MSH Flybarless system test flights - Dario Neuenschwander

[brtmac]

Here's how I see it. 3-axis mems gyro silicon is like <$10/unit at volume. Some of the new stuff does all of the hard work in silicon and just gives you outputs of the angular motion. Translating that to servo movements to match helicopter motion to transmitter inputs is relatively simple, not much harder than doing it for just a tail gyro. If you can constrain the software to a fixed set of parameters, such as a few specific helis (ie. Mini-Protos, Protos-500, Maxi-Protos), and don't worry about configurability beyond what a typical flybar setup gives you, it shouldn't be hard to do the programming. Really, the programming is easy, the testing is hard. By limiting it to a specific set of heli's you reduce the testing immensely.

So, consider a FBL unit tailored to a specific heli, or family of helis, similar to a heli head that's designed for a specific heli. The parts are cheap. The programming and testing should be fairly simple. So, the idea of a $50 FBL unit that works perfectly for the specific heli's it's designed for doesn't seem unreasonable. No more so than designing a specific motor or head geometry.

If you add to it the ability to accept satellite receiver data and put $2 worth of servo connectors on it, that $50 pretty much vaporizes for a lot of users who can now just buy a couple of satellites instead of a full blown receiver. Thrown in some diversity circuitry and you have the perfect setup for almost $0 incremental charge.

One thing you have to remember in the digital world is that all costs tend towards $0. Which is great because we can then concentrate all of our addiction money on the actual hardware.

[Otatiaro]

Hello,

@brtmac : I have some projects for you if you want ... what about a 3$ coffee machine? a 10$ PC? a 100$ car?

Just one little thing to begin with ... you say "<$10/unit at volume", ok, but then you talk about a very specific FBL unit, only spektrum sats, only MSH helis, where is the volume then? Paradoxical ...

About price, 10$ for the sensor, ok, but a fbl unit is not just a sensor ... add a microcontroller, connectors, crystal, a PCB (can come handy ...), a case (don't want a shrink wrapped unit, no? and ho by the way, first hire a designer for the case), a manual (or you can guess how to set it up by yourself ...), cables (RX to unit, etc), some LEDs and at least one button for setup (if not a usb cable).

Ok now you have a bunch of parts, for more than 20/25$ I assume.

Next step, get all the things put together ... pick'n'place the parts, solder them, do inspection / quality control, have some defects of course (add the price of the defect to the price of the good ones). Pay someone to get the board in the case, screw it, put everything in the packaging, etc.

We're now at least to 30/35$.

Plug it in ... nothing happens? Sure, you need a software, specifically developed for your product. Pay someone for months to get a good firmware, plus at least one pilot to do testings, plus test helicopters, parts, etc. Divide this cost by the - assumed - number of units you will sell during the lifetime of your product (a few years in this case).

Bare minimum here 50$. Of course you want periodic updates with new features, so you need to continue paying the developer(s). Of course I assume your developer(s) are good enough so that there is no additional dev costs (1st version is fully functional, including EMC, EU compliance, etc).

Then add thermal calibration (you don't want your fbl unit to drift with temperature changes ...).

Now you have what seems to be a functional FBL unit, you already paid a lot more than 50$, and it's time to put it on the market.

Add VAT, company taxes, shipping charges, marketing charges, stock costs, loan interests, etc, etc, etc. PLUS, of course, the margin of the distributor and/or shop (40, 50% of the final cost?).

With absolutely no margin for yourself, you already are well over 180$ at the BARE minimum, and you did all this - hours and hours of works, stress, etc - just for fun, and you still need to find a way to pay your bills.

So yes, sure, on volume sensors are quiet cheap, totally agree, the rest of your thinking is totally wrong, you CANNOT sell a FBL unit for 50$ today (maybe in a few years, who knows). I mean a sellable FBL unit, not a cheap, non usable Chinese unit (no R&D cost, just copy someone else, no test cost, just don't test it, no quality control cost, just don't do it, etc) that will burn like a CC ESC ...

