The AM Forum
March 28, 2024, 07:24:56 AM *
Welcome, Guest. Please login or register.

Login with username, password and session length
 
   Home   Help Calendar Links Staff List Gallery Login Register  
Pages: [1]   Go Down
  Print  
Author Topic: Audio Processing for AM and SSB  (Read 5337 times)
0 Members and 1 Guest are viewing this topic.
KI4YAN
Member

Offline Offline

Posts: 161


« on: September 18, 2017, 05:32:01 PM »

I'm into the audio stage of my transceiver project now, and am looking at audio processing, something that seems to get left out of a lot of newer homebrew designs, but was often included in the older designs. The power supply is just about done and ready to go, and the microprocessor control of the VFO and LO is done and works correctly-so now on to the audio!

I am using the MC1496 balanced modulator to generate SSB and AM, and am going to be generating AM by unbalancing the modulator. The idea is to use a DPDT relay to select one of two balancing networks, so that the SSB mode can be adjusted for best carrier null, and the AM mode can be adjusted for proper carrier level. Hopefully this will allow convenient modulation.

That seems to have the AF-to-RF conversion covered, now the audio input/output is next.

For the Mic, an electret unit will be used, so +5v phantom power will be supplied by the mic amp (MC4558), then another 4558(I have a pile of these pulled from old stereo gear that got junked) will first high-pass the audio, cutting everything under about 100hz, then a low-pass filter to cut everything over 3Khz. Into another 4558 configured as an all-pass network/phase rotator, (Will two stages of this be enough? I see 4 and 8 stages in the wild...) and I think that should be enough processing to present a good signal.

So, Mic amp -> bandpass filter -> all-pass filter -> (do I need mic AGC?) -> Modulation input to MC1496

Does anyone here have any suggestions or improvements? Maybe a good method of mic AGC if the group thinks it's needed or even just extremely handy?
Logged
Steve - K4HX
Administrator
Member

Offline Offline

Posts: 2722



« Reply #1 on: September 18, 2017, 06:02:18 PM »

Good tutorial at the link.

http://amfone.net/ECSound/processing101-1.htm

Be aware that processing requirements for AM and SSB are different. Standard audio processing generally doesn't work well for SSB. That's why many SSB processors work at RF.
Logged
N1BCG
Member

Offline Offline

Posts: 836


« Reply #2 on: September 18, 2017, 06:07:47 PM »

You're certainly on the right track. Ideally, the phase rotator/all-pass filter comes first before any gain control or EQ. A Mic to Line level amp can precede it of course, as long as it doesn't alter the waveform.

The only "must have" is peak control. Negative peak limiting to prevent overmodulation and positive peak limiting to prevent circuit voltage peaks and to reduce distortion in most receivers. Clippers produce the loudest audio for this but require a thoughtfully designed LPF to filter out the harmonics caused by clipping a waveform.

An AGC will provide far more perceived loudness than any amount of positive peaks can, so that would be a great addition for keeping the average modulation consistent and optimized.

You might consider having adjustable bandwidth. 3kc audio (6kc bandwidth) works fine for SSB, but for AM on an uncrowded band, a bit wider will be impressive. A popular preemphasis is 75uS made from a 7.5K resistor and 0.01 uF cap in parallel. The loading resistor value (L network) will determine the boost. 470 Ohms give an 11dB boost which is very close to the NRSC AM broadcast specification in the U.S.
Logged
KI4YAN
Member

Offline Offline

Posts: 161


« Reply #3 on: September 18, 2017, 07:07:45 PM »

Does the negative peak limiting apply when running low-level modulation? remember, I'm not plate modulating here, it's a double-balanced gilbert cell modulator.

I'm aware of the differences between SSB and AM processing-I just want to be able to do what I can do to produce a good signal. Not trying to go overboard or emulate a full rack of gear in this little box. That's part of why I'm asking-some stuff probably just can't be done without seperate dedicated audio chains.

Peak control is something I haven't worked out yet-I see most clipper/limiter circuits essentially are FM IF strips, audio is mixed up to some IF and then back down to audio after the limiter. Not something I want to try to do on this one, but I have seen back-to-back diode clipper/limiters used in other audio realms-maybe one can be adapted here.

Adjustable bandwidth may not really work here-I am almost out of front panel room and the filters are 6Khz wide and 3Khz wide-and have nice steep symmetrical skirts too. I plan to make the high pass and low pass filters adjustable with a twiddlestick, but probably not from the front panel. Been listening to recordings of my voice and other voices that I know well through a graphic equalizer program called PEACE, which is a front end for a program called Equalizer APO. Lets me set up various passbands and listen in real time, and there are a lot of other processeing effects I can try out. So far, a filter curve that cuts all frequencies under 200hz and all high frequencies over 3200hz sounds pretty good-minimal intelligence is lost. I do get a bit of that SSB sound though, but it seems to be the best compromise. I also looked at 400-3400hz, which sounds OK but a bit pinched off on most male voices, including mine, that I listened too.

