Date   

Re: OFDM modes using simplex FM on 03-06-21

Frank Olaughlin
 

Barry,.

Thanks for the info. That good reporting. We are still having trouble with OFDM-2000F with checksum errors. I wonder if it's because the other tester has an older win 7 computer. We had a similar issue when 8PSK first was developed and we were using Intel Atom netbooks. CPU usage was just too high and we had to kill a bunch of processes to make it work. We ended up getting a bunch of Dell Centrino laptops with win 7 that solved that issue. I need to look at that now and see if we have that issue again with OFDM. Radios and signalinks are identical. The only difference is that I'm running a Dell i5.


Re: OFDM modes using simplex FM on 03-06-21

K3eui
 

Via FM local repeater, we were able to get OFDM500F, OFDM750F, and OFDM2000F  to work flawlessly.  These are not terribly wide FM modes. Our top mode prior to OFDM  had been  8PSK1000F  (3000 bit/sec).

It is now a test to see if  OFDM2000F  beats out  8PSK1000F in terms of reliability and speed, as both are about  3000 bit/second  and both work great simplex and via a analog voice repeater.

We normally run our tests with a sample  FLMSG  and  FLAMP.

I have an older USB SignaLink before the newer model came out  (2018) with the wide-band transformer and  "LINE" ID chips.

All I did on the SignaLink was to wire up the SPKR jumper block  to the  9600 baud pin (with a wire) rather than to the 1200 baud pin.  I did not use that jumper plug-in BLOCK that is so popular.   The cable itself has all six lines on the  6 pin MINI DIN jack.

On the Icom 706MIIG  I changed MENU 29  from  1200 baud  to 9600 baud (remember we are not really sending PACKET).


That change to both SignLink and Icom radio widened the audio bandpass a bit, out to about 3300 Hz, and removed the "de-emphasis" receive audio which is part of every FM radio with speaker or headphone audio. The typical "de-emphasis" on receive audio is about 6 dB/octave, which reduces the higher-pitch audio compared to the audio at around 500 Hz.   I can see that when I look at speaker or headphone output on background FM noise (radio unsquelched).


There are a few in the Philly region who are now set up with the wider bandwidth  DRA modem  (masterscommunication) and with the radio set to the 9600 baud to experiment with the much higher speed OFDM3000 and above, and the the  "wide"  VARA FM  modem used for Winlink. Remember - these super wide modes cannot be used on an fm analog voice repeater.

I will soon have access to the newer SignaLink (wider bandwidth transformer) to add to the mix.

We are finding that in transmit, the  9600 baud pin is likely going to need a higher transmit voltage than when using the 1200 baud pin. Read your radio manual carefully.  This is no secret. My oder SignaLink BARELY  had enough audio to drive the Icom 706MIIG  at the 9600 baud pin.  So it will work.

The question for us now is will modes like  8PSK1200F  and  OFDM2000F  work more reliably with  ALL  operators using the FLAT audio available only on 9600 baud pins or our radios.

We have to remember -  we are not talking about the real 9600 baud packet signals.  We are talking about VARA  and  OFDM  modes that have nothing to do with the structure of  9600 baud modulation packet.

We'll share what we learn over the next few weeks.

Remember - get permission to test these hi speed data modes over any FM (voice) repeater.

De  k3eui  Barry


On Mar 7, 2021, at 10:09 PM, Frank Olaughlin <wq1o.frank@...> wrote:

Barry,

What were your signalink settings on your old one? We have a few in our group as well.
Were you just conversing or using flamp/flmsg? Was 1500 your best center on OFDM-2000F?

73
Frank
WQ1O


Re: OFDM modes using simplex FM on 03-06-21

Frank Olaughlin
 

Barry,

What were your signalink settings on your old one? We have a few in our group as well.
Were you just conversing or using flamp/flmsg? Was 1500 your best center on OFDM-2000F?

73
Frank
WQ1O


Re: OFDM modes using simplex FM on 03-06-21

K3eui
 

In Chester County EPA  three of us got  OFDM200F  to work fine using the latest  FLDIGI 4.1.8.16  into our local  UHF  FM  analog repeater.


Two of us had the DRA  (wide band) sound card and Knwd V71A radios running on the 9600 baud port.

I had an old Icom 706MIIG  and an (old) SignaLink but wired up for the 9600 baud pin and the radio on 9600 baud.

Worked like a charm, as did  OFDM500f  and  750F.

We managed for force the center frequency to  1500 Hz  as we had our CW ID right at the 450 Hz level which clobbered the  OFDM signal.

It would be nice to have the OPTION  to keep the OFDM2000F center frequency back to 1500 Hz  (when on FM radio this makes more sense)

We are comparing the faster OFDK2000F  with the  8PSK1000F  mode that had been our record for speed for  2m/70cm  FM.

Barry
Cc   N3FLL,  NY3J


On Mar 7, 2021, at 3:25 PM, Frank Olaughlin <wq1o.frank@...> wrote:

 

John,

 

We did a two person test on OFDM on Saturday March 6th using your Fldigi 4.1.18.16

Tester 1: Win 10 Laptop, FT-8900 radio, Newer Signalink, 2m VHF FM
Tester 2: Win 7 laptop, FT-8900 radio, newer signalink, 2m VHF FM

Method was FM Simplex with 10 over 9 signals

Used Fldigi text and used Flmsg with a standard full ICS-214 form

Observations:

OFDM-500F and OFDM750F worked fine on FM simplex. We passed several ICS-214 forms (good size ones) without error. Center was the default 1500hz.

