I've got, and love, an RFSpace Inc. SDR-IQ software-defined receiver.
It covers 100 Hz to 30 MHz, and lets me eyeball any 190 KHz (or
narrower) chunk all at once, which is rather particularly nice. I'd
like to be able to use it on 30 to 60 MHz and higher, which means that
I need a downconverter. It seems to me that a suitable downconverter
would let me set a switch to choose which 30 MHz band got mixed down
to the roughly 0-30 MHz baseband.

I'd like to have this work up through the 70 cm. ham band for sure,
and as much higher as I can get. I know that construction techniques
change as one gets above HF, and change _radically_ at UHF and above.
How hard is this going to be, overall?

The general setup I envision is something rather like

LO
n*30MHz
(n to |
(n+1)) |
* 30 MHz 30 MHz wide V 0 to 30 MHz
input -- bandpass filter -- mixer -- output
to SDR-IQ

Some questions for the (perhaps not older: I'm 60) wiser heads with
more design-and-construction experience: is this at all a reasonable
thing to do? Does anyone have any ideas on suitable mixers? Am I
looking at a disparate-enough set of frequency bands that I will wind
up having to use multiple mixers, possibly one per 30 MHz band, or
per octave? It is trivially obvious to me that I'll need different
antennas.

Is it reasonable to look at using a comb generator which is driven by
a 30 MHz reference, so that I just take the appropriate harmonic and
amplify it to a power level that suits the mixer I have to use for
that particular 30 MHz band?

--
Mike Andrews, W5EGO

Tired old sysadmin

Mike Andrews wrote on Apr 9, 6:26 am:

I've got, and love, an RFSpace Inc. SDR-IQ software-defined receiver.
It covers 100 Hz to 30 MHz, and lets me eyeball any 190 KHz (or
narrower) chunk all at once, which is rather particularly nice. I'd
like to be able to use it on 30 to 60 MHz and higher, which means that
I need a downconverter. It seems to me that a suitable downconverter
would let me set a switch to choose which 30 MHz band got mixed down
to the roughly 0-30 MHz baseband.

I'd like to have this work up through the 70 cm. ham band for sure,
and as much higher as I can get. I know that construction techniques
change as one gets above HF, and change _radically_ at UHF and above.
How hard is this going to be, overall?

Offhand, NOT hard at all...with the aid of some MiniCircuits
components such as the mixer and perhaps an input low-pass
filters. The down-converter's LO is going to be the bear
if you want frequency accuracy.

The general setup I envision is something rather like

LO
n*30MHz
(n to |
(n+1)) |
* 30 MHz 30 MHz wide V 0 to 30 MHz
input -- bandpass filter -- mixer -- output
to SDR-IQ

That's the general block diagram of every down-converter. :-)
With perhaps the exception of the "30 MHz wide BPF." Nearly all
wideband spectrum analyzers use a lowpass filter there with the
LO riding ABOVE the input frequency; image will be on the "high
side" and that is blocked-out by the LPF.

Some questions for the (perhaps not older: I'm 60) wiser heads with
more design-and-construction experience: is this at all a reasonable
thing to do?

Absoutely reasonable.

Does anyone have any ideas on suitable mixers?

I would suggest going to www.minicircuits.com and seeing what they
have...a rather large selection. Those are ready-built in various
basings and all cover an octave or more in both input and LO ports.
MiniCircuits has some nice tutorials on various aspects of Mixers.
Price is reasonable for ready-built components.

Am I
looking at a disparate-enough set of frequency bands that I will wind
up having to use multiple mixers, possibly one per 30 MHz band, or
per octave?

I would say a single Mixer rated at LF to 1 GHz Input will do the
job. Passive (diode) Mixers are as broadband as their coupling
transformers and diodes allow, something now quite wide in frequency.

It is trivially obvious to me that I'll need different
antennas.

Not necessarily. A discone can do nearly a double decade in
frequency range but at a slight loss (1 db or so) less than a
dipole. For more gain and directionality, a log-periodic for
VHF-UHF can be used for an easy decade bandwidth.

