K3: Peculiar BW behavior?

[Walter V. Gilles]

Greetings,
 
I just came across what seems to be a rather peculiar operational quirk in my K3/100 while reducing the BW on CW.  I have 13, 2.8, 1.8 and 0.4 roofing, running current firmware release suite.
 
On 20M CW for example, as I reduced the DSP BW down, and as I hit 0.20 the signal dropped significantly; at 0.15 it completely cut out; at 0.10 it popped in again at a normal volume but shifted enough in freq such that it took a goodly RIT input to find signal pitch again, but stayed there at 0.05.  Going back up in BW, it just reversed the process, going out at 0.15 and coming back in at 0.20 weak and shifted back in freq, then back to normal on up in BW.  This behavior seems mostly to show up on 20M, and fairly repeatable.  Occassioannly it will show up on other bands, but not as severe.
 
Also, on all bands, I hear a distinct click as the BW hits 0.10 going down, then another click as it hits 0.15 on the way up.  No clicking sounds anywhere else on BW changes (i.e. no clicks as the roofing filters switch in/out). Clicks can be heard in the K3 speaker, and headphones.  This clicking is repeatable on all bands.
 
The clicking is mildly annoying but new, and the BW behavior begs some words of wisdom if anybody has some ideas.  I have not reloaded the firmware yet, as that would be my next step to see if something got corrupted.  Note that this peculiar behavior was not present after I loaded the latest firmware - that went without a hitch as usual.
 
Soooo.....anybody have any ideas?  Thanks.  73
 
Walter   WB2IDK


     
______________________________________________________________
Elecraft mailing list
Home: http://mailman.qth.net/mailman/listinfo/elecraft
Help: http://mailman.qth.net/mmfaq.htm
Post: mailto:[hidden email]

This list hosted by: http://www.qsl.net
Please help support this email list: http://www.qsl.net/donate.html

[pd0psb]

Hi Walter,

A few things to look at:
In CW mode check how you've configured your roofing flters.
Could it be you switch to the 0.4 roofer at BW 0.2 ? >> CONFIG: FLx BW
Could it be the 0.4 roofer needs some more filter gain ? >> CONFIG: FLx GN

The "click" you hear at 0.1 is caused by the FIR-IIR filter switch-over.
Under 0.1 BW K3 switches over to Infinite Impulse Response to prevent ringing.
NB: This switch-in of the IIR filter also bypasses the RX EQ.

So if you have RX EQ attenuating the CW freq ranges, you'll notice a gain-up when narrowing under 0.1 BW. You can check this setting RX EQ flat.
Note that EQ will also be flat for all other modes then.
It's also possible to disable IIR switch over CONFIG: FLx BW>tap 7>IIR on/off

I think you'll be able to solve by some reconfiguring.
The fact that you notice a difference on some bands will probably be caused by a varying noisefloor throughout different bands.

73'
Paul
PD0PSB

Walter V. Gilles wrote

Greetings,
 
I just came across what seems to be a rather peculiar operational quirk in my K3/100 while reducing the BW on CW.  I have 13, 2.8, 1.8 and 0.4 roofing, running current firmware release suite.
 
On 20M CW for example, as I reduced the DSP BW down, and as I hit 0.20 the signal dropped significantly; at 0.15 it completely cut out; at 0.10 it popped in again at a normal volume but shifted enough in freq such that it took a goodly RIT input to find signal pitch again, but stayed there at 0.05.  Going back up in BW, it just reversed the process, going out at 0.15 and coming back in at 0.20 weak and shifted back in freq, then back to normal on up in BW.  This behavior seems mostly to show up on 20M, and fairly repeatable.  Occassioannly it will show up on other bands, but not as severe.
 
Also, on all bands, I hear a distinct click as the BW hits 0.10 going down, then another click as it hits 0.15 on the way up.  No clicking sounds anywhere else on BW changes (i.e. no clicks as the roofing filters switch in/out). Clicks can be heard in the K3 speaker, and headphones.  This clicking is repeatable on all bands.
 
The clicking is mildly annoying but new, and the BW behavior begs some words of wisdom if anybody has some ideas.  I have not reloaded the firmware yet, as that would be my next step to see if something got corrupted.  Note that this peculiar behavior was not present after I loaded the latest firmware - that went without a hitch as usual.
 
Soooo.....anybody have any ideas?  Thanks.  73
 
Walter   WB2IDK


     
______________________________________________________________
Elecraft mailing list
Home: http://mailman.qth.net/mailman/listinfo/elecraft
Help: http://mailman.qth.net/mmfaq.htm
Post: mailto:[e-mail]

This list hosted by: http://www.qsl.net
Please help support this email list: http://www.qsl.net/donate.html

[w4nhj]

Walter,

This is probably not your problem, but might be worth checking.  Make sure
you have the filter BWs entered correctly.  I had a problem, I entered the
250 Hz filter as 250 and is should have been .250.  I got some crazy
results.  KG4YLZ who was here when I installed the filters caught my
mistake.

Frank - W4NHJ
----- Original Message -----
From: "Walter V. Gilles" <[hidden email]>
To: <[hidden email]>
Sent: Thursday, June 11, 2009 12:49 AM
Subject: [Elecraft] K3: Peculiar BW behavior?


Greetings,

I just came across what seems to be a rather peculiar operational quirk in
my K3/100 while reducing the BW on CW. I have 13, 2.8, 1.8 and 0.4 roofing,
running current firmware release suite.

