SmartSDR 2.3.9 Network Disconnect

Ed, K0KC

I appreciate the work that has gone into this latest release, but I can report that it apparently has not solved my random disconnect from the network problem. I have experienced this problem with multiple computers running SSDR on my LAN, with new CAT 5 Ethernet cables, and with a new network switch. I have also tried downgrading to earlier versions of SSDR without success. It looks like the problem is a hardware or firmware/software problem in the 6700. I normally operate my rig about 12 hours each day and I typically get several network disconnects during that period. When I get the disconnect message, I not only need to re-connect with SSDR but also I need to stop and restart DAX (and sometimes CAT) as I operate almost 100% on the digital modes.



Ed, K0KC


k0kc@arrl.net
http://k0kc.us/

Thurman Roberson

I leave .y flex on continuously and don't have that issue

Mike-VA3MW

Hi Ed,

I will defer to Tim on the final comments here, but if you have eliminated all the obvious, then it might be time to consider that the 6700 might have a hardware issue.   

All i can offer is that I do know of those that have continue crashes and then changed out a perfectly good working switch for another perfectly good working switch (so it seems) and their problems went away.    There is no way to test a switch that I know of.

Is there anyway you can test without a switch in place even if only for a period that exceeds your normal crash time?  Just a CAT5E cable between the radio and the computer running SSDR?

I am not sure that is your case, but I thought I would mention it.

Mike va3mw

WK2Y

Forgive me if I am talking about something that isn't relevant but one of the reasons I like my IQRouter is that I can log into the router and tell if there has been a network problem (drops, overload, etc.) or not.
73,
Bob, WK2Y 

Ed, K0KC

Mike,

I did buy a new switch with the same result and thus it seems like the only logical place to point is the 6700, if not software/firmware, then perhaps the Ethernet interface in the 6700. When the connection is solid, the network looks good with millions of packets transferred with no drops.

I have not yet tried a direct connection to the computer from the 6700, but I doubt if that will solve the problem given the new switch and cables, but who knows?

I am not quite ready to throw-in the towel yet, but this issue is pretty frustrating.

Ed, K0KC

[Deleted User]

That should have been the first thing you did . . . Direct Connection

I had an issue where I thought it was the radio but it turned out to be the USB/Lan adapter for my laptop. Replaced it and all is well.

Bill -VA3WTB

If your trying to solve something like this, start simple, direct connection, and work back...  As long as there is still something between the radio and computer then we really can't say for sure it is the radio failing.

Ed, K0KC

Unfortunately, I just tried a direct connection and sure enough, it very quickly disconnected, so it still points to a radio problem.

I appreciate all of the suggestions guys!

Martin AA6E

It could be the computer, too. You might check various tasks running in the computer, killing the optional stuff to see if that helps. Or try another computer. Because Windows!

Ed, K0KC

Martin,

I considered that possibility but I experience the same problem on two other computers on my LAN so unless it is something that is running on all three computers that cause the disconnect, it is not likely a computer problem. One of those computers is a business computer that only (now) has SSDR running on it with no ham radio-related applications.

Thanks for the suggestion, however.

Ed, K0KC

N8SDR

Most routers contain a log of events, as does Windows I'd look at both

Lawrence Gray

Same here.  I updated earlier this PM.  I was just in the middle of a QSO, heard a weird audio noise, saw that the panadapter was frozen, then SSDR closed and  I received the "connection to the radio lost" window.

I had the current version of Slicemaster running, as well as SDR monitor.  I use the Flex USB cables for the KAT-500 tuner and BCD data to the ACOM 600S.  I have a FlexControl connected via USB at the laptop.  Nothing else running on the laptop.

First time I have actually been using the radio when it has disconnected.  The long press reset brings the 6500 back to life. 

I really like the newest changes, but I don't think the disconnect issue is fixed?

Larry, W1IZZ

Eric-KE5DTO

Ed,

It sounds like you are trying all of the right things.  Do the disconnects happen when you are in front of the system or when you are away?  If the latter, have you already looked at the power management to ensure that the ethernet connection is not getting dropped when the system goes into standby or sleep state?

Ed, K0KC

Eric,

Thanks for responding!

