It's raining outside, so I maiden it in the underground parking lot. First impressions, it's a beast, the power to weight ratio is absolutely insane. Very little throttle is needed for normal flight, a tiny throttle bump shoots the quad forward. You need to reset your brain, it's really sharp. I have to do some reasonable throttle management, some throttle curve. I put 4 batteries in to calibrate the current measurement. The GR1404/3850 motors are incredibly efficient in this lightweight quad. Slightly faster cruising in underground parking lot with the GNB 4s660 HV battery used up 450 mAh, the final cell voltage of disconnected battery was 3.7 V and the flight lasted an incredible 10 minutes and 15 seconds. When the weather gets better, I'm going to tune the PIDs and give it a hard time...

I was a little worried about the naked Vista overheating, I tested at 700mW. Absolutely no problem, it manages to cool down nicely in flight, so thumbs up. So far very satisfied.

That’s a nice build.
It could be my imagination but it seems like 3.5in props work the best for the sub 250 class.
I’ve been looking for another 3.5” frame but nothing has really got my attention until I seen your build.
155 grams with a vista is amazing.
I hope to see some footage.

So 3 more batteries today, the rain has stopped. I did the initial tuning. It is necessary to completely turn off both low pass filters on the gyro. At first I just moved the cutoff frequencies up, but the best thing is to really turn them off. This alone improved the propwash. I moved the master multiplier to 1.3. I also set dynamic idle to 40 and enabled the voltage sag compensation. There is still some propwash left, but it is already very good. It will take some time to fully tune it.
As for the flight characteristics, so far the feelings are mixed, it behaves like a race quad. Huge acceleration and brutal cornering without inertia. So great for aggressive flying in limited space. Unfortunately, inertia is missing, so the quad cannot be thrown very well, at the same time it is very floaty. When flying hard, the flight time is just over 6 minutes for the 450 mAh. The unpleasant thing is that from the fourth minute of the flight, full throttle causes a voltage sag below 3.5 V. The truth is that it is not much needed, you are on the moon in no time. More to come ...

(14-Oct-2022, 02:16 PM)Rob3ddd Wrote: That’s a nice build.
It could be my imagination but it seems like 3.5in props work the best for the sub 250 class.
 I’ve been looking for another 3.5” frame but nothing has really got my attention until I seen your build.
  155 grams with a vista is amazing.
  I hope to see some footage.

The 4 inch props can not be driven without compromise to get the quad under 250g. The 3 inch props are just small, the quad flies too much like a micro quad. If the takeoff weight of the 3.5 inch quad is around 220g, it behaves very similarly to the freestyle 5 inch which weighs around 600 to 650g. The Scythe is too light, the 155g makes it something else, I have yet to understand. I also have a GNB 4s900 HV that I might try, maybe the extra weight along with the possible higher current will be a benefit... we'll see.

Holy smokes, that thing weighs nothing for what you put into it!

(14-Oct-2022, 02:55 PM)MomoBrut Wrote: The 4 inch props can not be driven without compromise to get the quad under 250g. The 3 inch props are just small, the quad flies too much like a micro quad. If the takeoff weight of the 3.5 inch quad is around 220g, it behaves very similarly to the freestyle 5 inch which weighs around 600 to 650g. The Scythe is too light, the 155g makes it something else, I have yet to understand. I also have a GNB 4s900 HV that I might try, maybe the extra weight along with the possible higher current will be a benefit... we'll see.

  I use BH 1504.5 3900kv and 4S 850 on my Nexa 3.5 and it weighs 235grams.
  I think the extra weight gives it just the right momentum to make it more locked in.
 I put on a Thumb Pro 4K over 250 grams now but it still catapults great.
  I do notice my voltage dropping below 3.5v per cell more than I like but I cam getting great flight time.

   It’ll be interesting to see if a heavier pack gives your quad benefits.

   I hope more designers try building 3.5” frames.
  I think 3.5” can be sub 250 and still be durable enough for freestyle or light weight racing.

(14-Oct-2022, 04:23 PM)Suros Wrote: Holy smokes, that thing weighs nothing for what you put into it!

