Shop our Airplanes Products Drone Products Sales
Thread Tools
Jan 22, 2014, 01:37 AM
Cá bhfuil an linn snámha?
spykez's Avatar
Discussion

Latency of radios today (2014)


Wanted to discuss this.

A new purchase of an rtf align got me thinking about radio latencies (or lag if you want to call it that). That's cos I am a spekky guy with a DX8 as a main radio and was astonished to see futaba's claim of a 6.8ms frame rate with S-FHSS.

I know the main source of latency in my case is my brains, but once a move becomes ballistic and more unconscious or 'trained' call it what you may, I wonder what latencies then mean.

Then there's the FBL latency as well...

This old RR thread has been revised relatively recently.

http://m.rc.runryder.com/helicopter/t172571p1/

This suggests the 14SG in specific conditions ie 12 ch FASSTest with the included receiver has the lowest latency, almost unbelievably low.

One main motivation for posting this thread is trying to figure out if an 8J is worth it (not sold by tower anymore it would seem) or just pay twice the money for a 14SG which is heavier and larger and probably has more stuff in it than I need.

Questions I have:
1. Anyone have any great ideas on how to measure latency in a whole system now that a lot of us have fbl's in?

2. Has anyone looked at the latency of S-FHSS, the rr figures do not include them...

I have since figured out S-FHSS probably takes 2 x 6,8ms frames to send 8 channels - that's 13,6 for a full refresh. But I wonder what that really means... In terms of latency... With fhss will this mean a latency of 13,6 all the time? If it misses a frame, just wait for the next?

3. Are there any quicker radios? Seems like freq hopping with a slightly broader band like FASST is the way to go rather than fhss, but I really dunno.

I'm quite sure a faster radio won't make a big difference to my flying but I'd sure like to try ; )

I should add that I'm SOLELY a heli guy, don't really need more than 6 channels, well, maybe 7 for a bailout...

Cheers gents!

Addendum: Just thought I'd add this video in

Latency in RC systems explained (10 min 51 sec)
Last edited by spykez; Jan 22, 2014 at 02:29 AM. Reason: Added in this video I found.
Sign up now
to remove ads between posts
Jan 22, 2014, 05:18 AM
Registered User
Futaba's claim (servo frame rate 6.8) it's kind of warning and propaganda at the same time.

It's the warning if you wanna use certain low tec analog servos. They won't work.

It's propaganda as long as the customers confuses the number with latency.

The benefit of a low servo frame rate while using analogs is that they behave like digitals. More holding power, more torque.

Also, the 6.8ms should be seen in context of PPM and PCM radios/receivers. Most PPM transmitter/receivers have/had a rate of round about 20ms (50Hz, 50 times a second). Most PCM radios/transmitters have/had put out 14ms servo frame rate.
Horizon/Spektrum's 11ms translate into 91Hz. For comparison, Futaba FASST in high speed mode claims 7ms servo frame rate
and Airtronics FH-3 6ms.

Typical specs for digital servos allow frame rates to the servo motor as fast as 3ms. That's a 333Hz update rate.

Taking about 2.4GHz radios, data rate (in Kbps) is another very important aspect...

That's radios. But another bottle neck is servos. There are laggy servos and fast servos...
Jan 22, 2014, 05:22 AM
Registered User
This video explains servo frame rate:

Testing an analogue servo with a lower PWM frame rates (1 min 18 sec)
Jan 22, 2014, 06:18 AM
Cá bhfuil an linn snámha?
spykez's Avatar
Agree about servos. Ok, let's assume I kit out a 500 flybar with I dunno, top of the line coreless jobs, MKS,BK/RJX what have you...

What radio you you reckon would be the fastest?

I already have a dx8 with a spare ar8000 somewhere. 11 msec frame. ?n Needed to send 8 channels.

I have a t6j that came with the align. 6.8 msec frame. 2 needed to send 8 channels.

I don't have a 14sg, but take a look at those rr figures...

Which is the fastest... I'm scratching my head. Questions... should I switch to s-fhss on my bigger helis or stay spekky. Or should it be.fasst or -est? Lol....
Jan 22, 2014, 07:57 AM
AndyKunz's Avatar
I have to ask this - what do you feel is your reason for needing lower latency? All but the very best pilots can't tell the difference (including some who think they can - we've experimented with that). There are other valid reasons, I'm just wondering where you're coming from with the question.

Andy
Jan 22, 2014, 10:50 AM
Registered User
Andy, you could ask the same about the difference in servo steps/resolution(1024, 2048 or higher). Lot´s of information are just for the "propaganda leaflets"...

The real benefit of nice latancy numbers comes to light in car racing but not while flying EPO planes.
Jan 22, 2014, 11:01 AM
AndyKunz's Avatar
I know that, that's behind the asking.

