Elecraft › [KX3]

Why is KX3 phase noise so much less than the K3 ?

Classic

List

Threaded

32 messages Options

Wes (N7WS)

Re: Why is KX3 phase noise so much less than the K3 ?

First of all, the ARRL is showing (as you note) transmitted phase noise. Rob
lists receiver phase noise. Ideally, they should be the same, but are they really?

Second, there can be significant test sample variation and variation between
phase noise test sets.*

Third, phase noise measurements aren't trivial. Witness the dozens of
application notes and articles dealing with the subject. ARRL's own Test
Procedures Manual states: " *NOTE: Composite noise setup is very sensitive to
environment noise. Use of common-mode ferrite chokes may be required if test
fixture problems are encountered as described in the procedure for this test."
In my working career I have seldom seen or made, a phase noise plot that wasn't
corrupted by 60Hz spurs, for example.

* Rob himself summarizes this on his website:
http://www.sherweng.com/images/MeasurementAccuracy&SampleVariation.pdf

And last, but certainly not least, why get bent out of shape over one number.
Rob does at once a service and disservice by publishing his charts. They are
extremely handy but too many people jump on one number to rate (rank) equipment
and totally disregard things like ergonomics (If they didn't the K3 would be at
the bottom of the list). He recognizes this in his closing paragraph of the
above referenced work. We should pay it more heed.

Wes N7WS

On 4/2/2014 10:09 AM, dave wrote:

>
> Why does the ARRL lab test show the KX3 TX phase noise as about -124 dBc/Hz
> (Fig 5 of their report) and Sherwood show this as -144 dBc/Hz, both at 10 kHz
> spacing? That is a huge difference.
>
> The ARRL report shows that the KX3 never goes below about -135 dBc/Hz, even 1
> MHz away.
>
> Is there an update to the ARRL lab test? Did the ARRL miss this one?
>
> FWIW, ARRL shows the K3 at about -142 and the FTdx5000 at -135, at 10 kHz
> spacing.
>
> 73 de dave
> ab9ca/4
>
>

______________________________________________________________
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]

Grant Youngman

Re: Why is KX3 phase noise so much less than the K3 ?

In reply to this post by george fritkin

The Sherwood list ranking isn’t about how well radios “hear” in general. Under the right conditions, most of the radios on the last (and a whole lot of radios that aren’t) can hear everything our most beloved radio (whatever that is) hears. The list is ranked in order of decreasing narrow spaced (2Khz signal spacing) dynamic range, which is something that makes a difference in extreme conditions. The rankings don’t say a thing at all about MDS, or how the AGC behaves, or internal phase noise, or many other factors (although Sherwood does indicate when phase noise gets to be limiting in the measurements). Radios even close to the bottom of the list can hear everything just fine that a radio at the top of the list can — most of time. It’s those “not most of the time” situations where narrow spaced dynamic range matters that the ranking matters …

Grant NQ5T

On Apr 2, 2014, at 12:55 PM, george fritkin <[hidden email]> wrote:

> No! And I have them all. My TenTec Eagle hears as good as my FT5K and our beloved K3 and KX3
>
> George, W6GF
> On Wednesday, April 2, 2014 10:02 AM, "Charlie T, K3ICH" <[hidden email]> wrote:
>
> Better hope there's no Flex's transmitting too close to your location
> either.
>
> Chas
>
>
>
>> , Bill W2BLC wrote:
>>> Can anyone really hear the difference between any of the top rated rigs
>>> (Sherwood list) in the real world (not in a lab)?
>>>
>>> Bill K-Line
>>
> ______________________________________________________________

Edward R Cole

Re: Why is KX3 phase noise so much less than the K3 ?

In reply to this post by Larry Lopez

Larry,

Not knowing much about the R390A, but Sherwood ranks it very low in
phase noise. Xtal oscillators are easier to design with low phase
noise than synthesizers. This is primarily why later radios have a
problem with noise as most use synthesized LO's, and why the old
radios with xtal oscillators hear so well.

As to why it would matter, phase noise will limit the sensitivity of
a receiver. HF sky noise masks most receiver noise issues, but if
you are pursuing weak-signal contacts is matters a lot. HF receivers
used with transverters on VHF/UHF/mw frequencies are way more
sensitive if they have low phase noise, especially due the fact that
sky noise drops of very quickly above 6m.

Phase noise was one of the prime reasons I chose the K3 for my eme
station which detects signals in the -180 dBm level.

