Elecraft

[KX3] JT65-HF KX3 PA Temp Testing

Classic

List

Threaded

7 messages Options

Gary Hawkins

[KX3] JT65-HF KX3 PA Temp Testing

Last night I did some monitoring of the KX3 PA temp. My setup as
follows: Barefoot KX3 fitted with VE3FMN heatsink. PSU 13.8V, 14.076MHz
operation into dummy load, ambient air temp 21C. Set ALC = 4 occasional
flicker 5. WSJT-X S/W and Signalink

The test consisted running 6 CQ cycles of JT65-HF and noting PA temp at
end of RX and TX segments. Therefore lowest and hottest points in cycle
assuming no thermal overshoot. The TX signal is approx 47 secs long and
consists essentially of a CW carrier of varying frequency (less than
200Hz bandwidth). The RX cycle is one minute and 13 secs long, so one
complete TX/RX cycle takes exactly 2 mins.

Results as follows:

0.1W TX, min temp 29C, max temp 38C, typical temp variation of PA temp
in TX cycle 8 - 9C
0.5W TX, min temp 29C, max temp 40C, typical temp variation of PA temp
in TX cycle 8 - 9C
1W TX, min temp 29C, max temp 40C, typical temp variation of PA temp in
TX cycle 8 - 9C
2W TX, min temp 30C, max temp 41C, typical temp variation of PA temp in
TX cycle 8 - 9C
3W TX, min temp 29C, max temp 41C, typical temp variation of PA temp in
TX cycle 8 - 9C
5W TX, min temp 30C, max temp 48C, typical temp variation of PA temp in
TX cycle 13 - 15C

I also did a long term JT65_HF CQ sequence over 12 cycles at 3W to see
the where max PA temp stabilized. It appeared to be 42C, with typical
min to max temp PA range of 34 to 42C over TX cycle.

I was surprised not to see more degradation on going from 2 to 3W as
this takes the PA from high to low efficiency mode. At 5W the KX3 PA is
showing a considerable increase in thermal stress, sweet spot of my KX3
appears to be 3W or below.

I'd be interested to hear any comments/thoughts on these findings.

73's Gary K6YOA
______________________________________________________________
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]

Gary Hawkins

Re: [KX3] JT65-HF KX3 PA Temp Testing

I did a little additional PA temp testing today, this time focused on
simulated JT65-HF 10m operation. Setup the same as below except
transmitting on 28.076MHz operation into dummy load. Results as follows:

2W TX, min temp 29C, max temp 43C, typical temp variation of PA temp in
TX cycle 11 - 12C
3W TX, min temp 29C, max temp 43C, typical temp variation of PA temp in
TX cycle 14 - 15C
5W TX, min temp 30C, max temp 54C, typical temp variation of PA temp in
TX cycle 19 - 20C

The KX3 is clearly under more thermal stress operating on 10m rather
than 20m, and the sweet spot appears to be in the 2 - 3W range. 5W
operation showing a marked increase in cycle-to-cycle temp variation.

Does anyone know how I can estimate KX3 freq drift before and after temp
compensation as I don't have an easy way of measuring it? Again, I'd be
interested to hear any comments/thoughts on these findings.

73's Gary K6YOA

On 5/13/2015 9:09 AM, Gary Hawkins wrote:

