That would be my guess too.

Yes, for now I am shielding the input leads and output leads with aluminum foil (grounded) until I find some real shielded wire. I now have the MPF102 and other hardware to build the source follower. Its been a busy weekend, so I will get back to modifying this thing tomorrow.

Update: Today I built the voltage doubler and source follower. Preliminary tests show a significant improvement. I can now get stations across the entire band. I am getting some distortion in the audio however. I will be experimenting further with the resistor at the detector output. I also know the alignment is slightly off, so that needs to be touched up. And I will also be trying an external antenna. I think I will also box the whole decoder unit in foil and ground that. Its almost there, just needs some fine tuning.

I didn't see this in your replies, but does the LED light up on all FM stations you can
get?
Also, if you have an AM signal generator capable of some output at 100 mHz, try
and see if the un-modulated quieting you hear, when you tune in the generator
can also make the LED light come on.

Just a random thought but how large is the signal going into the LM1800? Their separation spec is given with a nominal 800mV peak to peak, or 282mVrms.

Just wondering if the distortion is coming from overdriving the chip.

Wouldn't you also need to modulate the RF generator with a 19 kHz tone? And shouldn't it be FM not AM?

Dave

I think that might be to check if there is a false stereo indication.

Sorry, I misunderstood his point. Yes, the LED should not be lit with a totally unmodulated carrier. And these days, it's tough to find an FM station that is not stereo.

Dave

Luckily there are two FM sports stations in my area that are mono. Neither of them light up the stereo LED.



That could be a problem. I haven't tried higher value resistors between the detector and decoder input yet.

Today I built a shielded box to put the decoder in. I haven't finished mounting the decoder in it yet. I've also noticed that if I take the metal cover off the ratio detector, the FM dies. And Telefunken used shielded wire between the FM tuner section and the FM switch and volume switch. I'm thinking shielding is very important here. It's also shown on the schematic for the decoder input and output wires.

Distortion is most likely caused by too much signal. My suggestion is to put a variable resistor (whatever value works well with the circuit) between the output of the buffer and the decoder. That way you can set the level easier versus having to try different resistors.

Have you taken the output from the stereo decoder out of the system, and connected it to
a different amplifier/speakers ? That is using your radio and decoder as just a tuner?

If the chip was producing a mono signal, then the left and right signal would be identical. Stereo
broadcasts are sometimes difficult to judge as the old ping-pong stereo, that is one channel
dropping distinctly in level, i.e.20dB in the L to R tests.

Do you have another LM 1800 to try?

I'm still tinkering with some things. Overall, it seems that the output is weaker when the signal is run through the decoder than without it. (Remember, I am now running the detector into a source follower.) That is, you have to turn the volume higher to get the same loudness. But also there is more distortion too. I have also hooked up a larger dipole antenna, which improves the signals across the board substantially. I don't have a 1M pot to stick in there, so I am reduced to experimenting with different resistor values between the detecter and decoder.

But how do you accurately measure the voltage going into the decoder? I tried a multi-meter and it shows AC voltage jumping around between 2.9 and 3.5 VAC when I have a 1.5M resistor hooked up to the detector without the decoder hooked up to it. That doesn't sound right too me.

It might help to have a schematic of how it's actually wired up now.

A decent follower should have insignificant loss, less than 1 dB, but reducing signal level to the LM1800 very well might, depending on how much you have to cut it down, because the LM1800 is essentially 1-1, plus or minus a couple dB. For example, if you cut the input to it by half you lose 6dB of volume so, to keep the same level, you'd have to gain it back up after the decoder.

That is, assuming the LM1800 ends up driving the same audio amp the original circuit did but where is the second channel going?

A pot to stick where? What kind of distortion?

The input impedance of a multimeter is probably too low and it's 'AC' measurement 'presumes' a 60Hz sine. In short, you need an oscilloscope.

I now have 90% of the kinks worked out. 1.5M ohms between the detector and decoder is working well at eliminating distortion. I probably could go lower on the resistance, but thats all I have that works right now. Using a dipole was also a very good suggestion. I also ran the outputs through the phone jack instead of the tape jack, to take advantage of the pre-amp stage. The output is much improved.

Aside from the JFET and LM1800, I built the entire decoder out of junk parts I scavenged off old TV boards found on the side of the road. The shielded box was made out of a cut up Corn Flakes box, aluminum foil and gluestick. There is still a slight but noticeable hum when the decoder is hooked up. I thinks its from the power line coming off the 6V dial light bulb circuit. I will try shielding that and rerouting it away from tubes and stuff. I'm almost ready to make a more permanent setup using real shielded wire instead of alligator clips covered with tin foil .... wish I had a better camera,

I have decided that the hum produced using a voltage doubler is too distracting. For now, I'm switching back to battery power.

Increase the size of the capacitors in the voltage doubler and the hum should go away.