So at this point, the chain looks like:

Mic Preamp -> All-pass -> Bandpass filters -> Limiter -> Balanced Modulator

The limiter is what is holding me up now-I'll need to work out a good circuit to fit that slot.
Logged
N1BCG
Member

Offline Offline

Posts: 836


« Reply #4 on: September 18, 2017, 07:16:50 PM »

The chain looks good, but the limiter will be a critical component even with low-level modulation as excessive audio creates a secondary waveform.
Logged
KI4YAN
Member

Offline Offline

Posts: 161


« Reply #5 on: September 19, 2017, 06:23:45 PM »

Below is a limiter circuit that may work out for us-it's designed for audio recording in a studio environment. it seems a bit complex but if it does what it needs to do, that's OK. The circuit is from Rod Elliot's ESP audio pages, Project 67.



"As described above, the attack time with the values shown is 5ms, with a release time of around 1 second. This is a good compromise for most audio material, but is readily changed by altering the values of R14 (attack) and R12 (release). Be careful of values for R14 of less than 1k, as the opamp will be unable to supply the current needed to charge C5. Ideally, C5 needs to be a low leakage capacitor - either a low leakage electrolytic or a tantalum if you must (although I never recommend tantalum for anything). A standard electro is probably inappropriate for this circuit, especially if longer release times are desired. Having said that, most are better than you may have been led to expect.

In addition, keep R12 a minimum of 10 times R14 ... for example, if R14 were to be 1k, the minimum value for R12 will be 10k. This would be a very fast limiter indeed! Faster decay is possible, but it doesn't sound nice. The circuit has been changed so that R12 (decay control) is connected to the diode side of R14, so if a fast decay is needed the control voltage is not attenuated."

Maximum Attenuation             40dB
Noise Level (unweighted)     -80dB (ref. 1.65V RMS output)
Typical Max. Output Level     1.65V RMS
Gain                                     6.8 (16dB)
FET Voltage (at max. o/p)    < 45mV typical)
Distortion                            < 0.5% typical

Is this overkill limiting, or fairly decent performance for a low-level audio limiter? I have used the circuit a long time ago, and it does "sound" ok for audio work-but I have never tried to use it in a radio transmit chain. Only big bugaboo I'll have to work out is that I have +12V and -8V supplies right now-I could split them at +/-10V though. Biasing may need to change a bit, although it may not need to change at all. I'll ask Rod if I can get an email out to him.
Logged
KI4YAN
Member

Offline Offline

Posts: 161


« Reply #6 on: September 22, 2017, 05:39:33 PM »

I started in on this and got the audio chain built into what I thought I needed on the breadboard-it all works. Doesn't sound tea bag either, but getting enough poles in the bandpass filter seems to be a problem-I am only down 4.4dB at 100hz and 3600hz, with a filter set up to pass 200hz-3200hz.

But do I really need an audio bandpass filter, if I am running the DSB modulated output through an 8-pole crystal filter?

I assumed I would/should need to do this, to cut out rumble and low frequency garbage on the bottom and pops and essssesss on the top, but do I need a low-pass filter to cut highs, or will the bandwidth of the crystal filter chop the highs down sufficiently?

Logged
Steve - K4HX
Administrator
Member

Offline Offline

Posts: 2722



« Reply #7 on: September 22, 2017, 08:10:32 PM »

What's the bandwidth of the crystal filter?
Logged
KI4YAN
Member

Offline Offline

Posts: 161


« Reply #8 on: September 22, 2017, 08:38:54 PM »

The filters are KVG brand, XF-9 series. 8 pole filters.

S42: +/-3.6K bandwidth at -3dB measuring points, center frequency 9.000000Mhz (7.2Khz bandwidth)
S43: +/-1.55K bandwidth at -3dB measuring points, center frequency 8.998200Mhz (3.1Khz bandwidth)
S44: +/-1.55K bandwidth at -3dB measuring points, center frequency 9.001800Mhz (3.1Khz bandwidth)

I've gotten two different listings for these filters as far as impedance goes, some places show 560ohms in parallel with 25pf, some places show 1000ohms in parallel with 25pf.
Logged
Steve - K4HX
Administrator
Member

Offline Offline

Posts: 2722



« Reply #9 on: September 22, 2017, 11:38:24 PM »

Those 8-pole filters will roll of the highs on the audio. I wouldn't worry too much about the roll off on your audio filters.
Logged
KI4YAN
Member

Offline Offline

Posts: 161


« Reply #10 on: September 27, 2017, 10:28:08 PM »

Short of having a spectrum analyzer, which I don't have functional yet, what's the best way to tune and adjust this DSB-SC/AM generator? I want to get carrier reject/carrier insert levels correct, signal levels correct etc. before I start on the next module, which will be the bidirectional IF/Filter strip. Trying to use each module to test the next module, and since I have the detector and detected audio finished and working, and now the generator assembled, the next bit is the IF strip.
Logged
Steve - K4HX
Administrator
Member

Offline Offline

Posts: 2722



« Reply #11 on: September 28, 2017, 10:50:21 AM »

If you have a scope, you can check and adjust the carrier balance. You could also do this with another receiver using the S-meter.
Logged
Pages: [1]   Go Up
  Print  
 
Jump to:  

AMfone - Dedicated to Amplitude Modulation on the Amateur Radio Bands
 AMfone © 2001-2015
Powered by SMF 1.1.21 | SMF © 2015, Simple Machines
Page created in 0.055 seconds with 19 queries.