[Elecraft] [OT] Dayton- new hf rigs?

[Joe Subich, W4TV]

Terry,

 > Your second assumption brings up a valid point, but not the intended
 > one. Yes, a Mac Mini probably won't work well.

My point there was the lack of multiple monitor support - Apple is a
closed system and the Mac Mini has only one video output.

> Not the addition of both signals, as you seem to imply. Think vector
> sums of AC voltages, not DC summing as a battery.

That is exactly my point.  An ADC must be able to handle the absolute
largest peak voltage for all signals at the same time.  If it can not
handle that peak voltage without overflow, the result is IMD that no
amount of decimation will correct.

> Granted, there may be instantaneous flashes where signals within the
> passband combine, but it's not constant like DC voltages, as your
> math implies. Than goodness we aren't all on the same frequency and
> the same phase.

While the chance of every signal adding to the maximum level is small
and the duration of any peak is certainly short, it never the less
represents a finite probability. It is exactly the reason that HDTV
television transmitters and cell phone "base station" transmitters have
power amplifiers that are rated for significant peak to average ratios
and generally need to run much higher standing currents (much closer to
class A than class C) to produce acceptable IMD/distortion/linearity
specifications.

If the receiver designer chooses to use AGC to prevent the ADC from
overflow, the AGC must have a very fast attack and will suffer from
the very issues of "window desensitization" we see with conventional
DSP radios when the DSP bandwidth is less than the roofing filter
bandwidth.

> Once again, go do research on Digital DownConversion, decimation,
> SDR, processing gain.

I have been ... I'm stuck in Charleston, WV due to car trouble on my
way home from Dayton.  I've been using the time to reread much of the
information on DDC, Decimation, etc. as well as discuss the issues off-
line with some of the smartest DSP people out there.  Decimation can
certainly improve Blocking Dynamic Range beyond that provided by the
"number of bits" ... higher sample rates and increased processing can
add to the effective dynamic range *IF* the ADC is never permitted to
overflow *or* AGC doesn't drop the weakest signals below the composite
noise floor.

I'm not questioning whether the Flex is interesting technology, I am
questioning whether the claimed 150 dB [blocking] dynamic range (no
claims about IMD DR) is sufficient to maintain sensitivity to below
-140 dBm if the ADC can be expected to see peak signal levels
approaching 20 dBm under normal (contest and/or low band) operating
conditions.

> And, once again, go to the Flex web site on-line store. Don't
> believe me. In black and white (not a tee shirt), it says the 6500 is
> $3,999 as a pre-order, with a $1,200 deposit required to pre-order.
> Click the button yourself. You might like the rig.

I'm said I like the concept and $3999 for the 6500 makes it more 
interesting - I was not able to find that price when I looked at the
web site (and downloaded the pdf brochure) before leaving Dayton.  I
do think there are some significant values in the "dual front end"
(6700) radio - like null steering, true duplex operation for "SO2R
in a box", and diversity.  However, I think the Flex products would
make more appropriately priced at $2999/$4999 when compared to other
amateur products than even the current $3999/$6999.

If I could get the 6700 for $4999 with updates when desired for < $150
and I could be sure that the front end would not fall apart or there
wouldn't be a blocking issue with multiple strong signals (or 5V from
the local 50KW AM station) the 6700 would be on my short list of
"future acquisitions".

 > Sorry for the off-topic bandwidth.

I don't know that issues of DSP and/or receiver performance are
off-topic but that's up to Eric.

73,

    ... Joe, W4TV


On 5/21/2012 3:07 PM, wb4jfi at knology.net wrote:
> First of all, I do not use every bell and whistle on my K3, and I don't
> have every option available on my K3, even though it's nicely loaded to
> me. For example, I do not have the second receiver. So, I do not use my
> existing rig to it's "maximum use". Therefore, to extrapolate that I
> would need to take advantage of every morsel in the 6700 or 6500 is
> apples to oranges.
>
> Your second assumption brings up a valid point, but not the intended
> one. Yes, a Mac Mini probably won't work well. But, not due to a lack of
> horsepower. Flex's initial software is for Windows. Yes, I know all
> about using Parallels and the other software thing that lets you run
> Windows programs on Mac. I have a Macbook here as well as Windows and
> Linux. Macs are notoriously expensive for what you get. I bet a
> $700-$900 (excluding monitor, since you priced just the Mini) Intel i5
> or i7 and Windows box would work very nice with a Flex 6000 series. Feel
> free to yell at Flex for not providing Mac support.
>
> Joe, your math below assumes that every signal is at the same frequency
> and the same phase in order to just add the levels, otherwise things
> don't just add up like that. For example, even if two signals (both
> +50dBm) are on the exact same frequency, but exactly opposite phases,
> they WILL CANCEL each other out, you will end up with NO dBm. Nada. Not
> the addition of both signals, as you seem to imply. Think vector sums of
> AC voltages, not DC summing as a battery. Granted, there may be
> instantaneous flashes where signals within the passband combine, but
> it's not constant like DC voltages, as your math implies. Than goodness
> we aren't all on the same frequency and the same phase.
>
> Back to the main point, the new Flex radios DO come with traditional
> filtering, if you need it. In Europe or near an AM broadcast station, it
> may be necessary. However, I am sure that the receiver will have
> significantly better dynamic range than the 96dB implied by a 16-bit RF
> A/D converter. THAT's the point. 16-bits only gets you 96dB minimum to
> maximum without "distortion". If there's any more dBs found, even in the
> harshest test, there is processing gain. And most of the DSP and SDR
> world seems to believe in processing gain due to decimation.
>
> Once again, go do research on Digital DownConversion, decimation, SDR,
> processing gain.
>
> And, once again, go to the Flex web site on-line store. Don't believe
> me. In black and white (not a tee shirt), it says the 6500 is $3,999 as
> a pre-order, with a $1,200 deposit required to pre-order. Click the
> button yourself. You might like the rig.
>
> I agree that $200 per year for software support is a bit steep. BTW, I
> don't think the rig stops working if you don't pay, you just aren't
> entitled to updates. Oh, wait. My Macbook requires me pay for EVERY OS
> upgrade, even in-between ones. I had to buy Snow Leopard just to have
> the right to upgrade from Leopard to Lion. And that's on a dumb Macbook,
> which costs a lot less that a Flex 6500. Maybe I should shut up now.
>
> Reflector, I am done with this. Long live my K3! I have, and will
> continue to defend my K3 on the Flex reflector just as vigorously. Sorry
> for the off-topic bandwidth.
> 73,
> Terry, WB4JFI
>