Skip to main content
Topic: VHF/UHF to HF down-converter (Read 29844 times) previous topic - next topic

VHF/UHF to HF down-converter

I've (relatively) recently added another item to my long list of obsessions - software defined radio - and have even more recently started ordering components for one of the probably most interesting (semi)open-hardware SDR projects out there: HPSDR.

While i'm waiting for the remaining boards (i'll need to assemble a working receiver) to be produced and arrive (HPSDR is a modular concept - you need a motherboard, a receiver board and a USB or Ethernet interface board to assemble a working RX), i've started toying with the idea of designing a board which would enable me to use it in the VHF/UHF part of the spectrum as well as its native HF. Firstly, i've put a few different design ideas on paper that can be binned in two broad groups:

A) Nyquist zone preselectors - since HPSDR's Mercury RX board samples DC - 61,44 MHz directly, this would allow me to selectively undersample each or just selected (interesting) Nyquist zones above the native range (after disabling Mercury's internal 55MHz BP filter)
B) Down-converters - down-conversion from higher RF frequencies to the native IF utilising a frequency mixer

I've rejected the idea of a preselector (designing filters above 200ish MHz would be problematic) and opted for Option B. I saw three possible  ways of building a down converter suitable and flexible enough:

1) using Si570 (high grade LVDS or CML version, if possible) as the LO to feed a high IP3 Mini Circuits frequency mixer, then filtering out the sum image with a low pass filter
2) same as 2, but with a DDS (Analogue Devices AD99xx)
3) same as 2, but with a bank of relay switchable fixed frequency oscillators

Option 1 seemed like the best bet, so that's what i've decided to try and design. To cut things short, here's an initial block diagram of what i've had in mind:



All the filters, amps, the balun and the mixer will be off the shelf components (Mini Circuits models, as indicated). The quasi-oven is there to keep the Si570 frequency as stable as possible. It will be based on the Genesis G59 design, except SMD-ified as much as possible. The board will be 2-layer if possible, with one side dedicated almost purely to ground (separate analogue and digital ground planes).

Omron SMD signal relays will be used for pre-amp and filter bypassing and oven control (control in the sense of having it completely on or off, not in the sense of PWM). An Atmega168 or 328 will control both them and the Si570 providing the LO frequency. The Atmega will, in turn, be controlled via an I2C or SPI line, by the other boards on the bus, or manually, via a front control panel consisting of a rotary encoder, LCD and a few buttons. The balun is there to convert the differential Si570 CML output and suck out as much decibels as possible to feed a high IP3 frequency mixer. The (optional) FM broadcast band-stop filter is there to kill the interference from strong nearby FM towers, if needed. There's a third bypass line missing in the schematics - the one bypassing the whole down-converter, for regular HF use.

So... That's it for now. Since this is my first time designing anything RF-related, any/all input will be muchly appreciated.

Re: VHF/UHF to HF down-converter

Reply #1
there are single chip quadrature decoders that will work on 144 and 440 these would be a more simple approach.

i am very interested in sdr and am willing to help you out.

Re: VHF/UHF to HF down-converter

Reply #2
Thanks a lot! :)

I'm not sure a quadrature demodulator would be of use in this particular case, as i don't need an IQ signal i'd feed into the rest of the chain. Furthermore, the Mercury board directly samples a pretty wide pass-band (with a pretty expensive 122 Ms/sec ADC) and it would be nice to utilise that as much as possible.

Lots of other simple ways to do it were suggested to me, such as using an "old" off-the shelf I2C controlled analogue TV tuner. Most TV's use an IF of 38.9 MHz, which falls inside the Mercury pass-band nicely. A tuner such as that would technically be all the components i'd need for a down-converter and it would be a pretty cheap solution. However, it would be prone to drift, would probably have bad IP3 figures and would have questionable bandwidth.

Which is all fine, but would be kind of a waste considering the performance of the receiver "sitting" behind the down-converter :)

Re: VHF/UHF to HF down-converter

Reply #3
i am interested in a single board solution, using an audio codec, and possibly a gumstix


Re: VHF/UHF to HF down-converter

Reply #5
that is a great combination, are there any modulators like that?

Re: VHF/UHF to HF down-converter

Reply #6
With an integrated LO?
There are but, unfortunately, not with matching specs (min/max RF and IQ frequencies).

TI has one too, but it goes from 300 MHz upwards.

Are you planning on building a transciever?

Re: VHF/UHF to HF down-converter

Reply #7
[quote author="orcinus"]With an integrated LO?
There are but, unfortunately, not with matching specs (min/max RF and IQ frequencies).
[/quote]

there are some quadrature converters that work both ways, but the high frequency (if i remember correctly) is up to about 250MHz

[quote author="orcinus"]Are you planning on building a transciever?[/quote]

my goal is to have a system able to transceve 0-30MHz + 140-144MHz + 420-450MHz

Re: VHF/UHF to HF down-converter

Reply #8
That would be a nifty little gadget indeed. Especially if made compact and portable (i.e. HT).

BTW, if you're aiming for a unit that could operate in a self-sufficient mode too (i.e. with on-board demodulation as well as the option to output IQ for further processing), this might be of interest too:

http://www.intersil.com/data/fn/fn3651.pdf

:D

Re: VHF/UHF to HF down-converter

Reply #9
Oh, and re: quadrature converters... If they go up to 250 MHz, you could use that and just add a down/up-conversion block for 420-450 that you can switch in and bypass as needed.

Re: VHF/UHF to HF down-converter

Reply #10
[quote author="orcinus"]Oh, and re: quadrature converters... If they go up to 250 MHz, you could use that and just add a down/up-conversion block for 420-450 that you can switch in and bypass as needed.[/quote]

should be easy enough just add a mixer

Re: VHF/UHF to HF down-converter

Reply #11
Quick update...

So, it turns out i've been neglecting a vital part of the problem of designing a downconverter - dealing with the image, a.k.a. image rejection.

I've sort of always assumed the input would be pre-filtered with a convenient high-pass, but didn't think far away ahead to realise that it's more of a problem than the downconverter itself. The thing is, just as with the preselector i've originally intended to build, as it goes farther into UHF and near microwave frequencies, building suitable, high-Q filters gets progressively harder.

There's an alternative, in the form of imager rejection mixers, but their performance (Image Rejection Ratio) depends on the amplitude and phase error of the mixer LOs. For 60 dB of rejection, i'd need to have the I and Q outputs matched to within 0.1 dB / 0.1 degrees. That seems pretty hard to pull off. For 40 dB, the figures are 1 dB and 1 degree.

That means that the IR mixing filter would probably have to be combined with a preselector or a tracking filter...

Re: VHF/UHF to HF down-converter

Reply #12
i have seen quadrature (de)encoders for UHF/VHF applications.

Re: VHF/UHF to HF down-converter

Reply #13
There are, unfortunately, their IF typically does not go under 50ish MHz. And i need DC-55 for my purposes.
There are also some nice single-package surface mount image rejection mixers available, but they too don't fit my application (they typically don't go below 400 MHz or so).

Did you come across any that go lower than 50, but still cover RF to about 1000 MHz or above?

Re: VHF/UHF to HF down-converter

Reply #14
i have seen a single chip that will work in both directions from 100K to 250M, http://http://search.digikey.com/us/en/products/RF2713TR7/689-1018-1-ND/1302354

i have not run across any that will do HF/VHF/UHF.

but separating the VHF from the UHF would work well for satellite operations.

i'd keep all three separated, the filtering will be better