My 3” flies significantly better after turning off the gyro low-pass fitters without an increase in motor temps. 1.3 on D-term slider. Thanks for the tip!
Thank you so much for your contribution to the community. I will become a Patreon asap. Following your guides make my quad fly better, Making me crash less, so it’s good business over all 👍🏻
I don't know what I'm doing wrong, but every quad I have put 4.2 on develops a crazy wobble at low/no throttle, and a ton of bounceback. Messing with I-term relax helped, but not to my liking. I've fallen back to 4.1.7 and it's fine. This happened on a Tinyhawk 2 freestyle (though it was the best of them all), a 3" armattan tadpole, and my 6" hyperlow build. Not sure if I just have trash mechanical builds or what I'm doing wrong exactly. Anyone else had a similar experience? Thanks again for your videos UAV Tech, I swear at some point I'm going to get the time and dedication to donate and bug you with logs, etc. Keep up the good work!
On these smaller quads, and ones with less authority, the low throttle wobbles are easier to fix with thrust_linear (in the 20-30 range), but you may need to add more tpa_rate at the top end to keep that thrust linearization from resulting in high throttle wobbles
@@tehllama42 Thank you, I was messing with Thrust linear on the micros, but not on the bigger one. I had all but eliminated the wobble on that one, but still had bounceback. I still think that could have been a bad gyro or motor noise, I'll find out if I ever rebuild anything on that FC I suppose. I guess I didn't consider the big guy a 'low authority' quad...? Heh. It's got 2207 lumineers @ 1750kv running 6x5x3s on 6S power. I'll start with it at 24 the next time I start tinkering. 🤣 I highly doubt I'd ever feel any 'high throttle' wobbles, It's rare I'm over 50%, I just kinda cruise. Maybe the "wrong setup" for that, but I'm enjoyin' it when it flies properly.
@@JerR22 2207's are stout, but 6" props are still big, I can get some aurhority limitations on 2208's even - maybe try 20, see if that's of any help. Dynamic idle can help also, especially if you're not bringing the dshot_idle value as low as possible (you get a bit more control authority overhead)
'brain talk processor paused' is a real thing, although some race heats I'm better while carrying on a conversation. Somehow more strangely, when I'm drumming, my only option to output speech is sync pulsing syllables with beats, which is hilarious but really awkward.
Hay MR UAVtech. Please could you adjust the audio of your intro? lower the volume or when you render a vid use an audio compressor or limiter. I set the volume on my amp so as not to wake the kids. You usually talk for a bit then play the intro, so the vid starts off at a nice level then suddenly gets loud. I have been meaning to write this comment for a long time. I watch all your vids even if it is a topic I'm not interested in, just to give you the views. Love your channel.
I use Mr Steele props to help keep the amps down but prefer the S4 (lemon & lime) - the 3.7 pitch gives decent punch whilst still pulling less amps than a typical 4.5.
Great video my friend. I did exactly what you just did and this video and I got no prop wash perfect, but now I got a Humming sound any Ideas my friend?
Dynamic notch filter is the most obscure and most prominent thing at the same time. It is so confusing. We had some big fuss with Kalmen filter which fizzled, and dynamic notch filter just sneaked in unannounced. It feels as if the dynamic notch filter is just another gimmic, but it just doesn't go way.
Because the dynamic notch is not bullshit like the "kalman". The fuse was manufactured by snake oil salesmen that wanted to sell FC boards (Helio). They had a cult like following from F1 where folks drink the kool-aid all the time. The "Ai" filter is the next B.S. thing that seems is coming. The "kalman" is just a PT1 filter. The PT1 filter equation is right in the code itself. And lowpass filters are NOT made to kill peaks of noise like notches are. So the difference is when you run logging A-B tests on one (the DN) you can clearly see "holy shit, look how effective it is at killing motor noise." And when you run the other (the "kalman"), you are like "wait, i don't see any improvement here vs a conventional PT1" that is why one is the gimmic and one is real. Don't drink the kool-aid.
Thanks Mark. I’m trying turning these filters off today. Question: Is this when running RPM filters?Never mind sorry I just re-watched you answered my question. So I have the mark 4 6 s would you try these settings? Starting at 1.2 on the gyro ?
