Lockup issue and V2.0 / V1.11

Andy M5ZAP

Hi,

Is there a fix for the lockup issue in the proposed service release V1.11. And if not  (presumably because route cause has still not been found) has somebody who experiences the problem been used to test V2.0 to determine if the lockup problem carries over to V2.0

Jim Gilliam

I am not familiar with the lock-up issue nor have I ever experienced it on any of several computers nor my iPad/iPhone. I own two 6500's. Have you tried different computers and experienced this same lock-up problem? It would make me think there is a glitch in your radio.

Jim, K6QE

Andy M5ZAP

Hi Jim

It is the flex locking up, it doesn't affect all users and has been discussed at great length in other threads.

Jim Gilliam

Maybe it is a heat issue on your radio. I wish I had followed this issue on other threads. Hope you get it resolved. Since it seems to be related to a few select radios, perhaps Flex would want to use your radio to help diagnose the problem to see why this occurs on a few select radios. Perhaps some of the chips are not up to spec and no software can fix this possible hardware problem.

Jim

Tim - W4TME

The truthful answer is "hopefully".  We have not definitively found the root cause for the relatively small percentage of radios experiencing the crashing problem (the numbers do not minimize the severity, it just makes it much harder to find root cause when it isn't systemic).  If we had found a root cause, the fix would have been released immediately.

Whenever a major feature is added to the software, like SmartLink, it touches many different areas in the code and through whitebox testing exposes weaknesses or issues we may have overlooked.

Through the course of adding the new feature, we address those areas and make fixes and optimizations along the way.  We affectionately refer to this as "under the covers" fixes; things that have no visual component but improves stability and performance.  It is the serendipity effect.  This happened a lot when working on contesting enhancements in the past 2 minor releases.

We have done a lot of "under the covers" type work for 2.0.  Recent alpha testing for 2.0 has provided some strong empirical evidence that one or a multiple of fixes implemented as part of the work done for SmartLink has greatly improved radio stability.  It is partially due to this fact that after 2.0 is released, we will use its code as the basis for the 1.11 maintenance release.  Essentially it will be 2.0 code fixes with the 2.0 specific features removed.  Our expectation is that this will improve the stability of the radio and be a remedy for those experiencing periodic crashes. I anticipate that we'll share 1.11 beta releases with some who have reported the problem and are good candidates for testing.

If the crashes (not disconnects, which has a different problem pathology) continue for some, we'll reopen our investigations and dig a little deeper on those specific issues. 

Andy M5ZAP

Has anyone who is experiencing the issue on 1.10.16 tested V2.0 ?

One of the main reasons I am worried is, if it is a hardware issue that is causing the lockup (such as a stack up of tolerances) and as your future software places more and more demand on the hardware I could be  stuck with a radio that wont run software beyond 1.9.13. If that was the case I would need to return it before my warranty runs out.

Michael Coslo

Have you tried running it on a different computer?

Andy M5ZAP

It's the flex that locks up not the computer! Plus we don't want to restart the old thread. I am just trying to ascertain if V2.0 solves the problem for the many users with the same problem. But thanks for the suggestion:)

Andy M5ZAP

Hi Tim,

Any chance of an official response on the part of my post asking if you have got a user who experiences lockups on 1.10.16 to test V2.0 to determine if the problem carries over. Even though your reply was quite detailed it did not answer that specific point.

If the answer is "Yes" then did the problem persist. If your answer is "no" which I presume it will be judging by your reply, then will you be planning to perform this obvious test as it will help me and many other people suffering the flex lockup problem make a purchasing decision.

Bill -VA3WTB

In Tim's reply to you he said they will likely have some of the ones having this problem beta test 1.11,,that was 7 hours ago...

Gerald-K5SDR

I also expect that we will also know more when some people who have the problem test v2.0, which will be sooner than v1.11.

Andy M5ZAP

Gerald,

Thanks for your reply, I appreciate you answering the actual question asked.

Have you supplied replacement hardware to an owner who suffers the lockup problem?

This simple test would allow you to determine if the flex hardware has an influence and also potentially give you hardware that reproduces the fault.