I seem to get these disconnects no matter if I am at the rig operating or if I am away from it. I normally keep it running about 12 hours per day and see maybe four random disconnects over that period. I did let it run overnight the other day and I did not get disconnected for over 12 hours, but later that next morning, the gremlin was back!

I am pretty sure that I have checked to ensure that no devices or the system itself is configured to go into a "power down" or "sleep" mode, but it is worth checking again. As I mentioned above, I have seen this issue on three different computers on my LAN and I do not think that any of them are configured to go into a sleep mode but again it is worth checking.

Thanks again!

Ed, K0KC

N8SDR

If your seeing this on your Lan computers, then I would suspect that you have either a "Green" switch or router , these Green switches or Green powered routers are garbage when you need a reliable always available connection. If it's a "Green" power saving model look in the setup and see if there is a way to disable any power saving options. Also check there are a few routers out now that have the ability to reboot themselves after a X amount of time again if so disable it as well. 

David-N5PSM

Only wired connection right?

Ed, K0KC

David,

Yes, I only have wired connections between my computers, the 6700, and the network switch. The switch is wired to the router.

Ed, K0KC

Richard (N8SDR),

I have tried two unmanaged switches which to the best of my knowledge are not "green". Even if they were, there is no way to turn that option off. The router is not green either.

Good idea...thanks, but apparently not my problem.

Lawrence Gray

I see disconnects with resulting lockups of the Flex box with my 6500, but not as often as you are apparently seeing them.  I just setup a network tap on the switch port that connects to the Flex. The tap computer is running WireShark setup with a ring buffer.  The tap computer allows me to continuously "sniff" any network traffic to or from the Flex without interfering with the traffic.  The ring buffer will allow me to  capture the last packets sent to or from the Flex when it disconnects.  Perhaps this will provide a clue.

Since my 6500 was fine before 2.3.7 - 2.3.9, it seems fairly clear that the issue is within SSDR or the related Flex firmware.  I've checked my network and computers, examined logs, know that there are no "green" or power savings issues......   I am also a little puzzled regarding the lockup of the Flex.  I can understand a network disconnect dropping the connection, but I don't understand why the box locks up and requires a "long press" reset?

Just for background: In a previous life, I was a CIO running a 14 location 10GB fiber WAN, with hundreds of switches and thousands of connected devices, so I have some experience with connectivity issues.

Larry, W1IZZ

Wayne

Maybe it has something to do with the new spots feature, have you tried disabling the spots and see if the result is the same.

Ed, K0KC

Early-on in my troubleshooting, I had replaced a CAT 5e ethernet cable from the 6700 to the switch with another cable that I had gathering dust here in the shack. Replacing this cable did not offer any improvement, however. After a lot of additional troubleshooting, including buying a new switch, I somehow thought to check that replacement cable and it turned-out to be an old CAT 5 cable. I purchased a new shorter CAT 5e cable, ran it from the 6700 to the switch and magically it reduced my disconnects from several per day to perhaps once per day. I still am not satisfied and will continue troubleshooting as I find time.
Thanks for all of the above suggestions.

Wayne

So I guess you need to replace all your network cables with cat 6, thats what flex provides now with the new rigs.

Lawrence Gray

For even better shielding, use CAT 7.

John WA7UAR

Cat 7 is what I’m considering too. Apparently Cat 6A is the minimal rating now considered for Power over Ethernet (PoE). My biggest limiting concern at the moment is the outside diameter of all the physical cables I’ll be using. I’m limited to an existing 1-inch diameter roof penetrating pipe from my condo electrical closet to the rooftop. I’m trying to fit 2 or 3 (different) RF cables and 2 (similar) Ethernet cables. I’m strongly considering CAT 7 for the extra shielding since it will be laying next to the RF cable(s).

N8SDR

CAT 7 in an RF environment may actually cause you more issues as the ground conductor can act as an antenna, same applies to CAT 6 with the grounding connection. Although each individual setup can vary greatly the picking up of RF in the network cabling. I think or seem to recall an article or post that Flex Radio Systems mentioned the possibility of this increased RF pickup using those types of cables if I remember correctly it was made by Tim.  

