Is the NEW Bosch BMI 270 Any Good????

Bosch BMI 270 | OSR Modes Latency: / 70822838
Bosch Q&S on OSR Modes: community.bosch-sensortec.com...
Want to support me?
---------------------------------
Sign up to be a Patreon: / uavtech
Donate to my Paypal: [e-mail]
Use this Amazon Affiliate Link: amzn.to/36okaf7 to purchase ANY products [hobby gear, household items, etc.]. As an Amazon Associate, I earn from qualifying purchases.
Use this Race Day Quads link to make a purchase: www.racedayquads.com/?aff=51
Free Resources:
------------------------------------
Presets and Tuning Training Content: theuavtech.com
Facebook: / uavtech1
Betaflight BlackBox Log Review Group: / 291745494678694
UAV Tech Discord: discordapp.com/invite/rCCzgeT

[@MCsCreations]

Pretty interesting indeed, Mark! 😃 Thanks for testing and comparing them! Stay safe there with your family! 🖖😊

[@MurdersFPV]

Thank you for this. I've been dreading getting into tuning a new stack I have because it's one of these BMI270 gyros.

[@AndyDreadFPV]

I have several builds with Radix2 and I just learned in this video that they have the hardware cutoff set higher than everyone else at 700Hz and even with the Radix2 bmi270 defaulted that high a cutoff they have been flying amazing for me so you have absolutely nothing to worry about.

[@danmocodean6463]

I just purchased a FC with BMI 270 and I’m actually impressed with how it feels. The quad feels more like kiss and I use no gyro filtering at all and was able to push pid far with no issue. I have the master multiplier at 1.45 and I can push it way higher. I love it so far

[@ChrisQuadFPV]

I'm sure glad you made this video. Because I'm sure everyone was wondering what the deal was.

[@PIDtoolbox]

Great vid! BMIs look pretty good to me. Very consistent and when filtered properly they are very clean, especially in the 50-150hz “quiet” region

[@Quad66]

How does the 3.2k of the BMI270 versus the 8k output of the MPU6000 factor in to the overall picture? Doesn’t the higher output give the opportunity to filter more?

[@ibnfpv]

@@Quad66 exactly my question Although we know that high sample rate also bring more noise to the table as it is collecting more data But I agree that the 8/8 was great balance

[@afcl1984]

I have 2 Foxeer f722 v3 on the way with that gyro. I’m glad it’s good. Thanks for the great video.

[@marlin-fpv]

I struggle with the f722 v3. Pretty much unflyable. Tested a broad range of settings but the quad remains to show flyaway tendencies.

[@afcl1984]

@@marlin-fpv I have them on 2 quads now and they are flying perfectly.

[@marlin-fpv]

@@afcl1984 Racing quads?

[@afcl1984]

@@marlin-fpv Nope, freestyle.

[@nuhbiwan]

Good to see that it not a bad alternative to the TDK ones. Btw:The „o“ in Bosch is pronounced more like the „o“ in Porsche.

[@IvanEfimovLimon]

Nice video. The plots at 3:35 are a little sus, it looks like PID toolbox didn't plot correctly BMI vs TDK because BMI had different blackbox recording frequency. You can see that noise peaks on BMI are shifted to the right, looks like 800Hz BMI range was stretched to 1000hz. I think Brian White fixed this in the latest PID toolbox when i found this little "bug", but im not sure if he checked the two plots of different gyros on the same plot. It does not change anything here tho, BMI noise floor is still lower.

[@dimeland2]

note those cutoff frequencies are only applicable when the output data rate is at 800Hz. changing odr will also change the cuttoff frequencies.

[@mouseFPV]

13:00 but what is the lower net latency? Does the hardware filter have more or less inherent latency to do the same amount of filtering? or is it the same? any benefits to doing it more up or downstream? I feel like a lower "internal" filter on the gyro and less software filtering makes more sense.

[@mallikarjunse]

Try ICM-42688P too.

[@AirKhanFPV]

Thanks for the review, but i think the MCU6000 are back in stock i ordered a Nazghul V2 Analog 4S Version (grey parts) and it came with a F7 and MCU6000 on it 3-4 months ago i ordered also one that one had a bmi270 (BLITZ FC) with usb-c the new ones are coming with the SucceX-E V1.4 with micro usb again.

[@chrisburke63]

Any idea if the BMI270 has fewer gyro issues (like noise at the end of flips or rolls / noise much worse on only one axis) compared to the MPU6000?

[@Sbangfpv_]

1:07 I used to use TDK audio Cassette tapes “90 mins” 🤯 technology. Year 2000-2001

[@Sbangfpv_]

Mark! Your BetaFlight Presets are amazing. Thank you for all that you do, sir.