 

Using OFDM-2000F and OFDM-2000. We were able to send 214 forms, but almost every form had a checksum error. We did get one clean, but it was the only one. Degradation seemed to worsen if the form was longer. Short text got through ok, just using Fldigi, but got worse as it lengthened.  I wished we had monitored the CPU usage, but forgot to do so. Not sure if it was even a factor without actually looking.

 

OFDM-3500 continued to be a no go. We would get part of a sentence here or there. Increasing the center frequency did not help.

 

We will try again this week.

 

Thanks for all the continued hard work on the new modes.

73
Frank
WQ1O
Cape Cod and Islands ARES DEC


OFDM modes using simplex FM on 03-06-21

Frank Olaughlin
 

 

John,

 

We did a two person test on OFDM on Saturday March 6th using your Fldigi 4.1.18.16

Tester 1: Win 10 Laptop, FT-8900 radio, Newer Signalink, 2m VHF FM
Tester 2: Win 7 laptop, FT-8900 radio, newer signalink, 2m VHF FM

Method was FM Simplex with 10 over 9 signals

Used Fldigi text and used Flmsg with a standard full ICS-214 form

Observations:

OFDM-500F and OFDM750F worked fine on FM simplex. We passed several ICS-214 forms (good size ones) without error. Center was the default 1500hz.

 

Using OFDM-2000F and OFDM-2000. We were able to send 214 forms, but almost every form had a checksum error. We did get one clean, but it was the only one. Degradation seemed to worsen if the form was longer. Short text got through ok, just using Fldigi, but got worse as it lengthened.  I wished we had monitored the CPU usage, but forgot to do so. Not sure if it was even a factor without actually looking.

 

OFDM-3500 continued to be a no go. We would get part of a sentence here or there. Increasing the center frequency did not help.

 

We will try again this week.

 

Thanks for all the continued hard work on the new modes.

73
Frank
WQ1O
Cape Cod and Islands ARES DEC


Re: [winfldigi] fldigi 4.1.18.16 posted

Haris SV1GRB
 

Hello Joe,

I had the same impression at first about the TxID not controlled but after reading a KL4YFD's post, this is intentional.
The TxID is included in the modem in order to have frequency accuracy between the stations. 

Haris
SV1GRB

Στις Σάβ, 6 Μαρ 2021 στις 12:02 π.μ., ο/η Joseph Counsil <counsil@...> έγραψε:

Dave,

That works much better.  The RSIDs seem to be decoded correctly.

However, the RSID is not controlled by the Fldigi buttons on either platform (Win10x64 and Linux Mint 20.1) for all modes except OFDM-2000.  On the other modes, the TXID and RXID are both active regardless of the button settings.  On OFDM-2000, the Fldigi buttons seem to work.  Also, I could not adequately test OFDM-3500 because the filters in my radios are too narrow (2500Hz) for that one to work.

73,
-Joe-

w1hkj wrote:
at http://www.w1hkj.com/alpha/fldigi/

Contains fixes to OFDM bugs reported for version 4.1.18.15.  Be sure that both Tx and Rx are using the same version for RsID to work corrrectly.

73, David, W1HKJ

Fri Mar 5 14:20:00 2021 -0600  - alpha  4.1.18.16

  WF-only mode menu
    * add missing modes

  Add OFDM modes for NBEMS - Author: John Phelps <kl4yfd@...>
    * OFDM-500F for HF SSB
      - survives HF NVIS/long path
      - 4-carriers of 62.5-baud 4PSK = 250 bps
      - 1/2 rate K=15 FEC with 4sec interleaver
    * OFDM-750F for HF SSB
      - survives HF NVIS/long path
      - 3-carriers of 125-baud 8PSK = 562 bps
      - 1/2 rate K=13 FEC with 3.2sec interleaver
    * OFDM-2000F for HF-SSB/FM
      - 8 carriers of 125-baud 8PSK = 2000bps
      - 2/3 rate K=13 FEC with 3.2sec interleaver
    * OFDM-2000 for HF-SSB/FM
      - 4 carriers of 250-baud 8PSK = 3000bps
      - NO FEC, meant for line-of-sight
    * OFDM-3500 for FM 9600-port
      - 7 carriers of 250-baud 8PSK = 5250bps
      - NO FEC, meant for line-of-sight
      - Requires Signalink made-after 2018 connected to the 9600-jack

  PSK-modes fix false triggers of DCD-OFF - Author: John Phelps <kl4yfd@...>
    * Certian character combinations were triggering DCD-OFF
    * DCD-OFF code now searches over 6-bit window
    * Prevents rare data-loss bug mid-transmission



--
Joseph A. Counsil
1310 Woodlawn Drive
Rolla, MO 65401
(573) 341-5186
Amateur Radio: KØOG


Re: new alpha FLDIGI available for testing new OFDM modes

 

Thanks Dave, I did see 100% but it looks like the majority is about 70% and is slowing things down. OFDM-500 and 700 look okay. I'm using Linux Mint 19.3 and i7 CPU. I did try 20.1 but some of my other Ham Radio programs wouldn't run because of dependencies. Fldigi ran like a charm on 20.1. I'll keep an eye on things.

73, Ron NY3J

On 3/5/21 5:14 PM, Dave wrote:
I am seeing high cpu usage on all of the OFDM-F modes with some small change between character tranmission and transmit idle.  None that reach the 100% mark.  I'm testing on an i5 4-core cpu running Linux Mint 20.1.

