Audio feedback problem

[Mikebits]

Order yourself a catalog from Mini-circuits. The app-notes within the catalog are very useful, and you may find a few substitutions for your current design.

Mini-Circuits : Request Product Literature

Don't get me wrong, I think it is great that you have made a working radio, but consider that there is always room for improvement. Your First RF stage is one.

 

[Space Varmint]

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.

Oh, I see. So you are not capable of debate because you are the "be all end all" of radio knowledge. OK.

Actually it is a function of mixer intercept points, and if you have not designed for short-wave, than you may not understand what is forgivable.

I was just listening to 80 meter appliance operators who own very expensive toys, complaining about a flying saucer noise in the back ground. All three people heard it. I did not. I used to, but got rid of it in the pre-selector.

 

[Space Varmint]

Bottom line, but I will discuss it with you.

IMD is a mixer product. All the front end does is increase the signal strength. Nothing magical about it.

 

[Mikebits]

Oh, I see. So you are not capable of debate because you are the "be all end all" of radio knowledge. OK.

Actually it is a function of mixer intercept points, and if you have not designed for short-wave, than you may not understand what is forgivable.

I was just listening to 80 meter appliance operators who own very expensive toys, complaining about a flying saucer noise in the back ground. All three people heard it. I did not. I used to, but got rid of it in the pre-selector.

I never claimed to be the end all authority on radio design; however, I do have experience in HF radio design, and one thing I learned while working in this area is that the RF front stage sets many parameters that will dictate the overall radio performance. IP3 is one, Noise figure is the other. If you ignore these then you can expect the sort of audio sample that you have supplied for us earlier.

My suggestion to you is this:

First change your RF front end to a part with Better IP3 specs, and NF, and in your case NF would be a big thing. Also if you have the gear, test your front end for these parameters.

A weakly designed front end will ripple on down to the rest of your already whimpy unit... and your output will just be droopy.

Have you ever done a cascaded gain distribution of your design which included NF and IP3.

There are Gain distribution spreedsheets available.

 

[Space Varmint]

These guys were just complaining about an oscillation in the back ground. He means heterodyne.

The first guy talking, down in the mud, is from Jamaica....attached

Also here is an AM recording....AM-Jones.wma

Signal is very strong. Rough on AGC. There is some clipping heard only as distortion.

 

[Space Varmint]

I never claimed to be the end all authority on radio design; however, I do have experience in HF radio design, and one thing I learned while working in this area is that the RF front stage sets many parameters that will dictate the overall radio performance. IP3 is one, Noise figure is the other. If you ignore these then you can expect the sort of audio sample that you have supplied for us earlier.

My suggestion to you is this:

First change your RF front end to a part with Better IP3 specs, and NF, and in your case NF would be a big thing. Also if you have the gear, test your front end for these parameters.

A weakly designed front end will ripple on down to the rest of your already whimpy unit... and your output will just be droopy.

Have you ever done a cascaded gain distribution of your design which included NF and IP3.

There are Gain distribution spreedsheets available.

First point out what it is you heard? The recording even though inline is still not as crisp. I don't know what that is.

But you need to pint out specifically what particular characteristic you found lacking.

I don't need you to bash my design using general high-tech lingo.

What didn't you like specifically?

A heterodyne?

I heard many of them complaining about heterodyne.

When you say "droopy"

Are you talking about loading from strong signals?

Yeah, there's a little of that. I have no buffer between the BFO and 2nd IF. It was done deliberately. I am striving for low cost, low power, convertable from portable to base station. It is intended for efficient battery usage.

Please be specific as I do appreciate your constructive criticism....general high-tech design jargan tells me nothing. I need symptoms.

Also remember, It is not fully complete as I have plans for pluggable PLL for specific bands. I have one ready to go for 40 meters.

I still disagree with you about rf front ends though. I do not know what P3 is. I have never encountered that parameter. Dynamic ratio revolves around the mixer.

