7000 and 8000 Protocol 2 users: what say you?

FIRMWARE TOPICS ONLY--non-firmware topics will be MOVED
Forum rules
Until such time as the New Protocol firmware goes into general release, all discussion will be concentrated here.
User avatar
w-u-2-o
Posts: 5525
Joined: Fri Mar 10, 2017 1:47 pm

7000 and 8000 Protocol 2 users: what say you?

Postby w-u-2-o » Wed Jan 09, 2019 2:28 am

OK, guys with the BIG radios: P2 version 1.5 has been posted for some time. Who's running it, what do you think? Stable? Throwing any Seq errors?

You can leave off anything that is obviously Thetis related.

Thanks!

Scott
Jay_N2MGA
Posts: 15
Joined: Sun Apr 09, 2017 5:26 pm

Re: 7000 and 8000 Protocol 2 users: what say you?

Postby Jay_N2MGA » Thu Jan 10, 2019 6:52 pm

Ok, here are some observations that may offer up some data as to how/what/where/why the Seq indicator blinks. I ran some long duration Wireshark captures to observe the UDP traffic going back and forth to my 8000 and the Win 10 PC running Thetis 2.6.2 and FW 1.5. I have to preface this by saying that in my setup I only occasionally see any Seq flashes, mostly in Recv. For the majority of all the UDP packets going back and forth, including other TCP packets being generated by other programs running on the same PC with the 8000, no Seq flashes were observed for a long period of just listening to some 40m traffic.

But, after running for many many minutes I did see the Seq flash and stopped the Wireshark capture to see if I could find the culprit. Turns out, each Seq flash appears to be correlated to a TCP error as flagged by Wireshark that was generated by another program running on the same PC (email, web browser, who knows what, etc...). The TCP errors flagged were to/from external IP addresses back and forth from my PC, errors seen were of the type "Previous segment not captured" and "Duplicate to the Ack in frame xxxxx". Some other errors shown within just a few packets of time were "TCP fast retransmission" and "TCP out of order" along with more "Duplicate to the Ack in frame xxxxx", and "TCP previous segment not captured". In total there were 10 of these packets that occurred in rapid succession amidst the usual clean UDP packets of radio traffic. Afterward no other Seq flashes were observed for the time I was capturing packets and letting the radio receive as usual.

Again this may be directly related to how my network is setup along with the capabilities of my PC and may not be of any importance to this discussion but I wanted to at the very least add some insight as to a possible cause for some of the reasons the Seq indicator flashes. A good test on my setup would be to disconnect any traffic from outside of the network to see if there are any other errors thrown in Wireshark that can be correlated to the Seq flashes.

As usual, your mileage may vary....

Cheers!
jay
n2mga
User avatar
w-u-2-o
Posts: 5525
Joined: Fri Mar 10, 2017 1:47 pm

Re: 7000 and 8000 Protocol 2 users: what say you?

Postby w-u-2-o » Thu Jan 10, 2019 7:33 pm

Thanks, Jay.

The streams from the hardware to the PC are UDP. Each packet has a sequence number. If one or more are missing then that will cause a Seq error. The numeric associated with the Seq error tells you what stream or streams are affected (there is one for each DDC, and for mic audio).

UDP is, of course, inherently lossy. So an occasional Seq error is nothing to worry about and very expected. If you have something that is hogging your NIC for any length of time then UDP is going to lose the battle.

Since you are only seeing Seq errors on network congestion this tends to support that the firmware is doing OK.

It's UDP in the other direction as well. Unfortunately the firmware does not track the same stat's nor report on them back up to the software.

73,

Scott
Jay_N2MGA
Posts: 15
Joined: Sun Apr 09, 2017 5:26 pm

Re: 7000 and 8000 Protocol 2 users: what say you?

Postby Jay_N2MGA » Thu Jan 10, 2019 10:23 pm

Hi Scott, completely concur re; UDP packets being on the low end of priority & network congestion although albeit in reality it appears to be maybe one or a few packets involved, question for you, do you know where the Sequence number resides in the packet, i.e.; is it part of any header or is it buried in the data portion past the header information?

