Digital Frequency Control

[Space Varmint]

Not for years, and only simple ones.

I would suggest you buy the ARRL or RSGB manuals, these are essential for anything radio based - everyone here should really have one or the other (or both).

Absolutely! I got my ARRL Handbook. WHat did you build? DC receivers? I've built a couple of dual conversions. Just never made a good full coverage digital synthesizer. I've made single band. I never put a transmitter on the air anymore without digital lock. But my VFOs never chirped. They may drift a little but no chirp! Drift is not tolerable for SSB reception.

 

[Nigel Goodwin]

I've built various single band DC and TRF receivers, there never seemed much point building superhets, as they are so freely available.

 

[Space Varmint]

I've built various single band DC and TRF receivers, there never seemed much point building superhets, as they are so freely available.

Oh Lord, not a TRF. Parish the thought. I love building receivers but hey can be time consuming. One day I will get one on PC board for the life of me. Man, there are so many things you can do. You know me. I look for rat under every wood pile. I keep thinking about that 2nd IF frequency. Now days I believe all 2nd IF's are 455 KHz. Man, too easy to jam the world! Hell, you could broadcast it over HAARP. I've had some fun playing with the 2nd IF. There are some neat things you can do. If the first IF is sharp enough, you got your selectivity. Now by shifting the 2nd IF you can move images to where they are non existent.

Another major factor. Other that a few low end radios like Grundig and Beagine which work fairly well, have you ever priced a good communications receiver lately? Man it's nothing to drop a few grand on one. We're talking the good stuff. Kenwood, Tentec, Icom etc.

 

[Nigel Goodwin]

Another major factor. Other that a few low end radios like Grundig and Beagine which work fairly well, have you ever priced a good communications receiver lately? Man it's nothing to drop a few grand on one. We're talking the good stuff. Kenwood, Tentec, Icom etc.

Well Grundig were never 'low end' radios, they made some really nice stuff - but are now long gone!

Lowe Electronics produced a fairly well received communication receiver, and I actually know the guy who designed it he's a genuine genius, bordering on insanity - a really, really clever guy.

 

[Tesla23]

Absolutely! I got my ARRL Handbook. WHat did you build? DC receivers? I've built a couple of dual conversions. Just never made a good full coverage digital synthesizer. I've made single band. I never put a transmitter on the air anymore without digital lock. But my VFOs never chirped. They may drift a little but no chirp! Drift is not tolerable for SSB reception.

Space Varmint,

The challenge of good HF receiver design is to maintain the snesitivity and selectivity you want in the mass of other (often much stronger) signals coming in the antenna.

Generally the most important thing is linearity, it's bad enough that there is a mess of other signals there, you don't want to mix them together to create something on top of your signal. So you will often see no LNA, a mixer running at obscene levels of LO power (100mW and up), followed by significant filtering. Often there is an absolute mmiminim of amplification before there has been significant filtering.

LO spurii are a no-no, they simply create spurious responses.

LO phase noise can limit your sensitivity in the presence of large interferers due to reciprocal mixing.

Don't make the mistake of going for sensitivity, below about 20MHz it is unlikely that you will be receiver noise limited and you are better off with linearity. So LNAs are not much use except at the higher end of the band and in quiet rural environments.

PLL design is a challenge, particularly if you want smooth (real time) tuning and frequency resolution suitable for SSB. One reasonably low cost way is to go for modest resolution in the first LO (say 1kHz), and instead of using a fixed second LO, provide +/- 500Hz tuning here. It does mean that the wanted signal moves around a bit in the first IF filter.

Providing the resolution in the second LO can be achieved using a second PLL and division, a second PLL using a n/n+1 divider, a DDS etc...

If you are serious check out Ulrich Rhode's books, there is one on Communication Receiver design I'm pretty sure. There is also a book called "SSB Design Handbook" or something similar.

Peter

 

[Space Varmint]

Space Varmint,

The challenge of good HF receiver design is to maintain the snesitivity and selectivity you want in the mass of other (often much stronger) signals coming in the antenna.

