Elecraft

K3 to IC-7800 Comparison?

Classic

List

Threaded

69 messages Options

1234

Jim Brown-10

Re: Analog vs. Digital Front Ends

On Tue,9/15/2015 1:52 PM, Tony Estep wrote:
> On Tue, Sep 15, 2015 at 3:36 PM, Joe Subich, W4TV <[hidden email]> wrote:
>
>> ....noise power ratio testing hides the fact by not providing MDS values under each test condition *and* fails to indicate that even with *no preamplifier* the total noise signal is more than 10 dB *less*...
> =========
> I take this to mean that the test results are not comparable across radios, and the SDRs are 10db worse than they show up on the comparison chart. If this is true, it ain't good.

You've got the concept right -- the only difference is that 10 dB is not
a "hard" number -- it could be anything from 5 to 30 dB, depending on
how much test signal strength has been turned down, or preamp gain
removed, to get the input of the digital system out of clip!

Here are some comments by an engineer i trust.

73, Jim K9YC

= = = = =

- According to the ARRL, the Flex 6700 has 14 dB worse preamp-off BDR than the K3.

- If you turn the Flex's preamp ON, the BDR is probably 24 dB (or more) worse than the K3's. But if you turn the preamp on the 6700 OFF, its 2-kHz IMDDR3 is 8 dB worse than the K3's. You can't win

This all stems from the poor NF of their ADC and its low input voltage range. The MDS of the Flex is only -118 dBm (according to Sherwood) with the preamp off. (A K3's MDS with preamp off is -134 dB.) So you have to turn the Flex's preamp ON to get good MDS. The 2-kHz IMDDR is then good, but the BDR is terrible. Sherwood was so baffled by the weird responses he saw that he simply put "A/D Limit" in the BDR column for the Flex. He's still thinking about how to test it.

ARRL measured Flex's BDR, but only using the conventional method. To properly assess the damage caused by an ADC at the front end, you have to inject MANY signals into its input to give them a chance to combine in phase and hit the ADC's clipping threshold. And ADCs do NOT gracefully degrade.

______________________________________________________________
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 (N7WS)

Re: K3 RX in real world

In reply to this post by Hisashi T Fujinaka

When you turn on the noise blanker, all bets are off.

On 9/15/2015 2:50 PM, Hisashi T Fujinaka wrote:
>
> Exactly. Get on during CQWW. I could hear interesting transmitter widths
> from people half a continent away and interesting effects in my receiver
> as well (mostly noise blanker effects). I heard this all just on my
> dipole with no hams close by. Actually, K7VIT is 5 miles away but I can
> hardly hear him.
>

______________________________________________________________
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-4

Re: Analog vs. Digital Front Ends

In reply to this post by Jim Brown-10

I am going to go off on a wild tangent as a method of testing all
receivers for the ability to copy signals in the midst of very crowded
band conditions and/or the presence of noise. Whether those receivers
have analog front ends or an ADC.

An extremely crowded band could be simulated by broadband noise.
So take a medium strength single signal (say S-5) and inject it into the
receiver under test.
Now add a low amplitude broadband noise signal, and increase it until
the signal is buried in the broadband noise.

The receiver that can withstand the greater broadband noise level while
still recognizing the single signal "wins".
That says nothing about the overload of the ADC for those receivers that
put the ADC at the antenna - that is something for other test parameters.

73,
Don W3FPR

On 9/15/2015 6:28 PM, Jim Brown wrote:

______________________________________________________________
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]

Mike Markowski-2

Re: Analog vs. Digital Front Ends

Don,

The broadband jammer is the most traditional of all in electronic
warfare. If a receiver can withstand substantial AWGN (additive white
Gaussian noise), chirp, and CW jam, then that's the receiver for me. :-)

73,
Mike ab3ap

On 09/15/2015 06:51 PM, Don Wilhelm wrote:

> I am going to go off on a wild tangent as a method of testing all
> receivers for the ability to copy signals in the midst of very crowded
> band conditions and/or the presence of noise. Whether those receivers
> have analog front ends or an ADC.
>
> An extremely crowded band could be simulated by broadband noise.
> So take a medium strength single signal (say S-5) and inject it into the
> receiver under test.
> Now add a low amplitude broadband noise signal, and increase it until
> the signal is buried in the broadband noise.
>
> The receiver that can withstand the greater broadband noise level while
> still recognizing the single signal "wins".
> That says nothing about the overload of the ADC for those receivers that
> put the ADC at the antenna - that is something for other test parameters.
>
> 73,
> Don W3FPR

Johnny Siu

K3 to IC-7800 Comparison?

