150KHz-30MHz receiver design

Hello,

In a previous discussion on a picaxe-controlled receiver (<A href='http://www.rev-ed.co.uk/picaxe/forum/Topic.asp?topic_id=6699&amp;forum_id=30&amp;Topic_Title=RF%2BProject%2BIdea%2B%252D%2Bsuggestions%253F&amp;forum_title=No+new+posts+please%21+22 ' Target=_Blank>External Web Link</a> ) I shortly mentioned that I have a long-term plan for the design of a general coverage receiver (150 KHz - 30MHz). Although I probably will start next year at the earliest (first some other projects have to be finished, notably the &quot;SerialPower&quot; network concept) and the project takes a few years to complete, I have already the general concept in my mind. As a few forum members indicated similar plans, I thought it would be nice to start a discussion here on such a design and see were the main challenges are.

My final, &quot;ideal&quot; goal is the following type of receiver:

- Double-Conversion Superhet
- Receiving range 150KHz - 30MHz (LW, MW, SW)
- Receiving modes AM, SSB, FM, DRM
- selectable IF bandwidths (PICAXE controlled)
- SA602A first mixer at 10.7MHz first IF
- TDA1572 (or TDA1072A) amplifier, second mixer + detector at 455KHz (single-chip AM circuit)
- PICAXE-controlled DDS circuit for receiver freq. oscillator
- PICAXE-tuned preselector stage
- PICAXE-controlled RF pre-amplifier (amplification control dependent of frequency and time!)
- Extreme user friendliness using PICAXE control circuit (fully automatic + full manual modes)
- Possibility to download and directly use SW broadcast schemes from internet
- Single PICAXE used (PICAXE-28X1)

This is a very challenging design, I am not a RF specialist (although I have some prototyping experience, see <A href='http://www.kranenborg.org/ee/ee2.htm ' Target=_Blank>External Web Link</a> ) and the risk of not arriving at the final stage is large, I want to stucture the design process in such a way that always some result is obtained:

- A simple design from the start (single-conversion superhet, functional even without a picaxe)
- Each incremental stage tested to give a better receiver (so that it is not a waste of time/money if one gets stuck at a certain level).
- Use of readily designed PCB project boards for RF part (Elektor).
- Use experiences from existing designs as much as possible (4 Elektor/Elektuur designs from Gert Baars)

So what are the main challenges, and how does the picaxe get involved?

A: DDS application for frequency generation:
--------------------------------------------
A DDS provides a lot of potential for direct frequency control, and I plan to use the AMQRP DDS-60 kit (<A href='http://www.amqrp.org/kits/dds60/ ' Target=_Blank>External Web Link</a> ) as this seems to me the only one that can be practically built and used. It can be interfaced via spi directly to a PICAXE, the latter generating the tuning word.

B: Preselector:
--------------------------------------------
The AD9851 DDS of the DDS-60 kit uses a 10-bit DAC (14-bit would have been much better) which implies that a lot of spurs (different frequencies at non-negligible amplitudes, typically -50 - -70dB of the main signal) are generated. Due to this the receiver gets tuned to several frequencies at the same time, causing an unacceptable quality degradation.

To solve this problem I opt for a very andvanced preselector and subsequent pre-amplification stage between antenna and first mixer. This is clearly the most important stage!

* 8-band preselector consisting of:
- - a double AM varicap
- - switchable coils (using mini-relais) to select each band
* tuning of the preselector:
- - 256 steps per band via DS1803 digital pot
- - each step is calculated via a 4th-order polynomial approximation
- - approximation is calculated on the fly during tuning using a uM-FPU
- - approximation polynomial is calculated using a PC and a measurement procedure
* double-gate mosfet as RF amplifier:
- - gate voltage controlled via DS1803 pot

The automated tuning of the preselector is complicated and requires a one-time manual initialization, but seems still feasible, i have worked out some ideas on how to do it.

