150% Positive Peaks

[ka1tdq]

I've been tweaking my home-brew linear/transmitter combo with audio and power drive to get the perfect combo.  I think I'm at a point where I've found the perfect match.  I'm using 200 watts carrier and my PEP-hold feature on my Daiwa wattmeter is showing that I'm at 1200 watts peak.  My math tells me that I'm at 150% positive peak.

I was talking with WV5G last night in NM who told me that my audio sounded great.  We ran a test and I increased audio to 1400 watts, but he started to notice distortion at that level.  In our discussion, he mentioned that he'd never heard a class E rig on the air, and that the audio level between my rig and an E rig would be equivalent.

I thanked him for the compliment, but that I had heard many E rigs on the air back east and the audio quality is unmatched.  

I would like to say thanks to all who have helped me get my station to where it is today, from parts to technical support.  As I gather parts for my 40 meter E rig, hopefully I can put super-efficient and quality sounding RF on the air from Phoenix as well.

Jon
KA1TDQ

[w1vtp]

Yup!

That's what my math shows.  Too bad we weren't closer.  Hope your new QTH is working out.  Congrats on your project

Al

[W3GMS]

I've been tweaking my home-brew linear/transmitter combo with audio and power drive to get the perfect combo.  I think I'm at a point where I've found the perfect match.  I'm using 200 watts carrier and my PEP-hold feature on my Daiwa wattmeter is showing that I'm at 1200 watts peak.  My math tells me that I'm at 150% positive peak.

I was talking with WV5G last night in NM who told me that my audio sounded great.  We ran a test and I increased audio to 1400 watts, but he started to notice distortion at that level.  In our discussion, he mentioned that he'd never heard a class E rig on the air, and that the audio level between my rig and an E rig would be equivalent.

I thanked him for the compliment, but that I had heard many E rigs on the air back east and the audio quality is unmatched.  

I would like to say thanks to all who have helped me get my station to where it is today, from parts to technical support.  As I gather parts for my 40 meter E rig, hopefully I can put super-efficient and quality sounding RF on the air from Phoenix as well.

Jon
KA1TDQ

Keep in mind Jon that when you run extensive asymmetrical modulation percentages, you will sound distorted on many receivers due to the conventional envelope detector.  So if you want to sound good on most people receivers turn it down a notch or two and run some voice processing to keep the average percentage up. 

Joe GMS

[steve_qix]

I've been tweaking my home-brew linear/transmitter combo with audio and power drive to get the perfect combo.  I think I'm at a point where I've found the perfect match.  I'm using 200 watts carrier and my PEP-hold feature on my Daiwa wattmeter is showing that I'm at 1200 watts peak.  My math tells me that I'm at 150% positive peak.

I was talking with WV5G last night in NM who told me that my audio sounded great.  We ran a test and I increased audio to 1400 watts, but he started to notice distortion at that level.  In our discussion, he mentioned that he'd never heard a class E rig on the air, and that the audio level between my rig and an E rig would be equivalent.

I thanked him for the compliment, but that I had heard many E rigs on the air back east and the audio quality is unmatched.  

I would like to say thanks to all who have helped me get my station to where it is today, from parts to technical support.  As I gather parts for my 40 meter E rig, hopefully I can put super-efficient and quality sounding RF on the air from Phoenix as well.

Jon
KA1TDQ

Keep in mind Jon that when you run extensive asymmetrical modulation percentages, you will sound distorted on many receivers due to the conventional envelope detector.  So if you want to sound good on most people receivers turn it down a notch or two and run some voice processing to keep the average percentage up. 

Joe GMS

This is true, although not because of the conventional (non-sync) detector, it's because the receivers are improperly designed (including in the detector!!).  If the envelope detector is properly designed, it can (in theory) demodulate unlimited positive peaks.  Of course this is not technically practical, but you get the idea.

[N2DTS]

Its great that you CAN do it, but it might not be best to run it that way.
Just what transmitter is it?

How do you limit the negative modulation?
I do not think 150% to 95% is normal for most voices.
I also hear lots of great sounding class E rigs but they are all very wide, and many have so much modulation that they sound distorted on many receivers.
What is the point of that?
SSB has a very high peak to average ratio, lots of audio, does not make me want to run it.

