455kHz SDR - a second thought

My previous blog was all about the idea of adding a softrock, or any other simple SDR-DC-RX, to a cheap (synthesized) AM radio. Well, honestly said, when thinking of it, this may be a totally unnecessary overkill.
Why? Well, very simple. The main purpose of all the quadrature stuff is to make the two sidebands that a DC-RX receives different. But, what if there is not other sideband? The following may not apply to the absolute cheapest of AM-receivers.
Concerning the ATS-404, I have ambiguous information. While some technical data mention the AM i.f. being 450kHz, the schematics diagram mentions a LT455H, which is a 455kHz ceramic filter having +/-3kHz 6dB bandwidth (+/-9kHz for attenuated bandwidth). If we tap before that filter, we definitely need quadrature, should we however tap the i.f. behind that filter, a non-quadrature SDR would be OK too.
Most of the better world-band receivers use a first i.f. somewhere high with a relatively wide crystal filter. Most of the narrow filtering is done at 455kHz. In this case, we probably wont need quadrature at all. All we have to do is to ensure that our SDR center frequency (or SDR-l.o.) falls close to but outside the range of the intermediate frequency range. In such a scenario, there would not be a second sideband to care about and also a simple mono-audio interface would already do the job.
The Target HF3 would be an example for such a receiver. The first i.f. is at 45MHz having a bandwidth of +/-3.75kHz. The second i.f. band would consequently be 455-3.75=451.25 to 455+3.75=458.75 kHz. In yesterday's example, using a 1.8432Mhz local oscillator, we ended up at an SDR center frequency of 460.8kHz, which is close but outside the HF3's second i.f. band. A regular direct conversion receiver with a local oscillator at 460.8kHz would therefore receive only a lower side-band, since there is no signal in its upper side-band.
My idea would be to try that out using a canned oscillator and two flip-flops for frequency division. With some isolation amplification a singled ended diode mixer and a cheap USB audio adapter should round up that experiment.

Clarifier Calibration Idea

Inexpensive receivers as the HF3 need occasional calibration of the clarifier since such receiver to exhibit some temperature drift. One option is zero-beat a known carrier, another option is adjusting a known carrier to create a beat close to the audio frequency we want to use/listen to, e.g. 2Khz, and display it on a spectrum display.

And here would be my planned approach to the second option. A micro-controller (PIC, PICAXE, ATMEL, etc.) programmed as frequency counter of particular kind. Lets count as precisely as possible around 2kHz.
Here are some ideas how to display the deviation from 2kHz:

1. 5 LEDs: red (+/-10 to -+/-5Hz), amber (+/-4 to +/-2Hz), green (from -1Hz to +1Hz)
2. 7 segment display: deviation in Hz, decimal dot a negative sign indicator

In both cases, one may want to indicate grater deviations. In the first case I would suggest having the left hand read and amber LEDs illuminated together with the green one when the frequency is below the range, right hand side for above. In the second case, a suggestion could be to have segments C,D, E and G  on for "too low", and hence, A, B, G and F for "too high".

The following idea is for the more advanced builder. It should also be seen as a modification to the receiver. The clarifier-potentiometer could be replaced by some circuitry creating/controlling the clarifier varactor. PWM could be an idea here. The concept would be to "visit" a known carrier, activate the counting and adjust the varactor voltage such that the counter counts a 2kHz beat. Now deactivate counting and keep the varactor voltage constant. Actually, in a grabber setup, one may consider to have such "calibration visits" periodically (maybe every 3h) executed by the micro-controller, e.g. by switching between memory channels.

HF3 vs FRG-100

Did some longterm test of the two receivers on 600m. Well, maybe it is more the combination of gear I am comparing here, but still interesting.
During daytime, G4JNT produces a steady WSPR signal in JO22. I used this to test the following two setup:

1. FRG-100 @ Octaplumb (24.02.2010)
2. HF3 @ E-probe (26.02.2010)

Both aerials were at the exact same position for the testing periode.
It seems that the FRG-100 @ the Plumbtenna yielded an average SNR of -19dB and the HF3/E-probe combination an average SNR of -22dB.

However, there is one additional change between the two setups. At February 25th, I added a 1:1 isolation transformer to the audio cable. I am not sure if that had a great influence on the narrowband signal, it had one at the audio from the speaker.

The test invisaged for the weekend: use the HF3 in combination with the Octaplumb.

