Thank you for the vote of confidence. :) Yes, I do strive fo
scientific objectivity as much as possible during these sorts o
discussions. Since we humans are subjective by nature complet
Spock-like objectivity may be forever beyond us (Malcom Gladwell's boo
"Blink" is a fascinating read on the subject), but all we can do is mak
our best effort, and I endeavour to do just that.
I really doubt that the reason the 2.4GHz TX has only 6 model memories - either
Futaba didn't have the flash space or else it is an intentional restriction to
position the TX to the low end of the market.
That's basically a stupid argument. It's 100 fold easier to move to 2.4GHz by
using other peoples chips than to have to grapple with the intricacies of a
completely new system. I'm prepared to bet that Futaba's comparatively late
entry to the 2.4GHz field has some dependence on their decision to go it alone
with the consequent huge development effort entailed by that.
They had a manufacturing problem where a process was not applied to part of
their poroduction run. I really don't see what that has to do with your
speculative "rush to get something on the market" ?
Me neither :-) In fact all the TXs I've owned in the past have been Futaba ones
but the lamentable failings of the 9C Super I own mean I am now looking to jump
ship when the 2.4GHz market settles. Ironically, Futaba's version of 2.4GHz
seems like it may have major advantages over the Spektrum / JR implementation
(in respect of x-country glider application) so I may have to go with them after
No. That sadly is a blind.
They have a hardware circuit that makes two fundamental errors.
- they store something that is supposed to be unique,and unalterable in
FLASH ram, which is very alterable.
- They fail to design around WELL KNOWN problems with FLASH RAM under
which power supply voltages outside of known limits cause completely
unpredictable behaviour in it. Because of the nature of Flash, you have
to erase and rewrite whole sections of it to change a single bit. Lose
power at that point, and you end up with a whole section erased, and not
The hardware workaround is a big fuckoff capacitor and some voltage
detection stuff that signals to the circuit that power is failing,and
holds enough voltage in the cap to complete a write cycle, and then
shuts down cleanly.
Me neither. But in this case they have made a big cockup, and haven;'t
yet gotten around to admitting the full extent.
Personally speaking, I think that with the bugs out, the FASST stuff
looks a market beater. That doesn't mean I am prepared to excuse the
COVER UP on the cockup. Altho the cockup itself, is - well
understanable, if not excusable.
Oh, aren't we impressed with ourself.
Stop it with the teenage sounding baloney. Who are you
trying to impress anyway? Me? That won't happen. You have
already proven yourself to be an obnoxious twit.
Ed, I'm truly sorry for you. You must be in a lot of internal pain t
lash out at a total stranger in this way.
Have a nice life.
Don't go there man or you may end up getting your hand bitten as
I wrote this summary of the issue in my club forum:
And started getting bashed by Loyalists when I asked if it looked lik
a good synopsis of the issue.
THEN, when I said, if Futaba does not like what is being said about th
issue in the Futaba created Advisory thread they can moderate and delet
unwanted posts, Krysta, from futaba slapped my hand for such
suggestion and closed the thread. (start at post #189 to the bottom.)
I keep falling back on their original claim of always getting it righ
the first time... heh... kinda like putting together a $3,00
transmitter and loading it with Windows CE :D :D :D :D :
| >> I just bought a 9C. What failings can I expect?
| >> mk
| > A fairly sizeable latency for one:
| I didn't see my TX there, are we on the same TX? Fut 9C ?
Futaba 9C - R149DP is the 2nd to the last item on the first page, and
the last item on the second page.
In any event, a few things to consider here --
-- this is CCPM, where the 9C is known to be slower.
In other modes, it's faster than this.
Also note that this is a helicopter mode -- for planes, things
are not so bad.
-- this is PCM, where everything is known to be slower.
Using a PPM RX will give considerably lower latencies.
(Though to be fair, the spread spectrum gear generally has
similar features to PCM receivers, without the additional
cost or latency.)
-- I doubt anybody who reads this group has a 120 ms reaction time
(approximately the maximum measured latency of the 9C + PCM RX) and
most of us probably couldn't tell the difference between the worst
case 9C at 85 ms and best case 12Z + R5114DPS at 20 ms.
Yes, it's a shortcoming of the 9C. But it's a pretty minor one. And
this chart is basically set up to give it's worst case scenario.
My 9C has a problem with V-tails and 4 servo wings. This problem was
fixed, but I have one of the first ones made, so it still has the
problem. Newer ones should have this fixed.