By you own admission in another post "All your software team is working on is V2.0"

My route forward with this problem is as follows

1 - Test my hardware setup using a different PC connected directly to my flex with a different power supply, Ethernet cable etc. Eliminating the PC, network and power supply.

2 - Listen only so RF cannot be blamed.

3 - If the problem still persists then test the radio on V2.0 when it is released.

4 - If the above does not fix the issue it has to be flex hardware related and the unit will be returned under warranty I would also expect the cost of V2.0 to be refunded. As my warranty will expire I am time constrained and if  V2.0 is not released within 2 months I will be forced to skip testing V2.0  

All the above are simple tests and all could of been completed by flex.

Now before I get slated and ridiculed by the loyal flex followers (I'm not a nasty terrible person just a guy trying to sort out a problem) please bare in mind, yes I have tried to sell my flex and I got one low offer from the forum (I wish I had taken it now) and a couple of people interested in the UK. But because I have integrity I explained it had the lockup issue and then there interest stopped abruptly. No I didn't buy the radio as an investment, I purchased it to use and enjoy, but I do like to upgrade and try new kit and I didn't expect to loose half the value of the radio in 8 months.

I am a person who likes anything I own or use to be setup and work perfectly (Perfectionist?) I get as much pleasure setting up my radio system (Hardware and software) as I do operating. Hence when part of my system doesn't work it spoils my enjoyment. Some people are happy to simply reset there system and carry on for me it really bugs me. And to make it worse I cant resolve it!!

I am not expecting an instant reply as I understand it is your holiday. Hope you have a good holiday and look forward to your reply on your return to work!

Gerald-K5SDR

Andy,

Not much time off today for me.  ;>)

Unfortunately, we have not been able to recreate this problem internally but we do believe it to be software.  It is a real problem for a small percentage of our customers - maybe 1%.  

Your test process sounds reasonable but you should also try testing with v1.9, which you can do immediately.  v2 is in the final test stage so it will be out very soon.  Once v2 is in the field we will evaluate to see if it has resolved the issues some customers are having.  If so, we can then move that code base into a v1.11 for those who do not wish to upgrade.  As far as I know, the people testing v2 are not seeing the issues.

I sounds like your radio is only 8 months old so you have another 16 months of warranty left so you don't have a problem on that front.  

It would be helpful if you would enter a Help Desk ticket so we can track your issue and the results of your testing.  You can enter a ticket from our website here:

https://helpdesk.flexradio.com/hc/en-us/articles/202118688-How-to-Submit-a-Request-for-Technical-Support

Thanks and 73,
Gerald

Andy M5ZAP

Hi Gerald

1 - I have already reliably run on 1.9.13 the problem was not present, the first 1.10 public beta was horrendous with many lockups which improved on the second public Beta and remained at that level in 1.10.16 release.

2 - I did have a help ticket open but after logging software was put on my PC by flex to try and capture the issue and the issue did not occur for two weeks while the software was installed the ticket was closed!? Do you want me to open another ticket ?

The problem is quite sporadic in nature, but often happens in clusters with 2 or 3 or more lockups in a short period of time. At one point I even thought the problem had gone away after the logging software was installed but it did return and has be sporadically present still.

Also for any contests i downgraded to 1.9.13

Again I feel that to recreate the issue obtaining hardware from a local owner in the USA experiencing the problem would potentially help you recreate the problem if there is a hardware influence.

Tim - W4TME

The plan is to validate the fixes in the 2.0 release.  Migrate them into a 1.11 release.  If at that time if anyone continues to experience the lockup issues, then we'll address those specifically via our HelpDesk ticket process.  Thanks for your continued patience while we work through our plan.

Paul M

Tim I have been a supporter of your efforts and appreciate your help in mitigating my crashes, but this post is tone deaf.  Let me tell you how it reads to me.  "We want to get the revenue generating software out there before we take care of our obligation to the existing base."   I am sure that is not what the company means but that is how it sounds, at least, to me.   