In reply to this post by Adam Farson

Yes, Adam did the test for both 2.7Khz 5 pole and 2.8Khz 8 pole.
73
Johnny Vr2XMC
寄件人︰ Adam Farson <[hidden email]>
收件人︰ 'Jim Bolit' <[hidden email]>
副本(CC)︰ 'Johnny Siu' <[hidden email]>; Elecraft List <[hidden email]>
傳送日期︰ 2015年09月16日 (週三) 12:03 AM
主題︰ RE: [Elecraft] K3 to IC-7800 Comparison?

#yiv3730855040 #yiv3730855040 -- _filtered #yiv3730855040 {panose-1:2 2 6 9 4 2 5 8 3 4;} _filtered #yiv3730855040 {font-family:PMingLiU;panose-1:2 2 5 0 0 0 0 0 0 0;} _filtered #yiv3730855040 {panose-1:2 4 5 3 5 4 6 3 2 4;} _filtered #yiv3730855040 {font-family:Calibri;panose-1:2 15 5 2 2 2 4 3 2 4;} _filtered #yiv3730855040 {font-family:Tahoma;panose-1:2 11 6 4 3 5 4 4 2 4;} _filtered #yiv3730855040 {panose-1:2 2 6 9 4 2 5 8 3 4;} _filtered #yiv3730855040 {panose-1:2 2 5 0 0 0 0 0 0 0;}#yiv3730855040 #yiv3730855040 p.yiv3730855040MsoNormal, #yiv3730855040 li.yiv3730855040MsoNormal, #yiv3730855040 div.yiv3730855040MsoNormal {margin:0cm;margin-bottom:.0001pt;font-size:12.0pt;}#yiv3730855040 a:link, #yiv3730855040 span.yiv3730855040MsoHyperlink {color:blue;text-decoration:underline;}#yiv3730855040 a:visited, #yiv3730855040 span.yiv3730855040MsoHyperlinkFollowed {color:purple;text-decoration:underline;}#yiv3730855040 p.yiv3730855040MsoAcetate, #yiv3730855040 li.yiv3730855040MsoAcetate, #yiv3730855040 div.yiv3730855040MsoAcetate {margin:0cm;margin-bottom:.0001pt;font-size:8.0pt;}#yiv3730855040 span.yiv3730855040BalloonTextChar {}#yiv3730855040 span.yiv3730855040EmailStyle19 {color:#1F497D;}#yiv3730855040 .yiv3730855040MsoChpDefault {font-size:10.0pt;} _filtered #yiv3730855040 {margin:72.0pt 72.0pt 72.0pt 72.0pt;}#yiv3730855040 div.yiv3730855040WordSection1 {}#yiv3730855040 Hi Jim, 2.7 kHz 5-pole. I have added this information to my chart. 73, Adam VA7OJ/AB4OJ

From: Jim Bolit [mailto:[hidden email]]
Sent: 15-Sep-15 05:13
To: Adam Farson; 'Johnny Siu'
Cc: Elecraft List
Subject: RE: [Elecraft] K3 to IC-7800 Comparison? Adam, Please confirm what roofing filter was used for your test on the K3. Tnx JimW6AIM .

-------- Original message --------
From: Adam Farson <[hidden email]>
Date: 9/14/2015 3:14 AM (GMT-06:00)
To: 'Johnny Siu' <[hidden email]>
Cc: Elecraft List <[hidden email]>
Subject: Re: [Elecraft] K3 to IC-7800 Comparison? Hi Johnny,

The K3 with the KSYN3A is in a closely-packed cluster of radios which score highest in the NPR test. Its exact ranking depends on which test frequency is being compared. The differences are very slight, mostly within a couple of dB.

In practice, I regard an NPR value of 80 dB or higher as excellent.

73, Adam VA7OJ/AB4OJ

From: Johnny Siu [mailto:[hidden email]]
Sent: 14-Sep-15 00:54
To: Adam Farson; Elecraft List
Subject: K3 to IC-7800 Comparison?

Hello Adam,

From reading of your NPR chart, could I understand that apart from IC7851, the K3 (with new KSYN3A installed) is the second best in terms of NPR test?

73

Johnny VR2XMC

_____

寄件人︰ Adam Farson <[hidden email]>
收件人︰ Elecraft List <[hidden email]>
傳送日期︰ 2015年09月14日 (週一) 2:51 PM
主題︰ Re: [Elecraft] K3 to IC-7800 Comparison?

For Jim K9YC:

Hi Jim,

Many thanks for the post on my presentations at the club meeting on Saturday
September 5.

To quote:

"There's an important caveat to his work. The NPR measurements require

very sophisticated band-stop filters in his instrumentation setup, and

based on the filters he has been able to source, that limits the

frequency range where he can do his measurements. An example is in the

footnote for the Flex-6700, which has no preselector for the range where

he had to do his measurements, which may have caused that radio to

measure worse than it would on the ham bands."

The instrumentation I use for NPR testing is re-purposed telecom test
equipment, as described in my Web article and also in my article in QEX for
March/April 2015. I rely on the surplus market for the test sets, and also
for the filter pairs (bandstop and bandpass). The bandstop filters typically
have 95 dB stopband attenuation and ~ 3 kHz stopband width.