> Last night I did some monitoring of the KX3 PA temp. My setup as
> follows: Barefoot KX3 fitted with VE3FMN heatsink. PSU 13.8V,
> 14.076MHz operation into dummy load, ambient air temp 21C. Set ALC = 4
> occasional flicker 5. WSJT-X S/W and Signalink
>
> The test consisted running 6 CQ cycles of JT65-HF and noting PA temp
> at end of RX and TX segments. Therefore lowest and hottest points in
> cycle assuming no thermal overshoot. The TX signal is approx 47 secs
> long and consists essentially of a CW carrier of varying frequency
> (less than 200Hz bandwidth). The RX cycle is one minute and 13 secs
> long, so one complete TX/RX cycle takes exactly 2 mins.
>
> Results as follows:
>
> 0.1W TX, min temp 29C, max temp 38C, typical temp variation of PA temp
> in TX cycle 8 - 9C
> 0.5W TX, min temp 29C, max temp 40C, typical temp variation of PA temp
> in TX cycle 8 - 9C
> 1W TX, min temp 29C, max temp 40C, typical temp variation of PA temp
> in TX cycle 8 - 9C
> 2W TX, min temp 30C, max temp 41C, typical temp variation of PA temp
> in TX cycle 8 - 9C
> 3W TX, min temp 29C, max temp 41C, typical temp variation of PA temp
> in TX cycle 8 - 9C
> 5W TX, min temp 30C, max temp 48C, typical temp variation of PA temp
> in TX cycle 13 - 15C
>
> I also did a long term JT65_HF CQ sequence over 12 cycles at 3W to see
> the where max PA temp stabilized. It appeared to be 42C, with typical
> min to max temp PA range of 34 to 42C over TX cycle.
>
> I was surprised not to see more degradation on going from 2 to 3W as
> this takes the PA from high to low efficiency mode. At 5W the KX3 PA
> is showing a considerable increase in thermal stress, sweet spot of my
> KX3 appears to be 3W or below.
>
> I'd be interested to hear any comments/thoughts on these findings.
>
> 73's Gary K6YOA

______________________________________________________________
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]

Tony Estep

Re: [KX3] JT65-HF KX3 PA Temp Testing

On Wed, May 13, 2015 at 7:47 PM, Gary Hawkins <[hidden email]> wrote:

> ...Does anyone know how I can estimate KX3 freq drift....

===============
Gary, if you have another receiver it's easy. Just set up the other rx for
JT65HF, transmit via the KX3, and look at the received trace. You can see
drift down to a couple of Hz.

I found the drift not to be too dependent on power level. Even at very low
levels of output power, my KX3 drifted about the same (about 18 Hz) before
the temperature compensation. This is too much for JT65HF communication.

73,
Tony KT0NY
______________________________________________________________
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]

Gary Hawkins

[KX3] JT65-HF KX3 Freq Drift Measurements

The only other working receiver I have at home is an aging (30 years at
least old) Yaesu FT-690R. So I switched it on for the first time in at
least two years and within 30mins I'd convinced myself its receive freq
stability was such (after a little warm-up time) that I could measure TX
relative drift of a receive 6m signal down to 1 or 2Hz accuracy. Thus,
with the FT-690R connected through my Signalink to the WSJT-X waterfall
I went about measuring my KX3 freq variation under TX conditions.

KX3 TX power = 3W to dummy load, freq 50.076MHz, FM modulation
continuous TX until PA temp reach 55C. Starting PA temp reported at 30C:

* total min-to-max freq drift averaged over four runs was a staggering
104Hz, occurring over approx 2 mins;
* minimum freq observed around 40C. Initially freq drifts down,
reaches minimum point and then drifts positive.

So my next thought was could I observed performance on a lower band.
Well, the only band that offered any possibility was operating on 30m
and looking for the 5th harmonic on the FT-690R. To my slight surprise
with the KX3 transmitting at 10.1MHz, I found a weak 5th harmonic that I
could observe on the FT-690R at 50.5MHz. Keying the KX3 resulting in
the 5th harmonic appearing and disappearing. KX3 TX power = 3W to dummy
load, KX3 freq 10.1MHz, FM modulation continuous TX until PA temp reach
40C (seemed to stabilize at this max temp). Starting PA temp reported
at 27C. FT-690R monitoring 50.5MHz:

* total min-to-max freq drift averaged over four runs 40.5Hz, _which
considering I'm looking at the fifth harmonic is really 8.1Hz_
occurring over approx 4-5 mins;
* minimum freq observed around 32C. Initially freq drifts down,
reaches minimum point and then drifts positive.

For those that have done similar tests on an uncompensated KX3, do these
results seem reasonable? If they are reasonable is it safe to say that
successful 6m JT65/JT9 operation even after extended temperature
compensation is unlikely?

73's Gary K6YOA

_Please note these results are not a KX3 that has not gone through the
Extended Temp Compensation. _ That being said, while I think extended
temp compensation would likely prove sufficient for JT65-HF or JT9 for
30m and surrounding bands, I will be very surprised when I do the temp
compensation over the weekend, whether it can sufficiently correct the
xxxHz drift I'm seeing on the 6m band.

