Going back to 1.416 from smartsdr1.5

clint yates

I want ot convert back to 1.416. is there anything special i need to do? i was going to uninstall 1.5 and reinstall 1.416. There are to many bugs in the 1.5 version with the transmit signal. It is wide and dirty. i can reboot and it clears up for a while then comes back with a vengance. Tired of fighting it.

Jim Gilliam

Which radio are you using?

Jim, K6QE

clint yates

6300

Jim Gilliam

I am sure Flex will rectify these 6300 bugs in short order. Sorry for all of your problems.

Jim, K6QE

WA6FXT Mike

To down grade, simply run the older install program. You may want uninstall the current version, reboot, then run the SSDR1.4 installer. Mike

Bob Hinkle

You will absolutely have to uninstall the 1.5 version, reboot, and then install the 1.4 .... 1.5 crashes the program about every 15 minutes on me... "downgraded" tonight and all is well on 1.4.....  KK8ZZ

Dave Dave

1.5 has given me a screen missing a lot of text 3 times..... never had problem with 1.4.

Anyone else have this problem? I talk very little so I have not not iced

clint yates

Yeah..I Uninstalled and reinstalled 1.4 all is well

Mark Griffin

I am presently not a flexradio owner. But it seems that the issues that people are experiencing are maybe related more to the computer they are using with the SDR program? I guess I just don't understand that if Flexradio is testing the beta version of an upgrade to SDR, that the beta testers are not experiencing the same issues that others are.

Maybe we should find out what computers the beta testers are using and if they are not experiencing the same problems, then perhaps they should actively market those computers at a reasonable price so that the Flex community could purchase them and not have all the issues that people are having now.

I think the Flexradio definitely is up and coming and has a great future, but just reading the reflector day by day it seems that the issues are with computers and software, not software and hardware. Am I wrong in that analysis, or not?
Mark Griffin, KB3Z

Bill -VA3WTB

It is interesting indeed. I have very little problems here if any with 1.5. I find the WNB a great tool and works very well. I wonder out of all the hundreds of 6000's out there how many have these problems, or is it just 6 or 7 of them here posting?

Dave Dave

Needing a specific computer for FLEX would be crazy.

Issue is in the software and/or the firmware written to the radio.

Its unusual that so many testers didn't have the problem and with so many users having same problem is something the testers somehow missed.

Jim Gilliam

Indeed, it is a difficult situation. Many are new to this type of radio as well as inexperienced with computers and the blame always seem to tend towards Flex. I have never had any problems other than having a corrupt version of Windows I didn't even know I had until a couple of problems cropped up which were fixed by fixing Windows. However, of late, it seems most if not all the problems are with the 6300. It would be so interesting to find out if it is a Windows problem or a 6300 problem.

Jim, K6QE

Bill -VA3WTB

Flex must be testing this all day long right now to see what they find, I have not heard anything yet

W7NGA

I have attempted to run the gamut of possibilities with v1.5 on the Flex 6300 using Windows 10 and have not had any issues. I've run AM, RTTY, SSB, and CW without problems. I am enjoying the new power-limit parameter .. as before I melted one of the very expensive 3CX800A7's in my Alpha 89 when the power persistence failed me! Lesson learned but it is nice to have a restriction applied .. just in case!

73's

W7NGA  dan
San Juan Island, Wa.

Walt - KZ1F

I do not mean to impune anyone's intentions her but I think there is a fundamental problem with the FRS qa process. The reason for qa groups are two fold, 1 its a full time job (if done correctly), and 2, developers make awfully bad qa engineers. Developers, well intentioned, have subconscious built-in bias. They know how a line of code is supposed to work and, if only subconsciously, the expect it to work. Both of these conspire to make their testing always validate their beliefs. Now, further purturb that by outsourcing the qa to a group of non engineers that, perhaps, are more interested in having the new bits before the public does. Plus, maybe they have their own day jobs. As well intentioned as these folks are this is a hobby to them not a vocation. Qa is not an afterthought. It's a full-time dedicated profession business function. QA people are involved from the outset, they are members of the scrum team(s). There is a metric called code coverage this measures the percentage of lines of source code actually ever being tested. While the ideal would be 100%, over 90% is doing really well. Generally, without real focused effort, the number is under 10%. There are, however, practicalities. We're talking about consumer products here. No one is going to die if it doesn't work. This is why I always advocate for FRS to not feel pressured, 'take all the time you need'. I feel quite alone in that,especially against the chorus of 'are we there yet' and 'is it soup yet'.