Yeh, it is light ... 

(14-Oct-2022, 05:07 PM)Rob3ddd Wrote:   I use BH 1504.5 3900kv and 4S 850 on my Nexa 3.5 and it weighs 235grams.
  I think the extra weight gives it just the right momentum to make it more locked in.
 I put on a Thumb Pro 4K over 250 grams now but it still catapults great.
  I do notice my voltage dropping below 3.5v per cell more than I like but I cam getting great flight time.

   It’ll be interesting to see if a heavier pack gives your quad benefits.

   I hope more designers try building 3.5” frames.
  I think 3.5” can be sub 250 and still be durable enough for freestyle or light weight racing.

I also considered the Nexa 3.5, finally decided on the non-replaceable arms and significantly lower weight, which led me to the Scythe.

For your interest, I am sending a photo of my current freestyle quads that I fly with. From left Apex 5 HD, Smart 35 HD and finally the new Scythe 35 HD. The Apex is simply incredible, there's no point in discussing it, and thanks to the fact that I'm flying naked Hero8 in a full case TPU, I got to an all-up weight to 603g, which really suits me a lot. I managed to tune the Smart 35 HD so it's amazing, I use really high PIDs to eliminate propwash, it flies very similar to the APEX in tighter spaces and all-up weight is 225g. The Smart is a squashed X. The Scythe is a true X, so it's actually a bit bigger than the Smart, while its all-up weight is only 155g, if I use the same battery as in the Smart, I'm at 188g.

I found time today and tried to tune the propwash. Well I have to say I'm a bit disappointed, even after 3 more flights I didn't manage to improve it much, in the last flight I activated the blackbox. While the yaw and pitch are relatively quiet below 100 Hz, the roll axis is noisy, quite a bit in the entire range from 0 to 200 Hz and above at more than about 40% throttle, below 40% throttle it is relatively quiet. I've never seen anything like it. The noise is not dependent on rpm, which would be filtered by an RPM filter, it is in the whole range. PIDs cannot be increased well, the maximum is somewhere around 1.3 on the master multi, more leads to hums and instabilities in the motors.

The first screenshot is the spectrum of gyro scaled, the second is gyro filtered both for roll axis. The third and fourth image is in the same order for pitch axis. You can see that the pitch is fain, no problem, the rpm filter will absolutely clean it up. On the contrary, the roll is a disaster.
Does anyone have any idea what it could be that the pitch and yaw are calm and the roll is so bad?

(16-Oct-2022, 03:40 PM)MomoBrut Wrote: I found time today and tried to tune the propwash. Well I have to say I'm a bit disappointed, even after 3 more flights I didn't manage to improve it much, in the last flight I activated the blackbox. While the yaw and pitch are relatively quiet below 100 Hz, the roll axis is noisy, quite a bit in the entire range from 0 to 200 Hz and above at more than about 40% throttle, below 40% throttle it is relatively quiet. I've never seen anything like it. The noise is not dependent on rpm, which would be filtered by an RPM filter, it is in the whole range. PIDs cannot be increased well, the maximum is somewhere around 1.3 on the master multi, more leads to hums and instabilities in the motors.

The first screenshot is the spectrum of gyro scaled, the second is gyro filtered both for roll axis. The third and fourth image is in the same order for pitch axis. You can see that the pitch is fain, no problem, the rpm filter will absolutely clean it up. On the contrary, the roll is a disaster.
Does anyone have any idea what it could be that the pitch and yaw are calm and the roll is so bad?

  How does it behave with default pid settings ?
  How squished are your gummies?
  Did you use a nut on your stack screws before sliding on AIO?
I sometimes wonder if Rosser uses theory to boost sales but he claims a nut under stacks or AIO help vibration.

I equipped the frame so that it will be easy to service and also so that the Vista will not be on the same screws as the AIO, I did not want to burden the screws with additional mass. For service, the bottom screws are loosened and the top plate flips open for clean access to everything. The top plate is connected to the bottom only by Vista wiring.