Dave

On 3/5/21 4:01 PM, Ron NY3J via groups.io wrote:
Thanks Dave, I was kidding. I'm the one who reported high CPU usage a while back. Thanks again for the hard work. I am having so much fun with Fldigi over the years.

73, Ron NY3J

On 3/5/21 4:56 PM, Dave wrote:
Thanks for the CPU usage report Ron.  Every report is taken seriously.

Dave

On 3/5/21 3:13 PM, Ron NY3J via groups.io wrote:
Dave,

Thanks for the quick fixes. I went into the RsID settings and turned back on the OFDM modes. Looks like the TxID is working okay now. I wouldn't think you would send a TxID to change from the narrow to the wide modes because of the change in center frequency unless the receiving station clicks RsID passband. Then receiving station will move and change modes.

So one more thing. Please don't get mad. With the faster mode, OFDM-2000F it must really take a lot of resources. Top reports CPU usage approaching 100%. I noticed this when my typing was coming to a halt. I'm ducking from the items you are throwing at me :-)

73, Ron NY3J

On 3/5/21 12:08 PM, Ron Wenig wrote:
Dave,

I was doing local testing between two machines with 4.1.18.15. It looks like you have to deselect the OFDM modes in the RsID configuration or it will send out the TxID. Also, when you change to OFDM it changes you to 217 Hz on the waterfall.

73, Ron NY3J

On 3/5/21 11:55 AM, Dave wrote:
Early testing of the OFDM modes in 4.1.18.15 has discovered some critical issues.  I will try to have these fixed before your Sunday net.

David

On 3/5/21 8:39 AM, K3EUI Barry wrote:
*****   QST  *****

If you would like to help test the new "alpha"  OFDM  modes  (HF SSB  and  FM)  then please download the latest ALPHA (test) version of FLDIGI  4.1.18.15
and read all about these new changes.

The PaNBEMS net this Sunday (3583.0 kHz  at  0730 hr)  will test the new mode  OFDM500F  with a sample FLMSG and a  FLAMP (open FLAMP on your own after FLDIGI boots)
As you will see, this new mode has a very very long delay (interleaver) of about 8 seconds to deal with the long deep fades common on NVIS  80m paths.

TU  to  KL4YFD  John  for helping create these new  OFDM  modes.
de  k3eui  Barry

=========================================
at http://www.w1hkj.com/alpha/fldigi/
Apple dmgs are signed, but not certified by Apple.

  . *bs.dmg   - universal Intel/M1 cpu for Big Sur 11.x
  . *hs.dmg   - Intel x86_64 build on High Sierra 10.13.6
  . *lion.dmg - Intel x86_64 build on Lion 10.7.6

Fri Mar 5 06:30:00 2021 -0600  - alpha  4.1.18.15

  WF-only mode menu
    * add missing modes

  Add OFDM modes for NBEMS - Author: John Phelps <kl4yfd@...>
    * OFDM-500F for HF SSB
      - survives HF NVIS/long path
      - 4-carriers of 62.5-baud 4PSK = 250 bps
      - 1/2 rate K=15 FEC with 4sec interleaver
    * OFDM-750F for HF SSB
      - survives HF NVIS/long path
      - 3-carriers of 125-baud 8PSK = 562 bps
      - 1/2 rate K=13 FEC with 3.2sec interleaver
    * OFDM-2000F for HF-SSB/FM
      - 8 carriers of 125-baud 8PSK = 2000bps
      - 2/3 rate K=13 FEC with 3.2sec interleaver
    * OFDM-2000 for HF-SSB/FM
      - 4 carriers of 250-baud 8PSK = 3000bps
      - NO FEC, meant for line-of-sight
    * OFDM-3500 for FM 9600-port
      - 7 carriers of 250-baud 8PSK = 5250bps
      - NO FEC, meant for line-of-sight
      - Requires Signalink made-after 2018 connected to the 9600-jack








Re: new alpha FLDIGI available for testing new OFDM modes

Dave
 

I am seeing high cpu usage on all of the OFDM-F modes with some small change between character tranmission and transmit idle.  None that reach the 100% mark.  I'm testing on an i5 4-core cpu running Linux Mint 20.1.

Dave

On 3/5/21 4:01 PM, Ron NY3J via groups.io wrote:
Thanks Dave, I was kidding. I'm the one who reported high CPU usage a while back. Thanks again for the hard work. I am having so much fun with Fldigi over the years.

73, Ron NY3J

On 3/5/21 4:56 PM, Dave wrote:
Thanks for the CPU usage report Ron.  Every report is taken seriously.

Dave

On 3/5/21 3:13 PM, Ron NY3J via groups.io wrote:
Dave,

Thanks for the quick fixes. I went into the RsID settings and turned back on the OFDM modes. Looks like the TxID is working okay now. I wouldn't think you would send a TxID to change from the narrow to the wide modes because of the change in center frequency unless the receiving station clicks RsID passband. Then receiving station will move and change modes.

So one more thing. Please don't get mad. With the faster mode, OFDM-2000F it must really take a lot of resources. Top reports CPU usage approaching 100%. I noticed this when my typing was coming to a halt. I'm ducking from the items you are throwing at me :-)

73, Ron NY3J

On 3/5/21 12:08 PM, Ron Wenig wrote:
Dave,

I was doing local testing between two machines with 4.1.18.15. It looks like you have to deselect the OFDM modes in the RsID configuration or it will send out the TxID. Also, when you change to OFDM it changes you to 217 Hz on the waterfall.