On 20M CW for example, as I reduced the DSP BW down, and as I hit 0.20 the
signal dropped significantly; at 0.15 it completely cut out; at 0.10 it
popped in again at a normal volume but shifted enough in freq such that it
took a goodly RIT input to find signal pitch again, but stayed there at
0.05. Going back up in BW, it just reversed the process, going out at 0.15
and coming back in at 0.20 weak and shifted back in freq, then back to
normal on up in BW. This behavior seems mostly to show up on 20M, and fairly
repeatable. Occassioannly it will show up on other bands, but not as severe.

Also, on all bands, I hear a distinct click as the BW hits 0.10 going down,
then another click as it hits 0.15 on the way up. No clicking sounds
anywhere else on BW changes (i.e. no clicks as the roofing filters switch
in/out). Clicks can be heard in the K3 speaker, and headphones. This
clicking is repeatable on all bands.

The clicking is mildly annoying but new, and the BW behavior begs some words
of wisdom if anybody has some ideas. I have not reloaded the firmware yet,
as that would be my next step to see if something got corrupted. Note that
this peculiar behavior was not present after I loaded the latest firmware -
that went without a hitch as usual.

Soooo.....anybody have any ideas? Thanks. 73

Walter WB2IDK



______________________________________________________________
Elecraft mailing list
Home: http://mailman.qth.net/mailman/listinfo/elecraft
Help: http://mailman.qth.net/mmfaq.htm
Post: mailto:[hidden email]

This list hosted by: http://www.qsl.net
Please help support this email list: http://www.qsl.net/donate.html 

______________________________________________________________
Elecraft mailing list
Home: http://mailman.qth.net/mailman/listinfo/elecraft
Help: http://mailman.qth.net/mmfaq.htm
Post: mailto:[hidden email]

This list hosted by: http://www.qsl.net
Please help support this email list: http://www.qsl.net/donate.html

[Jon Kåre Hellan]

Frank Ross W4NHJ wrote:
> Walter,
>
> This is probably not your problem, but might be worth checking.  Make sure
> you have the filter BWs entered correctly.  I had a problem, I entered the
> 250 Hz filter as 250 and is should have been .250.  I got some crazy
> results.  KG4YLZ who was here when I installed the filters caught my
> mistake.

and if it's a five pole filter, the offset has to be right.

Jon LA4RT
______________________________________________________________
Elecraft mailing list
Home: http://mailman.qth.net/mailman/listinfo/elecraft
Help: http://mailman.qth.net/mmfaq.htm
Post: mailto:[hidden email]

This list hosted by: http://www.qsl.net
Please help support this email list: http://www.qsl.net/donate.html

[Walter V. Gilles]

Thanks to all who responded.  The problem appears to be solved now.

I used the utility to check things out before making any changes, and all was as it should be.  However I reloaded the current release firmware, and increased the insertion loss compensation on the 0.40 roofing filter from 1db to 2db.  I don’t use any other compensation.  Upon checking on air, the freq shift and fadeout experienced previously on 20M at 0.20 BW is gone, and a quick check on a few other bands indicates that problem is fixed.  Now on to that “clicking” noise.

The clicking seems a bit reduced, or I think it is anyway, which is really all that matters.  I have heard that the human brain is the best filter one can have, and I believe that clicking noise has always been there, as suggested by some of the responses (posted & e-mailed).  I believe I (my brain) was discriminating (selectively filtering) the non-information (QRM, band noise, crashes, pops etc.) from the information (CW note) before.  Then when I was just mentally idling, working with the controls (NB, NR, BW, etc.) to get familiar with specific settings for different band conditions, I probably shifted attention away from the CW note and onto the background – hence discovered that “hidden” clicking.  Anyway, the K3 appears to be a happy camper again, and I learned a few things in the process.  Thanks.  73

Walter  WB2IDK

[Bill W4ZV]

Walter V. Gilles wrote

I used the utility to check things out before making any changes, and all was as it should be.  However I reloaded the current release firmware, and increased the insertion loss compensation on the 0.40 roofing filter from 1db to 2db.  I don’t use any other compensation.  Upon checking on air, the freq shift and fadeout experienced previously on 20M at 0.20 BW is gone, and a quick check on a few other bands indicates that problem is fixed.

A 2 dB gain setting for a 400 Hz filter sounds far too low.  I have two 500 Hz 8-pole filters in my base rig and one 500 Hz 5-pole in my portable rig.  The 8-poles are 6 dB and the 5-pole is 5 dB.   I would expect more like 6-7 dB for the 400 Hz 8-pole based on the Inrad/Elecraft filter plot below :

http://www.inrad.net/product.php?productid=150&cat=140&page=1

You can check your gain setting by applying a steady carrier, zero beat with your desired PITCH and gradually turn WIDTH from the widest setting to the set point where the 400 Hz engages.  There should be no change as you transition across the 400 Hz set point.  If you want to do this more accurately you can use the internal AFV/dBV meter (remember AGC must be OFF) to measure changes of 0.1 dB.  You should adjust CONFIG: FLx Gain so that the transition is smoothest across the set point WIDTH.

If you have other filters besides the 400 Hz, you should start with your widest filter, set that one to 0 dB gain and then adjust each filter, from widest to narrowest in order so that the transition is smooth across the set points.

It's been awhile since I looked at the manual and I see page 46 says:

"Use VFO A to set the gain in dB. In general,
you’ll want to add 1-2 dB for 400-500 Hz
filters, and 3-4 dB for 200-250 Hz filters."

In my experience (with several different 500 Hz and 200 Hz filters) the recommended settings are about 4 dB too low.  They also do not agree with published insertion loss figures by Inrad.  I would expect 5-6 dB is correct for 400-500 Hz filters and 7-8 dB for 200-250 Hz filters.

Someone should write up a better description for how to accurately set the filter gain settings using the internal AFV/dBV meter...hint hint to Ron AC7AC.  ;-)

73,  Bill  W4ZV