ARRL

Register Account

Login Help

Forum Home - Rules - Help - Login - Forgot Password
Members can access, post and reply to the forums below. Before you do, please first read the RULES.

connecting RF modules; impedance matching

Jul 27th 2020, 13:02

henryrichter

Joined: Oct 14th 2018, 20:05
Total Topics: 0
Total Posts: 0
I am building a superhet radio based on a 1994 publication, using Toko shielded transformers. This is my first radio project meant to educate as well as and function.
I have relied where possible on Experimental Methods of RF Design,the "book".
Each of these modules is in EMI protected die cast aluminum boxes:
1.Front end, consisting of tuned filter, BF244 RF amplifier and dual gate MOSFET Mixer ,40673.
2.Local oscillator
3.IF section
4.BFO/product detector.
My question is how to run the signal between the boxes:
LO to Mixer
Mixer to IF
IF to BFO
The options are to simply run a wire thru the boxes, to run a shielded wire, or to use 50 ohm BNC connectors and RG-59.

The latter is the question. As I understand it, using the 50 Ohm connector/ cable will require a step-down:step up toroid wound transformer with the proper turns ratio.
First, is the 50 ohm route necessary?
Second If so, am I correct in adding the toroids ?

The answer to this question will dictate the size of holes I drill in the boxes so the project is waiting for me to figure this out.

These are old parts with one page data sheets that don't include impedance. I would need to know the output impedance of the BF244 and 40673 transistors. The closest answer I can find online is 5000 ohms, all resistive to the GHz range.

A similar problem, the correct turns ratio, occurs if I insert a filter after the LO or on the Drain output of the mixer, before the first IF transformer.
The data sheet for a modern IF transformer, 42IF 101-RC,, similar in the older versions inductance and turns lists the primary impedance as 60K ohms. Thus to add a filter between the mixer drain and first IF transformer would require step down from 5000:50 , a 10:1 turns ratio, and set up from 50:60000 a 245:1 turns ratio!


The Toko coils are mounted on DB9 connectors for band changing. The construction is all " ugly style" on copper board.
Since I am drilling holes, should I include a "star ground" from the copper board and the variable capacitor to each box. Otherwise the electrical connection form the board to the box is the screwed on standoffs on the boards.


Jul 27th 2020, 16:03

W1VT

Super Moderator

Joined: Apr 4th 1998, 00:00
Total Topics: 0
Total Posts: 0
I preferred to use the 50 ohm route because I had 50 ohm instrumentation. It thought it would be easier to optimize the designs if I could verify the designs with accurate measurements. It is likely that a wire close to a ground plane will have an impedance of roughly 150 ohms, though it is best to confirm this via measurement, perhaps with an antenna analzyer.

At microwaves, where you are trying to obtain a maximum power transfer, you will want to conjugately match the output impedance. At HF, rarely do you want to do this. Rather, you will design the output for a particular power level. The supply voltage and power will set the impedance, not the device.
Similarly, you will often want to load down an IF transformer to obtain a particular bandwidth and loss curve.
A star ground may work. It may not if you have a high current, say from the final power amplifier, running near and upsetting a sensitive VFO circuit. Circuit design is as much an art as a science.

Zak W1VT
ARRL Senior Lab Engineer
Sep 3rd 2020, 09:32

henryrichter

Joined: Oct 14th 2018, 20:05
Total Topics: 0
Total Posts: 0
This receiver will have three bands operating from 1.7-30 MHz. I am asking about switching bands in the LO. The LO consists of a Hartley oscillator (page 250 Secrets of RF Circuit design, J.Carr 3rd ed) three bandpass filters and a post-filter amplifier.
I plan on using pin diode switches at the output from the filters to the post-filter amplifier, using the two post filter diodes only from ("A Diode switched band-pass filter" D.DeMaw, QST January 1991).
My questions is where to insert the pin diode switch in the oscillator, where three bands requires three tank inductors. I believe I need to switch between the hot end of the tank and the gate. So, where?
between C4 and C1? C1 and R2, or R2 and the gate.
The alternative is to build three oscillators , each with their own inductor and then I simply need to switch the filters at their outputs to the post amplifier.
I actually prefer this route, because in a mock-up of the pin diode switch the power loss is considerable, so I have to be aware of that. I think that is the main purpose of my question: where to accept the loss?
It would be tight, but I can build three oscillators, in a way to put brass shielding between them.
As for noise I will do what is suggested in the 1991 article: when a switch is OFF, supply reverse voltage (+12V) to the diode cathodes and ground the anodes, to fully turn them off.
Any advice would be appreciated.
I see there are modern ICs available that act as a SPDT switch, but not a 3P3T switch that I need.
The three filters are 5 pole Butterworth band-pass filters. They require 50 ohm inputs and outputs. (rf-tools.com). The filter schematic has 50 ohm resistors at he inputs and outputs, which I think is figurative.
Just speaking of the filter input: the output of the oscillator is the secondary of the tank inductor. I can give the secondary an impedance of 50 ohms by either
1. winding correct turns on the coil to make the secondary 50 ohms all by itself, or
2, accept an arbitrary Impedance of the primary say 10 Mhz, in the center of my 5-14 Mhz band, and use N^2 to pick the number of turns in the secondary.
What is the best route?

Sep 3rd 2020, 10:58

W1VT

Super Moderator

Joined: Apr 4th 1998, 00:00
Total Topics: 0
Total Posts: 0
That is interesting. I wouldn't think that PIN diode loss would be a problem when switching LOs, where a dB or two of insertion loss can often be made up for with more low power amplification.

It is really hard to generalize what to do as everyone has different design goals. What I did was I built a lot of stuff in modules, and kept what worked. The stuff that didn't work was tossed in a junk box, where it was available for other projects.

When homebrewing circuits for the first time it can be hard to figure why something doesn't work as expected. Sometimes it is merely a misinterpretation of what someone wrote, so no amount of engineering can fix that. Sometimes the best approach is to try different circuits until you find those that work, and not dwell to much on what you don't understand. But there are times in which spending a ton of time just understanding a concept will open doors to new opportunities. Tough call.

When I built microwave transverters I build a couple really good ones that had modules that tested out as good as predicted. Then I took a lot of the stuff that merely worked OK and assembled them into working transverters that I would lend out to my friends. Those loaner rigs worked remarkably well!

Zak W1VT
ARRL Senior Lab Engineer

Back to Top

EXPLORE ARRL

Instragram     Facebook     Twitter     YouTube     LinkedIn