[ka1tdq]

The transmitter is my single FET, class C rig.  I have a 2.5db attenuator in line to the linear which gives me 20 watts carrier to drive the linear.  The carrier output of the linear is 200 watts. 

I'm using a 3-diode keep alive circuit as a negative peak limiter. 

And, maybe >100% isn't for all voices, but I spoke with that station in NM and a guy in CA last night and both said my audio was awesome.  Listening with my headphones on and my ricebox receiver RF gain turned down, it sounds good to me too.

Before, I was running a big carrier and less audio and didn't get many hoorah remarks.  This change seems to be for the better.

Jon

[W3GMS]

This is true, although not because of the conventional (non-sync) detector, it's because the receivers are improperly designed (including in the detector!!).  If the envelope detector is properly designed, it can (in theory) demodulate unlimited positive peaks.  Of course this is not technically practical, but you get the idea.
[/quote]

Yep I agree, but probably most of the stations I talk to are using envelope detectors that are poorly designed.  I believe they are generally happy with their receivers so its very unlikely they will change.  So, based on the larger population that I want to sound good to, I have chosen to not run extended peak modulation.  It seems like on most receivers going up to 120% is ok, but much past that and you get the nasty peak distortion in the detected audio.     

I want to evaluate the Kahn symmetrical audio circuit and give that a try.  For those not familiar with Kahn's circuit here is a link.

https://www.google.com/patents/US3060389

Also, here is some older dialog from this site on this dialog:

http://amfone.net/Amforum/index.php?topic=25597.0;wap2

Saying that, whatever you choose to operate is as good as it gets.   The subject of good audio will never die.

Joe - GMS
   

[N2DTS]

I was very wide when I tried the 3 diode peak limiter.
Have you had someone look at your signal?

I think it had something to do with the diodes not being fast enough on the high frequency stuff.

[ka1tdq]

Nobody has taken a look at my signal as far as I know. 

If I were wide, then I could replace the diodes with IXYS fast recovery epitaxial diodes.  That might help. 

But, 'round these pah-ts, there's a group of sidebanders that get on 7.288 but clear out around 9am for the 7.293 AM crowd.  It works out really well.  I haven't come across swishing VFO's yet.  Being wide wouldn't be too much of a problem.  Nighttime turns into a power contest with radio China.

I had a funny QSO one time... I called CQ and this guy from CA came back to me on LSB, not knowing that I was on AM.  In fact, I don't think he knew what AM was.  Anyway, he'd give me quick one or two sentence replies and I'd old buzzard back on mine.  After about three exchanges, he had to go.

Jon

[w3jn]

his is true, although not because of the conventional (non-sync) detector, it's because the receivers are improperly designed (including in the detector!!).  If the envelope detector is properly designed, it can (in theory) demodulate unlimited positive peaks.  Of course this is not technically practical, but you get the idea.

The design goal for conventional diode detectors most certainly isn't 150% upward modulation, so I'm not sure it's accurate to say they are improperly designed - rather, they are a design compromise.  Where most diode detectors fall down is excessive capacitance downstream; ie to filter the carrier to provide DC, generally for AVC.  If a receiver has a separate AGC detector, an AM diode detector can often be improved by reducing the filter capacitance.

Hallicrafters SX-101s in particular can be vastly improved by removing the .01 uF capacitor (as I recall) in the ANL circuit.  Of course this renders the ANL inoperative but the peak clipping is improved considerably.

Bottom line, there's no really justifiable reason to run such large positive modulation, since you don't know what the other guys will be running for receivers.

[ka1tdq]

I backed off the audio last night and I don't think anyone really noticed.  I'm running 1000 watts peak now.  It gives me more room now anyway for hard "P's" or a laugh.

Jon

[VE3LYX]

Speaking of wide someone was on this week I think Wed on 40m and even on my digital rx I could copy them about anywhere between 7290 and 7282. And strong too. That is a bit much I think. I thought my RX was broke at first.
I didn't try going above 7290 to see how high it was but wish I had now. As someone running 7 watts and getting out I don't get it. However to each his own.
don

[N2DTS]

Some of the class E guys are 25 KHz wide, most run at least 20 KHz.
Some run processing that makes the signal dense out to 10 KHz each side.
When I run my screen mod setups, I likely am wide because the rig will pass high frequencies just fine and I have nothing to limit the high frequency stuff enough.
I have an EQ but that only cuts the audio 6 or 12 db (range switch).
Old rigs had lots of crappy transformers, lots of screen and plate bypass capacitance to reduce highs, even after mods to the audio.