A number of these filters are on (or near) amateur bands, e.g. 1940, 3886,
5340, 7600 and 11700 kHz. The first filter pair I acquired was on 5340 kHz,
so all the test data in my web article are on this frequency (which is in
the 60m band). The Flex-6700 does not have a preselector for this band, so
the noise loading will hit the front end and the ADC harder. This will
degrade the NPR reading by a few dB, but it will show how the receiver
behaves if heavily loaded on a band for which no preselector is fitted.

As I picked up additional filters, the number of frequencies on which I run
the test has steadily increased. Links to multi-frequency NPR data for
various radios (including the K3 with KSYN3A) are on my website:

http://www.ab4oj.com/test/main.html#NPR <http://www.ab4oj.com/test/main.html#NPR>

It was a pleasure meeting you guys over the Labour Day weekend.

73, Adam VA7OJ/AB4OJ

______________________________________________________________
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]

Kevin Stover

Re: K3 to IC-7800 Comparison?

In reply to this post by Guy Olinger K2AV

What???
Everybody doesn't get a trophy? ;-)

On 9/15/2015 1:31 PM, Guy Olinger K2AV wrote:
> A test designed to make everyone feel good about what they bought is
> like T-ball for the tots. T-ball is really great for the little kids.
> Love to watch 'em. But the top of the Sherwood list is the major
> leagues. There are those that will make the "post-season" and those
> that will go home. In the playoffs you make the pitches, you make the
> plays, you get the hits, or you go home. No underhand pitches because
> the batter is hurt and can't swing hard. 73, Guy

--
R. Kevin Stover
AC0H
ARRL
FISTS #11993
SKCC #215
NAQCC #3441

______________________________________________________________
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]

Guy Olinger K2AV

Re: Analog vs. Digital Front Ends

In reply to this post by Jim Brown-10

On Tue, Sep 15, 2015 at 6:28 PM, Jim Brown <[hidden email]>
quoted:

> This all stems from the poor NF of their ADC and its low input voltage
> range.

This just points out that given problems in hardware have capacity breakout
points. We are at a point where pure digital stuff to handle full range
alongside a K3 just isn't available at a price even remotely workable, even
for high end consumer gear. It might be "close", but it's not at the
breakout where those devices are at commodity prices.

I remember the days back in the 70's when for commercial computing reasons
we would have spent exorbitant money to get our hands on processor speed or
storage sizes that I currently have in my iPhone, which is commodity
pricing for processor speed and storage size and everything else. Price
point A/D range to match a K3S just isn't quite here. One of these years
everything will be direct sampling, and it will be very, very good. The
features and process simplifications that will flow off that will be
awesome.

At this point the top extreme range handling boxes are hybrid, both receive
and transmit. High end reliable solid state ham HF amps are just now
hitting close to 1500 watts. Still rumblings from the contest clubs about
solid state amp unreliability. Very touchy, if one exceeds power limits
(that are legal to hams) and limited impedance ranges you can blow stuff.
So we still have tube amps for brick on key 1500 watts. Our best RX are
still hybrid, if you throw ALL the testing at RX with no held back punches,
the direct sampling stuff still has issues at certain extremes. But it is
close enough that many could use them because their personal environment
doesn't tickle those susceptibilities.

What bothers me is trying to blow off those specific weakness as if they
weren't there.

I think that a direct sampling K4 is still pretty "out there". Don't think
that stuff is improving on a steep enough curve. But direct sampling that
works as well as a K3S *will* happen some day. Apparently needs another 25
dB of sampling range? But it will happen.

73, Guy.
______________________________________________________________
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]

Johnny Siu

K3 RX in real world

In reply to this post by Bill-3

Hello Bill,
Real world experience could be difference for different operator and I don't think someone is selling you away from the K-Line.
Radio is a science so that we can still begin with some numbers. Among the top notch radios, choice eventually becomes a user preference.
I once had the luxury to own all the real big Icoms and K3. I now downsize my station for other priority (getting funds to support the purchase of my Tesla Model S). I sold the big Icoms at good price and now just keep the K3 for occasional contest operation.
I am not the one who always clap my hands towards elecraft but I will keep Wayne and Eric busy in improving their products. In the past, I shout loudly about the artifacts and poor audio of K3. Now in the K3S, they get better audio section for the radio.
One thing, I do appreciate that elecraft gives us the chance to upgrade our old K3 (though not cheap). The same philosophy is also found in Tesla. The owners have the chance to update their cars.
Enjoy your K-line and so will I.
73
Johnny VR2XMC
寄件人︰ Bill <[hidden email]>
收件人︰ [hidden email]
傳送日期︰ 2015年09月16日 (週三) 4:29 AM
主題︰ Re: [Elecraft] K3 RX in real world

Real World - that is all that really counts. All the testing and arguing
that has been appearing of late is great, if you are a numbers cruncher
and chart maker. But, to my thinking - the technical performance
differences between the top rigs can only be discerned in a "copper" room.

