The AM Forum
March 29, 2024, 02:18:29 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] 2   Go Down
  Print  
Author Topic: Question on negative peak limiting  (Read 17425 times)
0 Members and 1 Guest are viewing this topic.
AB2EZ
Member

Offline Offline

Posts: 1722


"Season's Greetings" looks okay to me...


« on: November 15, 2008, 11:45:50 AM »

Hi

I am thinking about putting a negative peak limiter on my plate modulated GS-35b amplifier.

Putting aside the various views on whether negative peak limiters are a good idea or not...

... I have attached a diagram of a 2-diode negative peak limiter that appears to perform all of the functions performed by Steve's (WA1QIX's) 3-diode design.  However, perhaps there is something subtle that occurs at the transition to negative peak limiting (and back).... maybe having to do with the parasitic capacitances of diodes ... that makes this 2-diode approach a bad idea (versus the 3-diode approach).

Note that when negative peak limiting is not occurring, the resistor (R) across the forward biased diode is irrelevant. During negative peak limiting, both the resistor R and the output tubes are connected to the keep-alive supply through the other forward biased diode. So the subtle difference must occur at the transition.

Any insights or information based on past experience would be appreciated.

Stu


* 2-diode negative peak limiter.jpg (26.97 KB, 960x720 - viewed 595 times.)
Logged

Stewart ("Stu") Personick. Pictured: (from The New Yorker) "Season's Greetings" looks OK to me. Let's run it by the legal department
WA1HZK
Member

Offline Offline

Posts: 1104


WWW
« Reply #1 on: November 15, 2008, 06:36:40 PM »

Took me a while to get the smoke out of the shack when I tried that on the 833 rig. I think it's easier on smaller rigs. If you want to do that do it in the driver. Just my two cents. Hey, I had to try it too.
Keith
Logged

AM is Not A Hobby - It's a "Way of Life"!
Timmy, Sometime in 2007 on a Mountain Far Away..
www.criticalradio.com
www.criticalbattery.com
www.criticaltowers.com
www.criticalresponder.com
Official Registered "Old Buzzard"
kc2ifr
Guest
« Reply #2 on: November 17, 2008, 05:10:22 PM »

Stu,
I ALWAYS had better luck when the limiting was done in earlier audio stages. Of course it helps if the limiter is made for AM transmitters.......not a symmetrical limiter.

Bill
Logged
steve_qix
Contributing
Member
*
Offline Offline

Posts: 2599


Bap!


WWW
« Reply #3 on: November 17, 2008, 07:40:21 PM »

Hmmmmmmm... Well, from an electrical standpoint, I cannot see any difference.

You will need one less diode, but you will have to insulate the resistor at the full high voltage.  Probably a wash.
But, looking things over very carefully, I think it is electrically equivalent.

On negative peak limiting in general:

I have performed high level negative peak limiting (clipping) on all sorts of BIG transmitters (5kW BC rigs for instance)  and my own 1kW tube rig ("back in the days... etc.") and never had a problem with any of them.

That being said, these were well designed transmitters with sufficient modulators (and modulation transformers) to handle the significant extra power and voltage generated with high levels of positive modulation.  150% positive is 6.25 times the DC power input of the carrier.  200% positive is *9 times more* power!

I agree with Bill on processing at low level if If IF (and only if Cheesy  ) the transmitter is capable of passing a square wave (or close to one), and properly.  Most, if not all transformer coupled modulators will not do this across too much of the audio spectrum.  Contrast that with any of the PWM broadcast or amateur transmitters out there.  These transmitters will do that all day.

The same can be said  for a properly designed DC coupled grid modulated, low level modulated, series modulated (class A, G, H, etc.), phase to amplitude (Ampliphase), SDR or other similar transmitter which does not use transformer coupling in the high level audio path.

In these cases, low level processing (neg peak clipping, etc.) works just beautifully, and is in fact superior because the transmitter's audio filtering and low pass filters are involved AFTER the processing.
Logged

High Power, Broadcast Audio and Low Cost?  Check out the class E web site at: http://www.classeradio.org
w8khk
Member

Offline Offline

Posts: 1201


This ham got his ticket the old fashioned way.


WWW
« Reply #4 on: November 17, 2008, 09:14:35 PM »

Hmmmmmmm... Well, from an electrical standpoint, I cannot see any difference.

You will need one less diode, but you will have to insulate the resistor at the full high voltage.  Probably a wash.
But, looking things over very carefully, I think it is electrically equivalent.


Hi Steve,

Looking at Stu's schematic, it appears he was able to eliminate the need for D1 (reference id in your 3-diode schematic) by connecting R1 to the load side of the keepalive diode D3, instead of the source side. 

