Call Search

New to Ham Radio?
My Profile

Friends Remembered
Survey Question

DX Cluster Spots

Ham Exams
Ham Links
List Archives
News Articles
Product Reviews
QSL Managers

Site Info
eHam Help (FAQ)
Support the site
The eHam Team
Advertising Info
Vision Statement

   Home   Help Search  
Pages: [1]   Go Down
Author Topic: Help needed with a superhet design  (Read 3560 times)

Posts: 26


« on: June 01, 2007, 05:53:44 PM »

I've just recently built a receiver(80M) which uses two sa602an ic chips, one lm 386 ic, a murata cfwg 455 khz filter and 455khz ceramic resonator, one mk484 ic, lm7805 5v regulator, and one lm 317t variable voltage regulator.  I'll try to describe the setup of the circuit so that my question can be understood better. Here goes:

I wound a toriod transformer for impeadance transformation of a 50 ohm antenna input to ~2K ohm input into one of the sa602 mixers rf input. I used a l/c meter to measure the inductances so that when I paralled a 680pf cap across the 50 ohm portion I'd get a match(50 ohm) at a center frequency of 3579khz and I used a 6-26pf trimmer cap across the input of the first sa602 mixer to get an equivalent impeadance at ~2K ohm. The input is configured as a balanced input as suggested in the pdf manual online for the sa602. For the LO section I used two 150pf caps, one 10nf cap, one air variable cap with ~54.7pf paralled across it, and an inductor of which I could not measure the value as my meter would not measure anything below 5uH. I managed to configure a DC 80m receiver using this setup while using ~104.7pf paralled across the variable capacitor but when I setup the second sa602 mixer using a 455khz ceramic resonator, murata 455khz filter, two 100pf caps, a mk484 as an if amp, and a shifted LO(using the 54.7pf instead of 104.7pf paralled with the air variable cap) I cannot seem to get any stations that I heard while configured as a direct conversion receiver. Both circuits(DC and superhet) lead to a lm 386 as an audio driving circuit. I used the lm 317t to supply and regulate the 1.5v to the mk484 ic and both sa602 mixers as well as the lm386 are running off the 7805 5v regulator. I used my other hf rig to spot the LO and the frequency ranges for the DC setup are: ~3505khz-3705khz

and for the superhet setup:
(3495Khz to 3705khz when you subtract 455khz)

Are the any suggestions as what problem(s) there are with using the superhet setup? I have no problems pulling in stations using one sa602 in a DC fashion, the selectivity is actually pretty decent although the vfo drift is about ~900 hz. I barely can hear the sw and am bcb above the noise however when I add in the additional sa602, mk484, and lm317t I can't seem to receive the same stations as I could before, even the am bcb and sw stations. Perhaps there's too much conversion loss? Maybe the second sa602 serving as a 455khz mixer and bfo using the ceramic resonator isn't oscillating? I know the impeadance input of the mk484 is ~4M ohms - ouput is ~ 2k, the murata filter is 2k ohms for both i/o, and the sa602's like anything around 1.5k or above for both i/o. I think one problem maybe with the output of the second sa602 455khz bfo(if oscillating) - it needs to see a 1.5k or above for a load resistance. I have it coupled to the mk484 if amp which is ~4M ohms as its manual states. Then the output of the mk484 is coupled to a 10k pot that feeds into the lm 386. Maybe this is a problem with te 4M ohm to 2k ohm mismatch? All of this is powered by a 9v battery. Any ideas are welcome - I'll try to draw up a detailed schematic so that I can share it with anyone interested. Thanks for the help!

- Mike

Posts: 13018

« Reply #1 on: June 02, 2007, 07:03:41 AM »

I'd try it without the MK484 for starters.  I'm using SA602-SA602-LM386
and it seems to work well.  All I have between the mixer chips is a
ceramic filter, and some RC filtering (balanced) between the second
mixer and the audio amp.  That will eliminate any possible problems
with impedance matching in that area.  Oh, I do have a bipolar cascode
RF preamp that allows me to adjust the gain, but I'm feeding it from a
6" diameter loop antenna so signals are pretty weak to start with.

The MK484 is rather of an odd part and feedthrough from the first
oscillator may be causing the AGC to reduce the gain.