[@jejefpv]

Thanks Mark. I had a bit of strugeling because of the names of the options in BF and the names of the modes for the BMI270. BF Normal = OSR4 on gyro, BF Option1 = OSR2 on gyro, BF Option2 = Normal on gyro, if I understood Chris' PR right (especially his "Full Spectrum" diagram). It is good news the gyro itself seems good, but what about the down sides ? By default, the BMI270 frequency is 3.2Khz, which obviously limits the PID loop at 3.2Khz too. And under 4Khz, BF team recommends running ESCs in DShot300 (since DShot600 is useless and may cause issues). Compared to 8Khz with a MPU6000, we now have a PID loop running at 40%. And the speed of the ESCs is divided by 2. Does it significantly affect performance ? Oh, and since a simple F4 with basic BLHeli_S ESCs (flashed with Bluejay) will now be plenty enough computing power for runnning so slow, why buying high end FCs (F7-G4-H7) and monster ESCs (F3-F4) ? ^^

[@uavtech]

3.2Khz (3,200 samples of data per second) for sub-100hz motion tracking.

[@jejefpv]

@@uavtech Do you mean 3.2Khz is more than enough for motion-tracking and there is no difference beetween 1Khz, 3.2Khz or 8Khz ? Same about DShot300 vs DShot600 ?

[@SupaflyFPV]

Thanks for the data...the lower freq responses look encouraging....I didn't get to test one yet, but a couple of things to consider about the BMI270 is the gyro and pid rate are 3.2khz so there will be slightly more latency running betaflight filters compared with an 8k gyro....and I'm wondering if the gyro hardware low pass is running at 6.4khz internally which could make it less latent than using the equivalent Betaflight low pass filters at 3.2khz... Karatebrot thinks the OSR4 mode 300hz filter is third order which explains the steep attenuation (anything >650hz flattened to -50db which an MPU6000 doesn't do), and in this case perhaps it makes more sense to reduce significantly or remove the Betaflight gyro LPF.....and just use notches... A final point to consider about the 3.2khz gyro/pid rate is that Betaflight PT filter cutoffs will be lower than expected because of the anti-alias calculations in BF...worth bearing in mind....

[@evilc2048]

Certainly, with the FC I use (The JHEMCU GHF411 Pro), they just replaced the MPU 6000 version with a BMI270 version, and the Gyro loop freqency maxes out at 3.2k The MPU version did 8k/4k, but the BMI version is 3.2k/3.2k - are you seeing the same thing with your BMI FCs? Also, with the BMI version of the FC, the factory flash defaults to NORMAL mode, but if you flash a new version of BF (They say to use the same target as the MPU version), and apply custom defaults, it defaults it to OPTION2. In this instance, the gyro orientation is also wrong. I am having terrible trouble tuning this quad. I run the exact same build using the MPU version of the FC, and it flies super sweet, but with the BMI version, I get this odd rebound effect at the end of a snap roll and I just can't tune it out. Only just found out about NORMAL/OPTION1/OPTION2 tho, so will have to play with that and see if it solves the problem Also, another thought about the modes - does doing it in the gyro not make sense for lower powered (eg F411) FCs, as you are doing the filter in dedicated hardware, rather than in software on the MCU? So you will have lower CPU usage?

[@RosscoStixFpV]

You should add a bf preset for the Bmi270 . I've tried the one that's there which made my quad fly even worse, then through other videos I'm under the impression that the bmi270 preset is for 4.2 users? I'm not quite sure what I need to do with filtering this gyro any tips would be greatly appreciated

[@BrainDeadEngineering]

My brain exploded !

[@wind_wizard]

Just be aware that if you are using Betaflight presets for BMI270 it will change your lp filter to option 2.

[@Enginair]

How important is gyro sample rate and pid loop frequency? I think most people are scared off by the 3.2khz

[@DriftaholiC]

You could test the phase delay of gyro filtering vs betaflight filtering.

[@uavtech]

done -- www.patreon.com/posts/bosch-bmi-270-70822838

[@phiveone]

Off topic: Have you ever done anything with shielding the FC from ESC FET noise? i.e. EMI copper shields.

[@uavtech]

No, but when I can - and if the gyro and MCU are in the same side - i try to face it away from the ESC if I can. Then the board ground acts as a shield.

[@leflamantrosefpv1725]

MPU6000 for life. and it come back. for the better.

[@aphinion]

Tried a foxxer F722 mini with that BMI270 in my aos 5.5 - and the roll axis showed a horrible noise around 160-180hz with 3 spikes in there - solely on the roll axis. Guess also those BOSCHs aren't safe for dead-on-arrival units. Let's see how the other 2 in my freshly built race quads perform!

[@uavtech]

In Normal mode?