73, Ron NY3J

On 3/5/21 11:55 AM, Dave wrote:
Early testing of the OFDM modes in 4.1.18.15 has discovered some critical issues.  I will try to have these fixed before your Sunday net.

David

On 3/5/21 8:39 AM, K3EUI Barry wrote:
*****   QST  *****

If you would like to help test the new "alpha"  OFDM  modes  (HF SSB  and  FM)  then please download the latest ALPHA (test) version of FLDIGI  4.1.18.15
and read all about these new changes.

The PaNBEMS net this Sunday (3583.0 kHz  at  0730 hr)  will test the new mode  OFDM500F  with a sample FLMSG and a  FLAMP (open FLAMP on your own after FLDIGI boots)
As you will see, this new mode has a very very long delay (interleaver) of about 8 seconds to deal with the long deep fades common on NVIS  80m paths.

TU  to  KL4YFD  John  for helping create these new  OFDM  modes.
de  k3eui  Barry

=========================================
at http://www.w1hkj.com/alpha/fldigi/
Apple dmgs are signed, but not certified by Apple.

  . *bs.dmg   - universal Intel/M1 cpu for Big Sur 11.x
  . *hs.dmg   - Intel x86_64 build on High Sierra 10.13.6
  . *lion.dmg - Intel x86_64 build on Lion 10.7.6

Fri Mar 5 06:30:00 2021 -0600  - alpha  4.1.18.15

  WF-only mode menu
    * add missing modes

  Add OFDM modes for NBEMS - Author: John Phelps <kl4yfd@...>
    * OFDM-500F for HF SSB
      - survives HF NVIS/long path
      - 4-carriers of 62.5-baud 4PSK = 250 bps
      - 1/2 rate K=15 FEC with 4sec interleaver
    * OFDM-750F for HF SSB
      - survives HF NVIS/long path
      - 3-carriers of 125-baud 8PSK = 562 bps
      - 1/2 rate K=13 FEC with 3.2sec interleaver
    * OFDM-2000F for HF-SSB/FM
      - 8 carriers of 125-baud 8PSK = 2000bps
      - 2/3 rate K=13 FEC with 3.2sec interleaver
    * OFDM-2000 for HF-SSB/FM
      - 4 carriers of 250-baud 8PSK = 3000bps
      - NO FEC, meant for line-of-sight
    * OFDM-3500 for FM 9600-port
      - 7 carriers of 250-baud 8PSK = 5250bps
      - NO FEC, meant for line-of-sight
      - Requires Signalink made-after 2018 connected to the 9600-jack







Re: fldigi 4.1.18.16 posted

Joseph Counsil
 

Dave,

That works much better.  The RSIDs seem to be decoded correctly.

However, the RSID is not controlled by the Fldigi buttons on either platform (Win10x64 and Linux Mint 20.1) for all modes except OFDM-2000.  On the other modes, the TXID and RXID are both active regardless of the button settings.  On OFDM-2000, the Fldigi buttons seem to work.  Also, I could not adequately test OFDM-3500 because the filters in my radios are too narrow (2500Hz) for that one to work.

73,
-Joe-

w1hkj wrote:

at http://www.w1hkj.com/alpha/fldigi/

Contains fixes to OFDM bugs reported for version 4.1.18.15.  Be sure that both Tx and Rx are using the same version for RsID to work corrrectly.

73, David, W1HKJ

Fri Mar 5 14:20:00 2021 -0600  - alpha  4.1.18.16

  WF-only mode menu
    * add missing modes

  Add OFDM modes for NBEMS - Author: John Phelps <kl4yfd@...>
    * OFDM-500F for HF SSB
      - survives HF NVIS/long path
      - 4-carriers of 62.5-baud 4PSK = 250 bps
      - 1/2 rate K=15 FEC with 4sec interleaver
    * OFDM-750F for HF SSB
      - survives HF NVIS/long path
      - 3-carriers of 125-baud 8PSK = 562 bps
      - 1/2 rate K=13 FEC with 3.2sec interleaver
    * OFDM-2000F for HF-SSB/FM
      - 8 carriers of 125-baud 8PSK = 2000bps
      - 2/3 rate K=13 FEC with 3.2sec interleaver
    * OFDM-2000 for HF-SSB/FM
      - 4 carriers of 250-baud 8PSK = 3000bps
      - NO FEC, meant for line-of-sight
    * OFDM-3500 for FM 9600-port
      - 7 carriers of 250-baud 8PSK = 5250bps
      - NO FEC, meant for line-of-sight
      - Requires Signalink made-after 2018 connected to the 9600-jack

  PSK-modes fix false triggers of DCD-OFF - Author: John Phelps <kl4yfd@...>
    * Certian character combinations were triggering DCD-OFF
    * DCD-OFF code now searches over 6-bit window
    * Prevents rare data-loss bug mid-transmission



--
Joseph A. Counsil
1310 Woodlawn Drive
Rolla, MO 65401
(573) 341-5186
Amateur Radio: KØOG


Re: new alpha FLDIGI available for testing new OFDM modes

 

Thanks Dave, I was kidding. I'm the one who reported high CPU usage a while back. Thanks again for the hard work. I am having so much fun with Fldigi over the years.

