HobbyKing.com New Products Flash Sale
Reply
Thread Tools
Old Aug 02, 2012, 06:55 AM
Live for speed
GoFaster's Avatar
USA, CT, Bethany
Joined Mar 2004
1,371 Posts
Question
Nano outputting 16 ch in CPPM mode?

Jim -

I am trying to get PPM support working with a Nano for an AutoQuad multicopter FC. Despite lots of effort by the devs, I've been getting inconsistent or erratic results. This morning I put a scope on the Nano's output and it appears that it's sending out 16 channels per frame (max that the Tx is capable of)?
No wonder, since the FC is only expecting 12 channels max.

I thought the Evo only sent out the number of channels defined in the Servo Assignment screen. I have 6 channels defined in the first 6 slots. All the rest are blank.
http://www.rcgroups.com/forums/showp...postcount=3943
Is the above no longer correct for Evo firmware v3.41?

Config:
Multiplex Royal Evo12 upgraded to v3.41 firmware, defined channels are in UNI mode.
XPS Nano v.17 in PPM mode, connected to Pin 1 for positive going pulses

Am I misinterpreting something?
Thanks,
- Felix
GoFaster is offline Find More Posts by GoFaster
Last edited by GoFaster; Aug 02, 2012 at 11:20 AM. Reason: more info
Reply With Quote
Sign up now
to remove ads between posts
Old Aug 02, 2012, 11:02 AM
Live for speed
GoFaster's Avatar
USA, CT, Bethany
Joined Mar 2004
1,371 Posts
Some more scope captures and questions. All except the last is capture at the Nano's pins.

It appears that if a channel's minimum endpoint is set somewhere below -100%, the TX module or the Nano gets confused and the PPM stream is corrupted.

It seems OK if the max endpoint is set 100%-110%, looks like the max pulse is clamped to ~2.1ms

Also, if more than 6 channels are defined, it appears that the next undefined channel has a default pulse width of ~1ms, compared to 1.5ms of other undefined channels. Is is because although the Nano seems to be capable of handling more than 6 channels in PPM, this is unsupported?

I've also seen guard pulse lengths of ~2.5ms which is shorter than the 2.7ms you've mentioned. This is when multiple stick and switches are at full deflection.

PPM support was recently introduced for this FC and I'd am trying to get the Nano working in PPM mode. 10/11bit Spektrum and S.BUS is working. I'd rather not stick a PWM->S.BUS encoder plus hex inverter between an XPS rx and FC
GoFaster is offline Find More Posts by GoFaster
Reply With Quote
Old Aug 02, 2012, 11:50 AM
Live for speed
GoFaster's Avatar
USA, CT, Bethany
Joined Mar 2004
1,371 Posts
Even more captures :)

Even though this FC currently supports 8 channels, I need a minimum of 6 to operate it.
I redefined the Evo servo assignments with 6 channels and looked at the captures when all controls are at minimum (endpoint -100%) and maximum (+100%).
GoFaster is offline Find More Posts by GoFaster
Reply With Quote
Old Aug 03, 2012, 08:50 AM
Live for speed
GoFaster's Avatar
USA, CT, Bethany
Joined Mar 2004
1,371 Posts
Interim conclusion and workaround

My theory:

It appears that Jim has clamped the maximum PPM frame length to 30ms to accommodate a max of 12 channels. What ever time is left over after the +/-2ms max X 12 channel pulse widths is the guard pulse time. I am guessing he limited frame length to maximize the frame rate.

However with my Evo 12 upgraded to 3.41 firmware, the Tx(?) is now outputting 16 channels (EvoPro's max). Any undefined servo assignments default to a pulse length of 1.5ms (mid position) with one odd exception. You can see that in the scope capture from the prev post where only Ch1-6 are assigned and the rest blank.


The issue crops up when defined channel controls are moved such that their total pulse widths + the default 1.5ms x Undefined channel leaves not enough time for a valid guard pulse. In the other capture from my previous post, where I maxed the values for Ch1-6, the guard pulse is only 1.1ms long. This makes it look like a channel value pulse and impossible to reliably detect the start of a frame.