Generally the most important thing is linearity, it's bad enough that there is a mess of other signals there, you don't want to mix them together to create something on top of your signal. So you will often see no LNA, a mixer running at obscene levels of LO power (100mW and up), followed by significant filtering. Often there is an absolute mmiminim of amplification before there has been significant filtering.

LO spurii are a no-no, they simply create spurious responses.

LO phase noise can limit your sensitivity in the presence of large interferers due to reciprocal mixing.

Don't make the mistake of going for sensitivity, below about 20MHz it is unlikely that you will be receiver noise limited and you are better off with linearity. So LNAs are not much use except at the higher end of the band and in quiet rural environments.

PLL design is a challenge, particularly if you want smooth (real time) tuning and frequency resolution suitable for SSB. One reasonably low cost way is to go for modest resolution in the first LO (say 1kHz), and instead of using a fixed second LO, provide +/- 500Hz tuning here. It does mean that the wanted signal moves around a bit in the first IF filter.

Providing the resolution in the second LO can be achieved using a second PLL and division, a second PLL using a n/n+1 divider, a DDS etc...

If you are serious check out Ulrich Rhode's books, there is one on Communication Receiver design I'm pretty sure. There is also a book called "SSB Design Handbook" or something similar.

Peter

Hi, Nice to meet you Peter.

Well you seem like you have tried your hand a few times at receiver design. They are fun! I know exactly what you mean about sensitivity though allot of the documentation tends to down play it I have found that you still need at least a couple of stages of amplification. Too much of course will cause blocking and here is where we get into the all too intensive discussion of dynamic ratio. Of course the mixer can have allot to do with it as well. I have had allot of success with JFETs as mixers. Diode rings work good but as you said you usually need allot of power to bring those weak signals over the bias level of the diodes. That is why I was asking Nigel about germanium diodes. He says they are still available but I don't know where he gets them.

I would like to meet your friend Nigel. Is he a member? Would be nice if he has sold his own brand and written books and all. He sounds like a real hot shot. I was real lucky back when I first moved to Atlanta about 10 years ago I had been building transmitters down in Tampa then when I got here I started into building some DC receivers. I met this guy on 2 meters and he knew everything...lol. I mean everything! Didn't know the first thing about computers or programming so we swapped knowledge and he ended getting a job in computers with Johnson & Johnson. He had to move for the position but I will never forget the guy. He also taught me allot about building good rf power amps. Like I said. He knew everything.

 

[Space Varmint]

One reasonably low cost way is to go for modest resolution in the first LO (say 1kHz),

.

I have to take issue with that. I find the phase sidebands completely unacceptable at that resolution. Even if you attempt to filter it down to where it could be considered reasonable, you will drastically lose bandwidth and I would fully expect that a powerful station that comes up on the beat note of several of those sidebands would bleed through. Of course the closer to center freq. the worse the situation. I like to see the first phase sidebands down at least "full quite" and preferably 120 db down. At 120 db down you will actually have an improvement over VFO due to the sharp high Q resonance.

 

[Space Varmint]

Hey Peter? If your on tonight. Have you ever played around with noise limiters in the loop? I did once when I had to in a transmitter because one of the low frequency references coming out of the divider was adding some modulation to the carrier that was audible. My experience was that the loop was not as strong though I have read that many times noise limiters are used. Wonder what other advantages could they produce because it is the energy in the sidebands that steer the signal back on freq f0. Any ideas?

 

[Tesla23]

.

I have to take issue with that. I find the phase sidebands completely unacceptable at that resolution. Even if you attempt to filter it down to where it could be considered reasonable, you will drastically lose bandwidth and I would fully expect that a powerful station that comes up on the beat note of several of those sidebands would bleed through. Of course the closer to center freq. the worse the situation. I like to see the first phase sidebands down at least "full quite" and preferably 120 db down. At 120 db down you will actually have an improvement over VFO due to the sharp high Q resonance.

