K3 Watt Meter Calibration Question

[engineercm]

Noticed today 20m output measured by LP100A was 90w when K3 wattmeter set to 100w. Same with 10m; other bands measured close enough.  Updated firmware to latest and performed Transmit Calibration.  Calib:WM HP at 50w set to agree with LP100A.  In order to get 100w on LP100A, I have to crank the K3 output so that it reads 110 on the internal wattmeter.  Why?  Slope effect of internal WM?

ATU is in Bypass; tried two different dummy loads with same result.  Yes, I know 10% difference is barely noticeable at the other end; it's annoying.

What am I doing wrong?

Clark
WU4B

[Don Wilhelm]

Clark,

The detection diodes in the K3 (and the LP100A wattmeter as well), do
not have a perfectly linear relationship vs. frequency response.
The LP-100A is quite good, but can be in error by as much as 5% (Larry
will not guarantee it closer, I have asked).
The K3 wattmeter can also have up to a 10% error in its readings (may
vary with frequency).

So the net of what I am trying to say is that at the 100 watt level, a
wattmeter reading between 90 and 110 watts is within the range of normal
expectations.

We have come to be "accuracy obsessed" since the advent of digital
displays.   In the days of analog meters, there was always a region of
uncertainty due to the inability to read the needle on the meter to a
high degree of accuracy - parallax effects, the width of the needle, as
well as the inherent accuracy of the circuits the meter was indicating.  
In todays world of digital displays, we assume that all displayed digits
are true and accurate, when in truth, only the first two digits of the
display are valid in an instrument that measures to a 10% accuracy.

73,
Don W3FPR

On 6/5/2016 6:49 PM, engineercm wrote:
> Noticed today 20m output measured by LP100A was 90w when K3 wattmeter set to
> 100w. Same with 10m; other bands measured close enough.  Updated firmware to
> latest and performed Transmit Calibration.  Calib:WM HP at 50w set to agree
> with LP100A.  In order to get 100w on LP100A, I have to crank the K3 output
> so that it reads 110 on the internal wattmeter.  Why?  Slope effect of
> internal WM?
>
>

______________________________________________________________
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]

[engineercm]

Hi, Don.

I suspected as much.  Precision (number of digits) and accuracy (compared
to a known standard) are very different. I guess it doesn't matter what the
power level is except when running QRP where one must be below 5w. That's
why I always adjust my power output to be 4.5 w just to be sure.

I'm sure you remember slide rules used in engineering schools before the
pocket calculators came along in the 60's.  I still have mine and when I
look at remember some of my colleagues that would read to 5 digits from
their "slip-stick".

BTW:  it was amazing to watch the K3 run through it's TX Calibration.
After having the K3 for a few years, I'm still learning.  What an amazing
feat of engineering.

Clark
WU4B

On Sun, Jun 5, 2016 at 8:37 PM, Don Wilhelm <[hidden email]> wrote:

> Clark,
>
> The detection diodes in the K3 (and the LP100A wattmeter as well), do not
> have a perfectly linear relationship vs. frequency response.
> The LP-100A is quite good, but can be in error by as much as 5% (Larry
> will not guarantee it closer, I have asked).
> The K3 wattmeter can also have up to a 10% error in its readings (may vary
> with frequency).
>
> So the net of what I am trying to say is that at the 100 watt level, a
> wattmeter reading between 90 and 110 watts is within the range of normal
> expectations.
>
> We have come to be "accuracy obsessed" since the advent of digital
> displays.   In the days of analog meters, there was always a region of
> uncertainty due to the inability to read the needle on the meter to a high
> degree of accuracy - parallax effects, the width of the needle, as well as
> the inherent accuracy of the circuits the meter was indicating.  In todays
> world of digital displays, we assume that all displayed digits are true and
> accurate, when in truth, only the first two digits of the display are valid
> in an instrument that measures to a 10% accuracy.
>
> 73,
> Don W3FPR
>
> On 6/5/2016 6:49 PM, engineercm wrote:
>
>> Noticed today 20m output measured by LP100A was 90w when K3 wattmeter set
>> to
>> 100w. Same with 10m; other bands measured close enough.  Updated firmware
>> to
>> latest and performed Transmit Calibration.  Calib:WM HP at 50w set to
>> agree
>> with LP100A.  In order to get 100w on LP100A, I have to crank the K3
>> output
>> so that it reads 110 on the internal wattmeter.  Why?  Slope effect of
>> internal WM?
>>
>>
>>
>

