Spektrum DSM2 Details - Post Analysis of a Crash

Problems, experiences or just something to say about RC gear? Say it here.

Spektrum DSM2 Details - Post Analysis of a Crash

Postby skyhigh » Fri Jun 25, 2010 4:50 am

I abandoned my DSM2 DX6i and all receivers this month after a glitch crashed my heli last spring. It was a serious glitch and not pilot error. My T-Rex 500 was hovering in front of me with 3 other T-Rex 500's as we were doing a 4-way "hover show" for the camera. I had my DX6i and there were 3 other DX7s so there was four DSM2 systems running in close proximity.

Within one minute of hovering and without any warning, T-Rex 500's elevator servo twitched and 1/8th second later, the servo went full forward elevator. My heli crashed inverted into the weeds. Lucky it went into some thick weeds otherwise it would have been major damage. Luckier it didn't flip backwards otherwise my buddies would have been driving me to the nearest emergency room for sure.

I've seen the YouTube videos showing that DSM2 selecting two channels after the initial link from transmitter to receiver. Since the 2.4GHz band has 1MHz wide channels, does DSM2 do any sort of frequency hopping within this 1MHz channel?
skyhigh
 
Posts: 19
Joined: Wed Jun 23, 2010 7:15 am

Re: Spektrum DSM2 Details - Post Analysis of a Crash

Postby RCModelReviews » Fri Jun 25, 2010 6:07 am

The 2.4GHz band doesn't actually have any predefined "channels" as such -- it's just 80MHz or so of undefined space in which equipment can operate.

The DSM2 system chooses to break up that 80MHz into 80 channels of 1MHz width -- but other radios break it up into more or less channels of different width, depending on their design.

The DSM2 system is now getting quite long in the tooth and is not dynamically agile -- which is to say that once it has chosen its two operating frequencies, it stays on those frequencies and will not shift even if both are totally obliterated by noise.

Even worse, sometimes the DSM2 system will choose two channels that are virtually side-by side rather than well spaced within the band. This makes the system far more vulnerable to interference because it's then effectively operating as a single-frequency system with no redundancy.

Read this article for more information on that scenario.

It's worth noting that Spektrum is now the only brand-name 2.4GHz radio that doesn't use frequency hopping.
RCModelReviews.com, just the facts.
User avatar
RCModelReviews
 
Posts: 2120
Joined: Tue May 04, 2010 3:40 am

Re: Spektrum DSM2 Details - Post Analysis of a Crash

Postby skyhigh » Fri Jun 25, 2010 8:54 am

Just to test if I understand this: if two DSM2 transmitters just so happen to select the same two channels, would this result in certain lockout for both pilots? I think you mentioned in another post about "temporal density" in that transmitters are not always transmitting similiar to networking packets. Does this provide at least some measure of resiliancy for DSM2 in this extreme example?
skyhigh
 
Posts: 19
Joined: Wed Jun 23, 2010 7:15 am

Re: Spektrum DSM2 Details - Post Analysis of a Crash

Postby RCModelReviews » Fri Jun 25, 2010 8:59 am

In theory, several Spektrum radios could co-exist without too much trouble on the same two parts of the band. However, in practice there appears to be a marked degradation of the link when this happens, which is why you'll usually find that once there are 40 spektrum systems already working (on 80 different parts of the band) it becomes very hard to get the 41st system to actually link up (or so I'm told).
RCModelReviews.com, just the facts.
User avatar
RCModelReviews
 
Posts: 2120
Joined: Tue May 04, 2010 3:40 am

Re: Spektrum DSM2 Details - Post Analysis of a Crash

Postby skyhigh » Fri Jun 25, 2010 9:18 pm

I powered on my transmitter before the others powered on their's. I believe what happened was a 2nd DSM2 radio grabbed one of my channels and a 3rd DSM2 radio grabbed my second channel. Feasible? I have no way to check after the fact.

Warning to all: Never do a close range hover of 3 or more midsize helicopters using all DSM2 systems. Someone could end up in the emergency room!
skyhigh
 
Posts: 19
Joined: Wed Jun 23, 2010 7:15 am

Re: Spektrum DSM2 Details - Post Analysis of a Crash

Postby rpstar » Fri Jun 25, 2010 11:56 pm

Ok, now you are being ridiculous. You are completely and totally speculating about the cause being the DSM2 receiver. The most likely cause was a power system issue which caused brownout of the receiver. Yeah, maybe Spektrum receivers don't tolerate lower voltage as well as others (don't know as I've not seen any direct comparison on this) but if you give them good power they work just fine. How do you know your crash was not a power system issue? Perhaps your ESC overheated dropping the voltage to the receiver? Perhaps it's of lower quality than that on your friends heli's or a bum unit? Perhaps your servo's themselves wigged out? Your assumption that the Spektrum radio is crap without further due diligence to determine the true cause is patently ridiculous. However, I do encourage you to go ahead and sell your equipment on Ebay as the more used equipment we get out there the more the prices will drop for me. :)

I do agree with RC Model reviews that a simple firmware change is in order to avoid the rare issue of the transmitter choosing two channels next to each other. It's probably the transmitter making this decision on it's own so I'm guessing newer radios is where you will see this update since radios like the DX6i don't have an easy firmware upgrade path meaning you would have to send it back to the manufacturer to be updated. The DX8 looks like a good candidate for example since it has firmware update either via the SD card or perhaps a USB plugin, not sure.
rpstar
 
Posts: 2
Joined: Fri Jun 25, 2010 11:47 pm

Re: Spektrum DSM2 Details - Post Analysis of a Crash

Postby RCModelReviews » Sat Jun 26, 2010 12:15 am

It's true, the vast majority of 2.4GHz incidents relate back to bad installation and poor choice of power for the servos and receiver.

