Call Search
     

New to Ham Radio?
My Profile

Community
Articles
Forums
News
Reviews
Friends Remembered
Strays
Survey Question

Operating
Contesting
DX Cluster Spots
Propagation

Resources
Calendar
Classifieds
Ham Exams
Ham Links
List Archives
News Articles
Product Reviews
QSL Managers

Site Info
eHam Help (FAQ)
Support the site
The eHam Team
Advertising Info
Vision Statement
About eHam.net

   Home   Help Search  
Pages: [1] 2 3 4 Next   Go Down
  Print  
Author Topic: Thick-Pipe vs Thin-Pipe SDR Configurations  (Read 17038 times)
K9ZW
Member

Posts: 180


WWW

Ignore
« on: May 21, 2013, 02:49:55 PM »

Very interesting aspect of the presentation at the FlexRadio Systems 2013 Dayton Banquet was the distinction between the Thick-Pipe PowerSDR older configuration using the distributed processing of the SDR Device and PC in an intertwined fashion, and Thin-Pipe SmartSDR where all needed processing is undertaken in the SDR Radio Server and the PC is just an interface to the user.

Would seem that Thick-Pipe will remain a favorite of the deeply involved experimenter and hobbyist, while Thin-Pipe generally will appeal more to those seeking raw performance gains.

Lots of discussion going on about these two concepts with their pros and cons.

73

Steve
K9ZW

BLOG:  http://k9zw.wordpress.com/
Logged

NI0Z
Member

Posts: 583


WWW

Ignore
« Reply #1 on: May 21, 2013, 04:18:06 PM »

Honestly my own take is that Flex's primary objective behind moving processing back to the radio was around reducing support costs.  Combine that with the $200 maintenance fee and you reduce cost and increase revenue.  I hope people don't object to this post, but, it's not a bad thing for a company to make itself more profitable.  This is also why they were in a hurry to discontinue the Flex 5K.  Expect the other models to follow suit once the base software is stable and they can make scaled down versions of the 6K line.  This is just a common sense roadmap for them and I don't fault them.

Take the hobbiest line of current DDC receivers and you have the opposite cost strategy of moving cost to the PC and thus reducing the cost of the hardware or radio.  This makes great sense for the the hobbiest teams as funds can more efficiently be allocated to bigger badder hardware.  Also, it's easier or more cost effective to get multiple people writing software via Open Source.

What would really be cool is if a bunch of Hams crowd sourced offshore development of new SDR software for the hobbiest platform.  We could get really nice new software like Studio 1 but at a much more accelerated delivery pace.  Yes, we would have to pay for it, however, it could be made to,surpass what flex will develop and cost each ham far less in the long run.  IE, we pay for the code to become open source but accelerated.  That way it's freely updated, but developed more rapidly.  I forgot where I read that this technique was recently deployed and highly highly successful, but suffice to say it can be very effective and efficient.

There is no reason to say thin or fat would limit features, both ways are capable of delivering robust solutions!
Logged

K5TED
Member

Posts: 815




Ignore
« Reply #2 on: May 21, 2013, 06:49:42 PM »

It seems as if the whole move to FPGA based rigs is for Bazillioflops of processing to give the ability to look at huge swaths of bandwidth, zooming in and out of ten ham bands at once on a multi-color panadaptor with integrated youtube recording and posting capability, to create 72 or 1300 simultaneous receivers.

All of that is good and interesting and useful, but given the choice, I'd rather put all that processing towards a receiver that can pick a gnat fart out of a hurricane of QRM, reducing the extraneous noise to 0. Give me a radio that can be perfect on one frequency at a time.

Logged
NI0Z
Member

Posts: 583


WWW

Ignore
« Reply #3 on: May 21, 2013, 07:22:25 PM »

If I am correct, the FPGA actually has a lot of resources still available that can and to some extent can/will be programmed to do more processing in the radio.

Diversity reception in its future state will be able to use two signals to cancel noise.  I liken it to CCD cameras where they use dark frames to cancel the noise in the actual frames.

I think there is stil a lot that can be done to improve signals when they are converted to digital data, we are only probably half way there in having tapped its capabilities.  The first phase was to reach parity with knobbed based radios, now we need the outside the box thinking and paradigm shift in signal processing to take us to the next level.

I suspect these statements will be greeted with skeptisism, however, time has a funny way of bringing down the walls that block us.