Class E, broadcast transmitters, some flex radios, and screen modulation can have no limits below 20 KHz.

So you can have someone 20 KHz wide, running very high power, taking out 3860 to 3885.
There is no reason for that, no one is listening with a 20KHz passband, but some people would like to be even wider. After all, its all about them.

Its VERY good you can hit such good positive peaks, because at 130% positive you will be very clean.
Clean sounds good, and tends to not be real wide.

I guess you are doing the 3 diode limiter in the low power part, I did it at the high power part, with brick diodes rated at 14kv and 1 amp, maybe that was part of the problem, lots of slow diodes in series...

[WQ9E]

The design goal for conventional diode detectors most certainly isn't 150% upward modulation, so I'm not sure it's accurate to say they are improperly designed - rather, they are a design compromise. 

Bottom line, there's no really justifiable reason to run such large positive modulation, since you don't know what the other guys will be running for receivers.

John is spot on.  Receivers were designed based upon accepted standards and there was no reason for the detector to be set up for a modulation level beyond what they were expected to receive. 

Many listeners will experience considerable distortion at those high levels and although it is fine for groups interested in experimentation,  operating with that level of modulation is rather rude when checking into general AM nets and large round tables.

[WD5JKO]

   Seems to me that a high average percentage of modulation is desirable, whether it be from raising the density (processing), or high peaks (to no more than 140'ish %) increasing the average. Either method will enhance copy-ability at the receiver end, especially if the frequency response is Taylored to make the best out of 100 - 4K audio response.  The effect to a round table of hams where most are averaging 30% modulation is huge....but is that rude?

   Sure, some receiver detectors will tilt as they distort a high peak. The hint of fuzziness at that peak and a bit thereafter is usually not bad unless the transmitting station is broadcasting a highly distorted modulation to begin with.

   What really sounds terrible to me are those that run low level AM from a balanced modulator, or SDR exciter that is driven to the point of producing "wavelets" (> 100% downward modulation).

   I routinely run peaks at ~ 140% with my Gonset G-76 station. I always get good reports, and folks copy me under adverse conditions. Is it high fidelity, NO, but it works, is received fine, and with good reports.

   I contend those that run high peaks that end up sounding terrible are usually broadcasting more distortion than what the receiver will produce from an over capacitively loaded diode detector. The exception would be the "wavelet" guys that are not transmitting a distorted signal, but instead are transmitting a signal not compatible to be received with a mere diode detector.

   Jon, I'd like to hook up with you sometime, 7290 +/- ?

Jim
Wd5JKO

   

[WD8BIL]

Be careful about calling a station wide unless you provide full data. I've measured stations some call wide as having 5kHz products at -40 dBc. But they were 30 over S9!! That puts the 5kHz products at about S7. Yes, you are gonna hear them even in a 3.5kHz passband given the skirts on most BA receiver I.F.s.

[kb3ouk]

These stations that are supposedly 20 khz wide, how are you determining this? Using a spectrum analyzer or waterfall display on an SDR, or just guessing by tuning up and down with your receiver? The perceived bandwidth of a signal on a receiver is the sum of the transmitter's bandwidth and the receiver's bandwidth, so if you have a wide receiver, that's gonna make the signal seem wide.

[steve_qix]

Be careful about calling a station wide unless you provide full data. I've measured stations some call wide as having 5kHz products at -40 dBc. But they were 30 over S9!! That puts the 5kHz products at about S7. Yes, you are gonna hear them even in a 3.5kHz passband given the skirts on most BA receiver I.F.s.

Thank you!  The voice of reason  

I'm not sure why there is a discussion of class E rigs here with respect to bandwidth.  I suppose I am more than familiar with the design of most of the class E rigs on the air.  They *ALL* have a whole lot of filtering before and as part of the modulator.  How many DB down at the edges?  LOTS!  This is born out all the time when observing spectrum displays (which are now quite common) and you can see how many DB down the signals are at the filter's edges.  

6 poles at the modulator input with the 3dB point a little above 5kHz is, in my estimation, a pretty reasonable filter.  Then you put a 6 pole filter AFTER the modulator with its corner at around 9kHz, and you have a lot of filtering.  The 2nd filter is always higher than the first to avoid unwanted effects in the pulse width modulator itself.