On 5/13/2015 6:01 PM, Tony Estep wrote:

> On Wed, May 13, 2015 at 7:47 PM, Gary Hawkins <[hidden email]
> <mailto:[hidden email]>> wrote:
>
> ...Does anyone know how I can estimate KX3 freq drift....
>
> ===============
> Gary, if you have another receiver it's easy. Just set up the other rx
> for JT65HF, transmit via the KX3, and look at the received trace. You
> can see drift down to a couple of Hz.
>
> I found the drift not to be too dependent on power level. Even at very
> low levels of output power, my KX3 drifted about the same (about 18
> Hz) before the temperature compensation. This is too much for JT65HF
> communication.
>
> 73,
> Tony KT0NY

GM4JJJ

Re: [KX3] JT65-HF KX3 Freq Drift Measurements

Please see the work done by Lance W7GJ on this at

http://www.bigskyspaces.com/w7gj/KX3HeatSinks.htm

73

David Anderson GM4JJJ

> On 14 May 2015, at 23:06, Gary Hawkins <[hidden email]> wrote:
>
>
> For those that have done similar tests on an uncompensated KX3, do these results seem reasonable? If they are reasonable is it safe to say that successful 6m JT65/JT9 operation even after extended temperature compensation is unlikely?
>
> 73's Gary K6YOA
>
______________________________________________________________
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]

Edward R Cole

Re: [KX3] JT65-HF KX3 Freq Drift Measurements

In reply to this post by Gary Hawkins

Gary,

I made frequency drift curves simulating JT65 by keying my KX3 at 5w
CW on 50-MHz for 50-seconds and measuring frequency every ten seconds
with a EIB-538 counter that uses a TCXO reference or can use my
Rubidium standard as ext ref. This offers +/-5 E-11
accuracy. Greatest error was 108 Hz before I ran temp compensation
procedure. After ward the max drift was about 9-Hz. The data at
10-seconds in the tests after temp compensation are aberrant and I
suspect something happened in the temp comp procedure to record a bad offset.
http://www.kl7uw.com/KX3_FREQ_DRIFT_TABLE.pdf

I think this really supports how well the temp compensation procedure
performs. I think you will find much better drift characteristics
after running it. I have added a heat sink on my KX3 since making these tests.

I find that drift of up to 20-Hz in one minute is acceptable for
JT65B which I uses on eme. My K3+DEMI transverter drifts about +7-Hz
and then stays near that offset for running past 30-minutes of JT65
on 2m. The KX3 is not as good but is acceptable. Note I run my K3
with an EXREF on 28-MHz seeing about +2 Hz error (but am running at
0-dBm in transverter mode so no extra heat results when transmitting).

----snipped
For those that have done similar tests on an uncompensated KX3, do these
results seem reasonable? If they are reasonable is it safe to say that
successful 6m JT65/JT9 operation even after extended temperature
compensation is unlikely?

73's Gary K6YOA

73, Ed - KL7UW
http://www.kl7uw.com
"Kits made by KL7UW"
Dubus Mag business:
[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]

Gary Hawkins

Re: [KX3] JT65-HF KX3 Freq Drift Measurements

In reply to this post by GM4JJJ

Hi David,

Thanks for the reference - I'd not seen this article. While I've fitted
a heat sink, it is a commercially available 3rd party option and I've
not implemented the additional heat transfer techniques Lance references
with his home made version. This weekend I hope to do the extended temp
compensation routine so I'll have a better ideal on how much better freq
stability gets. I'm sure I'll be posting results.

73's Gary K6YOA.

On 5/14/2015 3:29 PM, David Anderson wrote:

> Please see the work done by Lance W7GJ on this at
>
> http://www.bigskyspaces.com/w7gj/KX3HeatSinks.htm
>
> 73
>
> David Anderson GM4JJJ
>
>> On 14 May 2015, at 23:06, Gary Hawkins <[hidden email]> wrote:
>>
>>
>> For those that have done similar tests on an uncompensated KX3, do these results seem reasonable? If they are reasonable is it safe to say that successful 6m JT65/JT9 operation even after extended temperature compensation is unlikely?
>>
>> 73's Gary K6YOA
>>