Once we greatly reduce the noise we will uncover and recover the signals that have been hidden in it!
Logged

W6UV
Member

Posts: 540




Ignore
« Reply #4 on: May 22, 2013, 06:52:38 AM »

IMO, the thin-pipe approach is where Flex should have started, not the thick-pipe used in the current models. Using an operating system not designed for real-time processing to process time sensitive data is just a bad idea. It's what lead to the lag issues with CW and the stuttering issues some people have seen with the Flex-5000.

The other big issue is that Flex has no control over the hardware used on the PC end. There are all sorts of bizarre combinations of hardware/software out there that are not conducive to smooth operation of what essentially is a real-time system, plus you have people who don't know what they're doing playing around with configuration parameters and driver settings with no idea of what they're trying to accomplish, or why.

Keeping as much of the core signal processing in the box and off the PC is a good thing.

Decoupling my criticism of how Flex handles product announcements from the product itself, I think the Flex-6x00 is actually a very interesting design with lots of potential. They just need to moderate their promises and deliver on the ones they do make.
Logged
SWL2002
Member

Posts: 374




Ignore
« Reply #5 on: May 22, 2013, 07:15:41 AM »

IMO, the thin-pipe approach is where Flex should have started, not the thick-pipe used in the current models. Using an operating system not designed for real-time processing to process time sensitive data is just a bad idea. It's what lead to the lag issues with CW and the stuttering issues some people have seen with the Flex-5000.

The other big issue is that Flex has no control over the hardware used on the PC end. There are all sorts of bizarre combinations of hardware/software out there that are not conducive to smooth operation of what essentially is a real-time system, plus you have people who don't know what they're doing playing around with configuration parameters and driver settings with no idea of what they're trying to accomplish, or why.

Keeping as much of the core signal processing in the box and off the PC is a good thing.

Decoupling my criticism of how Flex handles product announcements from the product itself, I think the Flex-6x00 is actually a very interesting design with lots of potential. They just need to moderate their promises and deliver on the ones they do make.

This is the problem with Flex.  Everyone with little technical knowledge thinks that how Flex implemented their SDRs is the way SDRs are/should be.  There is nothing farther from the truth.  There are many SDRs on the market (many of which I have personal experience with) that use the PC for DSP processing and DO NOT have the latency issues/hiccups that Flex had.  Flex chose a poor hardware interface as well as a kluged together combination of C for a DSP core and MS C#/NET for PowerSDR.  If you would venture out of your little Flex-colored world into the larger world of SDR, then you would see that there are many successful PC based implementations without the issues that Flex has had.

It does not help when people like you spread SDR folk lore on your blogs for the uneducated Ham population to read and spread the folk lore.

« Last Edit: May 22, 2013, 07:19:39 AM by SWL2002 » Logged
K9ZW
Member

Posts: 180


WWW

Ignore
« Reply #6 on: May 22, 2013, 08:34:15 AM »

IMO, the thin-pipe approach is where Flex should have started, not the thick-pipe used in the current models. Using an operating system not designed for real-time processing to process time sensitive data is just a bad idea. It's what lead to the lag issues with CW and the stuttering issues some people have seen with the Flex-5000.

The other big issue is that Flex has no control over the hardware used on the PC end. There are all sorts of bizarre combinations of hardware/software out there that are not conducive to smooth operation of what essentially is a real-time system, plus you have people who don't know what they're doing playing around with configuration parameters and driver settings with no idea of what they're trying to accomplish, or why.

Keeping as much of the core signal processing in the box and off the PC is a good thing.

Decoupling my criticism of how Flex handles product announcements from the product itself, I think the Flex-6x00 is actually a very interesting design with lots of potential. They just need to moderate their promises and deliver on the ones they do make.

Back in 2005 to 2007 when the PowerSDR Thick-Pipe approach was developed the processing power of PC was needed to make the technology accessable to amateur radio enthusiasts.

Much of the processing power needed for a Thin-Pipe approach was then either unavailable, wasn't capable enough or was too expensive for ham radio.  "Uncle" may have been willing to pay $50,000 for a radio, but very few hams were likely to do so.

The drop in costs has made the new products affordable.  When compared to $10,000 plus transceivers from a number of manufacturers the new Thin-Pipe products are a good value.

You're right that the external PC has been troublesome in the Thick-Pipe model.  This should be fairly close to a non-event in the Thin-Pipe model.