As an aside, I went into the network card properties to experiment with the idea of Buffering, one item changed for the nic card properties/advanced screen is labeled "Receive buffers", the idea being any oversize packet such as the ones mentioned earlier, perhaps an increase in this number may help. For giggles and grins I changed the rather low default of 256 all the way up to 2048 and I no longer see any Seq flashes even though the same types of TCP errors as captured by wireshark are still occurring. Who knows, very early in the testing but overall I think as you stated the latest firmware is performing as expected and maybe some tweaking can be done on individual setups to perhaps help if need be. Everyone's setup is different so what works for me may or may not help another setup, but it's worth exploring (and learning!).

Tnx agn for any info you can provide...

cheers!
jay
n2mga
sv1jso
Posts: 36
Joined: Wed May 30, 2018 7:06 pm

Re: 7000 and 8000 Protocol 2 users: what say you?

Postby sv1jso » Thu Jan 10, 2019 11:07 pm

Hello,

Running my 7000 with firmware 1.5 and thetis 2.6.2 for 1 day. It is big improvement from powersdr mrx 3.4.9. Now my vac sound is not broken even without resampler (forced to 1). I run bandwidth to max for testing and I saw no problems. In the same computer the max viable bandwidth for powersdr was 96khz.
I see some problems with skins and numbers not centered but it is not critical. The only thing I am missing now is smooth scroll for ctun.

Thanks everyone for the software and firmware.

73
Mike
User avatar
w-u-2-o
Posts: 5525
Joined: Fri Mar 10, 2017 1:47 pm

Re: 7000 and 8000 Protocol 2 users: what say you?

Postby w-u-2-o » Fri Jan 11, 2019 2:29 am

Jay,

I played with NIC setting, including buffer sizes, with PowerSDR with no substantive differences being noted. Maybe I should perform the same experiment with Thetis.

The packet structure is fully documented:

https://github.com/TAPR/OpenHPSDR-Firmware/tree/master/Protocol%202/Documentation

73,

Scott
Jay_N2MGA
Posts: 15
Joined: Sun Apr 09, 2017 5:26 pm

Re: 7000 and 8000 Protocol 2 users: what say you?

Postby Jay_N2MGA » Fri Jan 11, 2019 3:03 am

Perfect, that's exactly what I was looking for, tnx again!

jj
n2mga
liam
Posts: 35
Joined: Mon Apr 10, 2017 7:26 am

Re: 7000 and 8000 Protocol 2 users: what say you?

Postby liam » Mon Jan 14, 2019 3:03 pm

Hi I have been operating an Anan8k for the last five days on
mostly RX and some TX RTTY and found it stable System stayed up for
the full five days No observed glitches

Here using Win 10 PRO V1809 OS 17763 X64bit

CPU Xeon W3350 at 3 Meg

Thetis 2.6 FW 1.5


Thanks for all the hard work

Liam EI3GC
User avatar
n1gp
Posts: 175
Joined: Sun Apr 09, 2017 6:34 pm

Re: 7000 and 8000 Protocol 2 users: what say you?

Postby n1gp » Wed Jan 16, 2019 12:43 am

Hey All,

I've been watching the incoming rpt's from the v1.5 releases on the P2 MkII & MkI.
It seems the MkII has been experiencing some crashes on a shift from rx to tx while
using PS or Diversity where the MkI hasn't. At least thats what is seen in the reports.

I've come up with an update for the MkII that has a meld of the changes for MkI and
would like to present it to be tested against v1.5 for MKII. Please give this a try:

https://www.dropbox.com/s/3wjluftuaomg3um/Orion_MkII_Protocol_2_v1.6_pre1.rbf?dl=0

and report back on forum.

Tnx!

-Rick / N1GP
User avatar
NJ2US
Posts: 106
Joined: Sun Oct 29, 2017 9:39 pm

Re: 7000 and 8000 Protocol 2 users: what say you?

Postby NJ2US » Fri Jan 18, 2019 6:17 am

Thetis 2.6.2 P2 1.6, 7000DLE
2 days, very stable, no pure signal problems, no TX crashes. Only saw one seq error, but it was without consequence.
Amazing work Rick, thank you.

Jeff NJ2US
User avatar
w-u-2-o
Posts: 5525
Joined: Fri Mar 10, 2017 1:47 pm

Re: 7000 and 8000 Protocol 2 users: what say you?

Postby w-u-2-o » Fri Jan 18, 2019 11:33 am

Also working very well here, Rick, on an 8000. Absolutely nothing negative to report.