KY6LA_Howard

@Paul You are definitely misinterpreting what Tim said V2 has some entirely new code in it which may cure the problem of lockups entirely The best way to test this is to put out V2 and test it over a wide customer base to see if it does the trick If so, then the new code of V2 would be integrated into V1.11 as a fix for non upgrades

Michael Coslo

- Paul Computer based software, especially in the Windows 10 ecosystem, is an issue. Flex isn't remotely the only manufacturer to have these problems, especially with Windows updates. For myself, I had one lockup, then a re-install of the software, and after the next W10 update, the problem hasn't come back. What does that mean? I have my ideas, but attempts to get people with problems to post their hardware suites and not gettting any response, that troubleshooting effort went nowhere.

And this is based on a hunch ( technical term) that there is some hardware with an issue, that there might be a driver that works, but it gets replaced by an update with what Microsoft thinks should be there.  But I was a sample of one, and the software logs are inconclusive. So I move on to finding the next shared element. 

Windows problems, especially ones that are difficult to replicate, are a nightmare for Software writers. Especially when we want to blame the software writers and not the root cause of the problem. That would be Microsoft. 

Andy M5ZAP

Sorry Michael,

I think you mis-understand the problem. It is not the Smart SDR for windows which is locking up it is the Flex firmware itself. The radio locks and becomes unresponsive.It has been discussed quite a lot in a previously closed post.

Paul M

Howard and Michael - I do not think I misinterpreted Tim's comment at all.   Reading it again I will still stand by my comment that it is tone deaf.  The COM port issue has a purported fix but it has not been released, the DAX issue has a purported fix but has not been release to the general community.  I am  beginning to resent the fact that i have to use a third party program every day or so to clear my COM ports so that my FLEX software does not crash.

Many of us on this board my, self included, have either worked for, as in my case, owned hardware companies that have lived through similar problems with random support issues that seemed unsolvable.   I can bore you with 10's if not 100's  of cases of issues that cropped up in the late 80's through the late 90's with my own hard drive and tape drive subsystem company and troubleshooting personal configurations.  We supported Windows, Sun, MAC and AS-400 so I have had my share of it works here but does not work there and even we cannot duplicate it in our shop issues.  In our own shop our initial DLT implementation worked on one RISC AS-400 but and identical machine just a few serial numbers away with the exact configuration would not see the drive. Noodle that one for a while.  

In the 80's there was an awful book written, awful in my opinion, by Tom Peters titled 'In Search of Excellence'.  I had the opportunity to hear Peters speak at a bank function in Dallas where someone asked him why he thought IBM was such a good company.  I will never forget his answer  'When an IBM customer has a problem they send 30 people to fix it; 29 by parachute'.  Being a 3090 shop in 1985/6 I can attest to that statement being true.

It seems from Tim's comment and others we do not even have one person in a row boat working on the old code.  Our hopes are pinned to the possibility that 2.0 is the fix while I continue to clear my COM ports each day hoping that is a root cause of at least one of the crashes and for the record I believe it is since my crashes are down to one every few weeks instead of one every few minutes.

These issue exist on bare machines with only minimal additions.  I have had crashes on an Intel i5 and an AMD 9XX 8 core. The Intel had nothing but SmartSDR, DDUTL, Ham Radio Deluxe and WSJT-X software.  The only non-ham radio software was Chrome.  

I will say it again, I have been a supporter of Flex on finding this issue but Tim's  comment is tone deaf to the people that are having issues.  

(I use Windows 7 not Windows 10)

Steve Gw0gei

Let's hope v2.0 comes out soon, so this wider testing can commence and those of us not wanting wan can wait as little as possible for the subsequent 1.11 maintenance release. I will be travelling tomorrow to a host station with a pair of ft5k radios in antenna diversity mode for 40m cw band slot as part of the rsgb gr2hq team in iaru contest. My 6300 is only being used for casual operating and the odd domestic contest at present until the random crash issue is solved. Tried selling it over the last six months but no takers despite reducing the price several times. On the plus side my wife is happier as I am spending less time on the radio! That may change shortly though as I have just taken delivery of one of the first production models of Linear Amp Uk 's new 1.2kw full duty cycle solid state 2m pa - delivered on time , 24 hours after I paid the emailed invoice and three months after my no deposit order was taken. Will be visiting their workshop on 20th July to see the pre production model of the new solid state fully auto lan enabled 1kw hf/6m pa under test - with the state of the uk pound value it might turn out cheaper to put two of these new amps on my a and b inputs on my 4o3a antenna genius than a single power genius.