--
72,

Clark WU4B
QRPARCI #10815
SKCC #3892
Southeastern DX Club <http://www.sedxc.org>
North Georgia QRP Club <http://www.nogaqrp.org>
Kennehoochee Amateur Radio Club <http://www.w4bti.org>


*"It is vain to do with more what can be done with less."*
*Attributed to *William of Occam (1288 AD - 1348 AD)
______________________________________________________________
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]

[Jim Brown-10]

On Sun,6/5/2016 5:37 PM, Don Wilhelm wrote:
> So the net of what I am trying to say is that at the 100 watt level, a
> wattmeter reading between 90 and 110 watts is within the range of
> normal expectations.
>
> We have come to be "accuracy obsessed" since the advent of digital
> displays.

Another excellent post, Don. I strongly agree.

73, Jim K9YC

______________________________________________________________
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]

[briancom]

Don,

Can't argue with this.  A long time ago a wise engineer said: "Never
believe the last digit displayed".  He was talking about the newly
released 3 and 4 digit equipment display. Today it's almost as you say.
  Believe the first two.

There are some exceptions to this rule.  HP equipment (at least the
older stuff) would only digitally display the number of truly believable
significant digits.  For example, their counters would not display all
available digits but just those one can real.  So there are exceptions
to all rules of thumb.

One has to know their measuring gear characteristics to know what to
believe.

73 de Brian/K3KO



On 6/6/2016 0:37 AM, Don Wilhelm wrote:

> Clark,
>
> The detection diodes in the K3 (and the LP100A wattmeter as well), do
> not have a perfectly linear relationship vs. frequency response.
> The LP-100A is quite good, but can be in error by as much as 5% (Larry
> will not guarantee it closer, I have asked).
> The K3 wattmeter can also have up to a 10% error in its readings (may
> vary with frequency).
>
> So the net of what I am trying to say is that at the 100 watt level, a
> wattmeter reading between 90 and 110 watts is within the range of normal
> expectations.
>
> We have come to be "accuracy obsessed" since the advent of digital
> displays.   In the days of analog meters, there was always a region of
> uncertainty due to the inability to read the needle on the meter to a
> high degree of accuracy - parallax effects, the width of the needle, as
> well as the inherent accuracy of the circuits the meter was indicating.
> In todays world of digital displays, we assume that all displayed digits
> are true and accurate, when in truth, only the first two digits of the
> display are valid in an instrument that measures to a 10% accuracy.
>
> 73,
> Don W3FPR
>
> On 6/5/2016 6:49 PM, engineercm wrote:
>> Noticed today 20m output measured by LP100A was 90w when K3 wattmeter
>> set to
>> 100w. Same with 10m; other bands measured close enough.  Updated
>> firmware to
>> latest and performed Transmit Calibration.  Calib:WM HP at 50w set to
>> agree
>> with LP100A.  In order to get 100w on LP100A, I have to crank the K3
>> output
>> so that it reads 110 on the internal wattmeter.  Why?  Slope effect of
>> internal WM?
>>
>>
>
> ______________________________________________________________
> 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]

Not necessarily.  For example, an HP8970 Noise Figure Meter reads to 0.01 dB,
yet uncertainty (for just instrument error) is +/- 0.1 dB.  In the days of yore,
LNAs were typically highly mismatched at the input and this added further
uncertainty.  I once measured a friend's homebrew 1296 MHz preamp NF at 0.7 dB.
I calculated the uncertainty at +/- 0.7 dB!

I got a chuckle when HP gave an '8970 to the ARRL lab and they did a test on a
preamp (ARR if memory serves) and gave the NF to two decimal places.

Wes  N7WS

On 6/6/2016 4:11 AM, brian wrote:
>
> There are some exceptions to this rule.  HP equipment (at least the older
> stuff) would only digitally display the number of truly believable significant
> digits.  For example, their counters would not display all available digits
> but just those one can real.  So there are exceptions to all rules of thumb.
>
> One has to know their measuring gear characteristics to know what to believe.
>
> 73 de Brian/K3KO

______________________________________________________________
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]