73, Ron NY3J

On 3/5/21 4:56 PM, Dave wrote:
Thanks for the CPU usage report Ron.  Every report is taken seriously.

Dave

On 3/5/21 3:13 PM, Ron NY3J via groups.io wrote:
Dave,

Thanks for the quick fixes. I went into the RsID settings and turned back on the OFDM modes. Looks like the TxID is working okay now. I wouldn't think you would send a TxID to change from the narrow to the wide modes because of the change in center frequency unless the receiving station clicks RsID passband. Then receiving station will move and change modes.

So one more thing. Please don't get mad. With the faster mode, OFDM-2000F it must really take a lot of resources. Top reports CPU usage approaching 100%. I noticed this when my typing was coming to a halt. I'm ducking from the items you are throwing at me :-)

73, Ron NY3J

On 3/5/21 12:08 PM, Ron Wenig wrote:
Dave,

I was doing local testing between two machines with 4.1.18.15. It looks like you have to deselect the OFDM modes in the RsID configuration or it will send out the TxID. Also, when you change to OFDM it changes you to 217 Hz on the waterfall.

73, Ron NY3J

On 3/5/21 11:55 AM, Dave wrote:
Early testing of the OFDM modes in 4.1.18.15 has discovered some critical issues.  I will try to have these fixed before your Sunday net.

David

On 3/5/21 8:39 AM, K3EUI Barry wrote:
*****   QST  *****

If you would like to help test the new "alpha"  OFDM  modes  (HF SSB  and  FM)  then please download the latest ALPHA (test) version of FLDIGI  4.1.18.15
and read all about these new changes.

The PaNBEMS net this Sunday (3583.0 kHz  at  0730 hr)  will test the new mode  OFDM500F  with a sample FLMSG and a  FLAMP (open FLAMP on your own after FLDIGI boots)
As you will see, this new mode has a very very long delay (interleaver) of about 8 seconds to deal with the long deep fades common on NVIS  80m paths.

TU  to  KL4YFD  John  for helping create these new  OFDM  modes.
de  k3eui  Barry

=========================================
at http://www.w1hkj.com/alpha/fldigi/
Apple dmgs are signed, but not certified by Apple.

  . *bs.dmg   - universal Intel/M1 cpu for Big Sur 11.x
  . *hs.dmg   - Intel x86_64 build on High Sierra 10.13.6
  . *lion.dmg - Intel x86_64 build on Lion 10.7.6

Fri Mar 5 06:30:00 2021 -0600  - alpha  4.1.18.15

  WF-only mode menu
    * add missing modes

  Add OFDM modes for NBEMS - Author: John Phelps <kl4yfd@...>
    * OFDM-500F for HF SSB
      - survives HF NVIS/long path
      - 4-carriers of 62.5-baud 4PSK = 250 bps
      - 1/2 rate K=15 FEC with 4sec interleaver
    * OFDM-750F for HF SSB
      - survives HF NVIS/long path
      - 3-carriers of 125-baud 8PSK = 562 bps
      - 1/2 rate K=13 FEC with 3.2sec interleaver
    * OFDM-2000F for HF-SSB/FM
      - 8 carriers of 125-baud 8PSK = 2000bps
      - 2/3 rate K=13 FEC with 3.2sec interleaver
    * OFDM-2000 for HF-SSB/FM
      - 4 carriers of 250-baud 8PSK = 3000bps
      - NO FEC, meant for line-of-sight
    * OFDM-3500 for FM 9600-port
      - 7 carriers of 250-baud 8PSK = 5250bps
      - NO FEC, meant for line-of-sight
      - Requires Signalink made-after 2018 connected to the 9600-jack






Re: new alpha FLDIGI available for testing new OFDM modes

Dave
 

Thanks for the CPU usage report Ron.  Every report is taken seriously.

Dave

On 3/5/21 3:13 PM, Ron NY3J via groups.io wrote:
Dave,

Thanks for the quick fixes. I went into the RsID settings and turned back on the OFDM modes. Looks like the TxID is working okay now. I wouldn't think you would send a TxID to change from the narrow to the wide modes because of the change in center frequency unless the receiving station clicks RsID passband. Then receiving station will move and change modes.

So one more thing. Please don't get mad. With the faster mode, OFDM-2000F it must really take a lot of resources. Top reports CPU usage approaching 100%. I noticed this when my typing was coming to a halt. I'm ducking from the items you are throwing at me :-)

73, Ron NY3J

On 3/5/21 12:08 PM, Ron Wenig wrote:
Dave,

I was doing local testing between two machines with 4.1.18.15. It looks like you have to deselect the OFDM modes in the RsID configuration or it will send out the TxID. Also, when you change to OFDM it changes you to 217 Hz on the waterfall.

73, Ron NY3J

On 3/5/21 11:55 AM, Dave wrote:
Early testing of the OFDM modes in 4.1.18.15 has discovered some critical issues.  I will try to have these fixed before your Sunday net.

David

On 3/5/21 8:39 AM, K3EUI Barry wrote:
*****   QST  *****

If you would like to help test the new "alpha"  OFDM  modes  (HF SSB  and  FM)  then please download the latest ALPHA (test) version of FLDIGI  4.1.18.15
and read all about these new changes.

The PaNBEMS net this Sunday (3583.0 kHz  at  0730 hr)  will test the new mode  OFDM500F  with a sample FLMSG and a  FLAMP (open FLAMP on your own after FLDIGI boots)
As you will see, this new mode has a very very long delay (interleaver) of about 8 seconds to deal with the long deep fades common on NVIS  80m paths.

