Flex 6K ADC overload prevention?

  • 2
  • Question
  • Updated 3 years ago
  • (Edited)
There was a posting to the Elecraft list describing a Field Day operation that included K3, Orion and Flex 6Ks.  It said that the Flex 6K receivers (ssb) were impaired when the K3 or Orion transmitted cw on the same band.  I know many things could possibly explain this but I just wanted to ask: does the SmartSDR software or Flex 6K firmware automatically dial in some front end hardware attenuation as necessary to prevent ADC overload?  I know that the K3 and Orion do just that ("hardware AGC" on s9++ signals).

73, Barry N1EU
Photo of Barry N1EU

Barry N1EU

  • 495 Posts
  • 121 Reply Likes

Posted 3 years ago

  • 2
Photo of Tim - W4TME

Tim - W4TME, Customer Experience Manager

  • 9149 Posts
  • 3467 Reply Likes
Yep, I suspect the 6000 was adversely affected by the phase noise from those transmitters.  I'd need to see a picture of it to be certain.  We have this issue when we demo the radio at some hamfests where there is a special event station.  The radio can easily handle the signal level, but the phase noise is incredibly wide and that wipes out a portion of the band. Signal is signal, no matter how it is produced. 

The FLEX-6000 ADC can handle a very large amount of aggregate signal and will overload somewhere in the neighborhood of +9 dBm.   When this happens you will see spurs or "images" on the panadapter.  There isn't an ADC attenuator that automatically kicks in to reduce the when you approach the ADC saturation level.  The AGC only operates on the recovered audio inside the RX passband filter.
Photo of km9r.mike

km9r.mike

  • 425 Posts
  • 62 Reply Likes
This brings up a question that I meant to ask while talking to a sales rep prior to purchase and the question still remains. I understand the key importance of the ADC w/in the flex and also understand that technology advances rapidly but do not understand the impact that can have on the advancement of ADC capabilities. In other words, do we live in an age that tomorrow's ADC will make yesterday's ADC look like a neanderthal and would that require a whole new radio to be designed around the better capabilities of the new ADC or could older radios be upgraded with the newer more capable ADC ? I just do not have a better background of the world of ADC and hence a good understanding of how advances can impact end users of the tech.
Photo of Tim - W4TME

Tim - W4TME, Customer Experience Manager

  • 9149 Posts
  • 3467 Reply Likes
This is a good question.  The leaps and bounds in ADC component performance is not as great as it once was as we are starting to approach the limits of physics is some cases.

Everyday there are advances in chip performance.  However, in order to make a significant engineering change to upgrade to a "better" ADC (greater bit depth or faster sampling rate), the change has to meet two criteria; is it significantly better to provide a tangible performance differentiator and is the component significantly cheaper so that the NRE costs can be recouped quickly with additional sales of the new product. 

Whether or not an upgraded part is "drop in" replaceable or not has a lot to do with the electrical characteristics between the two.  In rare cases, there is a minimal amount re-engineering that has to be done.  In most cases, the effort is significant.  However this is probably not want you want to do because along with one new part upgrade, there are probably additional parts and circuit upgrades that should be done at the same time, so it is actually more cost effective to do a complete redesign than retrofit an older piece of equipment. 

We designed the 6000s to have a long production life span and an even longer operational life span.  I don't think a 6000 is going to be relegated to being classified as a Neanderthal any time soon.
(Edited)
Photo of km9r.mike

km9r.mike

  • 425 Posts
  • 62 Reply Likes
Yes I did not think neanderthal anytime soon due to it's present ADC capabilities as well but simply did not have the knowledge base about ADC advancement to make a educated guess even. I assumed that there might be compounded issues with drop ins as well but there too that was just assuming on my part. The neanderthal was just a hyper exaggeration wrt to ADCs and was in no way an impression of the very thoughtful work that has gone into the making of this awesome rig. Thanks for the great reply and learning along the way : )
Photo of Barry N1EU

Barry N1EU

  • 495 Posts
  • 121 Reply Likes
Thanks Tim!  If ADC Overload was occurring, is there a visual indicator on the SmartSDR screen like there is in PowerSDR?
Photo of Tim - W4TME

Tim - W4TME, Customer Experience Manager

  • 9148 Posts
  • 3467 Reply Likes
There is not.  We have a pending feature request to add this indicator, but it has not been scheduled.