Posts: 26


« Reply #2 on: June 04, 2007, 01:51:06 AM »

Thanks for the suggestions! I finally found out what one major problem was: the second sa602 serving as the bfo/product detector wasn't oscillating at all after checking it out on the scope. It seems the 100pf caps I had used across pins 6 and 7 and pin 7 to ground were too small in value. I switched them out with a pair of 1000pf caps and that seemed to do the trick. I'm using a murata ceramic 455khz resonator connected from pin 6 to ground. I guess I could rubber it a bit to play with the bfo pitch. It works pretty good with out the mk484 in the circuit but I put it back in as it gives it a little more gain and a chance to add in a tail end second if filter. I added in a larger variable cap to replace the fixed caps in the first sa602 mixer's vfo setup and kept the smaller air variable in the circuit. I plan to use the larger cap to set the desired center frequency of interest and use the smaller for actual tuning. I'm able to cover just about down to Canada CHU on I think 3.330Mhz to 4.505Mhz with the new cap setup in ~220Khz sections. I haven't heard any cw or ssb yet but I did pickup a few am stations on what I think to be ~3.892Mhz or 3.885Mhz as well as some image stations up near 5Mhz(LO max + 455khz = around 5Mhz or so). I may add in another mk484 IF setup after the second 455khz filter and have the whole thing cascaded with the first but I think I will see how testing goes with and without the mk484's and addition IF gain stages. This is my first attempt at making a fully self designed home brew superhet receiver. I've only made two other receivers - both direct conversion, a pixie II kit and the sa602 dc receiver I mentioned earlier. I'm certainly having fun fiddling around with these receivers and finding that its very educational along the way. I'm curious as to how you have your sa602-sa602-lm386 setup configured. What are the specifics like parts(filters, etc), design parameters like the IF and whatnot that you decided to choose? Are they parts that came from a "junk box" or stuff you happened to have laying around? I picked up my sa602 and mk484 ic's off ebay but much of the other stuff either came from mouser electronics, radio shack, or scraped material from around the house(toroids from computer ps's, broken am/fm/cassette radios, speakers, old crt monitors, etc.). Thanks again for the tips!!!

- Mike

Posts: 26


« Reply #3 on: June 04, 2007, 12:26:49 PM »

Just realized another error in my design - I had the IF amp after the bfo/prod det, too many things clustered on the protoboard and I got a little confused - I removed the whole mk484 setup completely and I'm going to see what I can get with the SA602-455Khz filter-SA602-LM386 setup. I think I'll fiddle around later with the mk484 setup and see if I can cascade another 455khz filter and mk484 for a sharper cutoff and a little more IF gain. I tweaked the oscillator circuits for both the LO and BFO by switching to one 100pf cap across pins 6 and 7 and one 50 pf cap from pin 7 to ground on the first sa602 LO mixer and that gave me about ~147mVp-p and then I changed out the BFO caps to one 10nf across pins 6 and 7 and one 1nf from pin 7 to ground giving me ~84mVp-p from what I could see on the oscilloscope attached to pin 7 of either sa602 mixer. I will post the results later on for any one interested in this simple superhet design.

- Mike

Posts: 13018

« Reply #4 on: June 05, 2007, 04:04:06 PM »

Hi Mike,

My primary interest is for hand-held receivers for radio direction-finding.
The circuit I described used a cascode RF amplifier with a pair of 2N2222s
followed by SA602, ceramic filter, SA602, balanced RC audio filter to the
LM386.  I get plenty of gain using a 6" loop antenna on 80m, and I can
reduce the gain of the cascode stage when I get close to the transmitter
(within about 10 feet of a 1 watt transmitter.)

The cascode stage drives an untuned transformer feeding the SA602
mixer, but I can't remember what turns ratio I used.  You can also
control the gain of the SA602 chips - they did this on the K1 rig and got
about 50dB of AGC if I recall.  Remember that the SA602 doesn't have
really great strong-signal handling capability, so you'll want to be able
to reduce the gain on strong signals to avoid overloading the BFO stage.

I used single-ended coupling between the two SA602s with a ceramic
filter.  Probably could have used balanced configuration with two filters
but that wouldn't have added much.  The second SA602 uses balanced
output to drive the LM386 through an LC network that I found in a
Swedish 80m DF receiver.  You may want to increase the gain of the
LM386 a bit - there are a couple of ways to do this.  I used a resistor
from pin 1 (?) to ground, which can increase the gain to 70dB, but
in practice I didn't need that much, and the capacitor from pin 1 to pin
8 (from memory - check the data sheet) usually gives enough gain
without causing too much noise or other problems.