Real world experience with QRM, QRN, good or poor antennas, band
conditions are all the great equalizers. Reported real world experience
- comparing one rig (A to B) - is of far greater value to me than a list
of numbers.

Lots of explanations why one rig is preferred over another: handles
QRM/QRN better, has more pleasing audio, better/worse ergonomics, more
??? for the money, too many menus, great audio reports, great
keyer/break-in, eye candy, etc.

Of course, all that said - it would be a very hard sell to get me away
from my K3-Line. But, it would be interesting reading never-the-less.

Bill W2BLC K3-Line

______________________________________________________________
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]

Lynn W. Taylor, WB6UUT

Re: Analog vs. Digital Front Ends

In reply to this post by Joe Subich, W4TV-4

On 9/15/2015 1:36 PM, Joe Subich, W4TV wrote:

I agree with you. What I'm saying isn't about why an SDR or an Analog
receiver might have problems.

If we're testing receivers, we should treat them as a black box and not
look into the ciruitry to see how they behave.

After the test, then we could look inside and say "oh, the ADC is the
limiting factor."

73 -- Lynn
______________________________________________________________
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]

Joe Subich, W4TV-4

Re: Analog vs. Digital Front Ends

In reply to this post by Jim Brown-10

> - According to the ARRL, the Flex 6700 has 14 dB worse preamp-off
> BDR than the K3.
>
> - If you turn the Flex's preamp ON, the BDR is probably 24 dB (or
> more) worse than the K3's. But if you turn the preamp on the 6700
> OFF, its 2-kHz IMDDR3 is 8 dB worse than the K3's. You can't win

Here are the comparisons I pulled from ARRL's reviews of the Flex-6700,
Flex-6300 (April 2015) and K3 (January 2009) and VA7OJ/AB4OJ's noise
power ratio testing (http://www.ab4oj.com/test/main.html#NPR):

ARRL Labs shows the *blocking dynamic range* of the Flex-6700 to be
much *worse* than the K3 at all spacings with the preamps off. I've
added the MDS for comparison. Note: ARRL tested the original K3 with
KSYN3 *not the new KSYN3A*.

(preamp off) 20 KHz 5 KHz 2 KHz MDS
Flex-6700 126 dB 126 dB 126 dB -119 dBm
Flex-6300 127 dB 127 dB 126 dB -119 dBm
K3 142 dB 140 dB 139 dB -130 dBm

(preamp on) 20 KHz MDS
Flex-6700 (+20) 130 dB -135 dBm
Flex-6300 127 dB -128 dBm
K3 (+10) 138 dB -138 dBm

IMDDR3 is difficult to compare since ARRL did not report 2 KHz IMDDR3 with
the preamp turned on. This means that the Flex MDS is more than 10 dB
worse
than the K3. Note that the Flex-6700 drops by nearly 9 dB at 20 KHz if
the preamp (+10 dB) is turned on.

IMDDR3 20 KHz/Off 20 KHz/On 5 KHz/Off 2 KHz/off
Flex-6700 103 dB 94 dB 103 dB 103 dB
Flex-6300 92 dB 89 dB 92 dB 92 dB
K3 106 dB 103 dB 105 dB 103 dB

Using a different test protocol (Noise Power Ratio), AB4OJ/VA7OJ reports:

Noise Power Ratio 160M 80M 60M 40M
Flex-6700 no preamp 71 dB 73 dB 75 dB 74 dB
Flex-6700 +20 dB 76 dB 71 dB 71 dB 70 dB
K3 no preamp 82 dB 82 dB 82 dB not measured
K3 Preamp On 83 dB 82 dB 80 dB not measured

Note: the Flex was tested with 10 to 15 less noise power signal than K3 to
prevent ADC clipping.

73,

... Joe, W4TV

On 9/15/2015 6:28 PM, Jim Brown wrote:

> On Tue,9/15/2015 1:52 PM, Tony Estep wrote:
>> On Tue, Sep 15, 2015 at 3:36 PM, Joe Subich, W4TV <[hidden email]>
>> wrote:
>>
>>> ....noise power ratio testing hides the fact by not providing MDS
>>> values under each test condition *and* fails to indicate that even
>>> with *no preamplifier* the total noise signal is more than 10 dB
>>> *less*...
>> =========
>> I take this to mean that the test results are not comparable across
>> radios, and the SDRs are 10db worse than they show up on the
>> comparison chart. If this is true, it ain't good.
>
> You've got the concept right -- the only difference is that 10 dB is not
> a "hard" number -- it could be anything from 5 to 30 dB, depending on
> how much test signal strength has been turned down, or preamp gain
> removed, to get the input of the digital system out of clip!
>
> Here are some comments by an engineer i trust.
>
> 73, Jim K9YC
>
> = = = = =
>
> - According to the ARRL, the Flex 6700 has 14 dB worse preamp-off BDR
> than the K3.
>
> - If you turn the Flex's preamp ON, the BDR is probably 24 dB (or more)
> worse than the K3's. But if you turn the preamp on the 6700 OFF, its
> 2-kHz IMDDR3 is 8 dB worse than the K3's. You can't win
>
> This all stems from the poor NF of their ADC and its low input voltage
> range. The MDS of the Flex is only -118 dBm (according to Sherwood) with
> the preamp off. (A K3's MDS with preamp off is -134 dB.) So you have to
> turn the Flex's preamp ON to get good MDS. The 2-kHz IMDDR is then good,
> but the BDR is terrible. Sherwood was so baffled by the weird responses
> he saw that he simply put "A/D Limit" in the BDR column for the Flex.
> He's still thinking about how to test it.
>
> ARRL measured Flex's BDR, but only using the conventional method. To
> properly assess the damage caused by an ADC at the front end, you have
> to inject MANY signals into its input to give them a chance to combine
> in phase and hit the ADC's clipping threshold. And ADCs do NOT
> gracefully degrade.
>
>
> ______________________________________________________________
> 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]
>

Guy Olinger K2AV

Re: K3 RX in real world

In reply to this post by Hisashi T Fujinaka

One thing I notice with better and better RX with less and much less RX
generated cr*p, is that TX/amp generated cr*p is presented quite a bit
clearer with much less disguise. Once upon a time that was all "band noise"
to us. Ain't clarity interesting?

What is really cool is to listen to an SSB pileup in correctly done K3
diversity and hear cr*p coming from the same point on the sound stage as
just one of the voices, not at all spread around like the true "band
noise".

73, Guy K2AV

On Tue, Sep 15, 2015 at 5:50 PM, Hisashi T Fujinaka <[hidden email]>
wrote:

> On Tue, 15 Sep 2015, Jim Brown wrote:
>
> On Tue,9/15/2015 1:29 PM, Bill wrote:
>>
>>> the technical performance differences between the top rigs can only be
>>> discerned in a "copper" room.
>>>
>>
>> That is definitely not true. Digging weak ones out in a big contest, or
>> with a close by ham neighbor, will tell you right away. An inferior dirty
>> radio a few miles away driving a power amp will make your life miserable.
>>
>
> Exactly. Get on during CQWW. I could hear interesting transmitter widths
> from people half a continent away and interesting effects in my receiver
> as well (mostly noise blanker effects). I heard this all just on my
> dipole with no hams close by. Actually, K7VIT is 5 miles away but I can
> hardly hear him.
>
> --
> Hisashi T Fujinaka - [hidden email]
> BSEE + BSChem + BAEnglish + MSCS + $2.50 = coffee
>
> ______________________________________________________________
> 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]
>

Eric Swartz - WA6HHQ

Re: Analog vs. Digital Front Ends

Administrator

In reply to this post by Joe Subich, W4TV-4

I think what everyone is missing here is that Adam has clearly stated that his
tests were not designed to be tests to directly compare analog and direct
sampling radios, and certainly they were not intended by Adam for direct
comparisons to the tests run by the ARRL and Sherwood. They were originally
intended as -additional- lab tools for technical people and those like us here
at Elecraft and Flex, ICOM, Yaesu etc to use for evaluation of our designs going
forward. We can certainly take additional data in the lab to make these tests
even more useful as comparisons between different technologies etc.

Your comment about including the mds of the rig and the A/D limit level at the
settings used for each of his tests will help a lot though in comparing the
results. I'd also like to see that.

One other note - the 2.7 kHz SSB filter he used on his test of the K3 (without
new synth) does not have as good of a stop band and shape factor as our 2.8 kHz
and narrower 8-pole SSB filters. Thereis a big difference between thge two for
this type of testing. I'd like to see the noise power test data for a K3 (or
K3S) with new synth and an 8 pole 2.8 kHz or narrower filter. We did
surprisingly well with the 2.7 kHz 5-pole in his tests. I know what he will see
with the 8-pole filter and new low noise Synth. :-)

73,
Eric
/elecraft.com/

On 9/15/2015 1:36 PM, Joe Subich, W4TV wrote:

>
>
>
> On 9/15/2015 3:39 PM, Lynn W. Taylor, WB6UUT wrote:
>> Maybe I'm wrong, but if the thing we're testing is supposed to be a
>> radio, and we want to compare how radios work under conditions we'd
>> encounter in actual use, it just seems intuitively obvious to the most
>> casual observer that the tests should be the same.
>>
>> 73 -- Lynn
>
>
> Therein lies the rub ... direct digital conversion SDR devices have
> their best performance (best sensitivity) at (composite) signal levels
> just below the clip (overflow) point of the ADC in use. However, at
> that point they can not withstand any overload. If one decreases the input
> signal (e.g., turns off the preamplifier) to provide headroom or
> avoid ADC overflow, the sensitivity (MDS) suffers in direct proportion
> to the decrease in signal level.
>
> ARRL's review of the Flex-6700 and Flex-6300 shows this inescapable
> fact of nature (law of physics) very clearly. AB4OJ's noise power
> ratio testing hides the fact by not providing MDS values under each
> test condition *and* fails to indicate that even with *no preamplifier*
> the total noise signal is more than 10 dB *less* that that used for
> testing with traditional superhetrodyne or downcoversion DSP designs.
>
> 73,
>
> ... Joe, W4TV
> ______________________________________________________________
> 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]
>

Joe Subich, W4TV-4

Re: Analog vs. Digital Front Ends

On 9/15/2015 9:13 PM, Eric Swartz - WA6HHQ, Elecraft wrote:
> I think what everyone is missing here is that Adam has clearly
> stated that his tests were not designed to be tests to directly
> compare analog and direct sampling radios, and certainly they were
> not intended by Adam for direct comparisons to the tests run by the
> ARRL and Sherwood. They were originally intended as -additional- lab
> tools for technical people and those like us here at Elecraft and
> Flex, ICOM, Yaesu etc to use for evaluation of our designs going
> forward.

When one makes the same tests and reports the data in the same format
the reader is naturally going to compare the numbers of one type of
radio (upconversion or downconersion superhet with analog product
detector or DSP demodulator) with others (direct digital conversion
DSP). It is up to the test designer to make those tests directly
comparable *or* very clearly state the difference and why the two can
not be compared directly.

There is certainly no indication of the differences/incompatibilities
in Adam's "Multi-Band NPR Test Results". The difference in test
methodologies is documented only in "Noise Power Ratio (NPR) Testing
of HF Receivers" and even that document fails to clearly show the
differences and *why* the test results of traditional receivers can
not be compared directly with those of DDC SDRs.

Simply providing MDS for each test condition in each test would go a
long way to establishing comparability. Identifying "Ptot dBm" and
"Noise Loading dBm" consistently across all receiver topologies
*along with* MDS would be even better to establishing a common set
of test parameters that would allow the various topologies to be
*compared directly* for the ability to copy a signal at the [internal]
noise floor in the presence of strong adjacent channel interference
which is really what *all* of the receiver comparison tests are about.

Real world performance is not about one parameter in a vacuum.

> One other note - the 2.7 kHz SSB filter he used on his test of the K3
> (without new synth) does not have as good of a stop band and shape
> factor as our 2.8 kHz and narrower 8-pole SSB filters.

Actually the tests of K3 S/N 5812 are indicated as using the KSYN3A.

However, even comparisons of "analog" receivers are murky ... one needs
to look carefully to see than Noise Loading of the Icom 7800 and 7851
*drops* by almost the same amount as the preamp gain (just like a DDC)
where the noise loading of the K3 varies by only a small part of the
added gain. I won't even point out that the 7800 beats the 7851 on
160 meters - another anomalous result.

73,

... Joe, W4TV

On 9/15/2015 9:13 PM, Eric Swartz - WA6HHQ, Elecraft wrote:

> I think what everyone is missing here is that Adam has clearly stated
> that his tests were not designed to be tests to directly compare analog
> and direct sampling radios, and certainly they were not intended by Adam
> for direct comparisons to the tests run by the ARRL and Sherwood. They
> were originally intended as -additional- lab tools for technical people
> and those like us here at Elecraft and Flex, ICOM, Yaesu etc to use for
> evaluation of our designs going forward. We can certainly take
> additional data in the lab to make these tests even more useful as
> comparisons between different technologies etc.
>
> Your comment about including the mds of the rig and the A/D limit level
> at the settings used for each of his tests will help a lot though in
> comparing the results. I'd also like to see that.
>
> One other note - the 2.7 kHz SSB filter he used on his test of the K3
> (without new synth) does not have as good of a stop band and shape
> factor as our 2.8 kHz and narrower 8-pole SSB filters. Thereis a big
> difference between thge two for this type of testing. I'd like to see
> the noise power test data for a K3 (or K3S) with new synth and an 8 pole
> 2.8 kHz or narrower filter. We did surprisingly well with the 2.7 kHz
> 5-pole in his tests. I know what he will see with the 8-pole filter and
> new low noise Synth. :-)
>
> 73,
> Eric
> /elecraft.com/
>

bs usb

Re: Analog vs. Digital Front Ends

In reply to this post by Don Wilhelm-4

On Tue, 15 Sep 2015 18:51:31 -0400
Don Wilhelm <[hidden email]> wrote:

That sounds like a bias free test to me.

