Audio feedback problem

[JimB]

I
There are many SSB radio projects listed in Google. One said it has a gain of 100dB and its audio amp is so sensitive that it picks up buzzing from nearby mains wiring.

There is a technique called Direct Conversion, effectively there is no intermediate frequency, detection is by using a local oscillator at the RF signal frequency feeding into the mixer, and out comes audio.
This only works with SSB and CW.
It has the advantage of being relatively simple, but the downside is that you get a lot of "image" responses, and, you need a LOT of audio gain, which can be a real pain.

JimB

 

[flat5]

Direct Conversion might work with FM
but we're talking W I D E B A N D fm

unless slope detection is completely impossible in such a receiver.

 

[audioguru]

I saw a direct conversion shortwave radio that still had a mixer so that 10.7MHz crystal filters could be used. It didn't have an IF amplifier, just some RF gain, the mixer, the crystal filters, the BFO, the AM detector then lots of audio gain.
The AGC was made by rectifying the audio from the power amp then it turns on a Mosfet at the input of the power amp to reduce the audio volume.

 

[Space Varmint]

Thanks Uncle Scroogie. I got it licked. I have seen the crim-da-la-crim of receivers and I am happy with my formula. Very happy. It does everything. I did change that chip driver the TDA2822. I pulled the voltage amp and took two 2N2222As and made a Darlington current amplifier. .......Beautiful! That was the ticket. I went back through the whole audio and ballanced the frequencies and gain distribution to where it will receive a monster signal intelligently in AM down to the weakest SSB signals.

I still need to pick up a mike to do it justice on my recordings.

A single conversion is good but it is my experience that at HF, dual conversion is best. The AGC you suggest is like mine except that I get my source before the volume control so that it does not effect the loop.

I'm actually enjoying my radio right now listing to Alex Jones on AM broadcast WWCR. It is sitting right next to my computer. I can flip over and hear Brother Stair too. He just said that Bush is being detained in Camp David. He was snagged trying to make a B-line for his massive estate in Paraguay!

I'll get a mike soon and post some recordings. I love this formula.

It suffers no blocking as of yet and with a long wire antenna of about 12 feet I only get very seldomly some local (frequency) intermod that doesn't last more than a minute or two. Of course I could trim the antenna to get rid of that.

If I do anything to it, I might work with the preselector a little more. I got two positions that work quite well. One for the upper half...15 to 30 MHz and one for the lower half...down to 160 Meters.

Got it tuned to WWV 10MHz now...no problem even sitting 4 inches away from my laptop which is really bad practice. That's with that noisy switch mode power supply plugged in.

I put in a couple of nice features like a 1st IF gain control & a 2nd IF gain control. The second IF gain control uses a variable capacitor for minimal skewing of the IF envelope. It was a very convenient place to add the BFO. I the BFO perfectly ballanced and loaded for minimal drift. I can adjust upper and lower sideband.

 

[Nigel Goodwin]

I still need to pick up a mike to do it justice on my recordings.

If you have some urge to record it, don't use a mike at all, use a direct line connection - take it from the top of the volume control.

 

[Space Varmint]

If you have some urge to record it, don't use a mike at all, use a direct line connection - take it from the top of the volume control.

Excellent Idea!!!

I will do just that. I need to make up a cable, shielded of coarse to prevent computer noise from getting in.

One thing that is really slick that I just lucked out on is the LED signal strength indicator. As I change band select filters and proceed across the dial there will be varying levels of IF. It's relatively flat but there are a few humps that will contribute to birdies. You do not hear any birdies because the signal strength is much greater.

At any rate, what's nice is I can use the LED to to reduce the IFs to where the LED is barely ON. Then it not only is a good indicator of when there is a station, even it is so weak I don't hear it at first. Better than that, by not driving the IFs to excess, being the more amplification the more transistor noise that is generated, I can keep a very clean level giving me maximum S/N.

 

[Space Varmint]

My schematic is still somewhat on paper but this is the first iteration that does work quite well. As I get it detailed I will post.

As you can see the audio is still not up to date.

Also I purposely erased the Crystal frequencies. That's my little secret to allow me to buy cheap off the shelf crystals to build my filters with. I'm not using standard IF freqs for two reasons. One is so that I will not be easily jammed and secondly the 2nd IF is adjustable LC filter instead of fixed ceramic as is usually done. I still have a reasonably high Q well suited for communication work.