TU  to  KL4YFD  John  for helping create these new  OFDM  modes.
de  k3eui  Barry

=========================================
at http://www.w1hkj.com/alpha/fldigi/
Apple dmgs are signed, but not certified by Apple.

  . *bs.dmg   - universal Intel/M1 cpu for Big Sur 11.x
  . *hs.dmg   - Intel x86_64 build on High Sierra 10.13.6
  . *lion.dmg - Intel x86_64 build on Lion 10.7.6

Fri Mar 5 06:30:00 2021 -0600  - alpha  4.1.18.15

  WF-only mode menu
    * add missing modes

  Add OFDM modes for NBEMS - Author: John Phelps <kl4yfd@...>
    * OFDM-500F for HF SSB
      - survives HF NVIS/long path
      - 4-carriers of 62.5-baud 4PSK = 250 bps
      - 1/2 rate K=15 FEC with 4sec interleaver
    * OFDM-750F for HF SSB
      - survives HF NVIS/long path
      - 3-carriers of 125-baud 8PSK = 562 bps
      - 1/2 rate K=13 FEC with 3.2sec interleaver
    * OFDM-2000F for HF-SSB/FM
      - 8 carriers of 125-baud 8PSK = 2000bps
      - 2/3 rate K=13 FEC with 3.2sec interleaver
    * OFDM-2000 for HF-SSB/FM
      - 4 carriers of 250-baud 8PSK = 3000bps
      - NO FEC, meant for line-of-sight
    * OFDM-3500 for FM 9600-port
      - 7 carriers of 250-baud 8PSK = 5250bps
      - NO FEC, meant for line-of-sight
      - Requires Signalink made-after 2018 connected to the 9600-jack





Testing OFDM500F mode on HF SSB this weekend

K3eui
 

If you are interested in testing these new  OFDM modes on 80m NVIS paths, you will need the latest version of  alpha  FLDIGI.
See note from W1HKJ  below.

I will definitely send out a sample FLAMP using  OFDM500F  on the Sunday PaNBEMS net  on 3583.0 kHz  at  0730-0900 hr.

Earlier versions of alpha FLDIGI are not compatible because of TxID and RxID conflicts.

The "center frequency" we hope will stay at  1500 Hz but it MAY force you down to  1325 Hz on waterfall.  The reason is that the mode has a VERY tight tolerance for frequency error on the WF.  So the TxID takes you to where you NEED to be  (for now).

These are still in the very very early testing phase.
Be prepared for multiple problems using these modes at first.

The OFDM500F  mode is about the same bandwidth as THOR22 and MFSK32, but is used for sending a FLMSG/FLAMP,  not for checkins.

TU
Barry  k3eui



Begin forwarded message:

From: "Dave" <w1hkj@...>
Date: March 5, 2021 at 3:24:43 PM EST
To: ham-software@..., ham-mac <ham-mac@...>, winfldigi <winfldigi@groups.io>, linuxham@groups.io, nbems@groups.io
Subject: [nbems] fldigi 4.1.18.16 posted
Reply-To: nbems@groups.io

at http://www.w1hkj.com/alpha/fldigi/

Contains fixes to OFDM bugs reported for version 4.1.18.15.  Be sure that both Tx and Rx are using the same version for RsID to work corrrectly.

73, David, W1HKJ

Fri Mar 5 14:20:00 2021 -0600  - alpha  4.1.18.16

  WF-only mode menu
    * add missing modes

  Add OFDM modes for NBEMS - Author: John Phelps <kl4yfd@...>
    * OFDM-500F for HF SSB
      - survives HF NVIS/long path
      - 4-carriers of 62.5-baud 4PSK = 250 bps
      - 1/2 rate K=15 FEC with 4sec interleaver
    * OFDM-750F for HF SSB
      - survives HF NVIS/long path
      - 3-carriers of 125-baud 8PSK = 562 bps
      - 1/2 rate K=13 FEC with 3.2sec interleaver
    * OFDM-2000F for HF-SSB/FM
      - 8 carriers of 125-baud 8PSK = 2000bps
      - 2/3 rate K=13 FEC with 3.2sec interleaver
    * OFDM-2000 for HF-SSB/FM
      - 4 carriers of 250-baud 8PSK = 3000bps
      - NO FEC, meant for line-of-sight
    * OFDM-3500 for FM 9600-port
      - 7 carriers of 250-baud 8PSK = 5250bps
      - NO FEC, meant for line-of-sight
      - Requires Signalink made-after 2018 connected to the 9600-jack

  PSK-modes fix false triggers of DCD-OFF - Author: John Phelps <kl4yfd@...>
    * Certian character combinations were triggering DCD-OFF
    * DCD-OFF code now searches over 6-bit window
    * Prevents rare data-loss bug mid-transmission



Re: OFDM Modes Testing

K3eui
 

Reminder

These modes will be for sending files, at high speed (FLMSG/FLAMP)
They are not going to play well for "chat" with a keyboard
Gibberish text runs by way too fast in your Rx window.

K3eui



On Mar 5, 2021, at 4:07 PM, Gary E. Kohtala via groups.io <gary.k7ek@...> wrote:

Would anyone care to play with OFDM in the afternoon and/or evening hours, say on
80m, 40m, 30m, or ???.