 

[audioguru]

I am an audio guy. I am not one of those weirdos who use expensive cables.
I just like good sound with a frequency response flat from 20Hz to 20kHz, a S/N ratio of at least 90dB and distortion of only 0.003%.

Your shortwave radio sounds horrible to me. Its frequency response is from 600hz to 1700Hz, its S/N ratio is about 6dB and its distortion is 50% (with most of the harmonics reduced by the severe treble frequencies cut) and would be close to 100% distortion if the upper frequency response was flat.
Your radio has some kind of oscillation during loud vowels in speech. It sounds like the person talking is two people talking in a musical Chord.

My grandmother's shortwave radio sounded much better 53 years ago. With a few meters long piece of wire for an antenna it received hams and stations from around the world. Its frequency response was from 100hz or 200Hz to 3khz or 4kHz, its S/N ratio was variable but averaged about 40dB and its distortion was not heard (at most 0.2%).

 

[Space Varmint]

I am an audio guy. I am not one of those weirdos who use expensive cables.
I just like good sound with a frequency response flat from 20Hz to 20kHz, a S/N ratio of at least 90dB and distortion of only 0.003%.

Your shortwave radio sounds horrible to me. Its frequency response is from 600hz to 1700Hz, its S/N ratio is about 6dB and its distortion is 50% (with most of the harmonics reduced by the severe treble frequencies cut) and would be close to 100% distortion if the upper frequency response was flat.
Your radio has some kind of oscillation during loud vowels in speech. It sounds like the person talking is two people talking in a musical Chord.

My grandmother's shortwave radio sounded much better 53 years ago. With a few meters long piece of wire for an antenna it received hams and stations from around the world. Its frequency response was from 100hz or 200Hz to 3khz or 4kHz, its S/N ratio was variable but averaged about 40dB and its distortion was not heard (at most 0.2%).

LOL...listen to the AM

 

[Space Varmint]

OK Mikebits. You've peaked my curiosity. Who did you work for in short-wave?

I never worked in short-wave but I worked for 3 different companies in radio. Most was VHF & UHF. I worked for Melpar E-Systems and Motorola and Teltronic industries.

 

[Mikebits]

I worked for a company called Cubic Communications, now part of Cubic Defense. The company specialized in HF to UHF receivers.
https://www.cubic.com

 

[audioguru]

AM-Jones is extremely distorted, there is buzzing in the background, some of the low audio frequencies are missing and all of the high audio frequencies are missing.
It sounds awful.

 

[Mikebits]

I worked for cubic in the late 80's but from what I remember IMD and Noise figure were key. IIRC a lot of the front ends we used were Lossless Feedback Amps, or a derivative of the Norton amp. They had the benefit of high dynamic range and low NF. At the very least I would suggest you change your RF front end amp to a 2N5109 or something like that along with a bypass cap on top of L4. It should improve your front end performance.

Assuming your preselector is designed for 50Ω you should delete R110, it is not needed. Also what is the reason for R111? It is probably causing a big impedance mismatch, not to mention it will look inductive.

Here are a few links about lossless feedback amps. Perhaps you can get a few ideas.

http://www.home.earthlink.net/~christrask/Lossless%20Feedback%20Amplifiers.pdf

https://www.electro-tech-online.com/custompdfs/2008/11/lwa0071.pdf

**broken link removed**

A little info on IMD
**broken link removed**

2N5109 data sheet.
**broken link removed**

 

[Nigel Goodwin]

I never worked in short-wave but I worked for 3 different companies in radio. Most was VHF & UHF. I worked for Melpar E-Systems and Motorola and Teltronic industries.

If you worked for them, what did you do? - sorry to be excessively harsh, but your radio design presented here shows you have no electronic design skills whatsoever, and not the slightest clue of how a radio (or indeed any other circuit) works.

I am impressed that you've managed to get any noises out of it whatsoever, by simply throwing loads of extra un-needed stages at the circuit - A* for perseverance and effort, but if it was an exam, you would be marked as ungraded.

Is this entire thread a wind up?, and you havent really built it at all? - you have answered a number of queries in other threads fairly competently, and yet this thread shows otherwise.