73, Ed
-----------------------------
I'm wondering why the KX3 Lo noise is so much better than the K3.
The R390A value strange since the first local oscillator is a quartz
crystal.

I've owned a Icom 720A and own a FT-1000D.
Ive never owned a K3, KX3 or a R390A.

Larry

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]

Bill Frantz

Re: Why is KX3 phase noise so much less than the K3 ?

In reply to this post by Don Wilhelm-4

Our field day has been all Elecraft for the last couple of years
with mostly K3s and a scattering of K2s. Last year we ran 9A QRP
battery + GOTA (+ satellite). We had antennas which allowed us
to run CW, SSB and Digital on the same band and had no
significant mutual interference. (We work hard to keep our
antennas in each others nulls, and being a west coast station,
always point them east.)

The bigest interference problem was phase noise from another
club's operation about 1/2 mile away. They were running 100
watts with I-K-Y gear.

Cheers - Bill, AE6JV

On 4/2/14 at 8:23 AM, [hidden email] (Don Wilhelm) wrote:

>That is why I try to encourage the local Field Day group to use
>only Elecraft gear. Even though we do not typically operate
>more than one transceiver on a single band, the YKI types can
>cause problems on other harmonically related bands. We used to
>have a Yaesu FT-900 on SSB, and it caused problems.
>Typically we are class 2A with a GOTA station.

-----------------------------------------------------------------------
Bill Frantz | QRP: So you can talk about | Periwinkle
(408)356-8506 | the ones that got away. | 16345
Englewood Ave
www.pwpconsult.com | | Los Gatos,
CA 95032

______________________________________________________________
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

Re: Why is KX3 phase noise so much less than the K3 ?

On 4/2/2014 1:42 PM, Bill Frantz wrote:
> The bigest interference problem was phase noise from another club's
> operation about 1/2 mile away. They were running 100 watts with I-K-Y
> gear.

My biggest problem is from a retired east coast lawyer 8 miles from me
who runs an ICOM 7600 to a solid state amp. He regularly wipes out 5-10
kHz on CW, twice as much on SSB.

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]

Larry Lopez

Re: Why is KX3 phase noise so much less than the K3 ?

In reply to this post by Larry Lopez

I'll try to summarize:

lo phase noise is good:

it reduces unintended radiation from transmitters.
it reduces reciprocal mixing on receiving.
surveys during contests are very revealing.

The KX3 and K3 have similar very low amounts of phase noise.

Unless I want to make a career of it I should stop at this point.
Nothing is perfect.

thank you very much everyone.
larry lopez N2CVS

Stewart Bryant

Re: Why is KX3 phase noise so much less than the K3 ?

In reply to this post by wayne burdick

On 02/04/2014 16:19, Wayne Burdick wrote:

> The K3 and KX3 both have very low phase noise. It can vary a bit from band to band and from one unit to the next. Sherwood measured one example of each, but typically both measure right around -140 dBc/Hz at 20 kHz.
>
> When we designed the K3, we were trying to improve on the phase noise numbers usually found in transceivers with synthesized local oscillators. Typical values were -110 to -120 dBc/Hz, and this definitely impacted stations within close proximity of each other. By using a very high C/L ratio in the K3's VCO, we achieved numbers in the -140 range, and our field testers in close proximity could no longer hear each other at all.
>
> This result was corroborated by the Ducie Island DXpedition, the first to use K3s. They told us that they had a 30-m CW station and a 30-m RTTY station both running at the same time within 1 kHz of each other.
>
> The KX3 uses a different type of synthesizer, but the result is the same. In this case an on-chip VCO runs at a very high frequency and is then divided down to get to the HF range.
>
> 73,
> Wayne
> N6KR

Wayne

My understanding is that the KX3 is worse than the K3 on TX
phase noise sidebands. Is that what you would expect?

This can be quite critical when choosing between the K(X)3s as
a transverter driver in a dense VHF/UHF environment (5 to 10Km
separation at 400W to high gain beams) such as you find in the
SE of UK.

Stewart/G3YSX

______________________________________________________________
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]

wayne burdick

Re: Why is KX3 phase noise so much less than the K3 ?

Administrator

On Jun 6, 2014, at 4:33 AM, Stewart Bryant <[hidden email]> wrote:

> My understanding is that the KX3 is worse than the K3 on TX
> phase noise sidebands. Is that what you would expect?

No. Both have very, very low transmitted phase noise.

73,
Wayne
N6KR

______________________________________________________________
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]

Ozy

Re: Why is KX3 phase noise so much less than the K3 ?