80m results (06.12.2009)

Colin and I did some tests on 80m lately. The following spectrum shows the final minutes, before Colin QSYed. Receiver setup on my side: NASA Target HF3 + E-probe. The computer used for analysing is a Packard Bell "dot" netbook with an Intel Atom N270 processor.

The darker regions on the spectrum are due to a local (The Hague) sideband station pulling the AGC to the max. But still, there is some F1A (FSK-CW) coming through... The cheap'n easy setup still allows for detecting weak signals...

HF3 clarifier adjustment second option

There are two ways, both not very difficult. Adjusting the clarifier takes some patience however, steady fingers help with an unmodified HF3. For the first option, please see the entry about HF3 grabber setup.

Here a second option that does not require a spectrum display, in this method you will be using your ears instead of your eyes. This is now it works, and also why...

1. pick a broadcast station or timesignal with proven stable frequency
2. switch to USB (you should now here a 2kHz tone, that would be the carrier)
3. set the dial 2kHz above the nominal frequency (this bring you close to zero beat)
4. adjust the clarifier for zero beat

This is why this works, the HF3 has got a very wide side-band filter, I guess that saves the maker some money. Hence, the filter's passband is broad enough to even pass the carrier and LSB contributions, allowing us to adjust the clarifier w/o any further tricks.

HF3 grabber setup

This is what the present grabber setup looks like. The difficult thing is to adjust the BFO such that it is clear where it is. This receiver will not be modified, however, there is a second of this make on the way into my shack. One of the two will be modified, severely ;-)

The grabber is set to listen to QRSS signals on 3,599,900.0Hz. This may look odd, but, the HF3 is not aimed at radio amateurs, it is aimed at yachtsmen (like me). It particular to receiver TTY weather reports, weather charts via FAX, NAVTEX and occasionally some news on HF-BC using A3E. In marine station listings, the frequency for teletype stations given is the frequency between mark and space. To help the sailors, the processor of the radio sets the radio's PLL to the dial frequency for A3E and 2kHz below for J3E/USB. I have not checked J3E/LSB yet, since it is of no importance to me.

When using the HF3 as a grabber receiver, I thus set the dial to the actual frequency I want to listen on, knowing about the 2kHz difference.

To adjust the clarifier, use a time signal, or a known to be stable BC station, to adjust the carrier such that it precisely results in a 2kHz tone.

I tell spectrumlab the dial-frequency minus 2kHz as offset, in this example, it would be 3,597,000.0Hz.

The FRT-7700 makes a good general purpose tuner. It helps to improve all sorts of mis-matches.

HF3 the Second

It was a toy in the beginning, the more I use it, the better I like it! Actually, I like it that much that I ordered a second one.

There are plans for this second one. Since there is a lot of space inside its' housing, there is room for modification and experimentation. Many years ago, I heard about a kit that converts a FRG-7700 into a transceiver. Cool, I thought, and forgot about it again. And now, the idea is back, with something that connects in either of the two, the FRG-7700 or the HF3. Both are double conversion superhets, both use a lower intermediate frequency of 455kHz. Using the generated frequencies, one could imagine to convert a single side-band signal all the way back from 455kHz to whatever the receive frequency would be.
I figure, it would be the question of just some NE612 mixers and two crystal filters, one for 455kHz (side-band) and one for the first intermediate frequency.

One further could consider to add some temperature control to the reference oscillator.

And, who knows, maybe this could be an easy way of producing some QRSS-signals too.

HF3, improved SSB filter

Was able to catch some Murata 455kHz ceramic SSB filters on *bay. Also looking into a CW-filter designs using ceramic resonantors. The RX, I believe, would suffer some improvement in terms of bandwidth...

HF-3W improved - long time drift

Had the receiver WSPRing all night. The result is pretty conclusive.

There is no drift as such. The receiver however shows significant thermal dependence.

• Window closed, everything stable, room temperature stable resulted in no drift.
• Window open, room temperature sinking, drift kicks in between 0340 and 0610 a 20Hz drift is observed (IQ4DJ MEPT).

The problem could probably be solved by adding an oven. For me this thermal behaviour is understood well enough to just don't care about. My HF-3 will remain as is.

BTW, the jumps did not reoccur.

HF-3W improved - first results

UPDATE!!!

The receiver is now hooked up to a short-wave antenna (G5RV-jr). Presently receiving the 30m QRSS/QSPR band. WSPR decodes with no drift on most stations. I therefore believe that the drift problem is purely cosmetic, maybe it has something to do with warming up, I would not be surprised. No jumps occured up to present.