Mostly I'm using parts from my junkbox, but then it is VERY well
stocked from hamfests and surplus parts sources over the years.  The
major design criteria for me are:  wide range RF gain control, so I have
good sensitivity while also being able to take bearings on very strong
signals when I get close to a transmitter;  small size, physically rugged,
and easy to operate with one hand, since I use it while running through
the woods;  and low power consumption so I can run it on a small
battery.  My latest version draws about 10ma at 9V and is built into a
length of 1" PVC pipe.  The hardest part is getting the circuit into and
out of the pipe, but it makes a sturdy and cheap housing that is just
about right for handheld use.  It is a fairly specialized application, but
most of the circuit components aren't that out of the ordinary.  Well,
not everyone listens to 80m on a 6" diameter loop with switchable sense
antenna, or intentionally puts the tuning control where it is hard to
reach.  One thing I still want to add is an S-meter output, which will
drive a CD4046 VCO chip to give a tone that changes with signal strength
(so I can take bearings by ear while watching where I am running.)  Also
I'm considering some 2m circuits, but haven't been able to spend as
much time playing with them as I would like.

Good luck!

Posts: 26


« Reply #5 on: June 07, 2007, 12:26:21 AM »

I'd really like to see a schematic of your design - sounds like a very interesting rdf project, I may want to build something like that now that I've got it rolling around my head. I put a 220uF cap across pins 1 and 8 of the lm386 to increase the gain somewhat but I think I should look into a pre-amp/bpf/lpf for cw and ssb before the audio amp - it works good enough to drive walkman type headphones but not much more than that. My goal was to keep my design simple enough for those that may be interested in building a superhet but maybe find it too difficult compared to a dc/regen/trf. I was hoping to be able to show the basic concepts of the superhet to a few of the local clubs that I go to with this design of mine and I figured being new to designing one's own equipment that I may be able to provide a unique point of view to those that are in the same situation I was in not too long ago. Only until recently did everything fall into place and started to make sense. I decided to remove the IF amp completely as it wasn't much of a difference except my SNR went down a tad bit(hear a lot more noise than signal than without it). I also left a lot of room for improvment. I too coupled the first sa602 mixer output via single end  to a cfws455 filter from pin 5 and left pin 4 disconnected. I was able to hear W1AW and few other CW stations this evening with the current setup I have here:

I didn't bother to balance the second mixer/prod det output to the Af amp and if you notice the front end is n't very tunable beyond the original center frequency design of ~3,579,000 hz as it started out as an xtal dc receiver originally just using one sa602 with an unbalanced coupling to the lm386. There's plenty of room to add a cw filter, if amp chain, agc, etc. to the basic design but being that I have limited experience in making my own equipment I opted for a simpler approach as well as the prior reasons above. I would definately like to build it in a reliable, sturdy, rugged enclosure like you mentioned. Do you prefer making your own or buying pre-formed cases? I like the idea of sticking it into a 1" PVC pipe - I know some of the people at the local club would definately get a kick out of that and would make for interesting discussion. Should I opt for a copper(shielding?) or pvc pipe? I can imagine sticking into a 1" diameter pipe not only the radio but the s meter circuitry to accompany it - definately doesn't sound easy but seems like a load of fun! Well maybe at least to some. I think you may have started a new fashion in the hobby - tubular radios.

- Mike

Posts: 13018

« Reply #6 on: June 07, 2007, 07:16:34 PM »

Unfortunately I don't have the schematic handy - may not even have it
recorded anywhere that I can find it!  But I can pull the receiver apart
and draw it out if needed.  Don't have a good way to send it by email
yet, either.

A couple comments on your schematic.  First, you'll get somewhat
better performance from the SA602s on 6V rather than 5V - sticking a
couple of series diodes in the ground return of the 7805 should do the
job.  (But I'm running mine on 5V and it seems to work OK.)

Another option would be to use an audio preamp.  I've used the common
base design in direct conversion receivers to provide a 50 ohm termination
to the mixer, but in this case you could run the balanced output from
the product detector to a low noise op amp (perhaps an LM308?) to give
more voltage gain driving the LM386.

You might also look up the circuit for the SW40+ QRP transceiver (or the
similar 40-40 rig).  The receiver section uses a pair of SA602s with
a crystal filter between them.  I've built one, and have the board around
here to build another.  It uses two op amp sections for the audio, and
I had to select a sensitivie pair of Walkman headphones to get enough
output from it, but it may give you some ideas.

Hmmm...  T1 doesn't seem right to me.  Are you designing this for a
standard 50 ohm antenna?  Is the secondary really 315 uH???  What sort
of former are you using to get that much inductance - a ferrite toroid?
If so, powdered iron may be a better choice for a tuned circuit.
I'll have to look at the calculations a bit further.  The turns ratio should
be 5 or 6 : 1 to match a 50 ohm antenna to the 1500 ohm input of the
SA602.  (I've rewound IF transformers for 80m, and the 10.7 MHz ones
often can be used simply by adding enough capacitance across the
secondary to lower the resonant frequency.)   Offhand I'd guess that
lowering the inductance by a factor of 10 (and changing the capacitor
to match) might be better.

