Ok thanks. Too windy today and I need some fresh props, I will post a log later on in the week.

I'm sorry to say that those Avan props are fairly terrible, and because of that they may be hard to tune, so I would recommend you swap them to something else.

Thanks, yes I think I've watched ALL your videos... I guess I stuck with the Emax as I mostly do slower flying and even enjoy 3S on the 2900kv motors (maybe I am imagining a little more low end and hangy feel), but will try out the HQs on this frame.

A question on your frame, noticed there are no dogbones at the camera plate joints, I think CNC recommends them for easier fitting. Is that something easy to add or is your design open source and available for tinkering?

Just an update on the frame cutting. I heard back from Nick and he kindly offered to replace the thin plate as that was out of their tolerance. And he does offer chamfering service, for the Scythe plate it would be an extra $3 for one-side and $11 for both-sides. I believe for the bottom side, they need to flip the cut plate into a wooden jig. So it is good to know there is that option for those who don't want to do their own filing.

(17-Apr-2023, 11:31 AM)MomoBrut Wrote: There is more, I have little time, so just briefly. Log a flight (blackbox debug mode GYRO_SCALED, 2kHz) in which at the beginning you add full throttle sharply and briefly (split of a second) from a hover, then of course take it back to the hover lower above the ground, then slowly increase from the hover to the maximum gradually for about 1s and finally do at least 5 times drop into the propwash, preferably perpendicular to the propwash from a drop down and then sideways in an extreme 180 degree turn. Upload the log, the whole thing should definitely be within 100 s.

My PIDs and filters are completely different, I have different motors too.

Ok switched to the HQ props, and they fly a lot better than the Avans at high speeds. I think this frame is very dangerous as it is such a blast to fly fast I am flying beyond my skill. I have crashed a lot more, and even forget to pay attention to low battery warnings... I noticed motor 3 had all 3 motor screws come loose, and it seems slightly rougher than the others so maybe I already banged it up somewhat. I added back some filter, not perfectly smooth, but still seems relatively well behaved.

I have run the log attempting the moves you described. But logging is at 1.6khz as apparently this AIO has a BMI gyro and not the ICM gyro listed on the spec sheet.

https://drive.google.com/file/d/1sGGCeuR...share_link

First, the unfiltered spectrograms from all axis look very good, there is no noise between about 50 and 160 Hz, which is excellent. Below 50 Hz it's noisy, but filters can't solve that. This must be solved by suppressing resonances on the frame. So, using any filters below about 150 Hz is completely unnecessary. I assume that you have bi-directional D-shot enabled, otherwise the motors noise would not be filtered out. The motors noise is perfectly filtered out by the RPM filters, the only thing left are the remnants above 60% of the throttle between 350 and 500 Hz, but these are so few that you rather don't recognize it than you do.

If you want to suppress propwash, you must first reduce or eliminate the unnecessary filters that introduce delays into the PID loop. Propwash cannot be resolved if the delay is too big. So I recommend to test complete remove of the Gyro LP Dynamic 221-425, it is probably useless there, because this part should be solved with RPM filters and this filter only introduces delay, the lower the cut-off frequency is, the more delay the filter introduces. Leave the Gyro LP Static 2 as it is for now (and this one is most probably also not necessary, you can test it later). RPM filters work great, leave them as they are. And D-term filters are the key, your range of 67-135 is very low, again in the range where you have no noise. My D-term filters start at 97 Hz, which is a fairly sharp setting, but the result is essentially no propwash. And only then solve the PIDs, I would start by PIDs defaults and just setting the master multiplier to 1.4 or 1.5, the difference will be huge compared to what you have got. And just for fun, my Roll PIDs are 90, 160, 63, 47, 240, but that's the extreme I'm comfortable with, I also have got other motors and all settings are for the T3.5x2x3 propellers, you have to find your PIDs.

And a note, what really bothers me about BF, if you set the filters in one profile, the filters will be set to the new values for all other profiles as well, so when tuning filters and PIDs at the same time it is necessary to keep this in mind. So when you switch a profile with other PIDs, you have filter settings for PIDs from original profile, watch out for that. It's a fairly common mistake. And the filter settings decide how far you can go especially with the D value and thus also with all the other PIDs values, because they must be proportional to each other. And propwash, if there is one, can only be dealt with on such small quads with much larger PIDs than is usual for 5 inches.

