Elecraft

K3 to IC-7800 Comparison?

Classic

List

Threaded

69 messages Options

1234

Phil Wheeler-2

Re: Analog vs. Digital Front Ends

Re GPS: Hardly a "low" orbit, Brian: 'GPS
satellites fly in medium Earth orbit (MEO) at an
altitude of *approximately 20,200 km* (*12,550
miles*). Each satellite circles the Earth twice a
day.'

Phil W7OX

On 9/16/15 5:54 AM, brian wrote:

> 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

Jim Bolit

Re: Analog vs. Digital Front Ends

In reply to this post by alorona

Interesting.
The government is a big driver for wide range receive capabilities / signal analysis, but cost is not the main driver there.
Jim
W6AIM

-------- Original message --------
From: Lyle Johnson <[hidden email]>
Date: 9/16/2015 9:40 AM (GMT-06:00)
To: [hidden email]
Subject: Re: [Elecraft] Analog vs. Digital Front Ends

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.

______________________________________________________________
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]

Jim Brown-10

Re: Analog vs. Digital Front Ends

In reply to this post by Lyle Johnson

On Wed,9/16/2015 7:38 AM, Lyle Johnson wrote:
> 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?"

Exactly right. Back in the '90s, innovative designers of breakthrough
systems in the pro audio world piggy-backed onto high volume computer
professional networking hardware to transport audio signals in very
large sound systems (arenas, theme parks, etc.).

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]

Alan Bloom

Re: Analog vs. Digital Front Ends

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

On 09/16/2015 05:25 AM, Joe Subich, W4TV wrote:
...

> 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.

The amplitude distribution of a large number of signals of different
frequencies and amplitudes closely approximates Gaussian noise (see note
1 below). As a rule of thumb the peak to RMS voltage ratio of Gaussian
noise is about 5 or 6. Of course, theoretically it is infinity, but
peaks over about 5-6 are statistically rare (note 2), so that's a good
practical rule of thumb.

A voltage ratio of 5-6 is 14-15.6 dB. ADC overload should be rare as
long as you keep the RMS power 15 dB or so below the ADC full-scale.

A lot depends on what happens when the ADC overloads. If it simply
clips the signal, an occasional brief overload would not be
objectionable. That's how the ADC in the P3 works. On the other hand,
if it "wraps around" (for example, 0x7FFF = 32767 wraps to 0x8000 =
minus 32768) then it's obviously much more of a problem. In that case,
you would need an external analog clipper to keep the ADC from overloading.

Alan N1AL

Note 1:

The Central Limit Theorem:
https://en.wikipedia.org/wiki/Central_limit_theorem

Note 2:

The cumulative Gaussian probability density function is given in the
Wikipedia article:

https://en.wikipedia.org/wiki/Normal_distribution#Numerical_approximations_for_the_normal_CDF

If I calculated right then for x=5, CDF = (1 - 3*10^-8). If the ADC
sample rate is 100 Msps, then on average there would be 3 overloaded
samples per second. For x=6, CDF = (1 - 2.8*10^-11) so there would be
one overload about every 360 seconds.
______________________________________________________________
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 9/16/2015 2:02 PM, Alan wrote:
> The amplitude distribution of a large number of signals of different
> frequencies and amplitudes closely approximates Gaussian noise (see
> note 1 below). As a rule of thumb the peak to RMS voltage ratio of
> Gaussian noise is about 5 or 6. Of course, theoretically it is
> infinity, but peaks over about 5-6 are statistically rare (note 2),
> so that's a good practical rule of thumb.

We know N > 3 based on NC0B's test with the Flex-6700. 5-6 may be
reasonable if all signals are the same strength but what happens when
a very strong signal (that neighbor half mile away) preloads the ADC?

> A voltage ratio of 5-6 is 14-15.6 dB. ADC overload should be rare as
> long as you keep the RMS power 15 dB or so below the ADC full-scale.

Unfortunately, with practical DDC transceivers like the Flex-6000
series, that 15 dB below ADC full scale does very bad things to
sensitivity. Based on ARRL's April 2015 test, turning off +20 dB
preammp reduces sensitivity (MDS) from -135 dBm to -119 dBm while
the 20 KHz IMDDR3 goes from 94 dB (preamp on) to 103 dB (preamp off).
In practice, the user is faced with a choice between sensitivity and
IMD.

The real issue is that the -10 to -17 dBm clip point for the current
class of DDC devices is simply too low for full performance from MDS
to clipping. Considering that K9ZOA's Norton pre-amplifiers are rated
for *+19 dBm* at the 1 dB compression point, a well designed "analog"
transceiver has nearly 30 dB more headroom (while maintaining full
sensitivity) than the DDC products. This is the issue that the DDC
proponents can't escape (and don't want to address).

