That is the way how it is added when you setup the combined curtain in the switchbot app. I had this at first, too. Went to add a group helper in home assistant instead which contains both curtains (so I am able to control both sites separately but also at once)
The quiet drift mode is not yet in the stable release of home assistant so you are currently not able to use the quiet drift mode ‘natively’ in HA.
pyswitchbot was bumped to 0.43.0 in 2024.01. Can we use quietdrift, or something else needs to happen still?
Edit:
So the pySwitchbot open and other actions now support the speed attribute. And ha/core is using the updated 0.43.0 version of the pySwitchbot library.
I’m guessing this method call has to incorporate the newly added speed parameter? Does anyone know of any other similar use-cases which could be used for implementing opening covers with different speeds? Or can anyone point to a PR that one could use as inspiration for adding e.g. another cover entity for each device for opening/closing using the speed parameter?
I apologise in advance for going a bit offtopic, but I wanted those of you who own the u-rail track version of the switchbot, whether you think it could work on my rails. The problem is that my rails are not U-shaped but oval. I created a cardboard cutout of the mimimum dimensions for the switchbot wheels and as you can see in the picture below, it barely fits, without any room to spare.
Then there’s the question whether at the bottom of the rail there is enough surface for the switchbot to grip onto.
I do not have Switchbot Curtain 3 but I have similar worries.
My guess would be in your rail the rollers may fit inside. But as they touch on slopes, the switchbot may turn at an angle, the rollers may touch the top and get friction there in the reverse direction. If it sits right, the rollers rest on a very small corner. If you are brave, you may consider filing the rollers at an angle to match the slope and fit better (the rollers can be replaced as there are different sizes, I do not know if they can be bought separate if it fails).
Personally, my rail is too slim (and square) so I was considering the Aqara which uses native rollers from the rail, so that would always fit. It would just be a matter of small grip area for the motor roller. I read sanding the bottom could help with that, or sticking door isolation foam at the side of the rail to add grip, but that would not look very nice i.m.h.o.
Downsides of the Aqara is it seems very bulky. Upside I think it integrates better with Zigbee v.s. bluetooth. I have not seen a position slider on the Switchbot, the Aqara has a position slider so it knows about intermediate posistions and can be set to it. I doubt the Switchbot can, unless you time the stop button?
Edit: Could the notch in the middle above the rollers cause problems? The rollers seem metal with a small bit of rubber at the end, so filing at an angle is probably not going to help out much.
Thanks a lot @Edwin_D You make some good points, it does look quite questionable whether it would work in my rails.
I didn’t know the Aqara Curtain driver E1 uses the native rollers. I had a look at their website and it looks like it should work on my curtains. however what for you is a benefit (zigbee) is a turndown for me.
I have in my house an extensive coverage for z-wave, wifi and bluetooth. I’m not looking to add zigbee to the mix (especially as it will be transitioning to thread/matter soon). So that’s quite a bummer.
Also not the biggest fan of Zigbee/Thread because it operates on the same frequency as 2.4ghz wifi so would cause interference, and I have a lot of concrete and metal in my house, which the 868.42Mhz Z-wave frequency can penetrate much better than the 2.4ghz zigbee/thread frequency.
Most my stuff is z-wave too, I share your preference. But I do have Zigbee too, my preference is Z-wave, wifi, zigbee, bluetooth for wireless in that order, bluetooth is in the same band as Zigbee and wifi 2.4. My main reason to call it an upside is that I’m still not sure if the Switchbot communicates its position.
But I get you don’t want to put up another mesh just for this.
I’m interested in this too. What’s the last step holding us back now? I’m even fine with a hacky solution like @Breph 's one but like @matata I found I had no such curtain.py to modify.
Does anyone have any ideas as to how to trigger quietdrift from Automations now we have the new version of pySwitchbot in Home Assistant?
@quizzical since the width of the wheels is the mentioned 10.5mm they’ll definitely lay in the rounded part of your rails. I’d guess that this could still work as the pressure from the switchbot should be enough to drive over the rail.
Hard to make any reliable statements, guess you have to try. The switchbot will come with two additional pair of wheels with different wheel size, so there is an increased chance one of them will work.
Currently, the code to utilize the added speed parameters in the switchbot calls in HA is missing (as per Added QuietDrift support for Curtain 3 by den-mko · Pull Request #223 · Danielhiversen/pySwitchbot · GitHub)
Can you tell me if the bluetooth version of the integration (without the switchbot hub) also has a slider control to put the curtain in a specific midway position (e.g. to leave a door uncovered)?
Yes, it has! 
Is has, but for me it’s not so good opening the curtains (group) works well and the percentage is updated till about 60-70%. Then not anymore. They are fully open but status stops arpund 60-0
70%. Then we opening again, it then shows 100. And this is always the same. So never good status in the first run…
An update from my side. I now own both a Aqara Curtain E1 and two Switchbot Curtain 3, so I can share my experience with both.