Jim, can you always ensure a valid guard pulse >> 2ms by either increasing max frame length or variable frame length?
Also, currently available failsafe settings are either last hold or fixed value. Please add another option to shut off PPM output entirely. The FC looks for a complete absence of pulses for x amount of time as a loss of signal/failsafe condition.
Also to recap weirdness from prev post, setting a min channel value to somewhere less than -100% corrupts the frame.
And if only a few channels are defined, the next undefined channel has a width of 1ms unlike the others at 1.5ms; deosnt really matter but it's unexpected.

Would be great if you could fix the Nano firmware but at the very least make sure your upcoming new Rx's works with a Evo 3.41 in PPM mode.

In the meantime, I've managed to come up with a workaround by defining assignments for the unused channels, assigning them to no widget and then setting their low/neutral/high positions in servo calibration to -100% (1ms). On my Evo12, it won't let me calibrate Ch 13-16 so they are stuck at 0%(1.5ms). This leaves me enough frame time to have 7 usable channels and still have a valid guard pulse length.

Thanks,
- Felix
GoFaster is offline Find More Posts by GoFaster
Last edited by GoFaster; Aug 03, 2012 at 08:57 AM.
Reply With Quote
Old Aug 03, 2012, 04:26 PM
Xtreme Power Systems
Lake Havasu, AZ
Joined Jun 2005
15,910 Posts
Quote:
Originally Posted by GoFaster View Post
Jim -

I am trying to get PPM support working with a Nano for an AutoQuad multicopter FC. Despite lots of effort by the devs, I've been getting inconsistent or erratic results. This morning I put a scope on the Nano's output and it appears that it's sending out 16 channels per frame (max that the Tx is capable of)?
No wonder, since the FC is only expecting 12 channels max.

I thought the Evo only sent out the number of channels defined in the Servo Assignment screen. I have 6 channels defined in the first 6 slots. All the rest are blank.
http://www.rcgroups.com/forums/showp...postcount=3943
Is the above no longer correct for Evo firmware v3.41?

Config:
Multiplex Royal Evo12 upgraded to v3.41 firmware, defined channels are in UNI mode.
XPS Nano v.17 in PPM mode, connected to Pin 1 for positive going pulses

Am I misinterpreting something?
Thanks,
- Felix
Reduce the number of channels by turning them off in the servo menu. After you do this, you MUST rebind the receiver. The binding tells the receiver how many channels to use and if you change that number afterwards, it won't matter - and odd things can happen. There is nothing that needs to be changed on the Nano side, it gets it data from the transmitter.
JimDrew is offline Find More Posts by JimDrew
RCG Plus Member
Reply With Quote
Old Aug 03, 2012, 04:32 PM
Xtreme Power Systems
Lake Havasu, AZ
Joined Jun 2005
15,910 Posts
One other thing... the EVO's standard response mode is 16 channels, and the Fast Response mode is 12 channels. So, that is a possible way to reduce the number of channels too.
JimDrew is offline Find More Posts by JimDrew
RCG Plus Member
Reply With Quote
Old Aug 03, 2012, 05:01 PM
Live for speed
GoFaster's Avatar
USA, CT, Bethany
Joined Mar 2004
1,371 Posts
Quote:
Originally Posted by JimDrew View Post
Reduce the number of channels by turning them off in the servo menu. After you do this, you MUST rebind the receiver. The binding tells the receiver how many channels to use and if you change that number afterwards, it won't matter - and odd things can happen. There is nothing that needs to be changed on the Nano side, it gets it data from the transmitter.
Thanks Jim, I have tried that already.
I only have channels 1-6 defined in the Evo's servo assignment screen. The rest (7-16) are blank.
Rebind the Nano with the Evo and then enable PPM mode. Power cycle Evo and Nano.
I am still seeing 16 channels at the Nano's PPM output pin. Just like the captures in post #3 above. I cannot seem to stop the Tx from sending more the defined number of channels.

What is the 12 channel Fast Response Mode?

Thanks,
- Felix
GoFaster is offline Find More Posts by GoFaster
Reply With Quote
Old Aug 03, 2012, 05:17 PM
Xtreme Power Systems
Lake Havasu, AZ
Joined Jun 2005
15,910 Posts
The fast response mode is a mode where the frame is shorter and uses a max of 12 channels.

