Aller au contenu
forfait montage drone fpv
DroneBuilds

debat moteur rosser vs kabab


nochamo

Messages recommandés

petite reponse de Kabab au post /test de chrissrosser plus que douteux:

au dela du debat, son point de vue est plus qu’interessant (ps : désolé pour la mise en forme en screen foireux, pour ceux qui n’ont pas facedebouc)

https://www.facebook.com/groups/379155946182689/permalink/1163688917729384/?app=fbl

Citation
A brief history of 5" FPV freestyle motor sizes, how we arrived at 2506.2 and the endless height vs width argument.
(Please correct inaccuracies or major items left out - work in progress)
I estimate the birth of 5" FPV was some time around ~2008-2010. Before that we thought we needed giant props just to hover. We started with 1304/1306/1804 motors around ~2008-2010, awful. Moved on to 1806 motors not specifically due to size but because the quality of the DYS motor was better at the time. Nearly nobody could tell any difference other than more power because the flight code was impossible to tune. The Sunny Sky 2204 2300kv was the next hot ticket and was a marked improvement across the board but then ran into huge QC issues and so everyone moved back to the DYS and other 1806 motors. The flight code got a little better and there was a large camp of people that swore the 1806 was a much better performing motor than the 2204. Some time around late 2014 or early 2015, Cobra comes on the scene with a "2204" which was magically performing a solid 20+% better than anything else in existence. ~8mo later the world found out it was actually a 2205 and not a 2204. Emax then comes out with the red bottom 2205 with stronger magnets which was probably the very first really great performing mass market motor to ever exist in the 5" sector but ended up having a bunch of QC issues in the long run and died off. Soon after, ZMX enters the scene and works with Brother Hobby to manufacture a 2205 with further improved high performance magnets and some other more modern bits. The resulting motor is an absolutely awesome 2205. Brother Hobby steals the design and replicates it across everything they ever made and then T motor and other manufacturers do the same and we all move to 2206 which was again a pretty huge bump in performance across the board. Hate him or love him, Richard Shelton of Multicopter Builders loved him some 7" and as far as I know was the first one to ask Brother Hobby to make a 2207 with updated construction for him when we were all still futsing with 2205's. Almost immediately after 2206's popped up on the scene, we all realized 2207 was extra awesome sauce for 5" and nothing smaller made sense anymore. About a year later, Mr. Steele works with TBS and tests a bunch of motors and finds that the 2306 motor size feels better to him with whatever level magnets and construction he was testing. Before he gets his motor out, there were some manufacturers already working on the 2306 motor sizes and the Steele motor with other 2306's all came out around the same time. The 2306 quickly became the default and the community was and is still split on 2207 vs 2306.
From here, everything starts getting super experimental. I start testing all sorts of weird sizes with funding from the shop I was working with at the time. 2404, 2405, 2208, 2207.5, 2307, 2308. 2404 was worse than 2204, 2405 was super jumpy at the low end and lacked top end, 2208 felt amazing until it got too hot so I figured just not good for racing, 2307 was overpowered and hard to control at the time, didn't even bother putting the 2308 on a quad because it was so heavy...at the beginning of 2019, T-motor made us our 2207.5 prototypes and they were incredible! (https://www.instagram.com/p/BnN82-2lavW/). T-motor couldn't get their act together (because they were busy stealing the design for themselves) and we took it to Brother Hobby because they were too slow (https://www.instagram.com/p/Bn858qPlyE6/). 2207.5 still overheated a bit during racing so most moved back to 2207 or 2306 soon after. Somewhere in there Zoe worked on a 2009 for her 3D flying but it had way too much height and too little surface area which resulted in cooked magnets. EFAW also came out with a super light 2407 around this time which was awesome and unique because it used weak magnets to taper the power of the huge motor but fragile and the company faded away due to poor QC and poor business practices. Brother Hobby and T-motor ripped some design elements from EFAW on their way out. Later on, Bardwell remade the 2407 with GetFPV and then there were many other 2407's which were great but considered heavy (they're not heavy at all in reality) and very very powerful. The resulting quad you run a 2407 on is usually overweight and needs the power so the overall flight performance is fine considering it's weight but if the AUW was too low and the Kv was high where you like it to be, the quad is too powerful and hard to control if the motor has been made with high performance components. Literall nerfing a motor just because we can't fix it's issues is how this size worked out. A very legitimate fix that many motor designers and manufacturers still take today but I don't agree with.
July 2019 RCinpower sends me the Smoox 2306.5 1880kv wich felt like the absolute perfect balance of everything! Soon after FPVCycle starts and we re-create the Smoox with some modifications to reduce cost....before that motor was done, Brother Hobby sends me the 2806.5 at the end of 2019 for 7". The first few seconds of flight I noticed that this motor had a very distinct quality/luxurious feel to it compared to any other FPV motor I had ever flown and I just could not put my finger on what it was. At first I thought it was the large 11mm bearing so we started experimenting with huge bearings. RCinpower somehow crammed an 11mm bearing into the Bison 22.5mm motor but it was so darn heavy for the motor size and you'd have to fool yourself into thinking the big bearing was making a difference.
I canceled the Smoox remake for FPVCycle and started testing tons and tons of motor sizes again for 5" in search of that luxurious smooth feeling. Starting with a 29g 2803, I learned about the limitations of ESC code and physics the hard way. As long as you didn't give it the full beans, oh yes that luxurious altitude control feeling was there. I tried to get it to work with stronger or weaker, thicker or thinner magnets, then a 2803.5, then 2804, then taller 2804.5 magnets and smaller 2803.5 stator, but now it was just getting way too heavy. Moved down to 2605 but the Kv still couldn't be pushed up to the power level desired without 'physics/ESC' issues and adding another mm would be too heavy so again went down to 25mm. Figured out that this was the threshold where that smooth luxurious low altitude feeling started to taper away from the flight quality. 2505 felt like a 2306, nothing special. Having all that past experience with motor height changes, I knew that too much height would end up being too overpowered in the low end. Went up to 2506 and holy moly that was IT! Unfortunately the very fist batch still had some minor 'physics/ESC' issues as a result of Brother Hobby changing the magnet thickness and quality on us without informing us.
Went back to RCinpower after Brother Hobby screwed us out of a LOT of money and that was the end of my ~7yr relationship with Brother Hobby. Redesigned and re-tested the motor from scratch with RCinpower and added 0.2mm height and dropped the Kv by 10 in order to improve the 'physics/ESC' problem. That was the last batch that recently sold out. The next batch will be the same motor with significant durability improvements and hopefully a hot new color.