Thanks,

Scott
User avatar
W2PA
Posts: 165
Joined: Sun Apr 09, 2017 6:34 pm
Location: LaGrangeville, NY
Contact:

Re: 7000 and 8000 Protocol 2 users: what say you?

Postby W2PA » Sat Jan 19, 2019 10:06 pm

Hi Rick,

Just thought I'd chime in on this latest. I'd been running 1.5 with Thetis 2.6.2 for a while - since you released 1.5. It was quite stable and PS was working consistently - the main difference from 1.4 where it was randomly quitting.

With 1.6 all seemed fine. But in an extended QSO on phone this afternoon, I started noticing Seq->43 errors. Somewhat later, the receive audio stream began to sputter. By this I mean that although the background band noise (80m) was constant and even, strong signals were noticeably broken up and discontinuous. I heard it on all SSB signals across the band. Re-starting Thetis cured the problem and it didn't recur for the remainder 30 minutes of the QSO. The Seq->43 were still there but no more audio artifacts.

I'll continue to test to see if the discontinuous audio happens again, and go back to 1.5 if so as a comparison.

73,
Chris, W2PA
73,
Chris, W2PA
User avatar
NJ2US
Posts: 106
Joined: Sun Oct 29, 2017 9:39 pm

Re: 7000 and 8000 Protocol 2 users: what say you?

Postby NJ2US » Mon Jan 21, 2019 2:05 pm

P2 Firmware V 1.6 Pre-release
Thetis 2.6.3 (1/20/2019) beta
Anan 7000DLE Serial Number 00019
384K Sample Rate

Since 1.6 pre-release, no SEQ errors, No crashes, no anomalies, stable Pure Signal. Smoothest operation yet...

Jeff NJ2US
SM4PEL.
Posts: 5
Joined: Thu Jan 10, 2019 5:11 pm

Re: 7000 and 8000 Protocol 2 users: what say you?

Postby SM4PEL. » Mon Jan 21, 2019 4:35 pm

jeffreydoran wrote:P2 Firmware V 1.6 Pre-release
Thetis 2.6.3 (1/20/2019) beta
Anan 7000DLE Serial Number 00019
384K Sample Rate

Since 1.6 pre-release, no SEQ errors, No crashes, no anomalies, stable Pure Signal. Smoothest operation yet...

Jeff NJ2US

May i ask where i can find Thetis 2.6.3 Beta?
73´s Peter
User avatar
NJ2US
Posts: 106
Joined: Sun Oct 29, 2017 9:39 pm

Re: 7000 and 8000 Protocol 2 users: what say you?

Postby NJ2US » Mon Jan 21, 2019 6:26 pm

These pre-releases are generally available at https://github.com/w5wc/Thetis/releases/ but it's important to keep in mind that they are not official, they don't always include the msi installer, and there is no warranty expressed or implied, and your mileage will probably vary.
Jeff NJ2US
User avatar
NJ2US
Posts: 106
Joined: Sun Oct 29, 2017 9:39 pm

Re: 7000 and 8000 Protocol 2 users: what say you?

Postby NJ2US » Tue Jan 22, 2019 12:46 am

Anan 7000DLE Ser Numb 000019
Thetis 2.6.3 1/20/19 version
P2 Firmware Version Non Pre-Lease 1.6 dated 1/21/1019

This release of the Firmware 1.6 dated 1/21/19 BREAKS RX 2 and BPF2.

I moved back to 1.6 Pre release.

Jeff NJ2US
NC3Z
Posts: 463
Joined: Sun Oct 29, 2017 8:57 pm
Location: Merritt, NC

Re: 7000 and 8000 Protocol 2 users: what say you?

Postby NC3Z » Thu Jan 31, 2019 1:04 am

So which version is currently posted on the repository?
Gary NC3Z
User avatar
n1gp
Posts: 175
Joined: Sun Apr 09, 2017 6:34 pm

Re: 7000 and 8000 Protocol 2 users: what say you?

Postby n1gp » Thu Jan 31, 2019 4:10 am

Hi Gary,

v1.6 is on the repo.

Jeff reports that 1.6_pre1: https://www.dropbox.com/s/3wjluftuaomg3um/Orion_MkII_Protocol_2_v1.6_pre1.rbf?dl=0
works better than whats on the repo.