Certainly C1 doesn't resonate with the primary of T1 on 80m, which
will reduce your signal strengths.  I'd remove C1 to start with and see
if that makes a difference.  Also make sure that C2 really can tune
T1 to resonance (it should peak in two different locations as C2 is
rotated.)   My rough calculations show that 320uH would require about
6pf to resonate, and your trimmer would be marginal even without
considering the self-capacitance of such a large coil and circuit strays.

Meanwhile, I'll see if I can find my circuit or draw it out.  Who knows, I
may even get some photos of the receiver in the PVC pipe in the process.

Posts: 26


« Reply #7 on: June 08, 2007, 12:20:45 PM »

I rewound T1 and replaced C1 and C2 values with ~890pf on primary with 2.2uH and C2 was shifted up by 7pf to give 13-33pf where I calculated that ~30pf would yield a reactance of ~1500 ohms at Fo = 3,579,000Hz and the secondary @ ~67uH gave the same 1500 ohms going into the first sa602 mixer. Reason behind such high inductances was that my cheap L/C could barely get accurate readings under 7-10uH so I figured I could possibly somehow get a impeadance match to first sa602 mixer using higher impeadance in parallel to a capacitive impeadance that was lower such that the total impeadance seen by the sa602 would be around 2K ohms. This brings me to a question that I should have asked earlier but forgot: If the Xl and Xc of the L/C tank circuit are the same impeadance then is the total impeadance at a given frequency Fo looking into that L/C circuit X/2 or simply X where X = Xc = Xl? I calculated new values based for 50 ohms and 1500 ohms but that brings me to the question above, is it a total of 25ohms and 750 ohms looking into the primary and secondary sides of T1 or since its resonant @ Fo 3579000Hz is it 50 and 1500 ohms respectively? The eqs I used were for Xc = (1/wC) and Xl = (wL) where w = (2 * pi * Fo), Xl and Xc are in ohms, C is in farads, L is in henries and Fo is in hertz. I'm still working on a audio preamp/bpf/lpf before the audio amp but I need to brush up on some stuff like hybrid pi models, op amp designs, fet analysis and miller capacitance effects to name a few. I might also look into a rf pre amp but I'll save that for later. When you put in the diodes on the ground pin of the 5v regulator, does it pull up the regulated voltage to that of the required forward bias votlage of the particular diode used? EX: if I put in a diode that required say 1.2V to forward bias would it raise the regulated 5v output to 6.2V?

- Mike

Posts: 13018

« Reply #8 on: June 11, 2007, 06:44:28 PM »

Yes, a series diode in the ground lead to lift it up by 0.6 or 1.2V will
increase the output voltage by the same amount.  (A standard diode has
to be forwarded biased by the current flowing from the regulator to
ground in this application.)  Or you can connect the ground lead to a
pair of resistors connected in series from the output to ground just like
the standard circuit for an LM317.  Either will increase the output voltage,
just make sure that you aren't trying to heatsink the tab to the case
without an insulator!

No, that isn't how impedance matching works.  The actual impedance
of a parallel-tuned circuit isn't due to the reactances of the individual
components.  Two reactances of equal but opposite sign in parallel give
a theoretically infinite impedance (that is parallel resonance).  In practice
it is limited by the Q of the parts, but you want the impedance to be
set by the load, not by the parts.  I don't have time to go into a full
explaination, but a couple hundred ohms for the reactances should be
fine.  Presuming that he circuit is tuned to resonance, the only critical
thing is the turns ratio which should be chosen to step up 50 ohms to
1500 (or 2000) ohms - I don't remember exactly what the data book
says but either is close enough.  I'd probably choose an appropriate
coil based on what is physically practical (perhaps a convenient number
of turns on whatever size toroid I have handy, or a coil from the junkbox
in the 5 to 20uH range, or whatever looks good), choose a variable cap to
resonate at the operating frequency, then calculate the number of turns
required for the 50 ohm winding.  One problem with trying to tune the
input winding also is that it will require a very large capacitor, and it is
harder to find variables that size.  Perhaps a better approach would be
to use two identical parallel-tuned circuits with a small coupling cap
between them and the 50 ohm winding on just the first one.
Pages: [1]   Go Up
Jump to:  

Powered by MySQL Powered by PHP Powered by SMF 1.1.11 | SMF © 2006-2009, Simple Machines LLC Valid XHTML 1.0! Valid CSS!