It also raises a question. If the ADC performance can be so completely destroyed by overloading why would you put it at the antenna where it most likely could be overloaded?

With regard to analog front ends, I used to think that antenna tuners were crutches for folk not capable of building resonant antenas. In fact they may serve as analog front ends. Of course they need to be in the signal path to serve on receive. Not all rigs include the ATU in the receiver signal path.
______________________________________________________________
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: Analog vs. Digital Front Ends

In reply to this post by Lynn W. Taylor, WB6UUT

On Tue,9/15/2015 5:48 PM, Walter Underwood wrote:
> This might be an orthogonal opinion, but I think it is awesome that high-end direct sampling receivers are competitive with mid-range superhets.

But they are NOT competitive in strong signal environments! That's the
point of this discussion -- the measurement system must give them a
10-20 dB handicap to make them LOOK competitive.

Someday they will be, and many of us will be using them.

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]

Wes (N7WS)

Re: K3 to IC-7800 Comparison?

In reply to this post by Jim Bolit

I'll be. "You are digging a deeper hole," is now a critique by a peer.

On 9/15/2015 1:31 PM, Jim Bolit wrote:

> Yes, but many can be misled with "official looking" information that has cover
> statements buried in the content.
>
> Anyone who publishes information of this nature should be prepared to have it
> critiqued by their peers.
>
> Based on this mail string, it is clear there are numbers and methodologies
> that are in conflict, raising questions by peers.
>
> Jim
>
> W6AIM
>
> .
>

Alan Bloom

Re: Analog vs. Digital Front Ends

In reply to this post by Jim Brown-10

> Someday they will be, and many of us will be using them.

I think it was back around 1993 that there was a discussion on the
Internet about using a wide-band ADC to replace the front end of an HF
receiver. I commented then that you just couldn't get good enough
performance with affordable ADCs available at that time, but someday you
might.

Time marches on and now ADCs are available that give usable performance
at a reasonable price. They still can't compete with a state-of-the-art
receiver like a K3, but they are as good as many analog radios of a few
decades ago and are quite reasonable for casual operation or even for
contesting or DXing if you don't have a very high-signal environment.

So how long before ADC technology catches up to the K3? If Moore's law
applied (doubling of performance every couple years) it wouldn't be
long. Unfortunately Moore's law applies mainly to digital circuitry but
the key parts of an ADC are analog (the "A" in "ADC"). Unless there is
a big theoretical breakthrough in ADC architecture, I think we still
have some years to wait.

Alan N1AL

On 09/15/2015 08:49 PM, Jim Brown wrote:

> On Tue,9/15/2015 5:48 PM, Walter Underwood wrote:
>> This might be an orthogonal opinion, but I think it is awesome that
>> high-end direct sampling receivers are competitive with mid-range
>> superhets.
>
> But they are NOT competitive in strong signal environments! That's the
> point of this discussion -- the measurement system must give them a
> 10-20 dB handicap to make them LOOK competitive.
>
> Someday they will be, and many of us will be using them.
>
> 73, Jim K9YC

Joe Subich, W4TV-4

Re: Analog vs. Digital Front Ends

On 9/16/2015 12:43 AM, Alan wrote:
> So how long before ADC technology catches up to the K3? If Moore's
> law applied (doubling of performance every couple years) it wouldn't
> be long. Unfortunately Moore's law applies mainly to digital
> circuitry but the key parts of an ADC are analog (the "A" in "ADC").
> Unless there is a big theoretical breakthrough in ADC architecture, I
> think we still have some years to wait.

The question really boils down to when will we see affordable high
voltage ADCs. The problem is that every signal in the input passband
is in series and the [peak] voltage adds as a vector sum. So if one
S9+63dB (-10 dBm) signal causes the ADC to overflow (consistent with
the noise loading limit in the AB4OJ/VA7OJ tests of the Flex-6000
series) the ADC limit is 0.1V (peak) at the antenna. Spread that 0.1V
across multiple [equal strength] signals and you see the following:

N "S"
----------------------------------
1 S9 +63 dB (-10 dBm)
3 S9 +53 dB
10 S9 +43 dB
32 S9 +33 dB
100 S9 +23 dB
316 S9 +13 dB
~450 S9 +10 dB
1000 S9 +3 dB
~1400 S9 +0 dB (-73 dBm)

Since it is the instantaneous peaks that cause problems, increasing the
number of signals decreases the frequency of the ADC overflows. There
is certainly analysis that can be done to compute the probability of a
peak given a specific number of signals and frequency distribution but
my best guess is that the number of signals involved will be somewhere
between 10 and 100. Of course, if one of those signals is very strong
- a 1.5 KW neighbor 0.5 miles away or a broadcast station that puts
a few volts on any wire in the air - it does not take many (if any)
additional S9+10 dB signals to push the system over the edge.

