Need help with Yaesu FT-840 DRM mod

[todd_terry]

Hi everybody,

I have a FT-840 and I'd like to enjoy DRM reception too.
I've seen in the Google cache (here (http://64.233.183.104/search?q=cache:lSONaFLgrYcJ:www.drmrx.org/forum/showthread.php%3Fs%3Dab5cca670249ad0873fe966d0bb98 565%26postid%3D6539+ft-840+12khz&hl=fr&gl=fr&ct=clnk&cd=2)) that some members of this forum already did it but I can't access to the attachements (and I didn't receive any reply to my PMs yet).

Could somebody tell me how I can do it please?

My idea was to use 2 NE612 (or SA612A) + a frequency divider:
the first would add 2 fixed frequencies: 4.406MHz + 12 MHz = 16.406MHz
16.406MHz/2 = 8.203MHz (= the 2nd IF 8.215MHz - the DRM bandwidth 0.012MHz)
then the second would substract 8.203MHz from 8.215MHz to get the whole DRM signal.

Would it work?

[Edited]
The modification I've made is explaned here: http://yaesu.chez-todd.com/DRM.html

[dk8cb]

Hi Terry,

but why do you want to build the downconverter in such a complicated manner when it is a lot easier to downconvert from the last 455 kHz IF with just a single NE612 (and a ceramic resonator) as written in the post which you gave the link to?

Roland

[todd_terry]

Hi Roland,

unfortunately, the 455kHz IF is only available if the optional FM board is present in the transceiver.

[FritzWue]

Hi Terry,

thanks for pm, sorry, had my FT840 not on the table, so late answer here.
The original post and several other modifications were lost when the forum moved some time ago I think.

I did it like Roland suggested. My FT-840 has the optional 455kHz FM IF board installed. I am using FM mode to receive DRM. It is not the optimum as the agc is reduced in function and you have to keep a certain IF signal level manually with an antenna attenuator.

I used the Sat-Schneider converter board with 467 kHz crystal:
http://www.sat-schneider.de/DRM/DRM.htm
You have to use the option "Flip Input Spectrum" in DREAM.

This converter board is mechanically fixed on the backside of the FM board with three wires:
+9Volt supply to pin3 of board connector (FM 9V)
GND to the grounded end of the 8.6708 MHz quartz
Input to the pin of the 455kHz ceramic filter that is connected to pin 5 of the MC3357

The original ceramic filter is too wide for good results. I found optimum performace with a filter that is specified a little too narrow, but results were best with a muRata CFW455HT.
CFW455F, G or H should also work.

I connected the 12 kHz output to an additional cinch connector at the FT840 backside via a short piece of RG174, see attached fotos.

[todd_terry]

Thank you very much Fritz for those informations.
I think I need to find a FM-747 then...
Or I can still use the 2nd IF but with only one downconverter since the bandwidth is not critical, according to Dream's website

Since our receiver scans the entire sound card bandwidth for DRM signals, no special IF (e.g. 12 kHz) is needed.
A 8.192MHz crystal should be enough, I hope.

I'll give you more informations about the results as soon as possible.

[dk8cb]

A 8.192MHz crystal should be enough, I hope.

An 8.192 MHz crystal would yield an IF of 23 kHz, ie an IF range of 18 to 28 kHz in case of a 10 kHz wide DRM signal. This IF range is too high since the upper frequency limit of a soundcard is 24 kHz at a sampling rate of 48 kHz. In practice, it is even lower because of lowpass filtering on the soundcard.
Perhaps you can tune the crystal a bit higher but this will require a specially designed oscillator. Even with such a specially designed oscillator, the possible tuning range may still not be sufficiently large.

Roland

[FritzWue]

"since the bandwidth is not critical"

Besides the lowpass problem of the soundcard I always found that a good IF selectivity improves reception as it reduces the interference from adjacent frequencies.
DRM is very susceptible to AM interference.

[todd_terry]

Well, I couldn't find a FM-747 so I've build the downmixer with what I have: 2nd IF & a 8.192MHz crystal. This is not the best solution, I know, but I have pretty good results with that.
I've made some screenshots here (http://yaesu.chez-todd.com/DRM/screenshots/) to promote DRM around me. As you can see, the SNR is not that bad (and I can receive TDF's transmission @ 25775kHz (http://yaesu.chez-todd.com/DRM/screenshots/DRM_TDF_nosound.png)).

