Delayed Resistance Changes when Virtual Shifting

[kcknuckey]

Issue Description

When shifting to a higher gear in Zwift, trainer resistance takes approx 10 seconds or so before the resistance changes take effect. When changing to a lower gear, resistance changes are somewhat quicker, but still delayed.

The attached debug file should illustrate the situation:

• I selected 'Tick Tock' as the route in Zwift in order to use a flat route where the only resistance changes should be from shifting into a higher or lower gear (assuming a consistent cadence)
• I start riding at gear 1, which obviously is very low resistance, and choose a cadence that I try to maintain through the test
• I then rapidly change all the way up to gear 24. Both qdomyos and Zwift instantly show the gear change.
• Resistance from the trainer does not change for 8-10 seconds
• Eventually the resistance increases to the appropriate level
• I then take the gearing all the way back down to 1. Both qdomyos and Zwift instantly show the gear change.
• Resistance from the trainer again takes a long time to decrease to low resistance, though a little quicker to reduce than to increase

...so the setup is functional in that gear changes show on both my phone and Zwift as soon as I shift gear, but the changes in resistance from the trainer take a long time to take effect. This makes it very difficult to time gear changes for sprints or when hitting a steep hill.

Expected Behaviour

The trainer is clearly capable of changing resistance a lot quicker, because if I'm pedaling on the flat and suddenly hit a hill, the trainer resistance changes very rapidly. I'm not sure why gear changes should take so long to take effect..

My Setup

Hardware

• iPhone 17 Pro
• iPad Mini A17 Pro (current 2024 model)
• Wahoo Kickr 2018

OS

• Phone OS: iOS 26.2
• iPad Mini OS: iOS 26.2
• Wahoo Kickr Firmware: 3.5.2

Append a debug log

debug-Tue_Dec_23_11_20_55_2025.log

Additional context
Add any other context about the problem here.

[cagnulein]

[kcknuckey]

By the way, I am running version 2.20.17 of qdomyos

[cagnulein]

In the above screenshot you can see that the whole process takes 200 milliseconds from the input request to Qz to the answers of the trainer.

The only sospicious thing is that when you can changing more than 1 gears qz is sending more than 1 command and the trainer could go crazy.

Did you try to check if you have this also with one gear only?

[kcknuckey]

I haven't specifically tried that, since the resistance change associated with one gear can be small enough to be harder to see, but I will try that as a test. The thing is...sometimes we need to change gear from a very high gear to a very low gear (or vice versa) very quickly for hills or sprints, which is why that ability would be important..

The trainer itself seems capable of rapidly varying resistance changes, since these can be seen when the terrain changes rapidly on some of the Zwift routes..

[cagnulein]

Yes but my idea is: if the problem is the fast sequential speed changes I can add a filter to send only one command when the gears value stabilizes. Otherwise you can use the classic resistance protocol enabling the check box "without wheel Diameter" in the wahoo options, so qz will use only resistance commands to the trainer Let me know Roberto Viola I sometimes trade time with my daughter to help. If this was valuable, please support my time: https://buymeacoffee.com/cagnulein https://www.patreon.com/cagnulein/join https://paypal.me/cagnulein Software engineer and open source enthusiast, QZ Creator http://robertoviola.cloud http://qzfitness.com/ Il giorno mar 23 dic 2025 alle 18:49 kcknuckey ***@***.***> ha scritto:

…

*kcknuckey* left a comment (cagnulein/qdomyos-zwift#4018) <#4018 (comment)> I haven't specifically tried that, since the resistance change associated with one gear can be small enough to be harder to see, but I will try that as a test. The thing is...sometimes we need to change gear from a very high gear to a very low gear (or vice versa) very quickly for hills or sprints, which is why that ability would be important.. The trainer itself seems capable of rapidly varying resistance changes, since these can be seen when the terrain changes rapidly on some of the Zwift routes.. — Reply to this email directly, view it on GitHub <#4018 (comment)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AAALYWD4HRYXTNUG72Y4WX34DF6DFAVCNFSM6AAAAACP3YPJ5OVHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHMZTMOBXGQ3TIOBXGQ> . You are receiving this because you were assigned.Message ID: ***@***.***>

[kcknuckey]

Hi Roberto

So just to make sure I understand the test - if I check the "without wheel diameter" and change gear rapidly I should see if this gets a different result - correct?

In terms of the filter, it sounds like it could potentially work if the rapid commands are the issue... Anyway, I would be happy to try any potential solutions. If that didn't work, I also wonder if a long hold on the up or down shift could enable a fast change between high and low gears to see if that shortens the delay in changing resistance...

Anyway, as I say I would be happy to test any potential solutions to speed up gear changing. I'm going to be away on a trip until the end of the year, but would be able to do any testing at the beginning of the new year if any potential solutions are available

Thanks again for your help on this

Keith

[cagnulein]

Hi Roberto

So just to make sure I understand the test - if I check the "without wheel diameter" and change gear rapidly I should see if this gets a different result - correct?