I would give it a shot. Personally I don't go blind and I look at raw noise performance in blackbox, but give it a good for gyro lowpasses off. Remember, with gyro off, D-term slider is 1.0 to start.
@@uavtech I tried it yesterday it made a weird noise when I gave a throttle blip. So I put everything back the way I had it. So I have some type of motor de sinking going on I shouldn’t have any motor desinking. I put everything back to normal and I realize that I’ve been having this problem the whole time ever since I got the quad copter. It doesn’t fall from the sky but you can hear it desink but it keeps running. I messaged you on messenger if by any chance you have time could you please look at it ? I gave a very good description as short as possible. I’m also doing black box and I want to look at the raw noise like you said. I already have it in my computer. I’m talking to the old lady about becoming a five dollar Patreon I need your help.
do you have a video or reference for how to set audio alarms ? trying to figure out if I have to set up a Special Function for each logical switch line. thanks
I've already used the gyro's low pass filter to filter a range of frequencies. Now I have activated the gyro notch to filter out a high frequency noise spike but the width in the center is not enough. could you widen the center of the notch instead of widening the cutoff frequency? widening the center of the notch using the cut-off frequency is a very limited and inefficient method of attenuation, in this way I go to filter even points where there is no high noise and only increase consumption and I lose thrust. couldn't we put two central frequencies to have a much wider center with a more precise and correct attenuation?
Ypu cannot widen a notch center. That would be called a band stop filter which as much more delay. You can widen the notch with a lowe Q factor and in BF 4.3, you can have up to 5 notches target amd crush peak frequencies.
What was the deal with the filter slider not disappearing. I had my slider at 2x and shut the gyros all off and it shows 1.05. Even in the lua script it shows the min at 0 and the max at 1000. But the 1000hz is grayed out. They are still off right? Or is there some kinda protection thing going on under the hood that you need to disable in cli or something?
Hi,I just started tuning using the blakbox and am working my way through all your tutorials. When I watched this one, I was a little confused: In your previous videos to filters tuning you explained how "inaffective" PT1 filters are compared to BIQUAD filters. But here u use the PT1 filter on the D lowpass filter. Could you please explain why u use the PT1 in this case? Many thanks in advance, Henri
BF dev group move away from BiQUAD to a dual PT1 approach. To keep a consistent message, i didn't buck that trend. A dual PT1 approach provides close to the same as a single BiQUAD.
@@uavtech And I just like to add: Many thanks for all your work and "follow along" videos, what they are if you watch them, consider fix flight manouvers fr your logs and then start with the filters videos. But everybody should be aware: Don't become a blackbox junky! I'm just getting a second Battery for my Renault Twingo with Converter to be able to fly, record logs, analise them on the Laptop on the flight field a whole weekend. - Force yourself: Also do some fun-only flights... ;-)
Hi! I disable all gyro lps filters on my quad. Flight very well with them off, but the escs burned after a few batteries. I didint crash, and props were new. Its a very clean build.... my gyro_scaled logs show a very tiny peak of noise on all axys. I don't know if the escs burned for having turned off the filters or if it was a coincidence.
I have an older flight controller with the "jittery" Bosch Gyro. Will the ability to bypass gyro low-pass filtering this make it a usable flight controller?
@@uavtech My bad - it's an ICM 20608 on the notoriously bad Omnibus F4 Pro Corner FC. What is BiQuad? Searching for it on YT brings up very old videos, and it looks to be inbuilt and enabled by default.
@@uavtech Got it. What I'm actually curious about is whether the PID loop frequency has any impact on the effectiveness of RPM and dynamic notch filters. In some of my setups post-filters noise level was lower with 4K compared to 8K, both with 2KHz blackbox log frequency. Just curious if you have seen any difference in filtering performance among different PID loop frequency
Ive watch a lot of videos and have noticed that you dont tune bigger quads. Does these tips for tuning PIDs and Filters the same for bigger quads? I have a iFlight XL10 and i cannot seem to get rid of low throttle bobbles, especially with doing a Power Loop. I've recently turned off all the extra features so i can properly tune my quad per your method. Thanks
9" tune - ru-vid.com/video/%D0%B2%D0%B8%D0%B4%D0%B5%D0%BE-MiBpVPRB-xw.html 7" octo tune - ru-vid.com/video/%D0%B2%D0%B8%D0%B4%D0%B5%D0%BE-UnZjL_PBR80.html I have tuned many 7" quads. I have another 8" octo coming in this week. The key with big quads is PD Balance 1:1 (generally) and pushing PIDs up (60's for D gain)
@@uavtech Thank you for the Links. I have a 10" quad, and for the filters i chose to not use the sliders as the values seems to feel better with lower values the sliders couldnt get to. i was kind of "Iffy" about messing too much with the PIDs, and after looking at the values they were lower than what you recommended, so i will fly some packs today in trying to tune the quad. you are the man!!!! thanks Alot
I was considering a bit the opposite: removing the dynamic filters, because the lowpass filters will still have a very good effect on higher frequencies. Is there something I overlooked with this idea?
