Splatter TX on 6500

  • 2
  • Problem
  • Updated 3 years ago
  • Solved


Looking for ideas about what could be causing such a bad transmit signal?  I got a complaint from another ham that I was splattering while using DAX and WSJT-X, the SmartSDR display was covered up so I didn't even notice it.  This is what my signal looks like no matter whether DAX is running or not.  This is a screen shot from a TUNE from inside SmartSDR.  I get the same if I manually key, if I transmit from WSJT-X, etc.
Photo of Keith Schlottman

Keith Schlottman

  • 16 Posts
  • 2 Reply Likes

Posted 3 years ago

  • 2
Photo of Jim Jerzycke

Jim Jerzycke

  • 93 Posts
  • 19 Reply Likes
How far apart are the individual spikes?

Is it like this on all bands?

What kind of power supply are you using?

73, Jim
Photo of Andrew Russell

Andrew Russell

  • 276 Posts
  • 43 Reply Likes
See DAX distortion on past posts.
Win 10 fixed it for me.
Andrew de VK5CV
Photo of K7NXT

K7NXT

  • 37 Posts
  • 27 Reply Likes
Do you mean Windows 10?  Very interesting.  Would like to know more about what in the OS would cause this.  73s.
Photo of Andrew Russell

Andrew Russell

  • 276 Posts
  • 43 Reply Likes
I worked with Tim on chasing down DPCs in Windows 8.1as I had with PowerSDR on the 1500 on Win7. Still DAX would go feral after a while. For instance WSPR had multiple spots at 50 and 100cycles on strong signals similar to when a psu filter cap went bad.
sfc/ scannow reported corrupt system files that could not be repaired.
instead of reloading or renewing Win 8.1 I chose the Win 10 upgrade.
Since the its been solid. Andrew VK5CV.
Photo of Al K0VM

Al K0VM, Elmer

  • 607 Posts
  • 102 Reply Likes
The spacing of the spikes (~ 60/120hz) would indicate that there may be a power supply regulation problem.  Is the power supply dropping out of regulation under load ?

AL, K0VM
Photo of WA2SQQ

WA2SQQ

  • 475 Posts
  • 122 Reply Likes
I really don't think the OS has anything to do with this as your PC is not doing any of the signal processing. The power supply theory seems feasible, and easy enough to check. I'd say, wait for Tim or someone from FRS to read this and see what he says. Does it do this on all bands? Just wondering if its RF getting into your power supply, causing some weird situation. How about trying a dummy load to rule out RF problem or antenna related cause?
Photo of Tim - W4TME

Tim - W4TME, Customer Experience Manager

  • 9186 Posts
  • 3539 Reply Likes
I recommend getting a little more information, because based on your problem description, you may have one issue that is not related to what you are seeing on the panadapter when you are transmitting.

What type of load are you transmitting into?  An antenna or a dummy load?  If it is an antenna, it will present the transmitter with a complex impedance and probably common mode currents that could account for that you are seeing.  If you are not using a dummy load, try that and repost your picture using the same scale (I need to see the peak signal strength of the carrier to determine dBc)

Al made a good point about the power supply regulation.  The 6500 transmitter needs a solid 13.8 VDC voltage source under load.

If you are getting distortion when transmitting on digital modes only, then I suspect you are experiencing DAX signal corruption due to long duration DPCs.  To verify this assumption, when you have digital mode corruption, close and restart the DAX control panel when reinitializes the driver and if the problem stops temporarily, then this is the issue and we'll need to help address your PC's DPC issues.
Photo of Bill W2PKY

Bill W2PKY

  • 501 Posts
  • 84 Reply Likes
I noticed the screen shot shows mode CW not DIGU. Could that be the issue?
Photo of George Molnar, KF2T

George Molnar, KF2T, Elmer

  • 1650 Posts
  • 600 Reply Likes
If restarting DAX doesn't clear the issue, I suspect RF getting back into things. Is the display different from band to band, or on dummy load? Regarding "splatter" on JT, more than a few people still can't figure how to set up their receivers properly, and that can make a clean signal look dirty on the WSJT-X waterfall. If your levels are good (never run the audio level into the red!), you are probably good once any RF intrusions are dealt with.
Photo of Rory - N6OIL