Bill -VA3WTB

If it is a problem in the software, why is not every one having the same problems? If it is hardware all radios would be effected. There are many here have no problems, how can that be?

rfoust

That's probably incorrect -- most of the work is done in the radio.  The computer is pretty much just a console that presents what the radio is doing.  The software that runs inside the radio is also called SmartSDR -- my guess is some bug in SmartSDR (in the radio) and how it is interacting with the hardware in the 6300 - not with SmartSDR for Windows.

I actually have a 6700 and love it, the software/radio is very stable for me.

Edit: also, I wonder how many of the beta testers are 6300 users, vs 6500/6700.  Maybe thats why it wasn't caught in beta?

Bill -VA3WTB

Then maybe different releases will be needed for the 6300? Will be interesting to see what they find. But I see W7NGA has just posted saying his 6300 is working very well on 1.5

Jean_Luc F5JRC

Hi Dan, is DAX working correctly?
Jean-Luc / F5JRC

W7NGA

Oxford English

It seems to me that some people don't have a firm grasp of how Smart SDR works.

My understanding is that all the processing is done 'in' the radio and the computer merely acts as a display system.

Having said that, I wonder if most folks have any idea how to test software? In particular Operating Systems? Because, as I understand it, Smart SDR is actually the 6 series OS.

To properly test and OS such as Windows 10 is virtually impossible. This is because the number of possibilities is enormous - for example think about the number of motherboard/processors/memory/graphics cards/memory/ hard disk/power supply/BIOS combinations that it is possible to put together. Even for Microsoft it would be almost impossible to make even one example of each of the permutations that could exist.

And when you have made one of everything, now install your OS and then start to install all the different software packages that users might want to have on their computer. Now start your testing - see you in about 200 years with the completed bug free product?

I know how difficult this task could be - I once had to write a paper on the exact subject. It was 25 years ago and the number of possible combinations at that time was in the hundreds, not the million upon million that exist today.

However, this is not the problem that FRS is facing.

Smart SDR only runs on the 3 radios in the 6 series. Essentially, there are only 3 OS's that it will 'talk' to, Win 7, 8 & 10. So it is not a major problem from a testing point of view.

Reading through all the posts relating to problems with SSDR 1.5, it seems to me that approximately 90% of the problems reported are from users of the 6300.

That leads me to suspect that the code in SSDR 1.5 is causing the problem, NOT the version of Windows that the user is running. 

I further suspect that the problem actually lies with the updating of the 6300 firmware. The reason for think this way is that, as far as I understand the situation, the 6300 is a 'cut down' version of the 6500 & 6700.

I have reverted back to SSDR 1.4.16 running on my trusty I7 and Win7 64bit Pro for the time being. 

I have every faith that at some point in the future, hopefully not to far away, that the guys at FRS will find the source of the problem and give us a fix.

For me, the 6300 is still the best radio I have ever had in over 50 years of playing radio and if I could afford it, I would love a 6500 but I can't see it ever happening.

I'm not holding my breath for 1.5.1 - my radio works just great as it is!

In the meantime, thank you Flex for producing a beautifully engineered radio from  both a hardware and software view.

Bob Hinkle

Add my call to those who had to go back to 1.4 and who have the 6300..... tried twice, on two different machines. The 1.5 runs fine for maybe 10 minutes or so then crashes. Something with 1.5, or the 6300?  What say, Flex-Guys?

Jean_Luc F5JRC

Thanks Dan, it seems the driver Dax issue has just be fixed in a different post.
Some problem here with Dax and SmartSDR 1.5 upgrade on my 6500, but certainly due to how I badly manage the uninstall of the old version.
73 Jean-Luc

Steve-N5AC

There is an intermittent issue in the transmit audio and the vast majority of the complaints are on the 6300, but it is a small group of FLEX-6300 owners that have the issue.  We take this kind of issue very seriously and we have a team looking at this in the lab trying to determine the cause.  If you are having the issue, I would recommend downgrading to v1.4.x until we solve the issue.  As soon as we have news, we will post an update here.