Actually with servos, if you have good servos and 6" long arms on them it's very easy to see the difference. You just need to watch deadband and PWM resolution.

Andy
Jan 22, 2014, 11:45 AM
Cá bhfuil an linn snámha?
spykez's Avatar
Sorry double post ; )
Last edited by spykez; Jan 22, 2014 at 11:56 AM. Reason: double
Jan 22, 2014, 11:53 AM
Cá bhfuil an linn snámha?
spykez's Avatar
Quote:
Originally Posted by AndyKunz
I hae to ask this - what do you feel is your reason for needing lower latency? All but the very best pilots can't tell the difference (including some who think they can - we've experimented with that). There are other valid reasons, I'm just wondering where you're coming from with the question.

Andy
4th last line of the OP. "I'd sure like to try". No less no more.

I know essentially it is neurons initially being the limiting factor, but any learnt skill can become ballistic or unconscious.

Then you're dependent on stuff that happens downstream, ie radio firmware delay, presence of module, receiver firmware, fbl if any, then servos.

Lots of variables I agree.

I am just asking because I have a cheap s-fhss radio and am confused by the stats manufacturs are giving us. Is my s-fhss radio as fast as FASSTEST? If it is just as secure and fast I'll keep using it, it's nice and small and light and feels great (T6 or I may go T8)

Bottom line, I'm curious.

Don't need to get defensive : )

PS, in 11msec dsmx via spm9645s to a v bar - 6 channels used, how many channels get sent in one frame? Or do all 8 get sent? Also I see no way of going 10 bits (1024) via dsmx 11msec ,seems to add an extra bit to 2048. Do two packets get sent one after another to transmit 6 (8?) Channels???

Thanks f or answering my questions : )
Jan 22, 2014, 12:01 PM
Cá bhfuil an linn snámha?
spykez's Avatar
PPS, you should know I am radio agnostic, been using spekky since join date mainly for bnfs, but man that futaba t6 feels comfy.
Jan 22, 2014, 12:32 PM
↓↘→ + (punch)
theKM's Avatar
Quote:
Originally Posted by AndyKunz
I have to ask this - what do you feel is your reason for needing lower latency? All but the very best pilots can't tell the difference (including some who think they can - we've experimented with that). There are other valid reasons, I'm just wondering where you're coming from with the question.

Andy
I'm curious how you ran the testing of the pilots... was simply a blind "pick the difference" and let them do whatever they wanted to determine?... I've found many (most?) don't even understand the maneuvers where latency becomes an important factor, and if you just left the pilots to figure it out on their own, your results don't surprise me much.

...I would also be curious as to some of the pilots you had access to who managed to tell the difference with consistency (no need to talk of those who couldn't, just curious who some of these laser accurate thumbs guys are).

...and also curious how bad the manufactured latencies were that managed to go undetected.
Jan 22, 2014, 01:49 PM
AndyKunz's Avatar
Quote:
Originally Posted by theKM
I...I would also be curious as to some of the pilots you had access to who managed to tell the difference with consistency (no need to talk of those who couldn't, just curious who some of these laser accurate thumbs guys are).
Well, we have some of the very best pilots on the planet working here. And they didn't know they were even being tested. As to the details, sorry, can't share.

Andy
Jan 22, 2014, 01:50 PM
Registered User
Quote:
Originally Posted by AndyKunz
I know that, that's behind the asking.

Actually with servos, if you have good servos and 6" long arms on them it's very easy to see the difference. You just need to watch deadband and PWM resolution.

Andy
6" ? Well then, have fun in the hobby or in the marketing department!

http://threadsmagazine.assets.taunto...ar_bill_lg.JPG
Jan 22, 2014, 01:53 PM
AndyKunz's Avatar
Quote:
Originally Posted by spykez
4th last line of the OP. "I'd sure like to try". No less no more.
I guess it's just a personal test to see if it's detectable then? I guess I'm too practical - I usually have a good reason for "like to try" when I do something. It'll be neat to hear your results.

Quote:
Bottom line, I'm curious.

Don't need to get defensive : )
I didn't think you were. It's cool.

Quote:
PS, in 11msec dsmx via spm9645s to a v bar - 6 channels used, how many channels get sent in one frame? Or do all 8 get sent? Also I see no way of going 10 bits (1024) via dsmx 11msec ,seems to add an extra bit to 2048. Do two packets get sent one after another to transmit 6 (8?) Channels???
11ms only applies to servo ports AILE, ELEV, and AUX1 in early radios (thru DX8) and adds RUDD in DX10t and above. The other channels are transmitted at 22ms refresh rates. It's also somewhat adaptive based upon bind mode.

Andy
Jan 22, 2014, 02:11 PM
Registered User
Andy, let's take about data rate. How about DSMX? This might explain the slow servo frame rate of 22ms for most of the channels.