One thing I failed to mention is that if the aggregate signal level rises to the point of possibly damaging the ADC, there is a protection relay that will disconnect the RF signal path until the signal level drops.  I have actually heard this relay actuate when I accidentally transmitted 100W in the near field of the 6700s RX antenna.  It gets your attention rather quickly ;-) 
Photo of Ross - K9COX

Ross - K9COX

  • 335 Posts
  • 101 Reply Likes
Tim, Is that feature present is the 6500 and 6300?
Photo of Tim - W4TME

Tim - W4TME, Customer Experience Manager

  • 9148 Posts
  • 3467 Reply Likes
Yes, absolutely.  The ADC is one of the most expensive parts of the 6000 and we take extra measures to protect it from damage.
Photo of Steve N4LQ

Steve N4LQ

  • 568 Posts
  • 92 Reply Likes
Barry
Did the Orion and K3 impair each other? 
Photo of Barry N1EU

Barry N1EU

  • 495 Posts
  • 121 Reply Likes
No they didn't.  

To me it sounds like lack of antenna isolation plus lack of operator smarts to dial in attenuation exacerbated the Flex 6K filter-less wide-open DDC front end feeding the ADC.

Barry N1EU
Photo of Bill -VA3WTB

Bill -VA3WTB

  • 2771 Posts
  • 610 Reply Likes
Berry there were several Flex used this year in Field day, Maybe we will have some reports as to how they worked out for them.
Photo of Tim - W4TME

Tim - W4TME, Customer Experience Manager

  • 9148 Posts
  • 3467 Reply Likes
The 6000 has a very clean signal when it comes to phase noise.
Photo of Bill -VA3WTB

Bill -VA3WTB

  • 2771 Posts
  • 610 Reply Likes
Of coarse they would say no.... I wonder too if the Flex caused the other radios trouble, I would think not
(Edited)
Photo of Steve N4LQ

Steve N4LQ

  • 568 Posts
  • 92 Reply Likes
Of coarse they would say yes. 
Photo of Barry N1EU

Barry N1EU

  • 495 Posts
  • 121 Reply Likes
They said they didn't.
Photo of Steve N4LQ

Steve N4LQ

  • 568 Posts
  • 92 Reply Likes
So are we saying the 6000 needs hardware agc? I always wondered why it didn't have it. 
Photo of Steve N4LQ

Steve N4LQ

  • 568 Posts
  • 92 Reply Likes
So the answer is "yes" since this involved a 6500. I assume the preselectors enable automatically when the receiver is tuned to a ham band.
Photo of Tim - W4TME

Tim - W4TME, Customer Experience Manager

  • 9148 Posts
  • 3467 Reply Likes
Depends.  There are certain conditions that have to be met, like the frequency spread of the panadapter(s) associated with the SCU and the frequency of any slice defined in those panadapters must be within the frequency boundary of an established ham band (excluding 60m).  You can tell if the preselector is enabled by the absence of the WIDE indicator on the panadapter(s).
Photo of Joe Moffatt

Joe Moffatt

  • 11 Posts
  • 1 Reply Like
The preselector was on. Is the preselector before the ADC? Is it hardware?
Photo of Paul Christensen, W9AC

Paul Christensen, W9AC, Elmer

  • 318 Posts
  • 135 Reply Likes
For the 6500 and 6700, I think it comes down to the bandwidth of the preselector - at least for activities like Field Day operations where antennas are often haphazardly strung close to others.  Contest stations present a more controlled environment since they're mostly permanent installations.  Even for FD, I'm not so sure this is a problem with the Flex 6500 and 6700, but let's say signal levels do exceed +9 dBm. 

Here's an idea:

Flex could offer up a tracking preselector (TP) as either an internal or external option to the 6500 and 6700 models.  A TP is essentially a narrow bandpass filter whose center frequency changes in sync with the rig's operating frequency. When I owned an ADAT-200A, I added the optional TP package.  ADAT's standard preselector module popped out and the tracking module was added in a few minutes.  The TP could be bypassed in the menu. 

For the Flex, the TP would connect in the Rx Loop but that assumes my understanding of the loop placement is correct as I have not seen a schematic, nor detailed block diagram of either Flex model.

For an external TP, it could be a device that communicates to SSDR via the Flex 6500/6700 USB port or over Ethernet.  Probably few owners would need a TP as it would almost always be in bypass to have full pan, but as Elecraft has shown us time and again, we often buy accessories we *think* we need if only to have a warm and fuzzy feeling that we own the best.  I'm guilty of this, as are probably many others here.  I have no doubt that if Flex offered a TP, many would be sold. 