I just looked at the 3.41 firmware, and I see that the EVO always sends 16 channels in normal mode (no way to stop it). So, the killing the channels trick does not work anymore. I am not sure when that stopped - probably with 3.x So, you can try to putting it into the fast response mode.

The problem is this - the new EVO firmware does not use or support PPM at all. So, the frame rate of <25ms is going to be too fast for outputting 16 PPM channels. The fast frame rate mode might work, but I think the frame rate might be too fast for 12 channels too.

What I probably should do is cap the number of channels in PPM mode to something usable, like 10 channels or less.

Remember, a PPM frame is calculated as: max PWM time x number of channels + 5.7ms. So, with 12 channels that would be:

2.1 x 12 + 5.7 = 30.9ms

That is the amount of time a 12 channel frame can take. If you have a 25ms frame, you are going to be way out of sync! 16 channels is even worse:

2.1 x 16 + 5.7 = 39.3ms

This is why you can't even fit 9 channels in a standard 22ms frame:

2.1 x 9 + 5.7 = 24.6ms

Graupner uses a variable frame rate to get 9 channels - and oh, what a nightmare it is to support this!

The 5.7ms is what you are calling the 'guard time', and is per the PPM spec created back in the 70's. You can reduce this to about 2.7ms and get away with it:

2.1 x 9 + 2.7 = 21.6ms

Keep in mind that the original PPM spec is 750us to 2250us, not the 900us to 2100us that is common. Our system supports the original PPM spec.
JimDrew is offline Find More Posts by JimDrew
RCG Plus Member
Last edited by JimDrew; Aug 03, 2012 at 05:26 PM.
Reply With Quote
Old Aug 03, 2012, 05:33 PM
Live for speed
GoFaster's Avatar
USA, CT, Bethany
Joined Mar 2004
1,371 Posts
Yes, sending 16 channels all the time.
How do I put it into fast response mode?
Was PPM support superseded by MLINK protocol in the Evo?

So is the Nano or Tx module actually building the PPM frames instead of 'copying' them from the Evo? Which as you said is not viable anymore with the 3.xx firmware.

I currently have a wacky workaround that lets me use up to 7 channels with valid guard pulse length.
Also wondering about this Evo 3.xx specific issue with PPM output in the new Rx's too.

Edit: Found the setting for Fast Response, never noticed it before. But yes, if the frame rate in this mode is max of 25ms, that's still too short.
GoFaster is offline Find More Posts by GoFaster
Reply With Quote
Old Aug 03, 2012, 05:50 PM
Live for speed
GoFaster's Avatar
USA, CT, Bethany
Joined Mar 2004
1,371 Posts
I would certainly appreciate that; so I don't need to have my workaround and also so the decode routine in the FC doesn't have deal with the extra useless/filler frames.

I noticed that the max frame length is ~30ms from the Nano, and the guard pulse is squeezed to accommodate whatever the channel pulse widths are.
With my workaround, the min guard pulse width is 2.8ms and I set the valid time in the FC to 2.7ms.

Regarding the min 750/900us width, if I set the endpoint on a channel to a few % less than -100%, I see frames like the second pic in post #2 above.


Do you happen to have a link to the PPM spec?

Thanks for looking into this!

Quote:
Originally Posted by JimDrew View Post
...
What I probably should do is cap the number of channels in PPM mode to something usable, like 10 channels or less.

Remember, a PPM frame is calculated as: max PWM time x number of channels + 5.7ms. So, with 12 channels that would be:

2.1 x 12 + 5.7 = 30.9ms

That is the amount of time a 12 channel frame can take. If you have a 25ms frame, you are going to be way out of sync! 16 channels is even worse:

2.1 x 16 + 5.7 = 39.3ms

This is why you can't even fit 9 channels in a standard 22ms frame:

2.1 x 9 + 5.7 = 24.6ms

Graupner uses a variable frame rate to get 9 channels - and oh, what a nightmare it is to support this!