It would be helpful to know how it is working for others.

Tnx

-Rick / N1GP
NC3Z
Posts: 463
Joined: Sun Oct 29, 2017 8:57 pm
Location: Merritt, NC

Re: 7000 and 8000 Protocol 2 users: what say you?

Postby NC3Z » Thu Jan 31, 2019 12:55 pm

Thanks Rick,

I have been using the 1.6pre and have held off using the latest on the repo due to reports, plus it was confusing which was which.

What does this version on 1.6 change vs the 1.6pre?
Gary NC3Z
User avatar
w-u-2-o
Posts: 5525
Joined: Fri Mar 10, 2017 1:47 pm

Re: 7000 and 8000 Protocol 2 users: what say you?

Postby w-u-2-o » Fri Feb 01, 2019 2:52 am

Gary,

AFAIK the only change involves slowing down the control lines to the step attenuator to obtain reliable operation on the very latest serial number Orion MKII boards that have recently shipped.

Unfortunately, when Rick (or anyone) makes a new build the free Quartus Lite tool does not guarantee all timing will be precisely identical to any previous build. Hence it is necessary to get a lot of people to test any new build no matter how small the change.

73,

Scott
NC3Z
Posts: 463
Joined: Sun Oct 29, 2017 8:57 pm
Location: Merritt, NC

Re: 7000 and 8000 Protocol 2 users: what say you?

Postby NC3Z » Fri Feb 01, 2019 3:14 am

Just loaded the repo version of 1.6 (from the 1.6pre) and I get continuous Seq errors when I TX. Also RX2's baseline jumps up to the max scale. Only seems to happen with PS on.

Went back to 1.6pre and those issues went away.
Gary NC3Z
NC3Z
Posts: 463
Joined: Sun Oct 29, 2017 8:57 pm
Location: Merritt, NC

Re: 7000 and 8000 Protocol 2 users: what say you?

Postby NC3Z » Fri Feb 01, 2019 3:16 am

TX spectrum on RX1 is also maxing out on the spectrum scale during TX with the repo 1.6.
Gary NC3Z
User avatar
NJ2US
Posts: 106
Joined: Sun Oct 29, 2017 9:39 pm

Re: 7000 and 8000 Protocol 2 users: what say you?

Postby NJ2US » Fri Feb 01, 2019 3:22 am

7000DLE Serial # 00019
Thetis v2.6.3 (1/25/19)
ADC1 and ADC2 both active

My third attempt to run the Firmware P2 v 1.6 presently posted in the repo, resulted in similar behaviour that NC3Z saw.
Upon activation of Pure Signal:
SEQ->2 errors
ADC2 Panadapter noise floor jumps to -80
Continuous audio glitching on transmit audio.
After a minute, ADC 1 Panadapter starting "jumping", meaning the noise floor rose to -90 and higher.
SEQ errors continued. Audio glitching on receive and transmit.

Reverted back to 1.6_prerelease, all operations normal, 1.6_prerelease has been very stable for me.

Jeff
NJ2US
Jay_N2MGA
Posts: 15
Joined: Sun Apr 09, 2017 5:26 pm

Re: 7000 and 8000 Protocol 2 users: what say you?

Postby Jay_N2MGA » Fri Feb 01, 2019 9:27 pm

I had the same result as others with the v1.6 repo and my 8000, reverted back to v1.6 pre and that setup with Thetis v.2.6.2 restores the stability regarding SEQ errors, Pure Signal etc... On a side note, v1.6 pre with Simon's latest v3.0.4 SDR Radio console is also quite stable...

jay
n2mga
User avatar
n1gp
Posts: 175
Joined: Sun Apr 09, 2017 6:34 pm

Re: 7000 and 8000 Protocol 2 users: what say you?

Postby n1gp » Sat Feb 02, 2019 1:21 pm

Hi Guys,

Thanks for the reports on v1.6

I usually choose the next version to come with a _preX appended for pre test images.

The difference between 1.6_pre1 and the TAPR release 1.6 is a change to the clock which feeds
the Attenuators:

"change clock for Rx Attenuator from CMCLK (12.288MHz) to CBCLK (3.072MHz)
since new version of DAT-31A-SP+ attenuator chip is not reliable at faster clock speed"

This is only for new production boards.