Is it reasonable to look at using a comb generator which is driven by
a 30 MHz reference, so that I just take the appropriate harmonic and
amplify it to a power level that suits the mixer I have to use for
that particular 30 MHz band? Would it be better, perhaps, to use DDS
to generate the LO frequencies I want?

I would suggest a form of PLL, similar to a TV set-top box converter
for down-converting old UHF channels to VHF. The LO needs to
increment
only in 30 (perhaps 15 for 'havsies') MHz increments. The direct LO
frequency sample can be fed to a prescaler whose output (at a lower
incremental frequency) can be applied to a digital counter whose
output
is phase-frequency compared to a stable reference frequency. The LO
output to the Mixer needs only to be above the minimum specified level
and below a maximum, also specified if a ready-built mixer is used.
At below-1-GHz frequencies you can get away with attenuating the LO
level with a carbon-comp T or H pad figured for the common 50 Ohm
system. MiniCircuits also has VCOs of many kinds in stock and most of
those cover a half to a full octave in electronic tuning range.

Layout of the UHF needs, realistically, only an approximation of the
PCB trace width for "50 Ohm lines" on whatever substrate is used.
For very short trace lengths the "mismatch" isn't noticeable and,
for under-1.2 GHz frequencies, I've used #20 AWG wire without
problem in performance! :-)

In thinking more about it, salvaging an old TV UHF channel converter
might be a better idea, modifying its output to go DC to 30 MHz
instead of stopping at Channel 2; a few for cable TV service go
lower in output frequencies for built-in cable service functions,
depends on the maker of the set-top and the cable service. All
of those seem to operate at 2 MHz increments but use "in-house"
PLL subassemblies, circuit diagrams printed on unobtainium. :-(

Prescalers - as stock items - are now scarce but exist mainly as
"house number" special orders. A few dealers have them as old
stock left-overs. Nearly all operate up to 900 MHz so their
ability to prescale a UHF LO is no problem.

The SDR-IQ is a fairly precision instrument as a spectrum analyzer
and it deserves to have a down-converter checked out with good
test equipment. Any one of the (now ancient) HP-608 signal
generator models can cover up to the 70cm band with known RF
output levels. That's another thing to keep in mind.

I hope some of this long text is a help to you.

73, Len AF6AY

[age 74 and been in the electronics design biz some...:-) ]

"Mike Andrews" wrote in message
...
I've got, and love, an RFSpace Inc. SDR-IQ software-defined receiver.
It covers 100 Hz to 30 MHz, and lets me eyeball any 190 KHz (or
narrower) chunk all at once, which is rather particularly nice. I'd
like to be able to use it on 30 to 60 MHz and higher, which means that
I need a downconverter. It seems to me that a suitable downconverter
would let me set a switch to choose which 30 MHz band got mixed down
to the roughly 0-30 MHz baseband.

I'd like to have this work up through the 70 cm. ham band for sure,
and as much higher as I can get. I know that construction techniques
change as one gets above HF, and change _radically_ at UHF and above.
How hard is this going to be, overall?

The general setup I envision is something rather like

LO
n*30MHz
(n to |
(n+1)) |
* 30 MHz 30 MHz wide V 0 to 30 MHz
input -- bandpass filter -- mixer -- output
to SDR-IQ

Some questions for the (perhaps not older: I'm 60) wiser heads with
more design-and-construction experience: is this at all a reasonable
thing to do? Does anyone have any ideas on suitable mixers? Am I
looking at a disparate-enough set of frequency bands that I will wind
up having to use multiple mixers, possibly one per 30 MHz band, or
per octave? It is trivially obvious to me that I'll need different
antennas.

Is it reasonable to look at using a comb generator which is driven by
a 30 MHz reference, so that I just take the appropriate harmonic and
amplify it to a power level that suits the mixer I have to use for
that particular 30 MHz band?

--
Mike Andrews, W5EGO

Tired old sysadmin


I know it it's not a perfect solution, but this could be a good starter:
find a selective Jerrold of OAK CATV converter. There must be thousands of
them around for free. Try to improve the stability of the local oscillator;
replace the jerky PLL oscillator with one or a few decent synthesizers, work
on the input sensitivity, etc..

Have fun!