We have a good team of dedicated testers that have been running v1.5 in various forms for months without issue.  The software we released had been in our Alpha team's hands for about three weeks without incident.  As anyone in the electronics or software industry will tell you -- we all wish that things would either always work or break hard.  The hardest issues to find are ones that appear rarely or are selective about where they appear.  Unfortunately, the luck of the draw on this one was that no one on the Alpha team experienced the issue.  But we will find it and get everyone an update just as soon as we can.  The good news is that downgrading is quick and easy.

As has been mentioned here and on the other post about this, the computer and computer OS is not and can not be involved.  The FLEX-6000 radios do all of their signal processing inside the radio -- at no time did your samples leave the radio ;-)  This means that a computer issue will not cause this to happen so my recommendation would be to not alter your operating system, reinstall Windows, etc.  We used to do a test in the early FLEX-6000 days where we would get the radio on frequency and in the right mode and start a QSO and then shut off the computer in the middle of the QSO.  It works just fine and you can complete the QSO, etc.  We got some feedback that the loss of the computer (especially in the remote case) should stop transmit so you can't do this today, but it does illustrate the point that the computer is not involved in your on-air signal.  These things will help other SDRs that depend strictly on the precise timings, buffering, jitter, latency, etc on Windows, but not the FLEX-6000.

Mike - N8MSA

This may be a good time to mention that the Signature Series is a Linux PC with an integrated RF deck. The behavior of the system is dependent upon the Linux OS, the Linux radio application, the DSP code and the FPGA application. This is a rather complex ecosystem making software validation, difficult under any circumstances, exponentially more difficult than just the Linux radio application on its own. There probably isn't a practical way of testing every line of code in this system...

I'm not suggesting that the software quality is perfect, but it's pretty good for what it is and it's getting better. I know that Flex has made a significant investment in improving software quality, but they're in uncharted territory and demanding appliance-like stability from an experimental platform is probably not realistic. 

I know that some people will take offense to this, but these are experimental recreational devices and will probably never reach the level of stability that we see from more traditional radio architectures. 

73,

Mike - N8MSA

W7NGA

I tend to blame all problems in my life on the full-moon and the eclipse. Surely there is a correlation here that also affects my Flex .. but nothing yet has manifest!

W7NGA  dan

Steve-N5AC

Testing is complex, but we have an advanced build and test platform.  Every time an engineer checks in new code, it causes a complete rebuild of SmartSDR and then the new code is loaded into each of the four FLEX-6000 radios and automated testing commences.  Today, most of the tests are control related (not RF), but we expect to add RF tests in the future.  We already have advanced RF test and calibration that occurs in the factory.  We collect hundreds of data points on every radio in the factory and we consult the data when we have an issue on a radio that we think could have been missed in the factory.

It takes a lot of work to build up a platform like this and we're very proud of what we've achieved and the resulting quality.  We also do static analysis with every check-in to locate coding issues.  As Mike mentions, it's virtually impossible to get complete code testing coverage with the size of codebase we have, but we're continuing to add tests and get more coverage as time moves forward.

Just to show how hard it is, Apple recently released iOS 9.0 which had a serious issue that bricked iPads in the field.  Gerald (our CEO) happened to have one of these iPads and his got bricked.  He had to wait for new software to be able to upgrade.  Presumably, Apple has all the money to prevent anything like this from happening, but with today's complex electronics things like this will happen from time to time.  As I said, we take it seriously and work hard to fix the problem as quickly as we can.

clint yates

Not the computer.

clint yates

I feel.like more ppl have the problem they just don't recognize it.

Walt - KZ1F

The issue you're addressing Mike is 'knowing ' or rather understanding precisely what it is you're testing. Testing the OS, where it be Linux or Windows is always out of scope for an application group. It's very easy to be thinking you're testing one thing when, in fact, that's not what you're testing at all. This is but one reason why code coverage software is important, it ultimately maps a given test to the related lines of source code and can color code what's being covered. Testing, not only is its own subject, it definitely is its own discipline. This is why qa engineers are differentiated from mere software engineers.

clint yates

I disagree. Consumer product or not. I paid good money for a product that is supposed to work as advertised. I get the fact that there will be bugs but I believe this should be a high priority. I am satisfied that flex HAS made this a high priority.