Rory - N6OIL

  • 301 Posts
  • 55 Reply Likes
Keith I sure hope that's only a test freq you're on for JT65. 
Photo of Keith Schlottman

Keith Schlottman

  • 16 Posts
  • 2 Reply Likes
Not sure what you mean Rory.  As noted in my post the screen shot was from a tune function in SmartSDR, which is a CW tx.
Photo of Keith Schlottman

Keith Schlottman

  • 16 Posts
  • 2 Reply Likes
Thanks guys for all the suggestions!  As I said in the post this happens whether DAX is running or not, so I don't think it has anything to do with DAX issues.  I'll work on testing out the power supply possibility.  I'm currently using a switching ps (Alinco DM-330) after I had problems with an Astron.  I'll check it out with a dummy load but won't be able to do that until next week.  As an aside, during my own testing of this problem, I played around with narrowing the tx bandwidth (using the lo-cut and hi-cut settings in SmartSDR).  It's a challenge to change them quickly enough when using JT and replying to a CQ, but it does seem to help to narrow it down to a tight bandwidth centered on the appropriate signal.  If there was some way to automate that, might be really cool - i.e. on digital modes to use a narrow tx width that matches the mode?
Photo of George Molnar, KF2T

George Molnar, KF2T, Elmer

  • 1643 Posts
  • 600 Reply Likes
Not so much an automated process, but you can adjust the digital mode (DIGU, etc) transmitted bandwidth that will stay once saved.

One thing worth noting, is that the transmit display is (unfortunately) not always a faithful indicator of actual emissions. You can take some solace, though in that it will almost always look WORSE than real life, and your "spurs" etc., are at least 60 dB down.

Good luck!
Photo of Keith Schlottman

Keith Schlottman

  • 16 Posts
  • 2 Reply Likes
Good points George.
What I was thinking of was mainly based on how WSJT-X works.  Say I tune the rig to 14.076 and rx a 5kHz bandwidth.  If I see a JT9 signal up around say 3500 that I want to QSO with, the current advice per forum posts seems to be to widen the hi-cut tx setting to 4-5000, otherwise the default usb settings which cut off at 2700 won't actually transmit any power at 3500.  But, the JT9 signal is extremely narrow so wouldn't it be nice to just use filters to produce a very narrow signal.  I can do that (assuming what SmartSDR shows me is reasonably true), but to do so I need to set the lo and hi cutoffs centered around my 3500 JT9 signal.   I think that currently, the only way to do that is to manually input those numbers?  And since each QSO is on a slightly different frequency, it becomes a cumbersome process to keep changing those settings, especially when you have less than 12 seconds to do so. If there was an automatic way to set a very narrow digital mode tx bandwidth that followed the desired tx shifted frequency that would be much better.  (As an aside, it seems that it is possible to change the settings even during transmit, but not sure if that's advisable).
Photo of George Molnar, KF2T

George Molnar, KF2T, Elmer

  • 1643 Posts
  • 600 Reply Likes
I see your point, Keith. If you have the levels set from WSJT-X properly, your Flex will be very clean. I run with WSJT-X also, using a 5 kHz wide RX and TX filter for DIGU, and have absolutely no problems. Just keep in the DIGU mode; running USB in that style will probably win you no friends :-)
Photo of Ken - NM9P

Ken - NM9P, Elmer

  • 4172 Posts
  • 1331 Reply Likes
Keith,
One thing that helps is to understand how WSJT-X, FLDigi and other programs generate their signals.  The Transmitted signals are made in the AUDIO spectrum and sent via sound card interface with traditional rigs and transmitted using SSB (Usually USB). The same is true for FLEX rigs, except that the audio is sent via a VIRTUAL Sound Card (DAX) and remains in digitized audio internally until it becomes RF via either the SSB (NOT Recommended) or DIGU (Recommended) mode on the 6000 rigs.

The "Audio" signal, since it is never converted to actual audio and then back up to RF as in a traditional SSB rig is VERY Clean, as long as you are not driving the level too high from the DAX TX slider in the DAX Control Panel.  You should not need any additional TX filtering to keep it clean.  Just keep the "Mic" level out of the RED.  Usually a notch or two below the red is OK.