[@ZylonFPV]

Are you going to try flying the new gyro vs tdk? That would be interesting

[@uavtech]

yes, this week's patreon.

[@Zentriefugal]

But how about it's lower sampling frequency of 3.2kHz?

[@glytch5]

So what i'm gathering from this video roughly (please correct me) is within betaflight 4.3 on a BMI270 the stock gyro "normal" mode sets the internal filters to be pretty close if not nearly identical to our MPU 6000 we are all used to? So I can stop fearing to purchase a newer FC?

[@uavtech]

Correct.

[@glytch5]

@@uavtech Did they fix the weird bug where you could brick the board by flashing with one? IF thats the case i'm probably going to pick up another FC

[@MouwDown]

We still need to run pid loop at 8k-4k with BMI270? If so, this is a negative thing or a neutral thing?

[@uavtech]

3.2khz is plenty fast to sample sub 100hz motion. I need to follow up how filters change at higher frequency sampling (cutoffs move up, which make it seem like faster sampling is better. But you can simply move cutoffs up for lower sampling).

[@UnmannedSky]

Good explanation of the BMI270. Thanks for the graphs too. May want to discuss the lower sampling rate of 3.2K since some see that and think it means it's worse.

[@AndyDreadFPV]

I think it's 3.2k no? ether way wasn't kiss flying on 1k for a long time and people were singing it's praises back then.

[@UnmannedSky]

@@AndyDreadFPV yes sorry my bad. 3.2khz. A lot of us have been told for years the faster the gyro sample rate the better so I understand why some are concerned. Also BF devs recommend dshot300 for sample rates 4K and under. I'm curious to know if the lower sampling rate and therefore slower pid loop refresh causes worse propwash handling or if it's really not noticeable.

[@butcher]

I had the same thoughts but after several tests it seems way better then old gyros and you can turn off most of the gyro filtering. I won't buy fc-s without this gyro anymore. :)

[@UnmannedSky]

@@butcher good to know. I ordered the speedybee V3 on UAVTechs recommendation so I hope it works out.

[@AndyDreadFPV]

@@UnmannedSky The biggest noticeable difference i've seen regarding prop wash is getting the noise down as low as possible per the frame then turning off/down filters and increasing P and D. Been able to do that running DHSHOT300 and 3.2k and get zero noticeable prop wash oscillation.

[@thsfl23]

Your tuning presets aren’t showing up in betaflight anymore. Any idea why?

[@uavtech]

They are there: www.screencast.com/t/uiDYsOgmAW

[@ibnfpv]

WDYT About the lower sampling dose it is a downside or do you see no disadvantages?

[@uavtech]

No disadvantage. Let me do a follow up on that to explain why.

[@ibnfpv]

@@uavtech aliasing and stuff (;

[@scrimdog3328]

Got two 30x30 mamba stacks and a 20x20 FC all with bmi270. I'm having zero issues with any of them. In fact with everything in account in bf and new quad set ups I must admit it's probably the best my fleet has flown ever. But if you get a shit stack don't come point your finger at me 🤣🤣🤣 👍

[@k4x4map46]

interesting

[@freedomcaller]

Were all these tests done with an identical odr? Because that will affect noise more than anything. Also as I understand it the osr2/osr4 are meant to be used with a higher ODR and then it averages down to a lower effective ODR, so it doesn't add any latency at all, but I could be using it incorrectly. The BMI docs are far from the worst imu docs, but also far from helpful.

[@uavtech]

Check out the PR for details on ODR testing.

[@freedomcaller]

@@uavtech was planning on doing that, but thanks for the nudge. Ill pull it up right now.

[@AndyDreadFPV]

So this leaves the question: Is there anyway to set the cutoff to 300Hz on Radix2? at all?

[@uavtech]

I think yes. Check out their flash page for the variable. Select BF 4.3.0

[@AndyDreadFPV]

@@uavtech Awesome thanks!

[@FPVian]

I think it would be best if manufacturers could settle on a “standard” gyro again with good availability, even if it’s not the “best” gyro. Having to tune the filters differently for each different gyro makes it harder to use presets and tune quickly.

[@uavtech]

It is really the same thanks to Chris and OSR4 mode (Normal) on the BMI 270.

[@FPVian]

@@uavtech Is this (generally speaking) true for all the new gyros? There are only subtle differences in frequency response from the user perspective?

[@uavtech]

Correct. The key now is what has the lowest noise floor (80hz to 200hz energy).

[@FPVian]

@@uavtech great! Looking forward to seeing your testing on that

[@SupaflyFPV]

There is a crew of Betaflight developers working on some guidance with regard to this....however it's not really in the spirit of the project to strictly limit to a single gyro...we can only recommend for best performance...