Scott's comment is spot on. Many times just a small change affects the overall timing of
a particular build. That appears to be the case here.

I'm glad that the 1.6_pre1 is working for most.

73's

-Rick / N1GP
User avatar
NJ2US
Posts: 106
Joined: Sun Oct 29, 2017 9:39 pm

Re: 7000 and 8000 Protocol 2 users: what say you?

Postby NJ2US » Sat Feb 02, 2019 1:57 pm

Thanks for the explanation Rick, many of us did see that release note from Phil about the clock speed, and it was a head scratcher.

The implication of that would seem to be a ding on the performance of later production Orion MKII boards, not being able to handle enhanced clock speeds. My guess is there is no real performance degradation, true statement? It also explains why some users like W2NER are not seeing the issue with 1.6 repo as he is using a high serial number later manufacture 7000 DLE.

Jeff NJ2US
User avatar
w-u-2-o
Posts: 5525
Joined: Fri Mar 10, 2017 1:47 pm

Re: 7000 and 8000 Protocol 2 users: what say you?

Postby w-u-2-o » Sat Feb 02, 2019 2:35 pm

Jeff,

There are two different issues at hand here:

1. Are the SPI signals used to control the step attenuator correct?

2. Does the firmware make timing in all FPGAs on all boards?

With regard to (1), I reviewed the timing and don't understand how the attenuators ever worked at all, going all the way back to the original Hermes. And with the slower timing for the newer revision attenuator chips on the higher serial number boards I still don't understand how it is working. The latch enable signal would appear to be coming out at the wrong time. I think we are just getting lucky in both cases.

With regard to (2), this goes back to the problems inherent in using the free Quartus Lite development environment. When you re-compile/re-build the code it comes up with a different metallization every single time. Even if you don't change anything at all, the new build is different. So even if you are changing something simple, like a SPI clock speed, or even just the version number of the code, it's all different.

BREAK

Rick: I have a new idea on how to get deal with the Quartus Lite issue. I will email you separately.

73,

Scott
cLicari
Posts: 109
Joined: Mon Apr 10, 2017 9:33 pm
Location: North Texas

Re: 7000 and 8000 Protocol 2 users: what say you?

Postby cLicari » Sat Apr 27, 2019 9:44 pm

Have been running P2 v1.9 and Thetis v2.6.6 on my 8000DLE for about five days now with generally very good results. I have seen a few, very brief SEQ 43, 42, 11 and 8 error message but they do not appear to cause any noticeable issues.
The last two days I have had two occurrences of where upon key down the Panadapter goes from my normal transmit signal to full gray and locks up without releasing. The only way to get out of is is to power off Thetis. Upon powering back up all returns to a normal state of operation. My description may be a bit cryptic as it happens unexpectedly and I'm just trying to get out of it. If it happens again I'll try to record it more accurately. Anyone else seen something like this? Scott, any thoughts?

Thx,
Carl
NX5T
Carl Licari
NX5T
Anan 8000DLE, FW v2.0 2b, Thetis 2.8.11, 21k9
Windows 10, dedicated network subnet, Intel i7 8700K, 32GB, NVIDIA GeForce GTX 1060, Samsung 40" 4K display
User avatar
w-u-2-o
Posts: 5525
Joined: Fri Mar 10, 2017 1:47 pm

Re: 7000 and 8000 Protocol 2 users: what say you?

Postby w-u-2-o » Sun Apr 28, 2019 2:48 am

I have not seen this, Carl. When I have a firmware related crash like that I will normally see a highly elevated noise floor. But what you describe might be firmware or software.

73,

Scott
User avatar
n1gp
Posts: 175
Joined: Sun Apr 09, 2017 6:34 pm

Re: 7000 and 8000 Protocol 2 users: what say you?

Postby n1gp » Sat May 04, 2019 8:25 pm

Hi Guys,

This build has some thermal parameter changes in the Quartus (FW Compiler) timing area
suggested by our resident Systems Guru Scott/WU2O. The idea is to tell Quartus to compile
for a wider thermal range and hopefully work better on more or most 7000/8000's

https://www.dropbox.com/s/ndahfr5zgh1l252/Orion_MkII_Protocol_2_v2.0_pre1.rbf?dl=0

Tnx

-Rick / N1GP

Return to “Protocol 2 Firmware (all radios)”