Pardon my ignorance, but can I get some clarification on a few things?

1) Is phase noise the measurement of the instability of an DDS? If so, does this mean a cheap part is being used? Does a GPSDO or a OCXO reduce phase noise?

2) How does this impact TX only, and why does it create such an issue to nearby listeners?

3) Does phase noise go down if you use a faster master clock? The Flex 6700 uses a 983.04mhz vs a 122.99Mhz clock in the 6300.

4) Do low phase noise radios allow in-band use, such as someone on CW on 20M and someone up on voice on 20M? What is considered a low value?

Sorry for so many questions, but this is still a very difficult concept for me to grasp.

Thanks

Chris

On Jun 6, 2014, at 8:18 AM, Wayne Burdick <[hidden email]> wrote:

>
> On Jun 6, 2014, at 4:33 AM, Stewart Bryant <[hidden email]> wrote:
>
>> My understanding is that the KX3 is worse than the K3 on TX
>> phase noise sidebands. Is that what you would expect?
>
> No. Both have very, very low transmitted phase noise.
>
> 73,
> Wayne
> N6KR
>
> ______________________________________________________________
> 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]

Alan Bloom

Re: Why is KX3 phase noise so much less than the K3 ?

On 06/08/2014 09:21 PM, Chris Johnson wrote:
> Pardon my ignorance, but can I get some clarification on a few
> things?
>
> 1) Is phase noise the measurement of the instability of an DDS?

Yes. To oversimplify a bit, if the frequency/phase of the oscillator is
varying back and forth at a 20 kHz rate then there will be phase noise
at a 20 kHz offset from the carrier.

> If so, does this mean a cheap part is being used?

Not necessarily. Phase noise can creep in at many places in a frequency
synthesizer. You have to get the entire design exactly right to have a
clean signal that has low phase noise and low spurs at all frequency
offsets.

> Does a GPSDO or a OCXO reduce phase noise?

Maybe. Normally it would only improve the phase noise within the loop
bandwidth of a PLL-type synthesizer. With a DDS-type oscillator it
would depend on whether the phase noise from the clock or the DDS itself
dominates (at any given offset).

> 2) How does this impact TX only, and why does it create such an
> issue to nearby listeners?

If the same local oscillator is used for both the receiver and the
transmitter, then the phase noise will be the same. In both cases it
causes the oscillator spectrum to "spread out". In the transmit case,
that can cause interference on nearby frequencies. In a receiver, it
causes "reciprocal mixing" which makes it sound like nearby strong
transmitters have excessive phase noise even if they don't.

> 3) Does phase noise go down if you use a faster master clock? The
> Flex 6700 uses a 983.04mhz vs a 122.99Mhz clock in the 6300.

Maybe. Every time you divide an oscillator's frequency you potentially
reduce the phase noise by up to 6 dB per octave (division by 2).
However, that only helps if the high-frequency oscillator has good phase
noise to begin with and the dividing process introduces no phase noise
of its own.

> 4) Do low phase noise radios allow in-band use, such as someone on CW
> on 20M and someone up on voice on 20M?

That's the goal. Almost all the early synthesized transceivers of 30-40
years ago had horrible phase noise. Hams soon discovered that they were
useless in multi-transmitter environments like Field Day. It was a
problem both on receive and transmit. People resorted to using old
non-synthesized tube rigs like Drake and Collins.

> What is considered a low value?

It's a matter of opinion. The top receivers in Sherwood's chart are on
the order of 140-ish dBc/Hz at 10 kHz offset.

http://www.sherweng.com/table.html

Alan N1AL
______________________________________________________________
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]

wayne burdick

Re: Why is KX3 phase noise so much less than the K3 ?

Administrator

In reply to this post by Ozy

Chris Johnson <[hidden email]> wrote:

> 1) Is phase noise the measurement of the instability of an DDS?

That's one contributing factor. But the DDS signal drives a mixer, and the other input to the mixer may be the larger source of phase noise in transmit mode.

> 2) How does this impact TX only, and why does it create such an issue to nearby listeners?

In the case of the KX3 and Flex radios (e.g. Flex 1500/3000/5000), the limiting factor is probably the transmit D-to-A converter. Its performance is limited by quantization and sampling noise, setting an upper bound on phase noise. The KX3 has an advantage in that its synthesizer uses a DDS-driven PLL, while the Flex radios are DDS only, without a following PLL. But the TX DAC is still a factor in both cases.

