Re: K3 / RF gain - S meter issue - Similar issue

[Elecraft mailing list]

I have a similar - but different! - issue on one of my Elecraft's.

I own both a K3 and a K3s.  Both are configured equally with the K3
having been upgraded to the highest level to achieve near K3s equivalency.

My K3s RF Gain control works exactly as expected, i.e. as the control is
rotated counter-clockwise the "S-Meter" increases to a point at almost
full scale.

On the K3, however, as the RF Gain control is rotated counter-clockwise
and the "S-Meter" increases to about the "40-over" point at the
control's 10 o'clock position the "S-Meter" then inexplicably drops to
the "S9" position where it remains for the remainder of the RF gain
control counter-clockwise rotation.  This shouldn't be but like N1LN's
issue, the radio does otherwise perform correctly.  I have done
everything that N1LN has done including the RF Gain calibration using a
signal generator.

Very strange, indeed.

Dan -- N3ND

*************************************

Date: Wed, 13 Dec 2017 16:25:20 -0500
From: "Bruce Meier"<[hidden email]>
To:<[hidden email]>
Subject: [Elecraft] K3 / RF gain - S meter issue


Looking for suggestions here from someone that might have experienced this
same issue - -

 

I have (2) K3 units (not the S model) that are acting different when the RF
gain adjustments are turned fully counter clockwise.

?  Unit one ? s/n 1062: When the RF gain is decreased to full counter
clockwise position the S-Meter registers full scale (60db over S9) as
expected.

?  Unit two ? s/n 1193:  When the RF gain is decreased to full counter
clockwise position the S-Meter registers only S8.

  Facts:

?  With the exception of the S-Meter readings with RF gain settings, both
units are working fine

?  Both units have ?almost? all mods installed.  If not all, they are at
least both the same.

?  Both units are running the same version of microcode (5.54) (2.88) (2.88)
(1.26)

?  Both units seem to register very close actual RF gain of stations when
connected when tested on the same antenna

?  Both units have the new synthesizers.

?  Both units have second RX

?  In both cases (both units) I can hear the audio level decreasing linearly
as expected as the RF gain is turned counter clockwise

I have tried to initialize the K3 and reload the firmware resulting in no
change to the RF gain / S-Meter readings.

 

Elecraft Tech Support suggested:

?  Calibrating the s-meter ? did nothing

?  Next suggestion ? send the rig to California.

 

Anyone have this issue and fixed it??

 

73,

Bruce N1LN

 

______________________________________________________________
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
Message delivered to [hidden email]

[Bob McGraw - K4TAX]

My concern would be "does the S meter indicate correctly?"

Thus with 50 uV applied to the antenna input it should read S-9, and
with 0.78 uV applied it should read S-3, and with 5000 uV applied it
should read S-9 +40 dB.    I've checked my K3S at these and other values
and I'm confident that the S meter is accurate. Hence, I'm confident I
can give accurate signal reports.  Thus unlike many radios and their S
meters, they are only good for a rather unscientific and largely
inaccurate signal report, somewhat like a SWAG.

73

Bob, K4TAX


On 12/14/2017 10:16 AM, Dan Atchison via Elecraft wrote:

> I have a similar - but different! - issue on one of my Elecraft's.
>
> I own both a K3 and a K3s.  Both are configured equally with the K3
> having been upgraded to the highest level to achieve near K3s
> equivalency.
>
> My K3s RF Gain control works exactly as expected, i.e. as the control
> is rotated counter-clockwise the "S-Meter" increases to a point at
> almost full scale.
>
> On the K3, however, as the RF Gain control is rotated
> counter-clockwise and the "S-Meter" increases to about the "40-over"
> point at the control's 10 o'clock position the "S-Meter" then
> inexplicably drops to the "S9" position where it remains for the
> remainder of the RF gain control counter-clockwise rotation. This
> shouldn't be but like N1LN's issue, the radio does otherwise perform
> correctly.  I have done everything that N1LN has done including the RF
> Gain calibration using a signal generator.
>
> Very strange, indeed.
>
> Dan -- N3ND
>
> *************************************
>
> Date: Wed, 13 Dec 2017 16:25:20 -0500
> From: "Bruce Meier"<[hidden email]>
> To:<[hidden email]>
> Subject: [Elecraft] K3 / RF gain - S meter issue
>
>
> Looking for suggestions here from someone that might have experienced
> this
> same issue - -
>
>
>
> I have (2) K3 units (not the S model) that are acting different when
> the RF
> gain adjustments are turned fully counter clockwise.
>
> ?  Unit one ? s/n 1062: When the RF gain is decreased to full counter
> clockwise position the S-Meter registers full scale (60db over S9) as
> expected.
>
> ?  Unit two ? s/n 1193:  When the RF gain is decreased to full counter
> clockwise position the S-Meter registers only S8.
>
>  Facts:
>
> ?  With the exception of the S-Meter readings with RF gain settings, both
> units are working fine
>
> ?  Both units have ?almost? all mods installed.  If not all, they are at
> least both the same.
>
> ?  Both units are running the same version of microcode (5.54) (2.88)
> (2.88)
> (1.26)
>
> ?  Both units seem to register very close actual RF gain of stations when
> connected when tested on the same antenna
>
> ?  Both units have the new synthesizers.
>
> ?  Both units have second RX
>
> ?  In both cases (both units) I can hear the audio level decreasing
> linearly
> as expected as the RF gain is turned counter clockwise
>
> I have tried to initialize the K3 and reload the firmware resulting in no
> change to the RF gain / S-Meter readings.
>
>
>
> Elecraft Tech Support suggested:
>
> ?  Calibrating the s-meter ? did nothing
>
> ?  Next suggestion ? send the rig to California.
>
>
>
> Anyone have this issue and fixed it??
>
>
>
> 73,
>
> Bruce N1LN
>
>
>
> ______________________________________________________________
> 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
> Message delivered to [hidden email]