Hi Space Varmint,

I have to say I really don't understand your requirements. On the one hand you talk of 120dB suppression which suggests you are interested in getting the sorts of dynamic range you need in a high performance receiver, but and on the other you don't like leaving out the LNA but prefer to add a couple of gain stages, and you like low power JFET mixers or germanium giodes, any of which would probably cripple a high performance receiver.

Performance HF receiver design is about linearity, linearity and linearity, then you can start thinking about selectivity, phase noise and sensitivity. For an idea of what sort of specs are achievable look at (I just grabbed these randomly)

**broken link removed**
**broken link removed**

Note that with the LNA off they offer a noise figure of about 17dB with a third order intercept point of 30dBm or greater. They don't guarantee a noise figure better than 10dB.

Do some basic system analysis to work out your intercept points, noise figure and headroom. If you want a simple system for comparison, try input sub-octave filtering, a minicircuits level 10 or 13 mixer with an LO of 40 - 70MHz, 1st IF around 40MHz, simple passive filter say 1MHz bandwidth then a lossless feedback IF amp (Norton style, say 10dB gain, 3dB NF, OPIP3 over 40dBm), then a 20kHz wide crystal filter. This is just off the top of my head, and I've been out of this game for quite a few years. Note that the mixer termination is critical to achieving the mixer linearity.

I agree that a first LO with a 1kHz channel spacing will not get you to state of the art performance, but it is more in tune with the other components you were suggesting.

Peter

 

[Tesla23]

Sometimes when I post here the posts go straight up, on other times they get put on hold for vetting. I haven't made many posts but I would have thought it was pretty clear they weren't spam. Sorry Space Varmint but it is a bit hard to have a conversation when there is a day or two delay in posting - so I'm off..

For some reason this one went straight through, whereas my reply to you was held for vetting - you work it out.

 

[Space Varmint]

OK, not sure I totally understood the last statement but will make general assumptions and work-arounds

I appreciate your input allot so don't disappear on me or I should say us. It sounds as if you have something to offer the group. First off I see I will have to review all the concepts of dynamic ratio again so that I can follow you. Of course I'm a fan of Tentec and so will definitely take a look at your links. I'm not sittin' up here pretending to know everything about radio. I'm just speaking from my own experience. Interesting schemes you propose still, the one 1KHz thing. It bothers me. Yes I like high performance. I'm an even bigger fan of Kenwood and in the older days I would say Drake was probably one if not the best. Maybe where I confuse you is I like high performance, low cost & low power. The best of all worlds for what is practical or doable.

ei the ultimate....Hand crank high selectivity digital tune.

huh? huh? That would sell with the right price tag. What do ya think?

 

[Space Varmint]

Ok Peter. I see what you got there. Yeah those are the cream of the crop. They do some amazing stuff with DSP these days. I am only interested in the bottom line. How well does it work? Just by looking at them I am sure they work quite well. I bet the price tag makes them work even better.

I suppose my objective is more toward the portable line. I mean we have yet to see 40 meter walkie-talkies. Why? Can you imagine how effective they would be? A good relay network and you don't need anything else.

 

[Space Varmint]

Well I got me a loop built over the weekend. Tomorrow I should have some time to start programming and testing it. I even put in a SIP socket for on board programming.

 

[neon]

2.5 mhz to 40mhz to 400mhz ? that is the question.

 

[Space Varmint]

2.5 mhz to 40mhz to 400mhz ? that is the question.

40Mhz is what I'm using right now.

 

[martin1]

Sir,
I read your explanations about Digital Frequency Control.This is wonderful explanation. Then how to control digital meters.

----------
martin

Advertising deleted - moderator

 

[Space Varmint]

Sir,
I read your explanations about Digital Frequency Control.This is wonderful explanation. Then how to control digital meters.

----------
martin

Advertising deleted - moderator

Oh why thank you. I told what I know. Digital meters are quite different. They use what is known as D to A and A to D converters. Essentially you will clock a D to A converter and use a comparator to test the voltage. Now days a processor will use a look-up table to give the voltage output though it is possible to do it without a processor by using the same counter output to be read on a 7 segment display. Much simpler with a processor.