73

Steve
K9ZW

BLOG:  http://k9zw.wordpress.com/
Logged

W6UV
Member

Posts: 540




Ignore
« Reply #7 on: May 22, 2013, 09:41:42 AM »


Back in 2005 to 2007 when the PowerSDR Thick-Pipe approach was developed the processing power of PC was needed to make the technology accessable to amateur radio enthusiasts.

Much of the processing power needed for a Thin-Pipe approach was then either unavailable, wasn't capable enough or was too expensive for ham radio.  "Uncle" may have been willing to pay $50,000 for a radio, but very few hams were likely to do so.

That is true to some extent, but a more modest implementation would have been possible, even then. The Flex-6700 is so overboard in processing power that really isn't necessary for most people -- something more modest could have been implemented that would have served the needs of the vast majority of users. Who really needs eight slice receivers anyway? Sure, you can come up with ways to use them, but the vast majority of the marketplace is still coming to grips with transceivers with a subreceiver.
Logged
NI0Z
Member

Posts: 583


WWW

Ignore
« Reply #8 on: May 22, 2013, 12:08:38 PM »

All will seem well with the thin pipe until the fat pipe takes the lead again. Then people will want to get fat again! Smiley. Lol, remember skinny people die of heart failure too!

Seriously, again, the reasons can't be quantified in terms of one being better than the other.  We could argue all day on it, provide examples both ways and never agree.  And the funny part about all that is none of it would have to do with the real reasons there are fat and thin clients being developed out there.

This debate so far is more about my radio is better than your radio than it is about the real merits and reasons for each. 

Flex, Apache, who really cares? Like the radio you own and if you don't like it then get rid of it and get another! It's not rocket science!
Logged

W6UV
Member

Posts: 540




Ignore
« Reply #9 on: May 22, 2013, 12:35:16 PM »

Flex, Apache, who really cares? Like the radio you own and if you don't like it then get rid of it and get another! It's not rocket science!

That's the position I'm in now. I'm trying to decide whether to get a Kenwood TS-990S or something else. Do I want to go out on a limb and get something like a Flex-6500? Maybe. Maybe not. My decision will be colored by my experiences with my Flex-5000A.
Logged
W4HIJ
Member

Posts: 367




Ignore
« Reply #10 on: May 22, 2013, 01:13:07 PM »

Hams in general are notoriously cheap. A lot of problems with Flex radios and the thick pipe implementation was that guys were trying to get by with the absolute minimum PC that they could. It was either some old one they had in their shack for years or some discount store package.  Add to it that a lot of hams don't really know jack about computers except running a logging program on them and you have the environment for there to be myriads of problems with PowerSDR. On the other hand, some of us who had experience with PC's and building custom machines for high end gaming and other applications knew from the get go what it was going to take to make PowerSDR run successfully and had very little, if any, problems.  With the thin pipe implementation Flex is able to protect the ignorant from themselves! Roll Eyes Roll Eyes
73,
Michael, W4HIJ
« Last Edit: May 22, 2013, 01:15:20 PM by W4HIJ » Logged
NI0Z
Member

Posts: 583


WWW

Ignore
« Reply #11 on: May 22, 2013, 01:50:12 PM »

Flex, Apache, who really cares? Like the radio you own and if you don't like it then get rid of it and get another! It's not rocket science!

That's the position I'm in now. I'm trying to decide whether to get a Kenwood TS-990S or something else. Do I want to go out on a limb and get something like a Flex-6500? Maybe. Maybe not. My decision will be colored by my experiences with my Flex-5000A.

Study them all well, SDR's are not for the faint of heart!  I'll soon have my 3rd and they all have their own challenges, most not actually being hardware issues.   That said, once Flex has all the bugs worked out the 6500 has the POTENTIAL to be a very nice radio.    I'll know more about the Anan later next week.  I think the KX3 is my fav right now, but it also has knobs! Smiley
Logged

K5TED
Member

Posts: 815




Ignore
« Reply #12 on: May 22, 2013, 02:28:29 PM »


There are many SDRs on the market (many of which I have personal experience with) that use the PC for DSP processing and DO NOT have the latency issues/hiccups that Flex had.....If you would venture out of your little Flex-colored world into the larger world of SDR, then you would see that there are many successful PC based implementations without the issues that Flex has had.



Please entertain us with a comprehensive list, or even two examples of mainstream SDR's on the market comparable in overall TX/RX performance to the Flex 5000, or even the 3000. In fact, if you know of one with the performance, and fit/finish of the Flex 1500, please document here.