______________________________________________________________
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
Message delivered to [hidden email]

[Wes Stewart-2]

Personally, I wouldn't go so far as to say the radio performs correctly. Without
writing a long history, let's just say that I discovered this RF gain issue 8 or
9 years ago.  I demonstrated it to an Elecraft engineer in my shack who had
brought along his fully up to date K3 which we compared.  They were widely
different.  Subsequently, I did a lot of beta testing of firmware that I believe
spawned the RF gain calibration program.

In my opinion based on inspection of the schematic, the issue resides in the fet
post-filter amplifier that is controlled by the "hardware" AGC (and manual "RF"
gain control)  The specs on this device are very loose, IIRC, Gm can vary 2:1. 
Thus the response of this fet to AGC/RF gain varies from radio to radio, and as
seen in another post, between the main RX and second RX in the same radio.  As I
understand it, and I would love for someone at Elecraft to correct me if I'm
wrong, the RF gain calibration program measures the radio gain control
characteristics and generates a lookup table of correction factors to apply to
linearize the control.

If I'm correct, I always been troubled by questions about the speed of this
process.  Being an analog guy I understand resistor-diode networks used for
similar purposes that operate essentially instantaneously.   The digital process
not so much.

Wes  N7WS



On 12/14/2017 9:16 AM, Dan Atchison via Elecraft wrote:

> I have a similar - but different! - issue on one of my Elecraft's.
>
> I own both a K3 and a K3s.  Both are configured equally with the K3 having
> been upgraded to the highest level to achieve near K3s equivalency.
>
> My K3s RF Gain control works exactly as expected, i.e. as the control is
> rotated counter-clockwise the "S-Meter" increases to a point at almost full
> scale.
>
> On the K3, however, as the RF Gain control is rotated counter-clockwise and
> the "S-Meter" increases to about the "40-over" point at the control's 10
> o'clock position the "S-Meter" then inexplicably drops to the "S9" position
> where it remains for the remainder of the RF gain control counter-clockwise
> rotation. This shouldn't be but like N1LN's issue, the radio does otherwise
> perform correctly.  I have done everything that N1LN has done including the RF
> Gain calibration using a signal generator.
>
> Very strange, indeed.
>
> Dan -- N3ND
>
> *************************************
>
> Date: Wed, 13 Dec 2017 16:25:20 -0500
> From: "Bruce Meier"<[hidden email]>
> To:<[hidden email]>
> Subject: [Elecraft] K3 / RF gain - S meter issue
>
>
> Looking for suggestions here from someone that might have experienced this
> same issue - -
>
>
>
> I have (2) K3 units (not the S model) that are acting different when the RF
> gain adjustments are turned fully counter clockwise.
>
> ?  Unit one ? s/n 1062: When the RF gain is decreased to full counter
> clockwise position the S-Meter registers full scale (60db over S9) as
> expected.
>
> ?  Unit two ? s/n 1193:  When the RF gain is decreased to full counter
> clockwise position the S-Meter registers only S8.
>
>  Facts:
>
> ?  With the exception of the S-Meter readings with RF gain settings, both
> units are working fine
>
> ?  Both units have ?almost? all mods installed.  If not all, they are at
> least both the same.
>
> ?  Both units are running the same version of microcode (5.54) (2.88) (2.88)
> (1.26)
>
> ?  Both units seem to register very close actual RF gain of stations when
> connected when tested on the same antenna
>
> ?  Both units have the new synthesizers.
>
> ?  Both units have second RX
>
> ?  In both cases (both units) I can hear the audio level decreasing linearly
> as expected as the RF gain is turned counter clockwise
>
> I have tried to initialize the K3 and reload the firmware resulting in no
> change to the RF gain / S-Meter readings.
>
>
>
> Elecraft Tech Support suggested:
>
> ?  Calibrating the s-meter ? did nothing
>
> ?  Next suggestion ? send the rig to California.
>
>
>
> Anyone have this issue and fixed it??
>
>
>
> 73,
>
> Bruce N1LN
>
>
>
> ______________________________________________________________
> 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
> Message delivered to [hidden email]