Paul, W9AC
(Edited)
Photo of Tim - W4TME

Tim - W4TME, Customer Experience Manager

  • 9148 Posts
  • 3467 Reply Likes
Is the preselector before the ADC? Is it hardware?

Joe - the answers are yes and yes.
Photo of Joe Moffatt

Joe Moffatt

  • 11 Posts
  • 1 Reply Like
I am the person who reported this on the Elecraft list.  I do not lack "operator smarts".   We dialed the attenuation WAY back to almost no RF Gain at all.  The K3 still affected it.  These were on two beams about 150 feet apart, 100 watts, beams pointed parallel to each other.    Now, according to Elecraft, the K3 has little phase noise.  According to Flex, it is has very little phase noise.  What is true?   I don't know.  I do know that K3s are famous for operating multi multi with little trouble inband.
Photo of Joe Moffatt

Joe Moffatt

  • 11 Posts
  • 1 Reply Like
Specifically, this was a 6500 I am referring to on 14.250 ish and the CW was on 14.030 ish
Photo of Joe Moffatt

Joe Moffatt

  • 11 Posts
  • 1 Reply Like
On top of that, this was never meant to be a bash Flex and totally praise Elecraft discussion.  I have both and both have their place.  It was meant to be a why did this happen, and how can we mitigate this in the future?    I had a lot of ops that came over to operate that are very interested in Flex purchases, and I want to be able to explain what our issue was with reasonable intelligence.
Photo of KY6LA - Howard

KY6LA - Howard, Elmer

  • 3541 Posts
  • 1396 Reply Likes
Joe

Still trying to diagnose your comments

k3 has lots of phase noise which is why they brought out K3S with the low phase noise board

Dialed back RF gain.?

Does not normally happen in 6000 as you can only insert -10DB attenuation
Photo of Barry N1EU

Barry N1EU

  • 495 Posts
  • 121 Reply Likes
Hopefully, nobody is perceived as bashing anybody and we're all about understanding why you guys were seeing what you were seeing.  Joe, if you do try the field exercise again, perhaps you should focus on whichever band had the worst interaction between the ssb and cw rigs and try all combinations of the 6500, 6300, K3, and Orion as cw and ssb rigs on that band and see how each rig affects the other.
(Edited)
Photo of Bill -VA3WTB

Bill -VA3WTB

  • 2771 Posts
  • 610 Reply Likes
Ya sounds like the K3 is a little dirty
Photo of Bill -VA3WTB

Bill -VA3WTB

  • 2771 Posts
  • 610 Reply Likes
It is very interesting, The flex has proven very good in close in transmitters, many use them in field day with no problems, a matter of fact out standing. So It is hard to know what their problems were.
Photo of KY6LA - Howard

KY6LA - Howard, Elmer

  • 3541 Posts
  • 1396 Reply Likes
Sounding more and more like a RFI issue.

I suspect that the Flex has more cables around eg Ethernet.

My best guess RF splatter from the K3 was getting into the cables going to the Flex which is why it was cutting out and the RF attenuators had no effect.
Photo of Joe Moffatt

Joe Moffatt

  • 11 Posts
  • 1 Reply Like
This is what I suspect too Howard!   I am going to retest this in the upcoming weeks when we can to see what is going on.  I don't necessarily think it is either radio.   This K3 has the new synth board in it as well.  
Photo of Bill -VA3WTB

Bill -VA3WTB

  • 2771 Posts
  • 610 Reply Likes
Joe it would be interesting, considering we already know how well the Flex works with close in signals, as they have been used in many field days so far. RF would explain what was going on.
Photo of Steve N4LQ

Steve N4LQ

  • 568 Posts
  • 92 Reply Likes
I noticed this:
With more than one slice open the preselector is disabled (WIDE). Obviously it has to be that way but good to keep in mind for FD operations.
---------------------------
This may be a bug..........
If you are on say .... 30 meters then open a slice for 6 meters which has +20 DB preselector engaged then change that same slice to 30 meters, both slices will then have +20 DB engaged. Then close the extra slice and +20 stays on. Doesn't seem right.
Photo of Tim - W4TME

Tim - W4TME, Customer Experience Manager

  • 9148 Posts
  • 3467 Reply Likes
Not a bug - this is by design.  All of the frequency changes you described are on one antenna (SCU) and the RF preamp is associated with the SCU, not the slice or the panadapter frequency spread.  The SCU (wide band digitizer) has no knowledge of the construct of a "band" this is defined from using a segment of the digitized spectrum. Therefore once you change the preamp gain for the SCU, it stays in that state regardless of the frequency of the slice or panadapter might be changed to. 

