Mtracer

I recently bought a set of Favero Assioma Pro MX power meter pedals. I've used their road version Assioma Duos for 4 years (with MTB body hack) without issue. In fact I have two sets of those that track well and also matched my Zwift Hub One indoor trainer well.

But the Pro MX versions are reading 8-10% high compared to the Zwift trainer. I'm using my Garmin Edge 530 to record and display the Pro MX power while the Zwift app is using the Zwift trainer power.

I have checked the static weight test of all of my Assioma pedals, including the new Pro MX and they are within specs (+/- 1%). I have triple-checked that I have the crank arm length set correctly in both the Assioma app and within the Garmin sensor settings. In my case 165.0 mm crank length. I have of course done the manual calibration of the pedals.

I noted today that Favero had an update for these pedals from version 4.27 to 5.04 (could have been 5.4). But I cannot find any information on what this upgrade may have addressed. Possibly it has resolved this issue. I'll see on the next indoor ride.

But I'm wondering if anyone has an experience with the new Favero Assioma Pro MX pedals or knows anything about this recent update. Or of course if there is an issue with the Zwift Hub One trainer that may explain this. Though I think the Pro MX pedals are reading too high based on how they feel on outdoor rides and my heart rate. But, with the variability of outdoor rides, I can't be sure.

RChung

I don't have any experience with this model but can you swap, say, the right pedal from your old set and pair it with the left, or vice versa?

You've done a static test (which is good) but the other two things to check would be that the imputed cadence is correct, and if it is, then to do a dynamic test.

Mtracer

You cannot mix the left and right pedals. They are two different pedal models. Though thinking about it, normally the right pedal data goes to the left, and from there the left pedal transmits both pedal's data to the head unit. But there is some mode that transmits each separately. But I think this is to work with some unusual head units.

The static test is per the Favero calibration procedure using a known weight etc. There is no dynamic test other than as I've done comparing it to another power meter.

A cadence error is certainly another possible error source. It's hard to see how that could be off. I would think that would require the timing in the pedals to have an error. I assume it uses the pulsing of the force data to detect the cadence. Of course a firmware bug is always a possibility. But that would be such an obvious bug, it's hard to imagine that getting out the door. But of course sometimes last minute software changes get made thinking they can't possibly affect something else, and they do.

I've sent a message to Favero support asking about this. I'm hoping they will come back with, "There was a bug, we fixed in in that latest update". Favero is known for making really good power meter pedals, and that's been my experience with two other sets. But this is a completely new model, so there may be some issue yet to be ironed out.

Also, I'm not necessarily sure that the resulting power data is just a number that comes out of the pedal. I think the head units get involved. For example, I have to set the crank length in the Garmin. So, I'm pretty sure the Garmin is doing some of the computation. So, it's possible there is some bug in the Garmin. Though why it would be specific to these pedals, I have no idea.

RChung

This is what I was wondering about. The left pedal is the "master," the right is the "slave," and I was wondering whether you could separately slave a different right pedal (or left).

Right, you'd have to do your own dynamic test. I've done that but I don't recommend it for the faint of heart. It's possible for the desperate, though.

The Assioma and other pedal-based PMs use accelerometers to determine cadence. Accelerometers have to know which way is up so they can determine when the pedal stroke is complete, and accelerometers can be noisy, so sometimes they can be off -- especially when the road is rough. (This, btw, is the reason why the original Garmin/Metrigear pedals had to be tightened to a specific torque). But if you have a dedicated (cheap) speed sensor, it's easy to figure out whether the cadence is off. That al said, this particular mode of failure tends to result in wider variance in power readings, not bias in the readings, so it's just for completeness' sake and if you need to do dynamic testing.

There are two different ANT+ standards for head units to talk to power meters: one sends power, and the other sends crank torque. When you enter the crank length into the Garmin, it transmits that info to the pedal so the pedal can do the multiplication and send back the power. A very rare problem is that the crank length doesn't get sent to the PM (or the PM doesn't receive it). One thing to check is to purposefully change the crank length to some other number and see if the power readings change by the correct magnitude.

Mtracer

Just thought I'd update this. The larger 8-10% error I was seeing simply disappeared. I strongly believe this was an error in the indoor trainer power (Zwift Hub One) rather than a change in the Favero Pro MX pedals. The Zwift Hub One auto calibrates anytime you coast for 5 seconds from some minimum speed. This is if you are NOT in ERG mode. I had done almost exclusively ERG mode workouts. So, it is very likely the trainer hadn't been calibrated in some time. This was made worse by having tested the Favero Assioma Duo pedals several months before I received the Pro MX pedals. So, there was plenty of opportunity for the Zwift Hub One to change calibration between the comparisons of these sets of pedals.

After intentionally calibrating the Zwift Hub One, the Assioma Duos still match more or less, though read 1-2% higher. This could happen and the units still be within their respective specs +/-1% Duos and +/-2/5 % for the Zwift Hub One. This difference also matches the cadence difference. The Zwift Hub One cadence reads 1-2 RPM (~1-2%) lower consistently. I also compared everything to a separate cadence sensor. The pedals matched this sensor very well and the Zwift Hub One reading 1-2 RPM low.

The Pro MX pedals are reading about 2% higher than the Duos, but I can only compare them indirectly. By this I mean I can compare each to the Zwift Hub One. I have done that with test rides that immediately following one and other using different bikes with the respective Duo or Pro MX pedals on them. Because of course I can't use two sets of pedals at the same time. But the difference seems to be consistent that the Pro MX read higher than the Duos by about 2%.

I'm looking to test the pedals on yet another bike that has a power meter spider on it. If the Pro MX pedals continue to read higher than the Duos, I'll be convinced they do in fact read higher. Unfortunately, this will require mounting and remounting pedals which may introduce other variables.

Favero has gotten back to me and requested I send them some test data for each. They want 5 minutes at 80 RPM. I plan to follow through on that, but want to make sure I have reliable data.

Anyway, I'm still concerned the Pro MX pedals are reading a little high, and even this difference may not be real. They certainly are not reading 8-10% higher as I originally thought.

RChung

Hmmm. 2% ought to be discernible if you do the right tests against the same standard, but you'd have to be reasonably careful.

Mtracer

That's the rub, the standard I have is the Zwift Hub One trainer. While I don't need it to be accurate, I due need it to be consistent and I simply have no idea how consistent it really is. But that's why I plan to try comparing it to a spider based power meter I have on an MTB. But as mentioned, then I introduce other possible variables having to swap the Assioma Duo and Pro MX pedals back and forth on the same bike.

But if I get similar results enough times, I'll start to get confidence in the results. Of course, what I'd really like to see is the opposite. I'd like to conclude the Pro MX pedals are fine. And honestly, if I hadn't messed up the Zwift Hub One calibration making me think I had a much larger issue, I likely wouldn't be looking this closely at this.

RChung

A penny weighs 2.5g, so $4 of pennies weighs 1 kg. I'm guessing that's a bit more than 1% of your total mass, so if you have a steepish hill you can climb it wth a few rolls of pennies in a water bottle, then without. Weigh yourself, bike, and pennies, then compare the virtual profiles you get from Golden Cheetah's Aerolab with and and without pennies. This method doesn't require that you compare with the Zwift Hub One. It will help if you know the total height of the hill.