I typically turn on rpm filtering and just ADD more filtering with the slider until it flies smoothly, but now I see the big benefit of properly reducing the filtering. Thanks for the detailed explanation and advise.
I too was very happy with the rpm filtering result seeing the all black gyro plot in the motor range
I will drop the dynamic gyro lp, move the d-filter, and then start playing with some higher pids.

The GHF420AIO-BGA should be arriving in the next days, so looking forward to putting together the second frame, but hopefully the ICM gyro will be fine... I don't have enough of the T-motor 1404s left, but have some larger 1504.5/1604 motors. Do you think they would work well with this frame, or just add weight and reduce flight times as it seems to handle so well already with 1404s?

OK, I'm happy to help. Give it a try and let us know how it goes. As for engines, this has already been addressed ... https://intofpv.com/t-3-5%E2%80%9D-scyth...#pid185235

What FC are you using now in Scythe? And when you have that new JHEMCU FC, I'm really interested to see how it behaves.

For the motors ok, I also had the same feeling. I might go ahead and install the bigger ones temporarily while I order more 1404s, maybe I should try 3800kv.

The board I am currently using was supposed to be this one JHEMCU GHF411AIO ICM 42688P F4 AIO Flight Controller 4 in1 40A, however the one I actually received has the BMI270 instead... maybe that was a good thing as the ICM seems not even listed on JHEMCU github.
https://www.aliexpress.us/item/2255800980488864.html

Fingers crossed for the 405BGA/ICM board, I will post some logs when I receive it.

Well the BGA board finally arrived, and it looks promising no weird default tune or filter settings (bd dshot and rpm filtering was also enabled). Unfortunately I must have toasted the MCU, probably applied too much heat trying to clean up the battery lead (this board has a very hefty ground plane). The MCU did power up after soldering, but seems it died when connected to USB. All voltage rails still working, but only solid red LED. I was looking forward to putting this together, but guess its not happening

On my first setup, I tried the tune changes but because I had some bent props, the motors got very angry even at low throttle. I think turning off the LPF1 may have been too aggressive. With it turned back on, on the most part it is not bothered by the imperfect props except at high throttle. When the weather clears up, I will try again with some fresh props.

If I built a 4" that was 82g, with LED and battery strap, surely you can do better!

(20-Apr-2023, 04:38 PM)mstc Wrote: For the motors ok, I also had the same feeling. I might go ahead and install the bigger ones temporarily while I order more 1404s, maybe I should try 3800kv.

The board I am currently using was supposed to be this one JHEMCU GHF411AIO ICM 42688P F4 AIO Flight Controller 4 in1 40A, however the one I actually received has the BMI270 instead... maybe that was a good thing as the ICM seems not even listed on JHEMCU github.
https://www.aliexpress.us/item/2255800980488864.html

Fingers crossed for the 405BGA/ICM board, I will post some logs when I receive it.

Most people prefer the BMI270. Its much less noisy. You can run more agressive filters and thus much better prop wash handling.

(20-Apr-2023, 04:38 PM)mstc Wrote: For the motors ok, I also had the same feeling. I might go ahead and install the bigger ones temporarily while I order more 1404s, maybe I should try 3800kv.

The board I am currently using was supposed to be this one JHEMCU GHF411AIO ICM 42688P F4 AIO Flight Controller 4 in1 40A, however the one I actually received has the BMI270 instead... maybe that was a good thing as the ICM seems not even listed on JHEMCU github.
https://www.aliexpress.us/item/2255800980488864.html

Fingers crossed for the 405BGA/ICM board, I will post some logs when I receive it.

Until there is a log, nothing is known. According to the default settings, I would guess nothing at all. My two pieces were absolute junk, even though they had normal default settings.

(01-May-2023, 07:50 PM)romangpro Wrote: If I built a 4" that was 82g, with LED and battery strap, surely you can do better!

If it is not for racing, I wouldn't compete on how much it weighs. Only one thing is important, to get under 250 g with everything that should be there and to have the best flying quad according to the category. Which is something different every time. 4 inch weighing 85 g will be a "specific" relatively fragile category.

(02-May-2023, 08:51 AM)MomoBrut Wrote: Until there is a log, nothing is known. According to the default settings, I would guess nothing at all. My two pieces were absolute junk, even though they had normal default settings.

The ones you tried were BGA version also? Then maybe I should not bother ordering a replacement and choose a different board.

(01-May-2023, 07:50 PM)romangpro Wrote: If I built a 4" that was 82g, with LED and battery strap, surely you can do better!

Really with a vista and carries a naked action camera nicely? I would guess there must be some serious compromise on the frame and motors. Please share some flight footage on how it flies.