It isn't intuitive because you are probably used to a single narrow band (final IF) receiver radio that has a hard construct of what a band is and changing the preamp gain when a band (frequency segment) change is a monolithic event and doesn't adversely effect other receivers or panadapters since there are none.
Photo of Steve N4LQ

Steve N4LQ

  • 568 Posts
  • 92 Reply Likes
Well the preamp came on automatically when I opened the slice set for 6m so I figured it should also go off automatically when I close the same slice. 
Slice turns it on so slice turns it off. So I don't understand the intent. But that's ok...I can kill it manually. 
Anyway I've reread your explanation a few times yet no progress. Sorry.
Photo of Tim - W4TME

Tim - W4TME, Customer Experience Manager

  • 9148 Posts
  • 3467 Reply Likes
Band persistence turned it on, not the slice.  persistence is an on state operation, not an off state one.
Photo of Al / NN4ZZ

Al / NN4ZZ

  • 1706 Posts
  • 580 Reply Likes
It might be more informative to change the "WIDE" message to say "Preselector ON" or "Preselector OFF-WIDE"    Besides giving a positive indication of both conditions, it provides a more descriptive message.  

If anyone thinks this is worthwhile, I can repost it as an idea and see if it gets any votes.  Sometimes it's the little things that make a difference.   Thoughts?

Regards, Al / NN4ZZ  
al (at) nn4zz (dot) com


Photo of KY6LA - Howard

KY6LA - Howard, Elmer

  • 3541 Posts
  • 1396 Reply Likes
Here are some simplistic calculations. As I do not know ground conditions, relative locations of antennas or types of antennas. And their gains

Average Power at the Antenna 1500 watts

Antenna Gain in dBi 2.2 dBi. dipole

Distance to the Area of Interest 200 feet 60.96 metres
Frequency of Operation 14 MHz

Are Ground Reflections Calculated? Yes

Estimated RF Power Density 0.0137 mW/cm2. Or 137mW/m2

Conversion to dBm. +21.4 /m2

Even withe 10dB attenuation. Overload limit =. +9+19= 19dBm

Therefor ADC likely overloaded.
Photo of Tim - W4TME

Tim - W4TME, Customer Experience Manager

  • 9148 Posts
  • 3467 Reply Likes
We use a directional RX antenna with 30 dB null we align towards the transmission source. so add 20 more dB of attenuation.  We are close to the overload point but not there.
Photo of Barry N1EU

Barry N1EU

  • 495 Posts
  • 121 Reply Likes
IMHO, adding the ADC Overload indicator should be moved up in priority.  Seems like a no-brainer but it obviously isn't.
Photo of Stan - VA7NF

Stan - VA7NF

  • 416 Posts
  • 92 Reply Likes

If this were FD then the power would be 150W max, or there is a letter to write.  On our FD (VE7SAR) we calculated using the side pattern of two tri-banders with 150' separation and the numbers came close to the limit.  This year we ran QRP - No problems with two on one band.

Photo of Lee

Lee, Elmer

  • 677 Posts
  • 285 Reply Likes
You can't "calculate a beam pattern" 150' from a yagi antenna on HF  you need to get a dozen wavelengths away from the beam for a pattern to form.  Basically at 150ft on 20m you are in the near field.  

My experience with a 6300 attached to a 43 ft vertical matched matched at the vertical base for RX, and a 1/2 wave 40M vertical 175ft away for TX is I could operate a CW signal up to about 800-900w on the TX vert before I had any interference in the 6300 on 40M with about 200 khz separation in freq.  Transmit phase noise was limiting up to +9 dBm where the ADC began to show signs of deterioration.  Note the 6300 has no pre-selector   The transmitter was a TT Paragon into a Alpha 78.

The question I was asking was how much power does it take to crunch the ADC and are there any HF or MF or BC signals at my QTH that will cause the ADC to overload given my antennas.  The answer is no there are no signals off my property strong enough to crunch my 6300 

73 W9OY
(Edited)
Photo of KY6LA - Howard

KY6LA - Howard, Elmer

  • 3541 Posts
  • 1396 Reply Likes
With 150w instead of 1500W you would get +11.4 dBm/m2 from a dipole within 200'. Too much for barefoot +9dBm but within limits for -10dB attenuation.

Of course these are only theoretical calcs Local conditions will dominate.