 

[Space Varmint]

I worked for cubic in the late 80's but from what I remember IMD and Noise figure were key. IIRC a lot of the front ends we used were Lossless Feedback Amps, or a derivative of the Norton amp. They had the benefit of high dynamic range and low NF. At the very least I would suggest you change your RF front end amp to a 2N5109 or something like that along with a bypass cap on top of L4. It should improve your front end performance.

Assuming your preselector is designed for 50Ω you should delete R110, it is not needed. Also what is the reason for R111? It is probably causing a big impedance mismatch, not to mention it will look inductive.

Here are a few links about lossless feedback amps. Perhaps you can get a few ideas.

http://www.home.earthlink.net/~christrask/Lossless%20Feedback%20Amplifiers.pdf

https://www.electro-tech-online.com/custompdfs/2008/11/lwa0071-1.pdf

**broken link removed**

A little info on IMD
**broken link removed**

2N5109 data sheet.
**broken link removed**

Forget Nigel...what a ______edy ____ _______

All I got to say to you Nigel is where is your design?

OK, Mike you sold me. Actually that schematic I posted is just the first iteration. It is far from complete but have not had time to update it. So by saying that I do already have the bypass cap on the front end. I have not had much of a chance to use it and last night and early morning I did experience excessive IMD. I'm gonna try the new front end.

What do you think about a JFET? I got some 2N5486 and I think I have some 2N5485s.

 

[Mikebits]

What do you think about a JFET? I got some 2N5486 and I think I have some 2N5485s.

The NF @ 100MHz is 2.0, so that should be a suitable device from a NF aspect. Can't tell if IMD is good though, but if your not near high powered radio interference it might not be such a problem. The thing is that a low IMD device will act like a mixer and sound crummy. Without an elaborate test bench it is best to base your specs by design. I assume you have a sig gen, so you should perform MDS (minimal disconcernable signal) test of your receiver, using different front end devices. Your MDS should be in the range of -105 dBm or better. I have tested radios with an MDS of -123 dBm. I think there is a theoretical limit of -127 dBm.

 

[Space Varmint]

The NF @ 100MHz is 2.0, so that should be a suitable device from a NF aspect. Can't tell if IMD is good though, but if your not near high powered radio interference it might not be such a problem. The thing is that a low IMD device will act like a mixer and sound crummy. Without an elaborate test bench it is best to base your specs by design. I assume you have a sig gen, so you should perform MDS (minimal disconcernable signal) test of your receiver, using different front end devices. Your MDS should be in the range of -105 dBm or better. I have tested radios with an MDS of -123 dBm. I think there is a theoretical limit of -127 dBm.

Oh OK Mike. cool. That's just it. I live about 0ne mile from an AM broadcast station on 750 KHz. I even pick there frequency up on my freq. counter. Any wonder I want a good prescaler?

As for Cubic. It was my understanding that Drake got the military receiver contracts.

 

[Mikebits]

As for Cubic. It was my understanding that Drake got the military receiver contracts.

I don't know about that. Cubic actually built a receiver used on the Hubble, however; I did not have the privilege of working on that project.

 

[Mikebits]

If you want some ideas for receiver design. I found this site that host one of Cubic Comms old receiver designs. It was the bread and butter for CCI for a long time, and it was our flagship radio back in the 80's.
**broken link removed**

Notice, the radio used switcher supplies.

 

[Space Varmint]

If you want some ideas for receiver design. I found this site that host one of Cubic Comms old receiver designs. It was the bread and butter for CCI for a long time, and it was our flagship radio back in the 80's.
**broken link removed**

Notice, the radio used switcher supplies.

You got some good links. Your OK.

Still got that radio bug, dontcha?

 

[Space Varmint]

I just caught that. Switcher???

I've got to see the filter on this. Of course you would not get away with that on any transmitter of significant power. Well maybe FM.

But I'm getting the pdf and I got the one in noise figure. Reading it now. I see what they are saying. Inner band IMD. I'll go along with that.