model C with a PM5 on it (hardware rev 502).

The previous owner replaced the sensor w/cable for no counting, apparently that didnt fix it.. but judging by the install job, they might not have installed it correctly hah. I was going to pull the PM5 off and take it into work on Monday, to do some bench testing..

I thought I'd do a little advance planning and ask what signal the PM5 is expecting.. so I can fake it with a signal generator and see if the PM5 or the unit is at fault..

thanks..

[Ombrax]

I thought I'd do a little advance planning and ask what signal the PM5 is expecting.. so I can fake it with a signal generator and see if the PM5 or the unit is at fault..

I"m 99.99% sure that Carl Watts has answered this question before - try searching for posts by him that contain "sensor" and you'll probably find it.

If you can't and for some reason you don't get an answer to your question, a simple way to test the PM5 would be to connect it to a different erg.

Finally, did you check to confirm that the PM5 was set to expect a Model C? That's easiest check of all...

Good Luck

[cakeman]

Thanks, will re-do a search a few times to find the post.. but..

.. your third point is an even better idea, as I seem to recall having quickly poked through the menus, it was set to Model D. Will try that tomorrow

thanks..

[JaapvanE]

I thought I'd do a little advance planning and ask what signal the PM5 is expecting.. so I can fake it with a signal generator and see if the PM5 or the unit is at fault..

I'm the lead developer of OpenRowingMonitor, and I looked at the C2 signals as a recipient of them (essentially doing what a PM5 also does) and several people have measured these signals. At ORM we can also record the data and replay it (as testing otherwise would be a physical nightmare). In general, decent rowing monitors require a pretty complex signal which is tough to simulate.

A PM5 in model C mode (without the generator) expects 3 pulses per full flywheel rotation. But it requires specific behaviour: an acceleration during the drive, a consistent deceleration during the recovery. The electrical properties are described here: https://github.com/laberning/openrowing ... ssions/157
We made recordings of several strokes here: https://github.com/JaapvanEkris/openrow ... odel_C.csv , each number represents the time between two consequtive pulses in seconds. Perhaps you can use this to simulate this.

A model D (or RowErg) with the generator expects a sinoid signal between 0 and 15 volt, with 6 sinoids per full rotation which are slighly modulated in such a way that at stable flywheel speed behaves like a sinoid per full flyheel rotation. And the flywheel has to accelerate and decelerate in a specific way during drive and recovery. For our use, we electronically transform the sinoid into a blockwave and measure the time between upgoing flanks. So our recording reports the time of each full sinoid. This recording can be found here: https://github.com/JaapvanEkris/openrow ... meters.csv

[Carl Watts]

You can fake the model C sensor with the sensor plugged into the monitor and just wave a magnet across the face of the sensor. A decent magnet will turn on the monitor on from about 25mm or 1 inch away from the face of the sensor. Make sure the PM5 is set to the Model C.

If you pull the handle on the model C and the monitor turns on but doesn't show the correct numbers of doesn't change then one or two of the three magnets in the flywheel are defective.

You are not going to fake the signal from the Model D sensor anytime soon, I didn't bother trying to rig something up like that, I kept a whole Model D rower to do that, its quick enough to swap a monitor over or do what I did run two cables from the one sensor and you can run two monitors at the same time to cross verify them as working.

The model C sensors seldom fail, its just a 105 Ohm coil of fine wire. The Model D sensor electronics never fails, its always the cable at the right angle connector end that fails when one of the 3 wires goes open or intermittent. The latest PM5 firmware helps a little, it can tell you the sensor is disconnected.