Either way, I wouldn't buy a 9C new now. But not because it's a bad
TX -- it's not -- but there are better ones out there now. And any
new gear I buy would be spread spectrum. My next TX will probably be
a JR X9303, unless something better comes out.
| > Let's not forget that Futaba was one of the first with 2.4
| > GHz gear for R/C cars, which preceded Spektrum by many
| > years.
And years before that, Futaba had spread spectrum industrial R/C gear.
I don't know why they took so long to get it into modelling R/C usage.
Perhaps because they already saw themsevles as a/the industry leader,
why try harder? I do not know ...
| True. It's also true that RC cars don't roll, loop, do snap-rolls, etc,
| and a glitch is rarely destructive as the model stays on the ground.
You'd be surprised how much damage an out-of-control R/C car can do.
Consider an 1/8th scale car running into your shins at 60 mph -- or
worse, hitting something and going airborne, flying at your face ...
| I attended a seminar given by Spektrum engineering manager John Adams
| back in 2006; during his talk he commented that providing a solid RF
| link to a model aircraft was an extremely difficult technical challenge,
| for the reasons above - the need for a continuous link without any
| dropouts at all, and the continually changing orientation of the model
| wrt the transmitter.
Cars have similar issues and needs, but the ranges are a lot lower.
| Also when the Tx frequency hops, the Rx has to do likewise, which
| means the two have to sync up again after every hop. And what
| happens when something goes wrong and the sync up doesn't happen
| immediately as intended?
This sync-up thing is pretty well worked out. People have been doing
spread spectrum like this for what, 50 years? They know how to make
it work, and it works. Basically, this is just FHSS FUD.
| Most forms of RF link can tolerate the brief dropouts that result from
| this sort of thing (FHSS sync gone wrong). Model aircraft cannot.
| Apparently this is why Spektrum tried FHSS, found it wanting, and went
| to a DSSS solution.
I know they weren't happy with FHSS, but I've never really seen a
technical explanation for why. If you have a reference, I'd like to
| > Statistically, Futaba's SS system is much more robust.
| I take shiny marketing pamphlets with a large grain of salt
You must have had lots of salt when you read Spektrum's original press
releases and FAQs then!
| (Not that Spektrum is perfect, for instance it would have been
| really nice to have two-way telemetry and eventually force feedback
| from the servos actuating the control surfaces on the model. :) )
The car people already have built-in telemetry from Nomadio and have
had it for a while. XPS has promised it for planes, and their
literature makes it sound like it's available now, but I can't seem to
find a place to buy it -- which suggests it's not ready yet.
Doug, I agree that a 120 mS reaction time to a sudden one-time externa
stimulus is exceptional - but you're overlooking a major factor. Ther
are many situations where we input periodic (rhythmically repetitive
control inputs for RC models, a classic example being a rolling harrier
In these cases, it may take a lot longer than 120 mS to implement th
first finger movement, but the subsequent rhythmic movements can easil
be timed to considerably greater accuracy than 120 mS by anyone - n
need for special super-fast reflexes.
There are also many situations when we anticipate an upcoming contro
input in advance, and therefore can time it to much higher precisio
than 120 mS. For instance, if you put your model into a snap-roll o
spin, you anticipate the point at which you need to input correctiv
control inputs to stop the roll or spin at the desired heading.
Several years ago I worked as an engineer on a team designing a hig
quality loudspeaker system with internal DSP processing. We wanted t
find out if such a speaker could be used as a live monitor by musician
while they played their music. In order to find out if the latency/dela
from the processing in the speaker would be an issue in this situation
I took one of my electric guitars in to work one day and did some quic
experiments playing scales and groups of sixteenth notes through a pur
delay line (with adjustable delay) to see how much delay mattered.
What I found surprised me. A mere 20 mS of delay was quite detectable
and had a deteriorating effect on my ability to play. 15 mS was abou
the threshold where I couldn't quite be sure if I could detect the dela
or not. Below 10 mS, I couldn't feel the effect of the added delay
Anything over 30 mS was really nasty - it became increasingly impossibl
for me to play the guitar at all as the delay was turned up beyond thi
Keep in mind I'm no Van Halen. My guitar playing is amateurish and m
top speed is not particularly fast as guitarists go. And even so, 20 m
was unacceptably large for me. Surely 15 mS would have been too much fo
someone with better guitar technique than myself.
So is 120 mS delay in an RC transmitter an issue? Perhaps not whe
we're using slow servos with 200 mS delay of their own, and almos
certainly not if the transmitter is being used to fly a sedate mode
with inherently slow responses. But the guys who fly twitchy 3
helicopters with lightning-fast 70mS tail rotor servos say they can fee
the difference. And I can see no reason not to believe them.