Lawrence Gray

“A common myth says that screens and shields can behave as antennas because they are long lengths of metal. The fear is that screens can “attract” signals that are in the environment, or radiate signals that appear on the twisted pairs. The fact is that both screens and shields, and the copper balanced twisted pairs in a UTP cable, will behave as an antenna to some degree. The difference is the noise that couples onto the screen or shield is actually 100 to 1,000 times smaller in magnitude than the noise that is coupled onto an unshielded twisted pair in the same environment. This difference is due to the internal pairs’ well-defined and controlled common mode impedance to the ground plane that is provided by the screen/shield.” The use of, and need for, shielded cabling has been researched at length, which is the reason CAT 7 cable exists. Needless to say, complete and proper grounding/bonding of all equipment, racks,... is required. Using one CAT 5 cable in your environment can negate all of the other noise reduction work.

L.Kubis

I have a similar drop out problem which has gotten worse with the latest upgrade v2.3.9. Although I can't say it's due to my 6500 at this point, my investigation is heading that way.

Living in a condo, I have to operate remote. I am fortunate to have a remote site at a nearby Marina about 1Km away over water to which I extend my LAN with a pair of Ubiquity loco Nanostation M5 radio's.
I have operated very successfully this way ever since Flex released the LAN connectivity feature some years ago.
Around the time of v2.3.7, I started to get the odd short drop out nothing serious, but as time went on it seemed to get worse with each new upgrade. Now it is really bad at times where the Maestro has to be rebooted even though it seems to pick up the connection and continue to play. 
Normally when working my latency as indicated by the Maestro is in the single digits, then every now and then it will jump to triple and quadruple digits. If it stays long enough Maestro will hang up and display a warning window --" The radio selected is in use. Plse try connecting again" even though the spectrum display is working in the background.
Looking at the Ubiquity unit it indicates normal received signal levels except for the data activity. When the latency jumps the data activity indicator slows or just stops, still showing a link connection but no data activity coming from the Flex radio.
Not sure what to make of this at this point but it shows some change in the data flow out of the Flex.
Perhaps someone might have an idea.
I still need to talk with a Ubiquity tech as to what this might indicate.
BTW -- I have the same drop out problem when using a i7 W10 desktop computer, so it doesn't seem to be computer related.
I have a new IQ router coming in hope it may help, but I'm afraid it's more likely the radio end.

Getting frustrating!

Lloyd
VE3ERQ

Mike-VA3MW

Hi Lloyd

That is pretty interesting.  Are you able to downgrade to a version prior to 2.3.7 long enough to test with and see if there is an improvement?

I can say that since 2.3.7, my remote has seemed to improved, but I have not done any scientific tests on it.

Mike va3mw

W9WLX_Chuck

Have you investigated that there may be issues with speed negotiation?  Try throttling your network card back to 100Mbps or even 10Mbps full duplex and see if that improves stability.  The default is usually Auto Negotiation.

Mike-VA3MW

Tim might comment on network negotiation, but he may be distracted for a bit.  However, years ago in the data backup world that I was in, we actually advised customers that the use of Auto Negotiation was expensive, so we had them hard code their speeds to something like 100mb full duplex.  Personally, I run very little Gig networks and the majour part of my network is a very very old $10 100mb/sec Cisco switch I bought at a flea market.  It does scare me with how many things I have plugged into it as it is a 24 port switch, and that is just in the ham shack. 

To 2 cents would be to try to do a Fix port speed if you can.  You can do this on Windows of course.  It is worth the try and you won't come close to using that much data.

Mike 

L.Kubis

I do have a switch at the radio site to switch between my Digital Logger IP Power bar and the Flex.
In my attempt to fix the problem I had replaced the switch with a commercial level D Link 1G switch thinking it was causing the problem. There appeared to be no change but I will see if I can find a new slower switch to try.
My router a Linksys E4200 has not changed since the installation. I have a new IQ one on the way but I doubt if that will make a difference.
I also have a Ethernet over Power Line connection between the Ubiquity at the house and the router. I tried replacing that but no change.
The most annoying thing when using the Maestro is that it wants a reboot if the connection is lost beyond a certain length of time even though the link recovered.
If this could be eliminated in the Maestro software it would make life more tolerable. Perhaps Flex can look at that behavior.
I still need to eliminate the possibility of interference to the 5G Ubiquity radio as the Marina has a lot of WiFi around. Maybe a Ubiquity channel change would help.
Very frustrating in that when it works, it's solid.

Cheers!
Lloyd
VE3ERQ