During my test, for a short periode, the MEPT of I0/N2CQR was visible.

WSPR decodes look like this:

2114 -23 1.4 10.140241 0 F6BIA JN18 33

2116 -13 -0.4 10.140189 0 DH8SA JO53 37

2118 -26 0.8 10.140207 0 PY2GN GG56 37

2120 -15 0.6 10.140106 0 DL0TFK JN48 30

2120 -14 1.2 10.140198 0 DH5RAE JN68 37

2122 -1 2.7 10.140226 1 DL2ZQ JO42 37

Not too bad for a €299.- receiver.

Conclusions would be the following:

I now really like the HF-3W. 1kHz steps are not ideal for ham-radio. The clarifier helps, but is hard to adjust. Using a reference broadcast-station, the carrier can be set to create a 1.3kHz beat tone, that's the offset required for WSPR operations.

The receiver primarily is targeted to yachtsmen. In naval frequency lists, the center frequency between mark and space is listed. The receiver is actually built to help sailors to easily find RTTY stations transmitting weather reports and thus, the frequency display in SSB is off by about 2kHz. Meaning, that the dial for WSPR needs to be set to 14.141MHz with some odd clarifier setting.

HF-3W improved

Just gave it a shot. The Target HF-3 VLF to HF receiver is certainly not the world's best receiver, however, it is not bad either. Actually, I like it!

Tests on DCF77 reveiled, that frequency stability is not one of the big pros of the receiver. OK, no QRSS, which was shown by Des (M0AYF).

Anyway, there was one feature I never understood, and I also disliked a lot. The clarifier position. Thus, I opened the (plastic) box to see what could be done. Next to the position where the clarifier poti is wired up to, a crystal and a varicap-trimmer are to be seen. The part is the local oscillator. Ideal, lets play with the varicap first. I got what I wanted.

Since the receiver was open anyway, I took the chance and carefully played with the inductors starting from the first IF, further to the second IF and the one in the LO. The factory settings were close, but on every single trial, a bit of improvement was possible.

At the very end of the exercise, I also tweeked the reference oscillator's crystal pull. That resulted in a short back and forth between this reference oscillator and the LO.

All in all I am somewhat satisfied with the result.

As an extra bonus, the erratic jumps of several 10 to 100Hz seem to have disappeared. Seems that the reference oscillator was slightly off, causing the PLL to fail locking from time to time.

Over the time scale of about half an hour, there still seems to be a 1-2Hz drift. I can live with that. I will try the radio as QRSS receiver, but hopes are not that high. Finally, the receiver was supposed to provide me with Navtex infos, nothing more.

(modified) PA0RDT-Mini-Whip

The "Target HF3/W", which is supposed to service me on the boat, offers switchable 12V supply via the antenna socket. Very nice, now for some active antenna to actually make use of the possibility.

Some research on the internet brought up a design by PA0RDT (pls search for it yourself). He call's the thing "mini whip". Some would call it an EH-antenne (haha, got you there), some others would call it an amplified E-probe. The latter it actually is. It is a capacitor probing the E-component of an RF-field.

However the design may be called, is it any good?

First to mention, there are many positive reports on this device. I had my doubts, but never the less, on a yacht, aerials are a compromise whatever one does. Thus, in particular inspired by the positive feedback found, I decided to give it a chance.

The key to the amplifier-circuit seems to be the 1MOhm resistor input network. I figure, 10MOhms would propably do a better job. However, not having 10MOhms-resistor in my stock, I went for 1MOhms. Which reduced my stock that much, that the next thing I'll do will be visiting my electronica-boer in order to buy said resistors.

Here comes the modification, unvoluntary, sort of... even though a 2N5109 seems widely available, it was not in my stock... I decided to replace it with a 2N918; this required to change the emitter-resistor to something higher-ohmic, in order to reduce Ic. I went for a conservative 470 Ohms. As for the DC-separating inductor, I went for what I found (whatever it was).

And this is, what I ended up with:

The result was pretty nice, although I tested indoors only. I am curious what this probe could do, when "flying" meters above my house's roof...

Should I ever build another one, that's what I would change:

• use a 2N5109 (will propably give better results... or not)
• increase the "ground-plane" part of the probe by reducing the "12V-part"

Conclusion: The probe will have to prove itself, when hoist to the top of my ship's mast. Presently it appears to be a nice alternative to something more vulnerable, since, rugged it is!