In a well-designed superhet like the K3, the TX DAC's noise floor affects only the in-band portion of the transmitted signal, i.e. the portion within the I.F. crystal filter passband. The K3 has two crystal filters in series in transmit mode, which results in very high rejection of noise outside the SSB passband (typically 100-3000 Hz). This means that the limiting factor on wideband transmit noise is not the IF injection into the mixer -- it really *is* the synthesizer.

The K3's synthesizer is extremely clean at wide offsets (another DDS-driven PLL), and that is why its transmit phase noise is so low.

> 3) Does phase noise go down if you use a faster master clock? The Flex 6700 uses a 983.04mhz vs a 122.99Mhz clock in the 6300.

That is a completely different design (direct digital up/down conversion), which requires very high clock speeds, very expensive ADC and DAC components, and high current drain in receive mode. It can be made clean in transmit and receive modes, although in receive mode this architecture will still have typically 15-20 dB lower blocking dynamic range than a well-designed superhet. None of this is applicable to the KX3, which is obviously intended to be an ultraportable radio with low current drain and low cost.

> 4) Do low phase noise radios allow in-band use, such as someone on CW on 20M and someone up on voice on 20M?

Yes. In general this is only a problem if you have stations in very close proximity. In this situation, the K3 is better than any other radio on the market. This is why the K3 is highly favored for Field Day and DXpeditions.

73,
Wayne
N6KR

______________________________________________________________
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]

Ozy

Re: Why is KX3 phase noise so much less than the K3 ?

Thanks Wayne and Alan, for that super detailed info. The radio side of things are still very magical to me! I'm a software guy, not a hardware guy. :)

I don't own a K3, but two KX3's for my "knob" radios. I use a Flex 6700 and 6300 for home use. I'm assuming the 6700's -147 dBc@10kHz, -152 dBc@100kHz phase noise values are pretty good then, if anything higher than -140 dBc is acceptable. Even the 6300 is spec'd at -140 dBc@10kHz. I do agree that my old (and sold) 3000 was messing up the bands during Field Day last year. This year we are using a K3, two 6300's, and a 6700, so I'm hoping for better performance all around.

Again, thank you for the info!

On Jun 9, 2014, at 9:08 AM, Wayne Burdick <[hidden email]> wrote:

> Chris Johnson <[hidden email]> wrote:
>
>> 1) Is phase noise the measurement of the instability of an DDS?
>
> That's one contributing factor. But the DDS signal drives a mixer, and the other input to the mixer may be the larger source of phase noise in transmit mode.
>
>
>> 2) How does this impact TX only, and why does it create such an issue to nearby listeners?
>
> In the case of the KX3 and Flex radios (e.g. Flex 1500/3000/5000), the limiting factor is probably the transmit D-to-A converter. Its performance is limited by quantization and sampling noise, setting an upper bound on phase noise. The KX3 has an advantage in that its synthesizer uses a DDS-driven PLL, while the Flex radios are DDS only, without a following PLL. But the TX DAC is still a factor in both cases.
>
> In a well-designed superhet like the K3, the TX DAC's noise floor affects only the in-band portion of the transmitted signal, i.e. the portion within the I.F. crystal filter passband. The K3 has two crystal filters in series in transmit mode, which results in very high rejection of noise outside the SSB passband (typically 100-3000 Hz). This means that the limiting factor on wideband transmit noise is not the IF injection into the mixer -- it really *is* the synthesizer.
>
> The K3's synthesizer is extremely clean at wide offsets (another DDS-driven PLL), and that is why its transmit phase noise is so low.
>
>
>> 3) Does phase noise go down if you use a faster master clock? The Flex 6700 uses a 983.04mhz vs a 122.99Mhz clock in the 6300.
>
> That is a completely different design (direct digital up/down conversion), which requires very high clock speeds, very expensive ADC and DAC components, and high current drain in receive mode. It can be made clean in transmit and receive modes, although in receive mode this architecture will still have typically 15-20 dB lower blocking dynamic range than a well-designed superhet. None of this is applicable to the KX3, which is obviously intended to be an ultraportable radio with low current drain and low cost.
>
>
>> 4) Do low phase noise radios allow in-band use, such as someone on CW on 20M and someone up on voice on 20M?
>
> Yes. In general this is only a problem if you have stations in very close proximity. In this situation, the K3 is better than any other radio on the market. This is why the K3 is highly favored for Field Day and DXpeditions.
>
> 73,
> Wayne
> N6KR
>
>
>