I CAN say for sure that I can feel the difference in latency between
model flown with 5 V DC power to the servos and the same model flow
immediately afterwards with 6 V DC to the servos, because I tried thi
on a .40 size electric model with a switching BEC that let me try bot
voltages by simply moving a jumper. If you look at typical servo specs
they often pick up about 40 to 50 mS of additional delay at 5 V compare
to 6 V. And I could easily feel this difference when, for instance
flying snap-rolls or spins: at 6V the model would stop snappin
noticeably sooner after I neutralized the controls. If in doubt,
highly encourage you to try it yourself - I was using a $8 switching BE
from United Hobbies, so there is not much cost outlay in trying th
I'll take your word for it. :) I have little personal experience wit
Doug McLaren Wrote:
A very good po>
I'm not so sure...most 2.4 GHz FHSS RF links we use have the odd shor
dropout that simply isn't of much consequence. Anyone who's used
cordless phone or a Wi Fi ethernet card has probably experienced thes
little glitches. They simply don't matter in the applications mentione
above. But even a very short glitch causes loss of pilot confidence whe
From memory, John Adams said that the FHSS systems had more shor
dropouts in the RF link, and the problem was never overcome. There wa
also reduced range and S/N ratio with FHSS due to the increase
Sure. When I worked as an engineer I quickly found out how wide the ga
is between what the engineers say and what the marketers say!
That was one of the reasons I attended John Adam's talk back in Ja
2006. This guy is an engineer and RC pilot (he brought a passel of hi
micro electric helicopters with him to the talk). I think he wa
Spektrum Chief Engineer at the time, now he's Engineering Manager o
something. Exact title notwithstanding, after his talk and the flood o
questions he answered from the very interested audience, I came awa
with a quite positive impression of the Spektrum engineering teams work
He didn't give us marketing BS - he gave us technical answers t
All I can say is:
1) Looks like a good summary to me
2) People who make the mistake of identifying their personal ego with
corporate product are easily upset - all you have to do to a confirme
Starbucks junkie is remind them that McDonalds coffee has been ranke
higher quality than Starbucks coffee by reviewers, and you have
Krysta here on RCG has been good. I guess Futaba reps are a littl
thin-skinned about this issue right now; I know I would be if I wa
spokesman for a product with an Achilles heel like this one!
Well, there are quite a few. Someone else has mentioned high latency (near the
worst of all txs available when used in PCM mode), a few of my pet peeves are
as follows :
1. Broken throttle-controlled airbrake design means trims are dangerous in acro
mode. (This was fixed in glider mode from the 9C to 9C Super so why they
couldn't have fixed it in acro mode too beggars belief). This means that an
intermediate plane like the H9 Ultrastick cannot be completely programmed on a
9C as it can on the JR equivalent. And before you point me at the howto on the
Futaba website, it doesn't work. And when I say the howto doesn't work I mean
hopelessly, uselessly, laughably doesn't work, and stands no chance being made
to work whatever you do.
2. Side-control sliders can't be set up to allow eg flaps to be controlled with
neutral flap at slider full up and full flap at slider full down. Best that can
be done is to have no flap movement from slider middle to slider up and full
flap at slider down. This is a problem with the sliders and no amount of mixing
will cure it.
3. The only servo that can be slowed is (bizarrely) the throttle servo. The
manual alleges this to be so an IC engined plane can be made to perform like a
turbine with a slow engine spool up. It's just plain weird IMHO, much more
useful would be a servo slow that can be used on other things like retracts.
4. Futaba's advertising states the 9C to be a 9 channel TX. It's not really :
even if you do splash for a Futaba PCM dual conversion RX it is not possible to
set the ATV for the 9th channel. And the key downside of a dual conversion RX
is if you stack the plane you can't just discard the crashed XTAL, you have to
send the RX away to get the internal XTAL replaced by a service engineer.
5. The rotary dial turns the wrong way cf the way the display selection changes.
Or maybe that's just me :-) I did ask Ripmax but they said there's no way to
6. There are others...
Having said that and got it off my chest, my 9C Super has been reliable in the
18 months I've had it and it's not a bad TX for all that. I just get more and
more tee'd off with the programming style that seems to be designed to permit
exactly those facilities that Futaba deem suitable for a mid range TX and to
prevent anything more adventurous from being tried. Oh for a mainstream TX with
the kind of flexible programming model shown by the Profi 4000, now demised.