MANY MANY things have been left out but I'll add in some honorable mentions: Lumenier had a decent basic motor with no obvious issues in a wide variety of sizes pretty much the whole way through. RCX became the budget KING even though they were crappy motors that they could never get right. Most have forgotten about the now discontinued RCtimer 5010 made by RCX that was the endurance king on ~15-17" props for a massive length of time by this industries standards. T-motor is straight up stole just about everything they ever made. Their signature move was to take on a new client and replicate their product and release it under their own brand before completing the clients product. Brother Hobby is similarly shady and their signature move was taking very large dollar amount orders from their biggest clients and using that money to build out a new manufacturing facility instead of making the clients motors. Then they would take tons of small orders and manufacture those to pay the bills before delivering the large clients order which is now extremely delayed. Xing motors came on the scene and were doing some nice things but the parent company iFlight is similarly hit or miss shady and hard to trust. I still like Xing motors. Garila made a 2506 motor WAY early on during the 2206 era which was totally overlooked by most because we were blinded by what was hot at the time. Though their 2506 motor was run of the mill low quality and construction, those that knew it absolutely adored it while the rest thought they were nuts. They were genuinely ~five years ahead of me but they didn't even know what they had. AMAXinno is genuinely a German company with German engineering and made some meticulously beautiful motors that would unfortunately fall apart unprovoked. They're still making stunningly beautiful motors but struggle to gain traction in the market and I'm not sure if their QC is actually better.
Next step is testing a new magnet structure never before tested in this industry. Not a new magnet structure in general but I never considered testing it before because it's so difficult and costly to manufacturer. I'm focused on manufacturing more than specific developments because I hate finding something I cannot produce. We've figured out a new manufacturing method for them so now it's worth testing to see if it's a suitable direction to move in. I personally think it's not going to make any difference...hopefully I'm wrong.
Basic things learned through this journey:
  • I start with my Kv range+voltage desired when choosing or designing a motor because that's what governs the power output more than anything else.
  • Too wide with not enough magnet height is pointless as you're limiting how much you can magnetize the stator with the small magnets which reduces the stator utilization and magnet interaction...very bad and very very significant to motor performance.
  • Too tall with not enough width is a significant overheating issue and that barrier is around 2207 for 5" and is partly why 2306 is actually a better motor and feels the same after the 2207 warms up a bit (warm magnets perform worse).
  • It doesn't totally make sense but a heavy motor at the end of the frame arm doesn't actually reduce control performance as long as the motor is producing the power to make up for its weight...within limits, you still want your AUW around 580-630g ideally on 4.9-5.1".
  • HYPOTHESIS: Increased width changes the torque curve of the motor such that you shift the peak lower on the RPM range of the motor. Typically, you have reduced magnet height to compensate for the added weight so your top end performance is also reduced slightly.
  • Taller magnets in general lead to FAR more torque creation than width however that performance is higher up in the RPM range and in some cases, beyond the max RPM of the motor because it'll just overheat, nerfing the magnets before it gets there.
I wrote this because I've gotten so many messages about a recent video where a bias and/or inexperienced author runs some very simple limited tests yet absolutely confirms truth when he hasn't had the breath of experience in the industry to be able to compare anything.
An F1 driver can explain how to win a race all day long but when it comes time for you to race, you will not have the experience or skill needed. Similarly, an F1 engineer can explain to the F1 driver how to drive the car but the F1 engineer does not have the experience or skill to come up with the feedback needed to push the development forward. The goal here isn't to compare anyone to an engineer or F1 driver. I write this to make a point that you need engineering AND the real world to 'win a race'. You cannot do it ONLY in a lab and you cannot do it ONLY in the field.

