I wanted to provide an updated review of the Sigeyi AXO Power Meter. I’ve had one on my Spark for a few months with flawless results. I recently purchased a second one for my commuter/HardTail for training at work on my dumb trainer. That also has been really consistent and reliable.
I recently did an accuracy test to my Direto X and both were within only a few Watts at all points up to about 380W. Admittedly I didn’t try any sprints or spend long above ~300W as it was on a rest day, but I was pleased with the results. (This doesn’t guarantee my accuracy I guess, but at least all my devices are as accurate as each other).
What follows is a brief comparison of power data from a recently purchased - the Sigeyi AXO (firmware - 3.802) paired to an Edge 130 head unit - with measurements from a fairly well regarded smart trainer, the Tacx Neo 1 recording direct to Zwift. Temperature of the ride was about 7 deg C. Data compared using Zwift Power website. Erg mode used in middle section of the ride, topped and tailed by free riding in Zwift - data for one ride.
There was generally reasonable agreement between the trends in the power readings. Curiously the cadence readings from the Neo were around 1 % lower than the AXO. For a “normal” power output range the AXO appears to read about 2 - 3.5% lower than the Neo. For higher power efforts there was relatively little difference between the two, so the calibration curve is somewhat non-linear. The power curve chart shows that for maximum average powers with durations > 1 min ranging from 421 W down to 271 W the AXO gave values between 13 and 7 W lower (generally in the 11 - 12 W range).
The Neo is considered to be a reasonably accurate trainer, given the power output is determined directly from electrical power dissipation (induced voltage * current) rather than relying on strain gauges and temperature corrections. Presumably there are some small additional mechanical losses from the bearing. Given the AXO is a spider based power meter, and the Neo measures power at the hub I would expect around an additional 2.5% loss after the drive train (the chain was clean and waxed). Consequently, my overall my impression is that this specific AXO reads about 5.5% lower than the power to the Neo if power were measured at the pedals, corresponding to around 3 % lower for comparable efforts neglecting drive train losses (in non sprint situations). This difference also appears consistent with a comparison of my power data with that from a friend’s quark on several climbs at matched speed accounting for fractional weight difference outdoors.
I am hoping Sigeyi will introduce a manual calibration feature to the software.
Just got the 8 bolt Shimano MTB version - not seeing the extreme spikes as mentioned on earlier firmware. But I am seeing 600 watt spikes if I pop a wheelie under power. Maybe that’s legitimate - just seems a bit high to me.
Does anyone with a dual source of data have any insight as to what kind of spikes happen if you mash the pedals to clear something?
A second calibration ride - same conditions as previously. This time a ramp test in erg mode on Zwift at 7 deg C again. Tacx-Neo vs Sigeyi AXO. Good tracking between the power meters during the ramp. This time critical power from AXO about 4% lower than Neo over a relatively wide range of powers (not including drive train losses). Some divergence between power numbers during cool down (still in erg mode) - perhaps related to cooling in the Neo?
Ater 2 months of utilization I connect my power meter in the app and he asked for a firmware update 3.92. After that he stops to work and I tried to contact the warranty, without many answers the SIGEYI request to me to sent the power meter back to China with no refund of ship charges and only after that they will send me a new one. Warranty services is a shame.
I don’t recommend the product. Pay a little bit more and buy a quality product avoiding headaches.
In a moment of weak judgment i bought a Sigeyi power meter for my mountain bike
I did a road ride with my Assioma pedals and logged both.
Here is the compare set: https://analyze.dcrainmaker.com/#/public/d48ea0a2-1da5-4ff1-4acd-bba5d1dfa5bd
The Sigeyi tracks a bit lower than the Assiomas. Would anyone more knowledgebale than me care to comment on these results?
As far as I can see this is acceptable and I think it will be good enough for me for MTB use.
Lets just hope there isnt any weird artifacts and spikes, and I hope it will last.
Road ride. I would also like a complete mtb test but I don’t like venturing to much into the wild with the road pedals and shoes. Maybe I’ll try on a rough but easy single track one day.
Thought it might be an idea to share my experience here also.
I’ve got a M7100 SLX crankset and I’d just bought an XT chainring for it which has 8 very small bolts
So I got the AXO-SHIMANO-MTB-8-100 model to suit this, which I think looks brilliant on the bike
To fit this you have to remove 8 very small Torx screws & there’s no clear indication on the torque spec required when you want to put the ring back on - I’ll assume the requirement is rather low. I ended up slightly rounding one of the heads fitting it, which really isn’t ideal! My assumption is that I don’t think these bolts are generally intended to be taken off. Have since bought a decent quality Torx head for my torque wrench, but I’m going to have a very nervous time when it comes around to fitting a new chain ring.
In hindsight if I’d been planning this better. I think the traditional style with a 4 bolt 104bcd i.e. this model AXO-SHIMANO-MTB-4-104 would be a much safer option!
It wouldn’t look as good on the bike, but you can use any 104BCD round chain-ring that you like & you aren’t limited to the model specific shimano rings with unique torx screws not seen anywhere else.
–
As for accuracy, not sure if it’s my model. I still get occasional spikes, most notably if you’re putting down sudden power to clear a log/step up type thing is when it will happen. Once a couple of times per ride, So if I look at the power curve for rides I’ll often have 1 to 5 seconds worth of 800 watt effort that are false.
Otherwise it works brilliantly. Very happy with the numbers. I think if this was on a road or gravel bike you’d probably not ever see a power spike.
I did also find that if you spin a very high cadence of say 110+ you can also get high power readings to happen. Both issues I guess are cadence related.
Thanks for your thoughts. I’m in the same boat as you. I have an SLX groupset but I’m leaning towards the 4 bolt BCD 104 but I also want to maintain the same gearing.
I went for the 8-bolt on my race bike (and 4-bolt on the trainer/commuter) and didn’t really have any such issues.
The chainrings are available mostly with the spider included so I imagine Shimano foresaw people having issues correctly replacing them.
Assuming you are able to undo it when it comes time (you should be ok unless you went to town on it as the edges should hopefully still have some shape in the opposite direction) then I would recommend a teeny weeny dab of low strength loctite to remove the feeling of needing to overtighten it.
I have a Sram Carbon DUB crankset with dm 30 tooth boost offset chainring. Ordered the axo spider with boost spacing and added a 32 tooth ring. I can cope with the 32 but have a race face 30 ring in the shed which might se use at a later time.
The axo still does a good job, but I have had one incident of which it did not shut off and then was low on battery at next use.
I have to check if I can add a battery status field on my garmin
285 Dollars is still a lot. A 4iiii double sided is 549 euros here in germany.
On 105 cranks. I think this is a safer option.
Maybe You can trade your Oldenburg cranks.
Will probably just note.
Video shows new app is released as of a week or two ago.
Also comes with a Firmware update to 4.
Will be interesting to see if it brings any improvements re:power spikes for those of us with earlier Hardware versions - mine is a 5 but GPlama has version 8.