Yes you shouldn't have the delay but the ratio should be different

Test also with the without wheel diamater unchecked doing 1 gear only at a time.

In terms of the filter, it sounds like it could potentially work if the rapid commands are the issue... Anyway, I would be happy to try any potential solutions. If that didn't work, I also wonder if a long hold on the up or down shift could enable a fast change between high and low gears to see if that shortens the delay in changing resistance...

Yeah it's what I would like to implement sort of.

Anyway, as I say I would be happy to test any potential solutions to speed up gear changing. I'm going to be away on a trip until the end of the year, but would be able to do any testing at the beginning of the new year if any potential solutions are available

Ok but I need first to understand if the issue was that doing the 2 tests. Are you able to?

Thanks again for your help on this

Keith

[kcknuckey]

Hi Roberto - Happy to do the tests as soon as I'm home (where my bike is..) I'll let you know what I find

Keith

[cagnulein]

Sounds good, thanks! Merry Christmas! Il giorno gio 25 dic 2025 alle 04:40 kcknuckey ***@***.***> ha scritto:

…

*kcknuckey* left a comment (cagnulein/qdomyos-zwift#4018) <#4018 (comment)> Hi Roberto - Happy to do the tests as soon as I'm home (where my bike is..) I'll let you know what I find Keith — Reply to this email directly, view it on GitHub <#4018 (comment)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AAALYWF2QEZ6IUOLRVIORQ34DNMCJAVCNFSM6AAAAACP3YPJ5OVHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHMZTMOJQHAYTKNBUGM> . You are receiving this because you were assigned.Message ID: ***@***.***>

[kcknuckey]

In the above screenshot you can see that the whole process takes 200 milliseconds from the input request to Qz to the answers of the trainer.

The only suspicious thing is that when you can changing more than 1 gears qz is sending more than 1 command and the trainer could go crazy.

Did you try to check if you have this also with one gear only?

Hi Roberto. Back from vacation and was testing the things you mentioned today

Re: one gear only shifting: I tried shifting up/down by only one gear, but the difference in resistance going up and down by one gear makes it difficult to judge how long the resistance change takes because the change is so small. In order to detect resistance change timing, I tried going up and down by four gears, and when doing so, from starting resistance to full resistance change takes around 10 seconds, with resistance changes happening in a somewhat linear manner over the 10 seconds.

Re: testing 'without wheel diameter': I tried this, and resistance certainly changes... basically to the point that I cannot pedal (even in gear 1). As such, I was unable to do this test.

Testing directly from the app: With 'without wheel diameter' set back to false (so I could pedal), I tried the test mentioned above with changing four gears then timing resistance change, but this time using the buttons on qdomyos-zwift app to change gear. Timing for resistance change was the same as using the Zwift Play to change gear. Not a surprise really, since any gear shift on the Zwift Play is seen on the qdomyos-zwift app immediately, so this makes sense.

Question for my curiosity: When I select my iPhone as the resistance source in Zwift, my understanding is that the iPhone app is telling the trainer what resistance to set. However, the overall resistance that the trainer sets is logically a product of the virtual gear, plus the gradient of the road someone is riding in Zwift... Am I right in understanding that during a Zwift ride, the Zwift program sends gradient-related resistance data to QZ, which in turn takes that into account along with the virtual gear selection to send an overall resistance setting to the trainer?

Testing to compare resistance changes from virtual hills vs. gear changes: I had wondered if any resistance change lag may be the older trainer physical design (mine being a Wahoo Kickr 2018) simply taking longer to physically change resistance once it receives resistance change commands. However, I had tried riding a course with a downhill section to a hill, and the resistance changed very rapidly based on hill gradient, so the trainer is clearly capable of rapid resistance changes... I've attached a debug file from the ride. Key points of the ride:

• Initial resistance changes due to gearing seemed sluggish at the start
• The hill started at approx 2 mins into the ride, and the resistance changed rapidly
• I did a number of U-turns on the hill. When turning down, resistance dropped immediately and went up almost immediately based on gradient of the hill
• Gear shifting on the hill (this time) seemed relatively prompt, though nowhere near as fast as the hill gradient changed resistance

Anyway, when you get a chance please could you take a look at these notes and my question

Thanks, and happy new year!

Keith

debug-Sat_Jan_3_17_06_56_2026.log

[cagnulein]

Re: one gear only shifting: I tried shifting up/down by only one gear, but the difference in resistance going up and down by one gear makes it difficult to judge how long the resistance change takes because the change is so small. In order to detect resistance change timing, I tried going up and down by four gears, and when doing so, from starting resistance to full resistance change takes around 10 seconds, with resistance changes happening in a somewhat linear manner over the 10 seconds.

yeah unfortunately in this way we are going into the scenario where qz is sending more wheel diamater commands at one time and i guess this makes confusion to the trainer. weren't you able to change the gear table under the wahoo options in qz to have large step in the gears so you can check that one gear is instantaneous or not?