KY6LA_Howard

@Paul We have similar backgrounds of running our own high tech shops except mine started in the mid 60's. Lost my appreciation of IBM when we caught them stealing our patents albeit we were ultimately successful in the courts in getting them to buy licenses ' I do think ur missing the point. The Lockups are a head scratcher Not reproducible or easily identifiable or for that matter only affecting a tiny percentage of the customer base. All attempts at fixes have failed so far as Flex does not know the cause or a possible fix The definition of insanity is to repeat the same thing and expect a different result.. So flex, in my opinion has come up with innovative solution Rather than wasting resources trying to fix something in v1 they can't find they decided to rewrite the CodeBase from scratch and see if the rewritten CodeBase fixes the problem. Hence version 2.0 which will have a completely new code base and hopefully after testing in a wider environment may have found to have a cure the problem Flex is very close to releasing version 2.0. Once it's released we'll see if it fixes the problem

Paul M

Howard - I appreciate your point of view and in the end it will probably workout the best for the community, but the fact that it only affecting a few machines is a fallacy in my opinion.  I have spoken with a number of Flex owners that do not follow this board that have not reported the problem.  How about Flex emailing all of the owners an asking if they are having the problem if they want to maintain that position?  My bet is that census will show a much different result than the company currently stands on. In lieu of that, if the software is nearly ready, send out a few timed copies to the people that are having the most issues (no not me I am not an early adopter)  and see how it works in the crashing environment.   

I was the Treasurer of an alternate long distance company in the early 80's (the era of 27 digit dialing to only pay $0.25 a minute for a long distance call) and I made the following comment in one of our board meetings.  There are two types of people that quit using us, those that call up, yell and scream, swear they will never use us again and those that just never use us again.   Flex's 'close to the vest policy' is not helping me stay out of the second group.   Many will leave and more important in the name of close to the vest policy they are losing customers who have decided to just not use them again.

I stand by my comment that telling people that you are going to release the revenue software before you address the non-revenue software issues is Tone Deaf.   It may be the truth but surely there is a better way to tell the masses that you are second fiddle. 

NM1W

Tim - Is Howard correct that 2.0 is a "completely new code base" ? And does that include a new (hopefully fixed) dax?

Tim - W4TME

Completely new as in we started from scratch?  No. SmartSDR v2.0 was built on 1.10.16.  What is new is that for the past several months we have put all our development effort into what has become the 2.0 code base.  It contains numerous fixes and improvements to existing features and capabilities in addition to the two new core features in 2.0 - SmartLink and Popout Panels.

As I have said before when you add a major feature like SmartLink, it ends up touching a lot of code.  During this process, you discover things that result in core blocks of code getting refactored (rewritten).  The end result is you come away with a software product that has better performance and stability.  

It would be resource wasteful to have our engineering team split in order to work on two code bases at the same time.  Instead, we address problems reported on 1.x by testing for them using our 2.0 code base and then address anything we discover in 2.0.  We'll then take the 2.0 code base and merge it into 1.x code base as 1.11 when 2.0 is released.  This way, the 1.x customer receives a more comprehensive set of code fixes and improvements.  In this particular case, it will take a little longer before we got a 1.11 fix out than is ideal, but it will be a better release than if it was released last month, as the fixes we implement today will be in the 1.11 code.

Case in point, while testing the 2.0 release candidate code over the holidays, we stressed an internal routine that was called from multiple locations in the code into an edge case condition that caused a radio crash.  We found the root cause and fixed it making the software better. Is this the crash that is affecting 1.10.16?  Maybe.

Now we could take the thousands of changes we have made for 2.0 and merge them into 1.10, but that is a lot of tedious work and provides multiple opportunities for making a mistake.  Instead, we will take the 2.x code base, remove the 2.0 specific changes and licensing requirements and build 1.11.  This way people choosing to continue to use 1.x will have all of the bug fix benefits as 2.0.

This was a business decision FlexRadio made that allows us to meet multiple goals simultaneously.  Was it a perfect decision that will please everyone?  No.  That is impossible, but as with any business, you have to balance multiple needs and we felt that this was the right thing to do.