73,

... Joe, W4TV

On 9/16/2015 2:02 PM, Alan wrote:

> On 09/16/2015 05:25 AM, Joe Subich, W4TV wrote:
> ...
>> 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.
>
> The amplitude distribution of a large number of signals of different
> frequencies and amplitudes closely approximates Gaussian noise (see note
> 1 below). As a rule of thumb the peak to RMS voltage ratio of Gaussian
> noise is about 5 or 6. Of course, theoretically it is infinity, but
> peaks over about 5-6 are statistically rare (note 2), so that's a good
> practical rule of thumb.
>
> A voltage ratio of 5-6 is 14-15.6 dB. ADC overload should be rare as
> long as you keep the RMS power 15 dB or so below the ADC full-scale.
>
> A lot depends on what happens when the ADC overloads. If it simply
> clips the signal, an occasional brief overload would not be
> objectionable. That's how the ADC in the P3 works. On the other hand,
> if it "wraps around" (for example, 0x7FFF = 32767 wraps to 0x8000 =
> minus 32768) then it's obviously much more of a problem. In that case,
> you would need an external analog clipper to keep the ADC from overloading.
>
> Alan N1AL
>
>
> Note 1:
>
> The Central Limit Theorem:
> https://en.wikipedia.org/wiki/Central_limit_theorem
>
> Note 2:
>
> The cumulative Gaussian probability density function is given in the
> Wikipedia article:
>
> https://en.wikipedia.org/wiki/Normal_distribution#Numerical_approximations_for_the_normal_CDF
>
>
> If I calculated right then for x=5, CDF = (1 - 3*10^-8). If the ADC
> sample rate is 100 Msps, then on average there would be 3 overloaded
> samples per second. For x=6, CDF = (1 - 2.8*10^-11) so there would be
> one overload about every 360 seconds.
> ______________________________________________________________
> 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]
>

Bob McGraw - K4TAX

Re: Analog vs. Digital Front Ends

In reply to this post by Jim Bolit

As a side note on this topic, there is a company in Tennessee that
builds some really nice specialty receivers for various Governments.
{notice I stated Governments, as in plural}. They also have built or
build ham radios such as the Orion, Orion II, Omni VII, Eagle, Argonaut
VI just to name a few. Many of the technologies from the commercial
products are spun off into the ham products.

73
Bob, K4TAX
K3S s/n 10,163

On 9/16/2015 9:59 AM, Jim Bolit wrote:
> Interesting.
> The government is a big driver for wide range receive capabilities / signal analysis, but cost is not the main driver there.
> Jim
> W6AIM
>
>

______________________________________________________________
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]

alorona

Re: Analog vs. Digital Front Ends

In reply to this post by Alan Bloom

This is really good information, Alan, and makes sense. Regarding the approximation to Gaussian noise... given that most signals on a crowded band during a contest are highly compressed (their peak-to-average ratios are much smaller) would this make matters better or worse for the ADC, or no difference? I think I read that some signals during a contest have PAR's of 1 dB.
Al W6LX

> As a rule of thumb the peak to RMS voltage ratio of Gaussian
> noise is about 5 or 6, or 14-15.6 dB.
______________________________________________________________
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: Analog vs. Digital Front Ends

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

On 09/16/2015 01:15 PM, Joe Subich, W4TV wrote:

> On 9/16/2015 2:02 PM, Alan wrote:
>> The amplitude distribution of a large number of signals of different
>> frequencies and amplitudes closely approximates Gaussian noise (see
>> note 1 below). As a rule of thumb the peak to RMS voltage ratio of
>> Gaussian noise is about 5 or 6. Of course, theoretically it is
>> infinity, but peaks over about 5-6 are statistically rare (note 2),
>> so that's a good practical rule of thumb.
>
> We know N > 3 based on NC0B's test with the Flex-6700. 5-6 may be
> reasonable if all signals are the same strength but what happens when
> a very strong signal (that neighbor half mile away) preloads the ADC?

In that case the peak to RMS ratio actually gets less. In the limit
where the strong signal is much greater than the rest, the ratio
approaches 0 dB.

...
> The real issue is that the -10 to -17 dBm clip point for the current
> class of DDC devices is simply too low for full performance from MDS
> to clipping.

Yes, the blocking dynamic range of direct-sampling receivers using
current state-of-the-art ADCs is 15-20 dB worse than a top superhet
receiver like a K3. In addition to that, the "soft clipping" of the
analog circuitry probably gives some additional advantage to the superhet.

On the other hand, for casual operation you don't really need all that
dynamic range anyway. A QRP rig based on a NE602 has horrible dynamic
range, but you can still make a lot of contacts with it and have a lot
of fun. (For a lot less money. :=)

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]

Alan Bloom

Re: Analog vs. Digital Front Ends

In reply to this post by alorona

Hi Al,

The Central Limit Theorem says that it doesn't matter what the
distribution function is for each individual signal. As long as there
are a lot of them, the total has a Gaussian distribution. In fact, it
doesn't take very many to get quite close to Gaussian - something like 4
to 6 is generally sufficient if they are roughly equal in amplitude and
uncorrelated (different frequencies).

Alan N1AL

On 09/16/2015 02:26 PM, Al Lorona wrote:

> This is really good information, Alan, and makes sense. Regarding the
> approximation to Gaussian noise... given that most signals on a
> crowded band during a contest are highly compressed (their
> peak-to-average ratios are much smaller) would this make matters
> better or worse for the ADC, or no difference? I think I read that
> some signals during a contest have PAR's of 1 dB. Al W6LX
>
>
>> As a rule of thumb the peak to RMS voltage ratio of Gaussian noise
>> is about 5 or 6, or 14-15.6 dB.
>

1234