Re: testing 'without wheel diameter': I tried this, and resistance certainly changes... basically to the point that I cannot pedal (even in gear 1). As such, I was unable to do this test.

that's strange. be sure to set the gears to 0 in qz. please retry this test because i guess on this setup you will have 0 delay for sure

Testing directly from the app: With 'without wheel diameter' set back to false (so I could pedal), I tried the test mentioned above with changing four gears then timing resistance change, but this time using the buttons on qdomyos-zwift app to change gear. Timing for resistance change was the same as using the Zwift Play to change gear. Not a surprise really, since any gear shift on the Zwift Play is seen on the qdomyos-zwift app immediately, so this makes sense.

yes this is makes sense. it's not a zwift play or a zwift channel communication issue, it's just the trainer.

Question for my curiosity: When I select my iPhone as the resistance source in Zwift, my understanding is that the iPhone app is telling the trainer what resistance to set. However, the overall resistance that the trainer sets is logically a product of the virtual gear, plus the gradient of the road someone is riding in Zwift... Am I right in understanding that during a Zwift ride, the Zwift program sends gradient-related resistance data to QZ, which in turn takes that into account along with the virtual gear selection to send an overall resistance setting to the trainer?

yes there are 2 ways:

• with "without wheel diameter" to false, qz is sending inclination from zwift to the trainer and also a special command to handle virtual shifting to the trainer
• with "without wheel diameter" to true, qz is sending a sum of inclination and virtual gearing to the trainer

Let me know and happy new year!

[kcknuckey]

Hi Roberto

Re: 'Without Wheel Diameter' test
Per your suggestion I tried testing again with 'without wheel size' set to true and gears set to zero, and this time I found that the trainer didn't respond to either gear changes or gradient changes. I tried it starting in gear 0 and also gear 1, with the same result both times... The gears shown on Zwift and the ones shown on QZ both changed on-screen and iPhone the way they should, but no matter what gear I put the virtual gears in, it made no difference to resistance. I then deliberately went up a steep hill to test hill resistance, and still nothing changed. Not sure why changing that setting stops all resistance commands... Debug file is attached:

After testing, I changed 'without wheel diameter' back to false, and the bike went back to its previous behavior, with sluggish gear resistance changing and fast hill resistance changing. For both tests, the main iPhone screen showed QZ attached to the trainer throughout the tests.

debug-Sun_Jan_4_14_22_16_2026 without wheel diameter set to true.log

Re: Large step in gears test
I configured my gearing such that gear 1 was normal and gear 2 was equivalent in gearing value to around gear 21, so that I could just change from 1 to 2 and see how fast resistance changed. This was interesting - resistance started to change almost immediately and went to full resistance over about 6 seconds (which I would attribute to how long it took the trainer to make such a large change). Debug file attached:

debug-Sun_Jan_4_15_04_25_2026 Gearing configured for huge change from gear 1 to gear 2.txt

This would suggest that your idea of having QZ wait a moment until gear changes stop changing before sending resistance changes to the trainer might help...

As a sanity check, I wanted to see if that delay would be the fastest way to a higher gear, or whether the resistance normally starts changing while I'm shifting up anyway, meaning you get to a high gear quicker by just letting the system start changing gear while I'm still shifting. However I tested that: I reset the gears to normal ratios then started from 1 and kept clicking fast to 21, and the trainer seemed to not want to do any resistance change until I stop clicking. As such, sending one command after the rider stops fast-clicking seems like it would be the fastest way to make changes when moving between a lot of gears.

Hope this helps!

Keith

[kcknuckey]

Hi Roberto

Another update: I just tried a longer ride with 'Without Diameter Protocol' set to false (default) and encountered much, much longer shifting delays (from 30 seconds to over one minute) on different parts of the ride... Very strange. Below is the debug from the ride - sorry about the large size of the file, but it was a 14 mile ride. Examples of large delays should be evident in various parts of the file, but I felt the delays most around the 20-35 minute mark where I was either climbing in a low gear with very high resistance because the downshifts didn't change resistance, or on the downhills where I could go up to gear 24 and not be able to get more than ~50W resistance.

debug-Sun_Jan_4_16_01_57_2026.zip

Hopefully the file is helpful in identifying if the trainer is not getting the resistance changes or just not responding to them until much later...

Per my previous note, as an alternate strategy to resolve this it would be nice to figure out why the trainer does not change resistance at all with 'without wheel diameter' set to true.. Apart from starting in gear 0, are there any other settings that I should be checking that might cause this?