______________________________________________________________
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
Message delivered to [hidden email]

[alorona]

This would be easy to test. Normally, 'hardware AGC' kicks in at an input signal level of -46 dBm or so. This point is essentially independent of the AGC THR and SLP settings. See if your K3 varies from this data point anomalously.

Al  W6LX



__________________________________________
In my opinion based on inspection of the schematic, the issue resides in the fet
post-filter amplifier that is controlled by the "hardware" AGC (and manual "RF"
gain control)  The specs on this device are very loose, IIRC, Gm can vary 2:1.  
Thus the response of this fet to AGC/RF gain varies from radio to radio

Wes  N7WS
______________________________________________________________
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
Message delivered to [hidden email]

[Don Wilhelm]

Al,

My understanding of the hardware AGC is that it is fixed and is not
dependent on the menu AGC parameters.  The AGC THR and SLP are strictly
DSP parameters.

The Hardware AGC is there only to protect the ADC from overload due to
strong signals.

There may some variation in the exact level that the Hardware AGC kicks
in form K3 to K3 due to component tolerances.

The fact that the manual RF gain operates at the Hardware AGC point may
be sufficient explanation that the RF Gain Control when at full
counterclockwise position does not drive the S-meter to full scale
(again due to component tolerances).

The real question in my mind is whether the K3 S-meter responds properly
at its calibration points of S-9 and S-9+30.  And those points are
correct with the RF Gain at full clockwise rotation.  Those calibration
points are determined in the DSP section, and not in the hardware part
of the K3.

73,
Don W3FPR

On 12/14/2017 6:40 PM, Al Lorona wrote:
> This would be easy to test. Normally, 'hardware AGC' kicks in at an input signal level of -46 dBm or so. This point is essentially independent of the AGC THR and SLP settings. See if your K3 varies from this data point anomalously.
______________________________________________________________
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
Message delivered to [hidden email]

[Lyle Johnson]

There are several things going on here.

1) The "hardware AGC" is derived at the 15 kHz IF and is designed to
prevent overdrive to the DSP's IF Analog-to-Digital Converter (ADC).  It
also serves to prevent the 15 kHz IF amplifier from being driven into
non-linear operation.

2) The "hardware AGC" voltage is fed to the 8.215 MHz IF amplifier.  By
placing the gain control here, it also helps to prevent overload of the
2nd Mixer (8.215 MHz -> 15 kHz).

3) The "RF Gain" control voltage is fed to the same point at the 8.215
MHz IF stage.

4) The RF ATTenuator (0 or 10 dB in the K3, more choices in the K3S) and
PREamp (one choice in the K3, up to two choices in the K3S) are applied
in the radio prior to the first mixer or crystal filter.  They thus have
an impact on the "hardware" AGC."

All of the above is independent of the DSP subsystem.

5) The "RF Gan Cal" works by injecting known voltages to the 8.215 MHz
IF amplifier control point and measuring the subsequent change in gain. 
The DSP measures this and creates a table.  Due to dynamic range
limitations of the measuring system, the calibration procedure requires
two signals levels to be injected. In the case of the higher level
signal not being available, the factory default values are used.  Due to
variations in the gain control led stage, there may be the change in S
Meter indication that some of you have observed at high signal (or
severely reduced RF Gain control) conditions.

In general, if you are going to do the RF Gain calibration, you really
should use both signal levels.  The XG3 specifically provides the
required levels.

6) The DSP regularly measures the control voltage to the 8.215 MHz IF
amplifier. The RF Gain Cal information is used by the DSP when it then
calculates the signal level in the IF passband.  It is not part of the
signal control loop; it is strictly a measure and report function.