[dk8cb]

Hi Terry,

since your screenshots show an IF of about 10 kHz, it looks like you have been able to tune the crystal from its native 8.192 MHz up to 8.205 MHz. Did you make any special provisions (like a parallel inductance) to accomplish that or did it work without by just using the standard circuit?

Roland

[todd_terry]

Hi Roland,
I haven't tried to shift the crystal's frequency by adding an inductor yet, the DRM signals are strong enough to be decoded even with the 23kHz output of my downmixer.
I've noticed a strange behaviour however:
when I increment the frequency, I see the spectrum going from the right to the left (but that's not the problem), I also see some signals going from the left to the right then going back to the left while I'm incrementing the frequency. If I uncheck "Flip input spectrum", I can decode these "inverted" signals.
It happens with Deutsche Welle's transmission @ 3995kHz for instance.
I'm sure I've made something wrong, but I have no Idea what it could be.

[dk8cb]

... even with the 23kHz output of my downmixer.

Hmmm... some of your screenshots show an IF frequency of 10 kHz. So perhaps your IF bandwidth is large enough so that you can shift the signal within it such that it moves to 10 kHz.
But with such a large IF bandwidth, there may also be spectral components below your crystal frequency, which will also yield a mixing product within the soundcard's audio bandwidth. These products move into an opposite direction when you tune the receiver.
They could also be unwanted higher order mixer products but this is easily determined: Detune your receiver by 1 kHz. If the unwanted product also shifts by 1 kHz, then it is not a product of higher order. If it moves by 2, 3 or n x 1 kHz, then it is a product of the respective (n+1)th order.

It might also be a frequency difference product caused by your soundcard undersampling some input components above 24 kHz. You can test this easily:
Tune the receiver to just a single carrier and adjust the frequency, such that it moves over the audio IF band and beyond 24 kHz. Do any unwanted products appear? Could their frequencies be the sum or difference of soundcard input frequency and the soundcard's sampling frequency ie 48 kHz? If you put in sufficiently large signals above 24 kHz into your soundcard, then they will be mixed with 48 kHz and multiples of it, resulting in products in the 0...24 kHz range, which - depending on the input frequency - may also move in opposite direction. The effect is called aliasing and is the result of undersampling, look here (http://en.wikipedia.org/wiki/Undersampling) for more information. An input lowpass filter limiting the input frequency range to 24 kHz will help in such a case.


Roland

[todd_terry]

Well, I've watched the downmixer's output with a spectrum analyser and I see exactly the same thing so this is not a problem with the soundcard (but what you wrote about sampling problems is very interesting, I wouldn't have think about that).
Your first guess seems to be the good one because the IF bandwidth is very large.
The only solution, I think, would be to put a bandpass filter between the IF and the downmixer's input to reduce the downmixer's input bandwidth.

The screenshots show a 10kHz bandwidth because there's no signal close to the DRM transmission, at least that my supposition. You know more about it than I do :)

[dk8cb]

The screenshots show a 10kHz bandwidth because there's no signal close to the DRM transmission, at least that my supposition.

The 10 kHz bandwidth is not the point, it's the 10 kHz centre frequency!

If your crystal oscillator really oscillates at 8.192 MHz then, assuming that the receiver is exactly tuned to the centre of the DRM signal, the audio IF centre frequency should not be 10 kHz but 23 kHz which is too high anyway, because it results in parts of the DRM spectrum located beyond 24 kHz and these won't be sampled properly.

With a very large IF bandwidth however, you can detune the receiver by another 13 kHz and come up with a DRM signal centred around 10 kHz, but then the receiver's frequency display won't display the proper centre frequency of the DRM signal but one that is 13 kHz apart.

Roland

[todd_terry]

Ok, I'm sorry I misunderstood what you wrote. You're right, I have to set the receiver's frequency 10.5kHz below the DRM signal to have the 10kHz center frequency on Dream.

[dk8cb]

Also keep in mind, that with "Flip input spectrum" selected, what is displayed as 24 kHz in Dream, in reality corresponds to an input frequency of zero and what is displayed as 0 kHz, is really an input frequency of 24 kHz!

I overlooked that point in what I wrote above, since I didn't look too close at your screenshots. So, with "Flip input spectrum" selected, you see a signal centred around 10 kHz but it's really on 14 kHz.

Roland