The "Audio" signals generated by WSJT-X are still limited by the bandwidth setting in the Mic Profile.  Signals at the far extreme right of WSJT-X (Usually JT-9 signals) can be as high as 3-5 KHz in "audio frequency" because they are 3-5 KHz above the base frequency to which the rig is tuned.  (Zero beat or carrier frequency on a SSB rig). 

If your TX Filters, which are set in the Mic Profile, are too narrow, the signals will not be passed, in the same way that you will not receive them if your receiver filter is set too narrow.

You can set a Mic Profile and name it "WSJT" with the TX Filters set to 25 Hz low and 5 KHz high, and with the DAX Button engaged. After you save it, shift to a different mode (LSB for example) and you will notice that the Mic profile changes to whatever Mic Profile you used for LSB.  When you change back to DIGU, it should remember your WSJT profile.  If not, manually select it again and it should remember the next time. 

Mic Profiles are persistent according to Mode.  So you can have different profiles for SSB, DIGU, DIGL, RTTY Mode, AM, FM, FDV, DSTAR, etc.  Handy, isn't it?

BTW... One additional source of some distortion can be a mismatch of the sampling frequencies in the setup files of WSJT-X (or other Digi programs) and the settings of the DAX sampling rates set in the Windows Record/Playback properties for each DAX channel. 

Check the properties in the DAX TX and RESERVED DAX TX virtual cables in the Windows sound application.  They should be set to 48,000 bps with 16 bit Stereo.  The past several versions of SSDR set these correctly automatically.  Before the rewrite of DAC & CAT (I think in v.1.5?) I often had to set them all up manually after a new install)   

Then check the sampling rate in your configuration area of your digital program.  It should be 48,000 as well, or at least a number that is easily divided into 48,000 - for example, 16,000 or 12,000 or 8,000 bps.  This will allow any internal digital conversion of sampling rate to be done cleanly without artifacts that may contribute to distortion.

If none of this answers your questions, you may indeed have some RF or ground loop problems, or other problems that a "help desk" with FRS should be able to clear up.

Good Luck,

Ken - NM9P
 
Photo of Ken - NM9P

Ken - NM9P, Elmer

  • 4172 Posts
  • 1331 Reply Likes
BTW, the reason you should NEVER use USB or LSB for Digi modes on a 6000 rig is that they may still have EQ or Processing activated, which will distort your nice, clean digital signals to the point of making enemies in the digital bands.  DIGU & DIGL take processing and TX EQ out of the line.

Always use DIGU or DIGL for most Digital modes.  Use RTTY Mode for best performance in RTTY, unless you want to use something like FLDigi's multi-frequency rtty receive.
(Edited)
Photo of Keith Schlottman

Keith Schlottman

  • 16 Posts
  • 2 Reply Likes
Thanks Ken for the detailed info - greatly appreciate it!  Yes I should have clarified, I only use DIGU on digital modes.  But now I understand a lot better why!
Photo of George Molnar, KF2T

George Molnar, KF2T, Elmer

  • 1650 Posts
  • 600 Reply Likes


Here's a screenshot of my 6500 to a dummy load.Tried to replicate your configuration best I could.
Photo of Keith Schlottman

Keith Schlottman

  • 16 Posts
  • 2 Reply Likes
Following up after I've done a few tests based on these comments (thanks again guys!)
I tried two different power supplies, and then direct to a battery and got the same results, so it doesn't appear to be power supply related.
Transmitting into a dummy load gave me a nice clean signal, similar to George's above - so it does appear to be antenna-related.  I have the same problem on two different antennas (one dipole, one magnetic loop), not exactly the same strength but the same spacing in the spikes.  I'm currently in an HOA which has a no-antenna policy, so both antennas are inside and fairly close to the rest of the station components.
I tried moving the antennas about 10 feet farther away from the rig, and the spikes dropped significantly.  So I'm going to try to come up with a way to better separate the station and antennas within the constraints of the home.  I'm planning to move the rig to another location in the fall but I've been enjoying the challenges of operating from a less-than-ideal space too.
I had basic choke baluns on the feedline for both antennas but after reading Tim's suggestion that common mode currents could be the culprit, I studied up on them and it looks like I need to do a bit more work on that.
I wonder if the antennas are overloading something in the computer system, which is leaking back in to the rig.  I'll put some more ferrite beads on various wires.
And I think I might build the counterpoise tuner from the June QST to see if that will help a bit.