73,

... Joe, W4TV

______________________________________________________________
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]

alsopb

Re: Analog vs. Digital Front Ends

Look at the A/D converter as a chip tool that could potentially be used
in ham RX design.

If one wanted a ham band RX using an A/D front end, certainly one would
add front end filters.

Of course, that would only help solve the problem with out of band
signals. Handling the vector sum of in band signals would still have to
be designed for.

The concept however is a good one. Who knows where these A/D chips will
be in 2, 5 or 10 years.

I remember when GPS was in its infancy. Many people thought it was
crazy to think that a bunch of low orbit non-geosynchronous satellites
could be used for extremely accurate positioning. Besides the ground
equipment required would be too expensive for most users. History
proves them wrong.

73 de Brian/K3KO

On 9/16/2015 12:25 PM, Joe Subich, W4TV wrote:

> On 9/16/2015 12:43 AM, Alan wrote:
>> So how long before ADC technology catches up to the K3? If Moore's
>> law applied (doubling of performance every couple years) it wouldn't
>> be long. Unfortunately Moore's law applies mainly to digital
>> circuitry but the key parts of an ADC are analog (the "A" in "ADC").
>> Unless there is a big theoretical breakthrough in ADC architecture, I
>> think we still have some years to wait.
>
> The question really boils down to when will we see affordable high
> voltage ADCs. The problem is that every signal in the input passband
> is in series and the [peak] voltage adds as a vector sum. So if one
> S9+63dB (-10 dBm) signal causes the ADC to overflow (consistent with
> the noise loading limit in the AB4OJ/VA7OJ tests of the Flex-6000
> series) the ADC limit is 0.1V (peak) at the antenna. Spread that 0.1V
> across multiple [equal strength] signals and you see the following:
>
> N "S"
> ----------------------------------
> 1 S9 +63 dB (-10 dBm)
> 3 S9 +53 dB
> 10 S9 +43 dB
> 32 S9 +33 dB
> 100 S9 +23 dB
> 316 S9 +13 dB
> ~450 S9 +10 dB
> 1000 S9 +3 dB
> ~1400 S9 +0 dB (-73 dBm)
>
> Since it is the instantaneous peaks that cause problems, increasing
> the number of signals decreases the frequency of the ADC overflows.
> There
> is certainly analysis that can be done to compute the probability of a
> peak given a specific number of signals and frequency distribution but
> my best guess is that the number of signals involved will be somewhere
> between 10 and 100. Of course, if one of those signals is very strong
> - a 1.5 KW neighbor 0.5 miles away or a broadcast station that puts
> a few volts on any wire in the air - it does not take many (if any)
> additional S9+10 dB signals to push the system over the edge.
>
> 73,
>
> ... Joe, W4TV
>
>
> ______________________________________________________________
> 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]
>
>
> -----
> No virus found in this message.
> Checked by AVG - www.avg.com
> Version: 2015.0.6086 / Virus Database: 4419/10649 - Release Date:
> 09/16/15
>
>

Lyle Johnson

Re: Analog vs. Digital Front Ends

In reply to this post by Joe Subich, W4TV-4

One has to ask, "Which large consumer of high dynamic range, high-speed
sampling ADCs is requiring more performance than presently exists? What
drove the market for the present ones?"

The driver for the current generation, based on limited information, is
the technical requirements for cellular base stations. Perhaps as we
move from the current generation to the next generation of mobile
devices, we'll see a need arise for better base station performance.

To get another 15 to 18 dB of blocking dynamic range (BDR), we need 3
more effective bits. This either comes from more bits at conversion
time (a 20-bit high-speed ADC instead of a 16-bit), increasing the
sampling rate dramatically while maintaining the same effective number
of bits (ENOB as it is called in data sheets), or a combination of
both. And of course you also need the downstream digital devices,
usually field programmable gate arrays (FPGAs) to implement what we
think of as the receiver (oscillators and mixers and filters) that can
handle the required interfaces at the necessary speeds, and the internal
resources to maintain precision).

Until there is a viable market for a large number of such ADCs,
semiconductor manufacturers are unlikely to invest a lot in R and D to
get there. At this point, they want to compete with each other for the
existing market, so they try to offer incremental advantages over their
competitors.

We may eventually be able to buy ADCs with the required performance to
obtain the BDR many of us want and some of us need, but unless there is
a large demand for the products whose design needs include ADCs that
will provide this performance, there is little incentive for their
development.

My personal opinion only,

Lyle KK7P

>
>> So how long before ADC technology catches up to the K3? If Moore's
>> law applied (doubling of performance every couple years) it wouldn't
>> be long. Unfortunately Moore's law applies mainly to digital
>> circuitry but the key parts of an ADC are analog (the "A" in "ADC").
>> Unless there is a big theoretical breakthrough in ADC architecture, I
>> think we still have some years to wait.
>
> The question really boils down to when will we see affordable high
> voltage ADCs.

1234