Best regards,

Gary, K7EK


Re: new alpha FLDIGI available for testing new OFDM modes

 

Dave,

Thanks for the quick fixes. I went into the RsID settings and turned back on the OFDM modes. Looks like the TxID is working okay now. I wouldn't think you would send a TxID to change from the narrow to the wide modes because of the change in center frequency unless the receiving station clicks RsID passband. Then receiving station will move and change modes.

So one more thing. Please don't get mad. With the faster mode, OFDM-2000F it must really take a lot of resources. Top reports CPU usage approaching 100%. I noticed this when my typing was coming to a halt. I'm ducking from the items you are throwing at me :-)

73, Ron NY3J

On 3/5/21 12:08 PM, Ron Wenig wrote:
Dave,

I was doing local testing between two machines with 4.1.18.15. It looks like you have to deselect the OFDM modes in the RsID configuration or it will send out the TxID. Also, when you change to OFDM it changes you to 217 Hz on the waterfall.

73, Ron NY3J

On 3/5/21 11:55 AM, Dave wrote:
Early testing of the OFDM modes in 4.1.18.15 has discovered some critical issues.  I will try to have these fixed before your Sunday net.

David

On 3/5/21 8:39 AM, K3EUI Barry wrote:
*****   QST  *****

If you would like to help test the new "alpha"  OFDM  modes  (HF SSB  and  FM)  then please download the latest ALPHA (test) version of FLDIGI  4.1.18.15
and read all about these new changes.

The PaNBEMS net this Sunday (3583.0 kHz  at  0730 hr)  will test the new mode  OFDM500F  with a sample FLMSG and a  FLAMP (open FLAMP on your own after FLDIGI boots)
As you will see, this new mode has a very very long delay (interleaver) of about 8 seconds to deal with the long deep fades common on NVIS  80m paths.

TU  to  KL4YFD  John  for helping create these new  OFDM  modes.
de  k3eui  Barry

=========================================
at http://www.w1hkj.com/alpha/fldigi/
Apple dmgs are signed, but not certified by Apple.

  . *bs.dmg   - universal Intel/M1 cpu for Big Sur 11.x
  . *hs.dmg   - Intel x86_64 build on High Sierra 10.13.6
  . *lion.dmg - Intel x86_64 build on Lion 10.7.6

Fri Mar 5 06:30:00 2021 -0600  - alpha  4.1.18.15

  WF-only mode menu
    * add missing modes

  Add OFDM modes for NBEMS - Author: John Phelps <kl4yfd@...>
    * OFDM-500F for HF SSB
      - survives HF NVIS/long path
      - 4-carriers of 62.5-baud 4PSK = 250 bps
      - 1/2 rate K=15 FEC with 4sec interleaver
    * OFDM-750F for HF SSB
      - survives HF NVIS/long path
      - 3-carriers of 125-baud 8PSK = 562 bps
      - 1/2 rate K=13 FEC with 3.2sec interleaver
    * OFDM-2000F for HF-SSB/FM
      - 8 carriers of 125-baud 8PSK = 2000bps
      - 2/3 rate K=13 FEC with 3.2sec interleaver
    * OFDM-2000 for HF-SSB/FM
      - 4 carriers of 250-baud 8PSK = 3000bps
      - NO FEC, meant for line-of-sight
    * OFDM-3500 for FM 9600-port
      - 7 carriers of 250-baud 8PSK = 5250bps
      - NO FEC, meant for line-of-sight
      - Requires Signalink made-after 2018 connected to the 9600-jack




Re: OFDM Modes Testing

Gary E. Kohtala
 

Would anyone care to play with OFDM in the afternoon and/or evening hours, say on
80m, 40m, 30m, or ???.

Best regards,

Gary, K7EK


Re: OFDM Modes Testing

K3eui
 

I just kept pressing the  LK  button on/off many times

It is going to be "repaired'




On Mar 5, 2021, at 3:16 PM, Frank Olaughlin <wq1o.frank@...> wrote:

Barry,

How did you move the center frequency on OFDM-2000F?  When I change it, it moves in the box, but the actual frequency doesn't change. Same if I adjust it in the WF.

73
Frank
WQ1O


Re: OFDM Modes Testing

Frank Olaughlin
 

Ron,

Good observation!      Thanks.....that works.

Looking forward to the new updates.


73
Frank
WQ1O


Re: interesting test today on 440 FM repeater

KL4YFD ham
 


Hi Barry,

The OFDM 500F, 750F, and 2000F modes should work over repeaters also.
Not certain if they will provide much more robustness, as line-of-sight-FM is already a very clean signal.
OFDM-2000F may work through multiple-linked repeater systems a bit better, but that would need to be tested.

OFDM-2000F and OFDM-2000 were designed to Just-Barely fit into a normal HF-SSB channel.
These modes were at the request of NBEMS-Hawaii mailing-list discussions, wanting a fast-as-you-can mode for HF line-of-sight (island to island)

The only way to get 2000 or 3000 bits/second over HF SSB with OFDM-2000/2000F was to really push the bandwidth-limits to their max.
This required locking the Tx frequency though to prevent signal-cutoff. Rx is still frequency agile to compensate for radio frequency offsets.

So, this is why OFDM-2000F and 2000 are transmit-center-frequency locked to 1325Hz.
 Adjusting even 10Hz either direction would cut-off portions of the signal.
At 2Khz of bandwidth, it is at the very-limit of what stock HF-filters allow (and a bit too wide for 1.8Khz filters).