One of the other things I want to do is hook up a scope to analyze the signal using an external device.  I'm not clear about the TX signal display on the SmartSDR waterfall - where is the data for this originating?  It would appear that it's not just displaying what you have set the rig to transmit, but rather there may be an actual module within the Flex that is sampling TX output?  How accurate should we expect the TX display to be?
Photo of Al K0VM

Al K0VM, Elmer

  • 607 Posts
  • 102 Reply Likes
Keith,
   The spectrum shown during transmit is infact the leakage path trough the TR switching.from the transmit feed line into the still active receiver.  It is a fairly good representation of RF going to the antenna. 
   As an experiment, you could put the radio into FDX mode and connecct a short antenna to the RXA input ( select RXA for receive)..  See the uses guide for limitiations of FDX signal levels.
   Thinking outside the box... I  suppose it might be possible for 60hz AC picked up  on the antenna to intermode with the RF in the transmitter... A DC path (RF choke) from center conductor to coax shield should help that  situation.

AL, K0VM
Photo of George Molnar, KF2T

George Molnar, KF2T, Elmer

  • 1650 Posts
  • 600 Reply Likes

Hi Al - do you know if the SCU receive is terminated or left open during transmit? Imagine that might change the scope representation some.

See you on the radio!

Photo of Al K0VM

Al K0VM, Elmer

  • 607 Posts
  • 102 Reply Likes
I don't know for a fact but suspect it is open unless an alternate RX antenna is selected..

AL, K0VM
Photo of Tim - W4TME

Tim - W4TME, Customer Experience Manager

  • 9186 Posts
  • 3539 Reply Likes
I talked to Keith on the phone yesterday.  It appears his issue is excessive common mode currents due to poor RF grounding for the return currents (lack of a counterpoise).  This is causing RF to get back into the radio on transmit.  He is going to try a few 1/4 wavelength counterpoises connected to the radio's grounding post and adding a 1:1 balun (commom mode choke).

I suspect what we are seeing in the spectrum is RF taking a path to ground through the power supply and then to the AC safety ground / neutral bond.
Photo of WA2SQQ

WA2SQQ

  • 475 Posts
  • 122 Reply Likes
Tim, how about one of those "artificial ground" tuners. I think MFJ offers one and there was also one featured in a recent issue of QST. Basically an antenna tuner for your ground cnnection
Photo of Tim - W4TME

Tim - W4TME, Customer Experience Manager

  • 9186 Posts
  • 3539 Reply Likes
I made one of those catch fire, so I can't recommend it. ;-)
Photo of km9r.mike

km9r.mike

  • 425 Posts
  • 62 Reply Likes
lol
Photo of Keith Schlottman

Keith Schlottman

  • 16 Posts
  • 2 Reply Likes
Official Response
Here's an update.  I think Tim did correctly identify the primary culprit as being excessive common mode currents.
I've added some ~1/4 wavelength counterpoise wires to the ground post on the back of the rig, and also added choke baluns in line with the antenna on the transceiver end.
This seems to help a lot - where the original spikes were as high as -70dB on the vertical scale in SmartSDR, they are now down around -90 to -95, which is fairly close to what I see driving into a dummy load.
I'm enjoying the process of trying to refine it to the best possible result so still have more tests to run.  But the majority of the problem seems to be solved at this point and hopefully my signals are a bit less obnoxious to others on the bands.
By the way I never thought this was possibly an issue with the Flex itself, which is why I posted it to the forum instead of as a support ticket.  I sure appreciate the community support!
Photo of Greg

Greg

  • 656 Posts
  • 63 Reply Likes

What kind of common mode chokes are you using?  take a look at this presentation from K9YC.  Great explanations on common modes and easy to build chokes that far out-perform store bought chokes.

http://audiosystemsgroup.com/NCDXACoaxChokesPPT.pdf


Greg


Photo of Greg

Greg

  • 656 Posts
  • 63 Reply Likes

Here's a source if you want to build your own....

http://www.kf7p.com/KF7P/Ferrite_chokes.html


This conversation is no longer open for comments or replies.