Eh, lowpasses are only really effective at attenuating noise at 8x the cutoff. So that is noise above 800hz. So they are just not effective at killing motor noise that exists between 200hz and 600hz.
@@uavtech But we have RPM filters for motor noise. My question is what are dynamic filters doing if you don't have a damaged propeller? I'm planning on doing a test with that specific situation to see how it's taken care of with the current tools we have in BF.
@@LenordeFPV it is not doing much either. Also RPM filters still hits bent props as they still spin at the motor RPM. Just a higher amplitude vibration. I think in my Patreon video this week we'll take a look at that.
what do you do for filtering if you dont have dshott escs, i have littlebee 20a escs, possibly old ones as they dont look like what i see online. they do multishot.. this is on a quad 4s with 4 inch props. no blackbox as fc doesnt have sd card logger.
@@uavtech i slved my issues turs out there was a bad esc that had 2 capacitors on th board broke. replaced with little bee spring escs. whole differet quad now even at defaults
remark - 8:40 there is no propwash at your logs at 80 hz. It is on 80% throttle, but 80hz and above are clean during both full flihts). That's cool! it acctualy means to push d lowpass cutoff much higher after turning off gyro LP and say goodbye to prowash, right? ;-) of cause I consider mechanicaly good quads first of all) Although I used this strategy even on mechanically very bad quads. and it works! Just use little bit biffier (2307 instead of 2205 for example) motors with same batteries and kv rating so they do not overheat even on bad gyro signal and very high D gains. Then set gyro LP off + (try) dyn dterm LP like 200-300hz region. Yep, d term and pidsum logs might look terrifying, but who cares when your motors can handle it and the (bad) quad still flys smooth). RPM and dyn_notch are amazing stuff!!
At 80% throttle to blob is going up to 80hz. Good to hear it works on bad stuff too. It should with the DN on duty for cleanup now. Have to always give a disclaimer of some type so folks don't think it is for every quad out there regardless of mechanical.
@@uavtech your 80hz blob is realy small) or I probably had too much experience with budget stuff giving THE blobs time to time. Anyway thank you for sharing this small things giving 90% of success on perfomance.
Wonder if there is any correlation between frame thickness and motor noise? I.e. do frames with thinner arms handle motor vibration better than maybe 3mm - 4mm arms do. Maybe something to look into.
Hey! Could you be please so kind and let us know how did yiu managed to have betaflight on your phone?! I got Speedybee but that is just not the right tool for everything 😕
No doubt, another great video. But hmmm, the thing that keeps me from disabling the gyro LPFs until now where to have some kind of "safety belt" in case of beaten up props or other things causing an imbalance. Do you think the RPM and dynamic notches are able to handle this scenario without the help of the LPFs?
Joshua Bardwell does a really good job of taking the salient points of tuning (and he uses Mark as a reference) and distilling those down - beyond some of the more basic stuff, it gets to a point where you have to get over that knowledge hump in order to understand why changing one thing results in a lot of other stuff being different.
Yes I watch both and many others. But still for some reason don’t understand the whole process. Plus with 4.2 coming out you can’t go to far back to understand it as things have changed since the update with more options and things to change or Bogarde
@@uavtech omg i need upgrade then. Been meaning since swapping to 3" prop probably will need retune.. Any patreon discount codes.. militsry discount lmao.. thinking im do the advance video one when i get paid at end of the month
A 1st pack isn't the best pack unless you cycled the pack 4-5 times!! Even tho after about 15-20 flights it's downhill fast & super fast if brought down past 3.0v per cell!!