As far as using OFDM-2000F vs the 8PSK modes... it depends on the repeater system.
When using faster-baud rate single carrier modems, there is a bit more robustness to phase-distortions than slower baudrates.
The faster the baudrate, the wider the signal-bandwidth, spreading-out the phase distortions and allowing decode.
The drawback though is faster baudrates need a higher signal-to-noise ratio to operate (less sensitivity).

Using a slower-baudrate multi-carrier mode like OFDM-2000F will work when there is less signal strength (more sensitive), but at the cost of being far less tolerant of phase-distortions. 
For HF we know the signal is most-likely going to be kinda quiet, and SSB-radios are very phase-stable, so multiple slower carriers win-out as the best solution.

The OFDM modes do spread the bits over multiple carriers, so if 1 carrier is cutoff by filtering or distortions, decode should still be 100%
This also nicely keeps us far-under the FCC baudrate limits for HF, while still allowing good datarates.


Thank you for all the testing,
John Phelps
KL4YFD



On Thu, Mar 4, 2021 at 2:05 PM Barry Feierman <k3euibarry@...> wrote:



John  KL4YFD

Updates....

Is there any reason why the  OFDM500F   and   OFDM750F   should not work on a standard FM analog voice repeater?

Afer all, it is just 'sound'.

We just tried OFDM500F  and it sent a FLAMP flawlessly on a  440 FM voice repeater (linked to two other 440 machines).

So the OFDM500F and 750F  are narrow enough for a typical SSB frequency (and filter) but I think should work fine on FM radios.


What if we were to try the  OFDM2000F   on  80m  SSB?

Obviously we'd need  2 kHz bandwidth  (on 80m in the morning, that's not a problem for a quick test).


I noticed the 2000F  wanted to "center" at   1325 Hz on waterfall.  Why?

We generally center all modes at  1500 Hz  on SSB and on FM.

Any reason why we should not center  2000F  mode at  1500 Hz on a FM voice repeater?


I modified my (old) SignaLink to use the  9600 baud Rx audio, and changed my Icom 706MIIG  to the MENU  9600 baud option,

and then modified the SignaLink to get a bit more Tx audio and everything works FINE on FM setting on the Icom 706.

Just exchanged some FLAMP  using  OFDM2000F  both centered at  1325  and  1500.

We even tried centering OFDM2000F  at   2000 Hz on WF, and it worked as well.

It sounds awful  (like VARA)  but it works!


So  OFDM2000F  is   8  "carriers"  or  sub-carriers, and each has a lower baud than the high values of single carrier  8PSK1000F.

My guess:  OFDM2000F  is going to work better than  8PSK1200F  on our typical radios and FM repeaters  ....

it spreads the information over  8  "carriers"  (lower baud each)  like having 8 wires carrying information  vs.  one wire.

Am I correct?


So far, this has been fascinating.

Oh...  and  NY3J  and  N3FLL  both have the RMA-70  new wide freq modems.  I'm jealous.

I'm still stuck with the old SignaLink, but maybe I'll break down and buy a  RMA modem.

I see the 'new' SignalInks have better wider audio transformers and better chips  (line input vs. mic input).


TU

Barry  k3eui



Re: OFDM Modes Testing

 

Frank,

Try this. After you move center frequency turn on LK then off again and it looks like it frees it up. It looks like Dave is working on a few things and an update is in the works.

73, Ron NY3J

On 3/5/21 3:16 PM, Frank Olaughlin wrote:
Barry,

How did you move the center frequency on OFDM-2000F?  When I change it, it moves in the box, but the actual frequency doesn't change. Same if I adjust it in the WF.

73
Frank
WQ1O


fldigi 4.1.18.16 posted

Dave
 

at http://www.w1hkj.com/alpha/fldigi/

Contains fixes to OFDM bugs reported for version 4.1.18.15.  Be sure that both Tx and Rx are using the same version for RsID to work corrrectly.

73, David, W1HKJ

Fri Mar 5 14:20:00 2021 -0600  - alpha  4.1.18.16

  WF-only mode menu
    * add missing modes

  Add OFDM modes for NBEMS - Author: John Phelps <kl4yfd@...>
    * OFDM-500F for HF SSB
      - survives HF NVIS/long path
      - 4-carriers of 62.5-baud 4PSK = 250 bps
      - 1/2 rate K=15 FEC with 4sec interleaver
    * OFDM-750F for HF SSB
      - survives HF NVIS/long path
      - 3-carriers of 125-baud 8PSK = 562 bps
      - 1/2 rate K=13 FEC with 3.2sec interleaver
    * OFDM-2000F for HF-SSB/FM
      - 8 carriers of 125-baud 8PSK = 2000bps
      - 2/3 rate K=13 FEC with 3.2sec interleaver
    * OFDM-2000 for HF-SSB/FM
      - 4 carriers of 250-baud 8PSK = 3000bps
      - NO FEC, meant for line-of-sight
    * OFDM-3500 for FM 9600-port
      - 7 carriers of 250-baud 8PSK = 5250bps
      - NO FEC, meant for line-of-sight
      - Requires Signalink made-after 2018 connected to the 9600-jack

  PSK-modes fix false triggers of DCD-OFF - Author: John Phelps <kl4yfd@...>
    * Certian character combinations were triggering DCD-OFF
    * DCD-OFF code now searches over 6-bit window
    * Prevents rare data-loss bug mid-transmission


1 - 20 of 17150