I quote you :

"Translating that to servo movements to match helicopter motion to transmitter inputs is relatively simple, not much harder than doing it for just a tail gyro. If you can constrain the software to a fixed set of parameters, such as a few specific helis (ie. Mini-Protos, Protos-500, Maxi-Protos), and don't worry about configurability beyond what a typical flybar setup gives you, it shouldn't be hard to do the programming. Really, the programming is easy, the testing is hard. By limiting it to a specific set of heli's you reduce the testing immensely."

I mean if it is so easy, I'm really looking forward to see your unit in flight, and buy it for 50$ (free shipping to France I guess ).

Already heard of PID, Kalman, FIR/IIR, DMA, USART, I2C/SPI, PWM/IC ... all (this list is not exhaustive, of course) these "details" you need to absolutely master to get something airborne without flybar?

I don't know (officially) anything about what MS Heli is doing ... but I'm in the consumer electronics market for some time and I'm really fed up with the standard "this chip cost only X$ and you dare to sell the whole product you built around for twenty times more???".

I used to sell wireless, universal trainer systems for 50€ all inclusive, +2€ shipping, and some people dared to tell me it was a lot too expensive for what it is (I had only a few percent of margin, and no distributor/shop costs, the only distributor interested laughed at me when he saw how much I was earning on each unit sold, and told me I was completely crazy).

Global market ...

Thomas.

[EDIT] BeastX cost is 155€ in Europe (220$?). See, I'm not that wrong, that's about 30/50$ of - raw - margin, except you'll pay taxes on this margin. That's not a lot, just enough to get the bills paid.

[Thunder Fighter]

Originally Posted by brtmac

The programming and testing should be fairly simple.

lol, spoken by someone who doesn't have much experience of writing software, let alone commercial firmware I would guess.

//Dennis.

[Otatiaro]

Hi again,

Just a question, how many persons in the world do you think have the skills to get this job done (I mean all skills in one person, not a team)?

Now how many persons have the skills to get the job done for a commercial product (means absolutely no bug, high speed, etc)?

Come on, do a guess !

Thomas.

[brtmac]

I am not saying that someone can buy all of the pieces and build this and program it for $50 for a one off unit. I'm saying that I believe it is possible to do in a commercial manner. Basically it's already been done, though for $70 instead of $50.

http://www.horizonhobby.com/Products...ProdID=BLH3501

[Otatiaro]

Yes,


But again that's different ... the software in the HH baord is a "lite" version of the BeastX.

No dev cost, no R&D.

Plus, there is absolutely not configuration because you know the helicopter AND the power system AND the servos. MSH couldn't sell FBL unit working only with a specific set of servos, that would mean selling only 10 units a year.

And this one is not the best of what can be done, slow servos refresh rate, etc.

This is just not the same target, I d'ont think people are waiting on MS Heli to release a degenerated FBL unit ... MS Heli is more on the "best quality/best performance" side, with affordable prices.

Thomas.

[EDIT] Plus, I assume HH is producing in asia to get the lowest costs. MS Heli is the European way of doing things

[brtmac]

If you re-read my post that's exactly what I'm talking about. I said a FBL controller that's made for a specific heli or heli family with very little configuration. I know the mCPX has no adjustment, but it wouldn't be hard to at least allow for gain adjustments.

I don't think that you'll be able to get something like a full blown V-Bar for $50 any time soon. But I think it should be possible to get a controller that's bundled with a specific heli for that kind of money. i was very surprised that the new Blade 450 didn't have an FBL option. I would not be surprised to see some sort of all-in-one unit for it (receiver, fbl controller in a single unit) with very little, if any, additional cost.

[exit44rc]

Should do a co-branded deal with BeastX and produce a BeastX - MSH version that ships with the ideal settings for the Protos line of helis.

Mikado is doing something similar with Henseleit.