C: Automatic Selection of IF bandwith (PICAXE controlled)

D: Control issues
--------------------------------------------
There are a lot of control issues, solved on the hardware level via I2C &amp; SPI communications as

follows:
* PICAXE DDS Frequency tuning word generation and communication via SPI
* Preselector band selection + tuning (see also item B above):
- - 8-band selection via MCP23008 I/O expander (I2C) and minirelais
- - tuning via DS1803 (1/2) and varicap
- - DS1803 control via interpolating polynomial evaluation via uM-FPU
* RF-amplifier via BF961 Mosfet and DS1803 (1/2) digital pot
* IF bandwidth selection via MCP23008 and PIN diodes
* Frequency display using 6*7-seg display controlled via MAX6955 (i2c)
* Separate LCD display for program information and control info display (i2c)
* EEPROM storage of configuration settings and program information (i2c)
* DS1337 clock (i2c) for program calender and RF amplifier control

It is very important to start simple and have a strategy for extending and testing, and I would propose something like the following sequence:

1) Start with a simple non-digital version for a AM radio circuit:
- 500 - 1800 KHz
- single-conversion superhet
- tuning of antenna circuit and oscillator using AM-varicaps

2) Replace oscillator circuit with DDS VFO and lCD circuit
- DDS is picaxe controlled via SPI
- LCD display for frequency display
- frequency change using potentiometer at PICAXE adc input

3) Add preselector and RF amplifier

4) Add 7-seg frequency display

5) Add IF band selection

6) Add PC - receiver interface for downloading of SW transmission schemes

I can add some more information on the various stages, but many details are still open to discussion and still unclear to me. Some other forum members with similar ideas?

Well, who wants to jump on the train?
Even issues on a simpler, limited design (direct conversion) may be discussed here as well

Regards,
Jurjen
http://www.kranenborg.org/ee/picaxe

Edited by - kranenborg on 15/05/2007 21:37:05

Wow nice little project that.. VK7HLE here. Sadly not too much experience with homne building recievers but would be very interested in buildng one up once you get the bugs sorted?? If you get real stuck there are a bunch of amateurs here that may be able to help?

 

Just found this thread and thought that others may be interested in my project to build a low HF (upto 10MHz) AM/DSB transceiver with, of course, Picaxe control!

The heart of the rig is a synthesiser which covers from about 1MHz to 7.4MHz in 1KHz steps. Not enough resolution for CW or SSB, but enough for my needs. The synthesiser has a programmable divider controlled by a 40X chip. I don&#8217;t need a 40X for the processing power, but it was the only Picaxe which would give me 15 outputs (enough for a division ratio of upto 8999). The synthesiser should work upto 20MHz with different components, but I haven&#8217;t tried it up there.

The 40X provides BCD encoding of the division ratio set via a potentiometer and LCD showing the division ratio. It doesn&#8217;t measure the frequency directly, but for my purposes, it is close enough. Using memories, I have two frequency working, with one frequency for the TX LO and one for the RX LO offset by the IF. So far, I&#8217;ve only used the RX in direct conversion mode into a MC1496 mixer. The switching time from one to the other appears to be fast enough for use with a superhet design. I&#8217;ve got a 6KHz 455KHz filter which just needs to be put in place.

The output from the synthesiser is filtered with low pass filters to produce a sine wave feeding the TX/RX stages. For amateur bands only, plus a few other spot frequencies I want to use it on, I can get by with just 3 filters, but for general coverage it would get a bit complicated. I think that the use of a DDS would make it easier, but that is beyond the current project.

The biggest problem I can see with a wideband HF receiver is with the amount of filtering and tuning required. If RX only, then tuning front ends could result in a lot of work. I&#8217;ve comtemplated it, but been put off by the workload! TX is slightly easier, in that you can use broadband amplifiers and just filter out the harmonics with pin diode switched filters.

Total cost so far is about &#163;20 for the synthesiser plus the 40X, board and LCD. The RF components have come from my junk box.

Hopefully this if of interest and if anyone wants more information, just ask.

Graham