For SmartSDR v2.0, DAX has not undergone any changes.  It is, however, one of the items we have scheduled to address in a future version of SmartSDR v2.x.  

Paul M

"It would be resource wasteful to have our engineering team split in order to work on two code bases at the same time.  Instead, we address problems reported on 1.x by testing for them using our 2.0 code base and then address anything we discover in 2.0.  We'll then take the 2.0 code base and merge it into 1.x code base as 1.11 when 2.0 is released.  This way, the 1.x customer receives a more comprehensive set of code fixes and improvements.  In this particular case, it will take a little longer before we got a 1.11 fix out than is ideal, but it will be a better release than if it was released last month, as the fixes we implement today will be in the 1.11 code."

Tim - contrary to your statement this is how companies write new versions.  One team keeps working on the old code so that the customer base stays loyal and the other team develops the new code  incorporating the changes that the maintenance team finds and fixes if appropriate.  
 
"... We'll then take the 2.0 code base and merge it into 1.x code base as 1.11 when 2.0 is released.  This way, the 1.x customer receives a more comprehensive set of code fixes and improvements.  In this particular case, it will take a little longer before we got a 1.11 fix out than is ideal, but it will be a better release than if it was released last month, as the fixes we implement today will be in the 1.11 code."

Not that I expect Flex to take my advice (or get an answer) but why not let the license manager decide what people have paid for and what they have not.  Then you will truly not have to release two versions,just download the current version, put in your license ( or no license) and you will have the options you paid for and nothing more.    

If you detect a change in my tone it is that I have reached my personal end of the rope after your post earlier today.    It just took me right off my support of Flex pedestal and sent me to the why do I want to give this company one more dollar side of the fence.  My 6500 is less than 4 months old, I have never complained about buying only to have it obsoleted two months later.  My choice to buy that close to Dayton and I knew the risk, but to basically be put in the category of 'We'll get to you later' is an insult.  

One last question that I do not expect to get a YES as an answer. Since you seem to have eradicated what is possibly the BUG are you willing to give a release date or reasonable window? 


Paul -W4PGM

Tim - W4TME

One last question that I do not expect to get a YES as an answer. Since you seem to have eradicated what is possibly the BUG are you willing to give a release date or reasonable window? 

This question was answered by Gerald - "We are not giving a date for v.11 - only it will be after we have broad field experience with v2.0"

Bill -VA3WTB

Paul, why, or what has made you 6500 obsoleted? I don't understand it.

Paul M

Obsoleted by the companies product announcement.    Obviously it still works if you ignore the crashing.   I made the decision for the 6500 close to Dayton knowing that there was a likelihood of a new announcement.   The trade that Flex gave me for the 6300 was fair at the time and amazing with 20/20 hindsight so I have no issues with the purchase.  

My issue is that the company is Tone Deaf to an issue that is over  6 months old.  Go back  and read the threads and there are complaints 7 months ago and probably longer if I took the time to read all the posts.  Not addressing the issue in 7 or more months and then saying to get a possibly stable platform buy our upgrade or wait until we decide we want to put the time and effort into the 1.11 is an insult to the installed base.   

Why should I support this company any longer if their answer to supporting me is downs grade, wait an unknown time, or pay the upgrade and in all three cases the unsaid Pray?   

Ria

I am running alpha software, and I have had the radio hang in the past but with the 2.0 code these are kept to a minimum now. So if that carries to 1.11 it should be fine.

No promises (I can't make them anyway), just my observation. I also hope that they have definitely found and corrected the root cause. 

Ria

Ria

Tim - contrary to your statement this is how companies write new versions.  One team keeps working on the old code so that the customer base stays loyal and the other team develops the new code  incorporating the changes that the maintenance team finds and fixes if appropriate.  

 
Are you familiar with the term "backporting?" This is exactly what Tim described, and is how a lot of companies such as RedHat actually take new fixes and add them to older versions. It's how a lot of dev teams I've worked with have done it. So essentially 1.11 will have backported fixes from the 2.0 code. 

Not every company has the luxury to work on a new version while dedicating staff to maintain an older one. Besides, it's not like 2.0 is really written from scratch anyway, so backporting should be much easier.