(16-Oct-2022, 05:08 PM)Rob3ddd Wrote:   How does it behave with default pid settings ?
  How squished are your gummies?
  Did you use a nut on your stack screws before sliding on AIO?
 I sometimes wonder if Rosser uses theory to boost sales but he claims a nut under stacks or AIO help vibration.
 

Default PIDs have nothing to do with how it resonates, moreover, they are optimized for a 5 inch quad that weighs over 650 g, so for a completely different quad.

What makes me nervous is the crazy difference between the roll and pitch axis, as I already wrote, I have never experienced anything like this before. And that's why I don't even think that it will be the way, how are the components installed in the quad. But of course I will try, but that's exactly what annoys me, I don't have time for this and this always takes a lot of time.

The screws from the bottom plate for the AIO are first secured with steel nuts, only on these secured screws is the AIO fitted through gummies and secured with plastic nuts. So I have it as recommended by Chris. Gummies are squished at medium hard.

The wiring form Vista and Rx are twisted, they are routed outside the gyro and when screwed together the PCB compresses them slightly so they don't move. The cable from the antenna is currently free, and leads again outside the gyro. These are all normal things, it's not the first time I've done this, it surprises me all the more.

The first thing I'll try is to tighten the screws further, then I'll try to fix the antenna cable, then I'll continue on the contrary by throwing out the steel screws and leaving it to the gummies only.

There will be nothing but trial and error here, unless someone comes forward who has dealt with a similar matter.

(16-Oct-2022, 07:42 PM)MomoBrut Wrote: Default PIDs have nothing to do with how it resonates, moreover, they are optimized for a 5 inch quad that weighs over 650 g, so for a completely different quad.

What makes me nervous is the crazy difference between the roll and pitch axis, as I already wrote, I have never experienced anything like this before. And that's why I don't even think that it will be the way, how are the components installed in the quad. But of course I will try, but that's exactly what annoys me, I don't have time for this and this always takes a lot of time.

The screws from the bottom plate for the AIO are first secured with steel nuts, only on these secured screws is the AIO fitted through gummies and secured with plastic nuts. So I have it as recommended by Chris. Gummies are squished at medium hard.

The wiring form Vista and Rx are twisted, they are routed outside the gyro and when screwed together the PCB compresses them slightly so they don't move. The cable from the antenna is currently free, and leads again outside the gyro. These are all normal things, it's not the first time I've done this, it surprises me all the more.

The first thing I'll try is to tighten the screws further, then I'll try to fix the antenna cable, then I'll continue on the contrary by throwing out the steel screws and leaving it to the gummies only.

There will be nothing but trial and error here, unless someone comes forward who has dealt with a similar matter.

  I hope you get it figured out.
  I have gotten FC with overly sensitive gyros and tuning did nothing but make it worse.Mine wobbled and shook.
  I tried everything manufacturer recommended but finally sent it in.It had a defective gyro or in their words a bit too sensitive and they replaced it.
    You could have gotten a AIO with bad gyro or chipset???

I was hoping it would be without a problem. It did not happen. My time for this project is currently exhausted. This is for a long time, the weather here is worth nothing. I don't know when I'll get to it.

It is difficult to judge if the gyro is bad. And that I would convince the seller from Ali Express about the defective gyro, I have no illusions about this at all, he is just a seller who does not know anything about it, gyro is a dirty word for him and usually does not even understand English.

So, I'll leave it at that for now, maybe someone will show up who has experience with a gyro that behaves absolutely differently in the roll and pitch axis.

I discussed this with the guys on the UAV Tech discord. According to them, it is probably a problem with the gyro itself or its power supply, which I did not believe, because the noise would be in all axes and this was also confirmed in the end. I only managed to add a 100 uF capacitor to the 3.3V rail and redo the log. Absolutely no improvement. I'll try the 5V rail again today, but I don't believe it if it didn't help at 3.3V. I'll have time to do some testing today. It is of course time-consuming, to disassemble the quad, change something there, put it back, log in and on and on... For me it is a problem, I cannot fly in the place where I change the configuration, I have to leave every time, time is running and there is no time... Very frustrating, no manufacturer selling a defective AIO can ever make up for this.