7) The DSP takes into account the PRE and ATT status provided by the MCU
in the S Meter ABS mode; it ignores such information in the S Meter NOR
mode.

8) The S Meter calculation is prior to the DSP's AGC algorithms so the S
Meter is not affected by THR, SLP or other AGC-related settings.

9) The DSP presents the results of the S Meter calculation to the
radio's MCU via a DAC voltage that is only routed to the MCU -- it has
no other function in the radio.

10) The S Meter OFS and SCL values are used by the MCU (not the DSP) to
manage the S Meter display of the DSP's DAC voltage on the K3's LCD.

Enjoy!

73,

Lyle KK7P
______________________________________________________________
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
Message delivered to [hidden email]

[alorona]

Wes stated that there's a component in the AGC circuit that changes the onset of hardware AGC due to component tolerances. I suggested putting that theory to the test by actually measuring it.

Al  W6LX
______________________________________________________________
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
Message delivered to [hidden email]

[Bill K9YEQ]

Isn't that answered by Lyle's point #10?

73,
Bill
K9YEQ

-----Original Message-----
From: [hidden email] [mailto:[hidden email]] On Behalf Of Al Lorona
Sent: Thursday, December 14, 2017 7:02 PM
To: [hidden email]; [hidden email]
Subject: Re: [Elecraft] K3 / RF gain - S meter issue - Similar issue

Wes stated that there's a component in the AGC circuit that changes the onset of hardware AGC due to component tolerances. I suggested putting that theory to the test by actually measuring it.

Al  W6LX
______________________________________________________________
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 Message delivered to [hidden email]
______________________________________________________________
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
Message delivered to [hidden email]

[Wes Stewart-2]

I don't think I said that.  I imagine the onset of hardware AGC is determined
elsewhere.  What I am saying is that the dB/V gain change is non-linear and
different from device to device.  I assumed, mistakenly it appears, that the
bias voltage to the device was tailored via a lookup table to create a linear
dB/V (if you can call dB linear) scale factor. Lyle's description, if I
understand it correctly, says that they measure the gain control characteristic,
but leave it be and correct the numbers via a table.

A much better choice here (IMHO of course) would have been something like an
Analog Devices AD600 (See "A High-Performance AGC/IF Subsystem" by Bill Carver
in QST, May 1996.)

Another place where the designers' QRP DNA got in the way of robust design was
the choice of the second mixer.  The oft repeated claim that hardware AGC is
protecting the DSP may be true, but the second mixer overload characteristics
are even worse.

Wes  N7WS



On 12/14/2017 6:01 PM, Al Lorona wrote:
> Wes stated that there's a component in the AGC circuit that changes the onset of hardware AGC due to component tolerances. I suggested putting that theory to the test by actually measuring it.
>
> Al  W6LX

______________________________________________________________
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
Message delivered to [hidden email]

[alorona]

You said, "In my opinion based on inspection of the schematic, the issue resides in the fet
post-filter amplifier that is controlled by the "hardware" AGC (and manual "RF"
gain control)."

You were making a relationship between hardware AGC and the loose spec of the FET in question. By saying that, I understood you to mean that the hardware AGC signal was being interpreted differently by the problem K3 compared to a 'normal' K3 because of this FET. At least this is what it seemed like. I will accept that I misunderstood you and that I have no idea what you were trying to say.

Al  W6LX
______________________________________________________________
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
Message delivered to [hidden email]

[Wes Stewart-2]

I've done a little playing around with this.  I made some measurements, using my
"old" K3, of the HAC1 voltage (applied to the FET i-f amp) vs. the RFG value
just to see if I could see the "correction" Lyle spoke of.  It's a
semi-interesting plot that I tried to post on the K3 Yahoo group. Unfortunately,
I have a hard time describing Yahoo groups without resorting to swear words, and
I couldn't accomplish the task. Someday, if I have the K3S out of the station
line up, I'll repeat the measurements.

Wes

On 12/14/2017 8:57 PM, Al Lorona wrote:
> You said, "In my opinion based on inspection of the schematic, the issue resides in the fet
> post-filter amplifier that is controlled by the "hardware" AGC (and manual "RF"
> gain control)."
>
> You were making a relationship between hardware AGC and the loose spec of the FET in question. By saying that, I understood you to mean that the hardware AGC signal was being interpreted differently by the problem K3 compared to a 'normal' K3 because of this FET. At least this is what it seemed like. I will accept that I misunderstood you and that I have no idea what you were trying to say.
>
> Al  W6LX

______________________________________________________________
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
Message delivered to [hidden email]