Brett,
In TUNE, the power control loop is not as tight in the firmware portion
of the loop as compared with normal keying. The reason for that is to
allow those using a manual tuner with their K2 to be able to adjust it
without the power jumping quickly all over the place.
I am glad to hear that you had good results with the direct connections
- that seems to verify that there is reactance (or other cause of a
non-50 ohm purely resistive) condition somewhere in your setup - whether
it is the coax, the external wattmeter or the dummy load itself we may
never know, but you can check each one easily at the frequencies of
interest with an antenna analyzer. You may be able to borrow an AA to
make those measurements if you do not have one.
You may find some frequency dependency between 1N5711 diodes (I have
occasionally noticed that), but I do not consider it a problem because
the power indication and control point are not more than a few tenths of
a watt off at any point if all is normal. Considering that most
wattmeter specs are in the range of 5% to 20% OF THE FULL SCALE READING,
the diode detector in the K2 is actually more accurate than most if the
load is 50 ohms resistive at the frequency of operation (measuring
with an ohmmeter is not suficient, resistance at DC is not likely to be
equal to the resistance at RF).
On your wattmeter accuracy, consider that even 5% of the 20 watt scale
can result in an error of 1 watt at any power level - and even the Bird
wattmeter is only that good immediately after calibration - you must
expect such errors in your measuring equipment (check the spec sheet and
verify the measurement by independent means if you really want
accuracy). The Diode Detector is actually more accurate than most
wattmeters, but it must be used with a known good non-reactive load.
73,
Don W3FPR
Brett gazdzinski wrote:
At lunch I did some more tests.
I made a dummy load using 4 200 ohm 2 watt resistors, measured
the resistance at 50 ohms, and plugged that right into the watt meter.
The other end of the wattmeter plugged into the K2 antenna jack
using adaptors.
I did the tx setup per the book, and now have at least 10
watts out on all bands, but it does roll off on the
higher bands. 10 meters might be below 10 watts, I don't
remember. 80 and 40 meters do about 13 or 14 watts.
What the 'tune' button seems to do is measure the rf out
per a set power (2 watts), display it, and when out of the tune mode,
corrects the power so its right.
You can really see it work on 80 and 40 meters, set the
power knob to 2 watts, hi the tune button and the meters read 3 watts.
Exit the tune mode with the power still set at 2 watts
and hit the cw key and I get 2 watts out on the bar graph
and my meter.
This works a treat on 80, 40, and 30 meters, the power output
matches the knob set point and the bar graph very closely
AFTER you do the tune setup, at least up past 10 or 12 watts.
Above about 30 meters, the bar graph and the tune position start
showing more power then the 3 watt meters I have do, I have some
diwa little thing, a Kenwood average/pep meter, and the heathkit
antenna tuner meter. All read the same.
The higher you go above 30 meters, the more the rig says its doing
more power out then my meters do.
The RF sense circuit is after all the stages, its at the
antenna output basically..
I checked all the components in the rf output sense circuit and
they are the correct values and measure correctly.
I did not mess with the diode (D9 1N5711), nor can I verify
its number without unsoldering it.
Maybe I should order a few 1N5711 diodes and see if there is
any difference in how they act.
This sure is a fun rig to trouble shoot and play with, although
I am still not over fond of computers in HF gear...
Brett
N2DTS
_______________________________________________
Elecraft mailing list
Post to: Elecraft@mailman.qth.net
You must be a subscriber to post to the list.
Subscriber Info (Addr. Change, sub, unsub etc.):
http://mailman.qth.net/mailman/listinfo/elecraft
Help: http://mailman.qth.net/subscribers.htm
Elecraft web page: http://www.elecraft.com