 

[Space Varmint]

Made a mistake. Where it says "1st LO in"...that is incorrect. That is the mixer output. It goes over to the attenuator switch which comes back to the IF amps just below.

 

[flat5]

Keep a diary and write a book about this adventure.
It makes for some crazy reading.

 

[Nigel Goodwin]

Excellent Idea!!!

Much as I would like to claim credit, it's just how it's always been done, and for good reason.

The circuit is far too full of design errors to comment on, but an obvious one is the almost complete lack of any supply decoupling - a couple on the 741's, but NOT on their virtual spilt supply, and one feeding the output IC with a 10 ohm series resistor - the ONLY place you didn't need one, and 10 ohms will cripple the output.

 

[flat5]

c50 might be backwords.
r68 is wired funny.
bypass the top of r108
make r760 two resistors and bypass the one connected to vcc

etc.

 

[blueroomelectronics]

Tell me you can get better opamps than the 741 in Galveston.

 

[Space Varmint]

Tell me you can get better opamps than the 741 in Galveston.

They do the trick. I've been using them for years.

 

[Space Varmint]

Much as I would like to claim credit, it's just how it's always been done, and for good reason.

The circuit is far too full of design errors to comment on, but an obvious one is the almost complete lack of any supply decoupling - a couple on the 741's, but NOT on their virtual spilt supply, and one feeding the output IC with a 10 ohm series resistor - the ONLY place you didn't need one, and 10 ohms will cripple the output.

What would you know about design flaws?

You couldn't follow the yellow brick road, much less a radio schematic.

 

[Mikebits]

With all respect, why are you using a general purpose switch transistor like the 3904 for your front end. This stage sets the IP3 and Noise figure for your whole rcvr?

The datasheet for the 2N3904 does not even spec NF above 100KHz, so I imagine it is not very good. The IP3 for that part is not very good either so your first stage will produce many Intermodulation products.

I would change that part to a more suitable device. I can look one up for you if you wish.

 

[Space Varmint]

I think intermod is more a function of the mixer. But I am using 3904s for my first mixer.

It's funny you mentioned that because tonight I am getting intermod which I haven't been getting. And it is none other than.....Brother Stair.

But face it. I heard a caller call in one time and tell him that the first time he ever heard him was on his guitar amplifier....lol.

Look, the dude simulcasts on about 5 frequencies up and down the band with some stations being as much as 9 million watts like WHRI who's final output tube has it's own building. Nothing an antenna trimmer won't fix.

 

[Space Varmint]

Another thing you have to consider is band conditions. I know they have been predicting some sun spot activity. I'm noticing a little heterodyne but this is coming off broadcast peak hours. That would be 12:30 AM EST.

OK now here I am ant just 12:50 AM and intermod is almost gone. Did not even shorten antenna. Fading out. That's with BFO on. Probably still Brother Stair.

Also I am in Atlanta and Brother Stair is also transmitting out of WWCR which is in Nashville,TN. So probably a strong ground wave. But quickly deminishing.

*edit* also can get rid of heterodyne by raising the noise floor by simply plugging in laptop PS. That should tell you something about commercial receivers with digital control. I'm not bashing digital. I build PLLs and have a single loop ready to install for 40 meters. I will be slinging an inverted V up in the trees for some 40 Meters work.

But there is something to be said for intercept points in regard to intermod and that is strickly a function of the mixer. I think diode ring mixer are very good in reguards to that particular characteristic due to the fact that they are lossy. But then, what are you loosing and can it be made back up in the front end.

I prefer the preselector approach. I have not detailed mine out yet.

 

[Mikebits]

Your sadly mistaken. IP3 is a function of distortion and it is of concern at many stages of a radio. IP3 products in the first stage means unwanted in-band products that are not easily filtered. Clearly, if your first stage amp produces unwanted in-band products then it will most likely pass though your mixer, and your rcvr will not be able to filter them out, and you will end up with a very noisy output. That is all I have to say about it. If you are too stubborn to recognize when someone is trying to assist you, then you are welcome to all the problems you encounter.

Your so busy trying to defend, that you are not able to recognize sound advice that may help improve your project. At this point, I put up the arm waving surrender sign.

 

[Space Varmint]

In-line at 1:08 AM, no antenna trimming.

 

[Space Varmint]

With some antenna trimming.