Logged
NI0Z
Member

Posts: 583


WWW

Ignore
« Reply #13 on: May 22, 2013, 02:38:43 PM »


There are many SDRs on the market (many of which I have personal experience with) that use the PC for DSP processing and DO NOT have the latency issues/hiccups that Flex had.....If you would venture out of your little Flex-colored world into the larger world of SDR, then you would see that there are many successful PC based implementations without the issues that Flex has had.



Please entertain us with a comprehensive list, or even two examples of mainstream SDR's on the market comparable in overall TX/RX performance to the Flex 5000, or even the 3000. In fact, if you know of one with the performance, and fit/finish of the Flex 1500, please document here.



The KX3, much nicer fit and finish and near the top of the charts for receive capability.  Fully decked out it only cost around $1500.

Anan 10
Anan 100
Anan 100D

All of which can out perform the Flex 5000.  Possibly may out perform the Flex 6K series as well.  Have to wait and see!

Quick Silver QS1R ... out performs the Flex 5000.

They are all cheaper new than a used Flex 5K.  In most cases even if you add a TenTec or Elecraft 100 watt amp they still end up being cheaper.

And here is the kicker, lol, they are all Fat!

Those are just a few off the top of my head.

 There are a few German ones that probably will measure up as well.  Flex 5K is old stuff.  Most of the new SDRs are going to make it pale in comparison.
« Last Edit: May 22, 2013, 02:47:38 PM by NI0Z » Logged

K5TED
Member

Posts: 815




Ignore
« Reply #14 on: May 22, 2013, 03:31:39 PM »


There are many SDRs on the market (many of which I have personal experience with) that use the PC for DSP processing and DO NOT have the latency issues/hiccups that Flex had.....If you would venture out of your little Flex-colored world into the larger world of SDR, then you would see that there are many successful PC based implementations without the issues that Flex has had.



Please entertain us with a comprehensive list, or even two examples of mainstream SDR's on the market comparable in overall TX/RX performance to the Flex 5000, or even the 3000. In fact, if you know of one with the performance, and fit/finish of the Flex 1500, please document here.



The KX3, much nicer fit and finish and near the top of the charts for receive capability.  Fully decked out it only cost around $1500.

Anan 10
Anan 100
Anan 100D

All of which can out perform the Flex 5000.  Possibly may out perform the Flex 6K series as well.  Have to wait and see!

Quick Silver QS1R ... out performs the Flex 5000.

They are all cheaper new than a used Flex 5K.  In most cases even if you add a TenTec or Elecraft 100 watt amp they still end up being cheaper.

And here is the kicker, lol, they are all Fat!

Those are just a few off the top of my head.

 There are a few German ones that probably will measure up as well.  Flex 5K is old stuff.  Most of the new SDRs are going to make it pale in comparison.

Starting at the bottom, There's a $17000 German SDR that comes to mind. Definitely not mainstream, nor are there any other German offerings in the world of SDR transceivers that can be considered "mainstream".

Quick Silver QS1R is not a transceiver.

The Anans are possibly the only contender for the coveted position of "mainstream", even though they've only been in production for months. That's not quite what I would consider "many" transceiver SDR options currently available.

As for the KX3, it's priced closer to a Flex 3000, and has 1/10 the output. To give it the full output, one would have to bump the price up at least $400 for a reasonably good amp. That's up to almost $2000 with TTL. I just saw a used Flex 5k with 2nd rRX and ATU go for $2375 right here on eHam. Have you ever compared the rearend of a Flex 5k with 2nd RX and ATU to the few available connections on a KX3? Hardly in the same class of radio.


So far, I haven't seen "many" other mainstream SDR transceivers comparable to the Flex radios in price point/feature set. There are a couple of dissimilar sets that offer some of the same functionality and features.

It's just a matter of time, of course, before some mainstream manufacturer unveils a true competitor. I suspect Flex has something in store to replace the 3000.

Oh, about that Flex 1500 comparison... Oh wait. There is no other mainstream manufacturer ready to go out of the box $700 5 watt 160-6m USB SDR with dual antenna ports and IF out. Is there? I may have missed one...

Logged
Pages: [1] 2 3 4 Next   Go Up
  Print  
 
Jump to:  

Powered by MySQL Powered by PHP Powered by SMF 1.1.11 | SMF © 2006-2009, Simple Machines LLC Valid XHTML 1.0! Valid CSS!