I first bought the Aqara. Pairing was a breeze with ZHA and Sonoff dongle E, but calibrating was hard to figure out. It took a while to realise that once positions are set, you need 5 clicks to reset before you can tripple click to set positions again. Also during calibration, you need to drag the curtain a bit, which made me fear it lost accurate positioning.
On my current rail, using gliders and not rollers (not officially supported) the old gliders were not smooth enough. The Aqara struggled and did not make it to the end.
I swapped out my rail for one that suits both the Aqara and the Switchbot better, it wasn’t expensive or hard to replace it, just difficult to find one with decent rollers. I settled on a cheap one with questionable rollers. The Aqara worked (2.7m wide curtain, 2.5m high), but sounded as if it still struggled. It made it almost to the end but only just failed to press all the rollers together when fully opening. Also, because the unit is big, especially wide, it didn’t quite fit between two curtain hooks.
I decided to move the device to my sons room with rollers and 1.5m high curtains and better rollers. It works well there, but it still sounds rather squeaky.
Then I bought the Switchbot. I like them way better. I use Esp BLE proxies with active scanning for other purposes too, they immediately picked up the Switchbots. Finding the best fitting rolls for the Switchbot was easy. Switchbot uses a magnet to mark the end position for closing. Though it is visible on the rail, I like the approach because the Switchbot automatically checks calibration with them on every use.
Calibrating was easy and fast. The switchsbots are a bit more powerful then the Aqara’s. It manages to fully close, even though the new rollers are still of rather questionable quality. Sound level is similar to the Aqara in fast mode, but less squeaky. Also, because the unit is longer, but not so wide, it hides better. The only downside I could find is they are less fast to respond to commands (could be difficult to stop midway with some level of precision), or in my case to avoid hitting the Christmas tree
So if I have to advise on a curtain bot, I advise to buy Switchbot. Even if it requires you to find another rail, should yours prove incompatible. The silend (snail speed) mode is an added bonus, should it make its way to HA. But it is said to use way more battery though, so I doubt I would use it on a daily basis.
Wow thanks a lot for the extensive reply!
Based on your testing I’m happy that I have ordered the switchbot and it is on it’s way (even though switchbot told me my rails are not compatible).
I have looked at more reviews showing also the smaller wheels and I see there is still margin to cut some material from the sides of those wheels, so I hope it will work 
I would not like to have to replace the rails because it would mean filling the holes in my wall and drilling new holes etc. possible of course but costs extra time (and money)bfor the rails and then I might as well look for motorised rails.
I’m hoping it works out. The small wheels are really tiny so there’s indeed hope left, you’ll feel immediately if it runs right by putting in the small side.
It didn’t work out unfortunately. The smallest wheels did fit into the rails, but it was the metal sides of the small wheels that contacted the oval shaped sides of the rails. So it wouldn’t move at all due to the huge resistance.
My spouse really wanted it to work in 1 room so we bought new rails for that one.
Too bad, as I would have liked to do more rooms…
I apologize if i missed it somewhere, but is there any news on the possibility of using quiet drift mode in Home Assistant?
Sorry this is very technical, but I too am anxious to see QuietDrift capability and have been watching the code. It looks like in order for this to work the pySwitchbot library needs to be modified and maybe homeassistant/components/switchbot/cover.py need to be modified.
Good news is the pySwitchbot library got the code for QuietDrift in version 0.42.0 on Dec 20, 2023.
This allows open cover,close cover and set cover position to utilize a speed constant.
Version 0.43.0 of the pySwitchbot library is getting merged into HomeAssistant with 2024.2.0b2, so expect the pySwitchbot foundational capability to be in the 2024.02 release. Bad news is, there doesn’t appear to be anything changed with cover.py yet. I think someone better at programming than me will need to change
https://github.com/home-assistant/core/blob/cd1ef93123f84f3adb54140c2c2d1c8e54b7e2b1/homeassistant/components/switchbot/cover.py#L76 (open)
https://github.com/home-assistant/core/blob/cd1ef93123f84f3adb54140c2c2d1c8e54b7e2b1/homeassistant/components/switchbot/cover.py#L85 (close)
https://github.com/home-assistant/core/blob/cd1ef93123f84f3adb54140c2c2d1c8e54b7e2b1/homeassistant/components/switchbot/cover.py#L103 (set to position)
to utilize the new pySwitchbot speed option of “00”/“ff” for normal or “01” for slow.
Architectural discussion here on speed attribute of cover Add speed attribute to cover integration · home-assistant/architecture · Discussion #789 · GitHub
Upvote it is you want a speed attribute (and therefore quiet mode)
I wonder what’s needed for this to actually happen and get into a release
It’s quiet around quiet mode