The 5.7ms is what you are calling the 'guard time', and is per the PPM spec created back in the 70's. You can reduce this to about 2.7ms and get away with it:

2.1 x 9 + 2.7 = 21.6ms

Keep in mind that the original PPM spec is 750us to 2250us, not the 900us to 2100us that is common. Our system supports the original PPM spec.
GoFaster is offline Find More Posts by GoFaster
Last edited by GoFaster; Aug 03, 2012 at 05:57 PM.
Reply With Quote
Old Aug 03, 2012, 06:12 PM
Xtreme Power Systems
Lake Havasu, AZ
Joined Jun 2005
15,910 Posts
I don't have a link to the spec, I had an original document from the late 60's/earlier 70's that was provided to me by one of Phil Kraft's employees from that era. I got that along with Phil's personal Gold Medal Series transmitter.

The spec just calls for 750us to 2250us with a neutral point of 1500us and a frame gap that exceeds one max channel time (2250us), using a frame rate of 22ms. Futaba used 22.050ms and JR used 22ms (or the variable frame, which ever is larger). This was never an issue because radios back then had a max of 6 channels anyways, so there were never enough channels to cause a problem.

If the channel PWM values exceed the frame time, you are going to have an issue. This is why your work around works - you set the unused channels to their lowest possible values so the PWM pulse requires the minimum amount of time (around 900us per channel).

I should probably calculate the frame rate and divide that by 2.250 to determine the maximum number of channels that can be allowed and use that for a cap with the PPM output for our receivers.
JimDrew is offline Find More Posts by JimDrew
RCG Plus Member
Reply With Quote
Old Aug 03, 2012, 06:44 PM
Live for speed
GoFaster's Avatar
USA, CT, Bethany
Joined Mar 2004
1,371 Posts
Super, especially if you can backport it for the Nano as well. Otherwise I'll live with my workaround.
Another thing for your ToDo list. Hopefully the guys in the XPS-3216 thread won't lynch me for diverting resources

I hope Phil Kraft's old Tx still gets to control a model and not just sitting in a place of honor.

Thanks again,
- Felix
GoFaster is offline Find More Posts by GoFaster
Reply With Quote
Old Aug 03, 2012, 08:13 PM
Xtreme Power Systems
Lake Havasu, AZ
Joined Jun 2005
15,910 Posts
No, Phil's radio is in a box put away. This is the very same transmitter he used to win one of the world championship events, so I don't want to break it.
JimDrew is offline Find More Posts by JimDrew
RCG Plus Member
Reply With Quote
Old Sep 13, 2012, 07:05 AM
Dan Thompson (MP8K developer)
Iflyj3's Avatar
USA, KY, Paris
Joined Dec 2002
329 Posts
Quote:
Originally Posted by JimDrew View Post
Keep in mind that the original PPM spec is 750us to 2250us, not the 900us to 2100us that is common. Our system supports the original PPM spec.
Jim, as an aside. A couple of years ago I tested the Futaba TM-8 RFdeck and if it sees any channel of less than 900 us, it locks up all channels. I found that by accident just before I was going to take off.
Iflyj3 is offline Find More Posts by Iflyj3
Reply With Quote
Old Sep 13, 2012, 11:35 AM
Xtreme Power Systems
Lake Havasu, AZ
Joined Jun 2005
15,910 Posts
Oooh.. that's a good find, and just in time!
JimDrew is offline Find More Posts by JimDrew
RCG Plus Member
Reply With Quote
Reply


Thread Tools

Similar Threads
Category Thread Thread Starter Forum Replies Last Post
futaba ic 4 ch reciever fp-r4f mode 2 MIKEYOUR Non R/C Items (FS/W) 3 May 16, 2012 07:08 AM
futaba x'tal 4 ch reciever fp-r4f mode-1 MIKEYOUR Non R/C Items (FS/W) 1 May 06, 2012 03:04 AM
Sold Futaba 7CAP 72MHz Mode-2 ch-30 ($100 incl S&H CONUS) Jhondra Aircraft - General - Radio Equipment (FS/W) 3 Mar 19, 2012 02:04 PM
Discussion How many channels does the A9 output in Pupil mode ? (trim and mix questions too) gyrokiteguy Hitec/Multiplex USA 6 Jan 26, 2011 03:06 PM
Reconfigure a 4 ch radio to output ch 5,6,7,8 nh401 Radios 4 Mar 01, 2005 06:16 PM