It is really obvious when a transmitters has filtering and when it does not.  The ones with filtering (such as most class E rigs) drop off pretty quickly at the filter edge.  The others?  Well, not so much.

The bottom line?  Absolutely none of this matters that much in the long run.  Frankly, it is VERY rare that a station's excessive bandwidth (translated: high frequency response or unmitigated excessive distortion) causes a problem.  

What is more likely to cause problems is a large disparity in the signal strength of the station you want to hear, and the station 10 or 12kHz away that you don't want to hear, but is 40 dB stronger than the one you're desperately trying to copy.  Been in that position many times.  The most useful answer?  Use a better receiver.  That's what works for me.

I've been on ham radio for more than 40 years.  I can only think of a handful of times in 40 years that someone being "wide" actually caused a problem with copyability.  Real strong station nearby?  Yup, that can be a problem, but not due to excessive bandwidth.  Sidebands? Maybe.  If it were just a carrier maybe I could copy the other guy.  Excessive sideband energy?  Very Rarely.  Annoying? Only if you let it.

<End Soliloquy>  

Regards,  Steve

[W3RSW]

Only once every 40 years?
This guy might have caught your interest.

Almost had to be a fellow AM'er to know what he was doing.

[N2DTS]

Class E:

http://n2dts.smugmug.com/Ham-radio/i-ZxTSnCR/X2
3865 to 3885.

Normal AM guys:
http://n2dts.smugmug.com/Ham-radio/i-JZp3kbg/X2

[steve_qix]

I can produce lots of examples too.  There is no point to it.

This finger pointing is not productive.  If someone doesn't like class E that's ok - everyone's entitled to their likes and dislikes.

The widest AM station I've ever heard, and the widest ones I hear now are not running class E rigs, but WHO CARES???  This is not relevant to anything practical or useful.

All of this talk about AM bandwidth is HIGHLY counterproductive and we should stop it, else we (as a mode) will get what the bandwidth police appear, by their constant talk about it, to be advocating.  An FCC regulation about bandwidth which we absolutely do not need and certainly do not want.

[N2DTS]

So you will support me running my new 5000 watt rig that is 100 KHz wide?
Surely that will not attract any unwanted attention by anyone, at least as long as we do not TALK about it on line.

And I never pointed anyone out (maybe you can tell by his signal?) and did not single out class E, I said like some other methods, it CAN be wide.
I think your wonderful dx60 modulator goes as wide as a class E rig...

[w1vtp]

Boy! this thread turned unpleasant in a hurry.  Please let's not start up a class "E" vs a class "Y" ( as in "classy" or class C) dialog.  A properly adjusted PDM AM transmitter that operates in class E has very low distortion hence very low distortion products that show up in the sidebands.

Brett, I looked at your first example and I cannot derive the carrier level so I cannot tell what the sideband levels are with respect to the carrier although I admit this might be a poster child for a poorly adjusted AM transmitter.

And please let's not get into a bandwidth confrontation.  In my view any type of AM modulation needs to be properly adjusted with proper gain distribution across all processing stages.  Properly designed and adjusted AM transmitters will not occupy an inordinate bandwidth but will be determined by the band pass of the audio that excites the modulator.  Improperly designed or inadequately monitored AM transmitters have the potential of occupying greater bandwidth than is necessary due to the distortion products of the modulating signal and the inability of the station operator to detect these distortion products.

We have the opportunity for these different modes of AM design and operation to make us good experimenters using good engineering practice..

Al

[kb3ouk]

The thing about that first picture is if that signal was only S9 and not S9+30 and higher, then that excessive sideband stuff might not actually be noticeable. Same with the second picture, add 10 dB to that signal and you might end up with something close to the first picture.

[steve_qix]

I think your wonderful dx60 modulator goes as wide as a class E rig...

Much wider actually.  The class E rigs and the SDR radios and most rice boxes have ACTIVE and purposeful bandwidth limits in their designs.

My dx60 modification has no such limitations, and therefore can be and probably will be much wider than any class E rig (well, at least any that I designed).

This thread has been derailed by a red herring, and I would like to bring it back to the discussion as it began:  Highly asymmetrical modulation and what that means at the receive end.

Thank you