If human names were just a string of numbers instead of words, it would be difficult/impossible to remember everyone's 'name' because there's little to no character or emotion attached to a long string of numbers. How can you 'name' a motors experience with just a single number? That's not going to get you very far.
I've never told anyone that they NEED to get my motor. In fact I often tell people to get a cheaper motor because they're too inexperienced and are unlikely to be able to capitalize on the benefits of a better motor or suggest a different motor that's better for the particular prop they're opting for....

 

 

 

 

 

Lien vers le commentaire
Partager sur d’autres sites

  • 3 weeks later...

J'avais pas vu ce post, juste le message de Kabab le mois dernier ^^

Les relations entre Kabab et Rosser ont l'air tendues ces derniers temps, c'est le moins qu'on puisse dire 🤣 

J'aime bien les 2, leurs idées sont tjrs intéressantes, mais ils sont à 2 extrêmes et faut toujours garder un peu de recul. Rosser est ultra-scientifique et comme Spatz, il se base sur de vraies données et sur la science. Il considère donc qu'il a tjrs raison, sauf qu'il peut oublier un paramètre ^^ Il a quand même assez souvent raison je pense, même si ça va à l'encontre de certaines idées bien installées. Kabab, c'est l'inverse, il fait tout à l'instinct. Il explore donc de nouvelles idées avec une approche plus "trial and error", mais y a donc une part de pifomètre et tire parfois des conclusions non étayées ou hâtives. Dans tous les cas ça apporte des débats intéressants.

Pour ces questions de moteurs, ces stators ultrawide ont forcément plus d'inertie en périphérie et ça les rend forcément moins réactifs. Et vu le nombre de soucis sur tous les moteurs de cette taille (qui ne touchent pas tous les setups, mais quand même), et même si l'expérience est intéressante, je ne compte pas m'y attarder davantage, les 22xx et 23xx on fait leurs preuves  en 5" 🙂 

Lien vers le commentaire
Partager sur d’autres sites

il y a 32 minutes, Jerome a dit :

ils sont à 2 extrêmes et faut toujours garder un peu de recul

c'est clair... et les deux n'ont pas forcément tords

je pense qu'effectivement les 25xx consomment plus de jus, comme le signale Rosser. mais il n'en est pas moins vrai qu'ils sont plus agréables (sous les 50% de gaz je trouve) . ils sont effectivement moins réactifs (tout comme du 48kHz vs 24kHz), mais tellement plus smooth..... mais la manière dont Rosser "analyse" le low control est un peu abusée (le quad pas tuné en stationnaire, c'est un peu n'importe quoi)

après c'est intéressant de voir que Kabab a testé toutes les tailles, et que le 2505 n'apporte rien par rapport a un 2306 (ce que tu avais vraisemblablement ressenti avec les Tmotor), et que c'est le fait de passer en 2506 qui procure ce bénéfice. il confirme aussi que le 2306.5 est très très bon.

Lien vers le commentaire
Partager sur d’autres sites

Créer un compte ou se connecter pour commenter

Vous devez être membre afin de pouvoir déposer un commentaire

Créer un compte

Créez un compte sur notre communauté. C’est facile !

Créer un nouveau compte

Se connecter

Vous avez déjà un compte ? Connectez-vous ici.

Connectez-vous maintenant
×
×
  • Créer...