If everyone flew 2S A123 packs I *bet* the number of "unexplained crashes" would drop significantly overnight.

It's unfortunate (for Spektrum) that one of the weakest points of the DSM2 gear so far has been its brownout performance. They have improved this by using fast-reboot and components that work better at low-voltages but they still lag a wee-way behind most other brands.

If I were head of development at Spektrum, I'd implement an SPC (supplementary power connector) like Hitec have done and I'd also implement a FHSS variant of their RF system. The underlying hardware *will* support it (albeit perhaps not as well as Futaba, Hitec or others) but it would effectively silence the critics by addressing the two key flaws in DSM2.

What's more, the DSM-FHSS could be made backwards compatible or user-selectable so that all those existing DSM2 receivers could be used with DSM-FHSS-capable transmitters.

Why don't they do it?

Lord only knows. Either they don't have the skills to design a system like this (rather unlikely) or the beancounters have just said "R&D reduces profits, stick with what we've got until sales tank and *then* develop a better system".

The average RC flying field is a *lot* different (in terms of 2.4GHz band use and noise) than it was when DSM was first introduced. What worked really well then may not work quite so well now that every man and his dog is filling the band with RC transmissions.
RCModelReviews.com, just the facts.
User avatar
RCModelReviews
 
Posts: 2120
Joined: Tue May 04, 2010 3:40 am

Re: Spektrum DSM2 Details - Post Analysis of a Crash

Postby skyhigh » Sat Jun 26, 2010 12:27 am

rpstar wrote:Ok, now you are being ridiculous. You are completely and totally speculating about the cause being the DSM2 receiver. The most likely cause was a power system issue which caused brownout of the receiver.


First, thanks for the suggestions. Yup it is speculation but I can't sit silently instead of bringing up the topic in an open discussion forum so lets discuss. But are you not also speculating yourself with the brownout? You have no evidence nor have I given you access to inspect my radio or helicopter. We have to start somewhere.

Another possibility is the DX6i transmitter itself. There was a recall on the DX6i for the potentiometer specifically on the elevator servo! My DX6i was not in the batch codes HorizonHobby listed but I sent the radio in anyways. They cleared it for operation and included a AR6100 receiver on the house.

I've since repaired that T-Rex 500. Cracked blades, two of the DS510 servos needed gear replacement and two linkage arms replaced. Damn lucky considering the nature of that crash -- I love weeds! I've flown this T-Rex 500 again on the DX6i with nobody else flying and also while wearing hockey equipment on :). My curiosity got the better of me. No problems! Same servos, same ESC, same batteries, same AR7000 receiver. Piro flips, inverted backwards, stall turns, tick tocks, rainbows and no glitches.
skyhigh
 
Posts: 19
Joined: Wed Jun 23, 2010 7:15 am

Re: Spektrum DSM2 Details - Post Analysis of a Crash

Postby skyhigh » Sat Jun 26, 2010 12:37 am

rpstar wrote:Ok, now you are being ridiculous. You are completely and totally speculating about the cause being the DSM2 receiver. The most likely cause was a power system issue which caused brownout of the receiver.


I just have to re-iterate the point that the incident occured in level hover at the top 95% of the main battery. There was no major servo stress pushing cyclic and collective at the same time like a piro flip tick tock or something crazy like that.

A brown out could happen anytime, yes. But during hovering? Oh, coincidentally, there were 3 other DSM2 AR7000's operating only a few feet away from me and I was the one to turn my transmitter on first. I was using new 40C LIPOs as well.
skyhigh
 
Posts: 19
Joined: Wed Jun 23, 2010 7:15 am

Re: Spektrum DSM2 Details - Post Analysis of a Crash

Postby skyhigh » Sat Jun 26, 2010 1:00 am

RCModelReviews wrote:If everyone flew 2S A123 packs I *bet* the number of "unexplained crashes" would drop significantly overnight.



Do you mean a 6V A123 receiver pack? Hmm, interesting. Maybe that will be my first use of a LiFePO pack.

RCModelReviews wrote:
What's more, the DSM-FHSS could be made backwards compatible or user-selectable so that all those existing DSM2 receivers could be used with DSM-FHSS-capable transmitters.

Why don't they do it?


Probably because implementing an additional DSM-FHSS would be an admission that there is a problem with DSM2. Which of course we know there is no problem with a non-hopping 2 channel system these days. :o
skyhigh
 
Posts: 19
Joined: Wed Jun 23, 2010 7:15 am

Next

Return to Radios and Servos

Who is online

Users browsing this forum: No registered users and 4 guests

cron