I would like to try this approach in my modified Valiant, instead of the 3-diode version, because space is at a premium.  My Valiant modulator has four 1625s in push pull parallel, driving a UTC S21 115W transformer.  No driver transformer, all tubes with cathode follower driving the 1625 grids.  I do not expect to need the capacitor input supply (with increased B+ for the modulator), because the parallel tube arrangement runs lower impedance and should handle peaks with standard choke input supply.  I can add cap input if necessary, after initial testing.  Initially I will configure 6K plate to plate impedance, with 1.5K or 2K output impedance. 

I was able to fit the original Valiant modulation transformer on the chassis, with primary and all secondary windings phased in series to use as a heising modulation reactor.  A 1uf 4kv oil cap from a microwave oven will be the coupling cap, from cold end of mod xfmr secondary to gnd.

I assume the voltage requirements are still the same, D2 series diode at least = to B+ voltage or higher, and D3 keepalive 4 times B+.  I will use 2 1KV 10A diodes in series for D2 (series diode), and 6 1KV 10A diodes in series for keepalive.  Resistor will be 2Kohms at 20 watts non-inductive.  I have a variable keepalive supply adjustable up to 60VDC.

Advice/Criticism is welcome, thanks!

73,
Rick


Logged

Rick / W8KHK  ex WB2HKX, WB4GNR
"Both politicians and diapers need to be changed often and for the same reason.”   Ronald Reagan

My smart?phone voicetext screws up homophones, but they are crystal clear from my 75 meter plate-modulated AM transmitter
AB2EZ
Member

Offline Offline

Posts: 1722


"Season's Greetings" looks okay to me...


« Reply #5 on: November 18, 2008, 06:14:50 AM »

All

Thanks for your inputs... which are very much appreciated.

During spare moments... when I don't want to think about other stuff... I've been thinking about some of the subtleties of a negative peak limiter. The fact that it is a non-linear circuit makes for a lot of subtle effects. Here are a few that might be important:

1. Diode switching transients

The diodes will generate some switching transients because the keep-alive diode won't turn on at exactly the same time, and with exactly the same speed as the other diode turns off (and vice-versa). Depending upon the details of the modulator (how the modulator responds to this change in load) and the capacitance across the output of the modulation transformer (see below)... this effect could add high frequency noise that is correlated with the modulation... and therefore perceived by others as "splatter". For some transmitters, the effect would be negligible. For other transmitters, the effect could be large. I think that this effect, if large enough, could also cause fireworks in some transmitters if one were to push the negative peak limiting too far.

2. Capacitive load on the modulation transformer

The capacitive load on the modulation transformer secondary will be lower during the keep alive period... because the plate choke r.f. bypass capacitor and the capacitance associated with the tank circuit (usually dominated by the plate DC blocking capacitor) at audio frequencies will not be across the modulation transformer during the keep alive period. Discontinuities in the load on the modulator (caused by discontinuities in the load on the modulation transformer secondary) can cause transients in the modulator that persist beyond the keep-alive period... and which are perceived as splatter.


For example, if the rf bypass capacitor on the modulated B+ side of the plate choke is .001 uF, and if the DC coupling capacitor leading to the output tank circuit is .0022 uF, (and assuming there is an RF safety choke at the output), then the total capacitance loading the modulation transformer secondary at audio frequencies is .0032 uF. At 5kHz, the impedance of this capacitance is ~ 10k ohms. Some transmitters have more capacitance bypassing the plate choke and/or have a larger DC blocking capacitor.

3. Is it necessary to keep a constant load on the modulator?

For some modulators, a discontinuity of the load on the modulator... during the keep alive period... will result in transients that persist beyond the keep alive period. In those cases, it is important to maintain a reasonably constant load on the modulator by using a resistor in the keep alive circuit that is matched to the load presented by the rf output stage during normal (non-keep-alive) periods.

However, some modulators... such as a modern audio power amplifier driving a backward-connected audio output transformer... may not produce much of a transient response if their load is removed during the keep alive period. They are more like ideal voltage sources. Therefore, the benefits of maintaining a constant load during the keep alive period may be much smaller in those cases.

Best regards
Stu
Logged

Stewart ("Stu") Personick. Pictured: (from The New Yorker) "Season's Greetings" looks OK to me. Let's run it by the legal department
w4bfs
W4 Beans For Supper
Contributing
Member
*
Offline Offline

Posts: 1433


more inpoot often yields more outpoot


« Reply #6 on: November 18, 2008, 08:18:20 AM »

hi Stu and others ... Astabula Bill reported to me that in his pair 4-250 mod by pair 833 transmitter that he had tried the all solid state 3 diode limiter and eventually had to replace one of the diodes with a thermionic rectifier (sorry can't remember tube type or placement) to get proper operation ... all who have talked with Bill running this circuit can attest to its effectiveness on the air ... seems that this circuit is placing similar turn on/off time constraints as a low freq switcher ... maybe FREDs or something similar are called for ...just my thinking ...73 ...John
Logged

Beefus

O would some power the gift give us
to see ourselves as others see us.
It would from many blunders free us.         Robert Burns
k4kyv
Contributing Member
Don
Member

Offline Offline

Posts: 10057



« Reply #7 on: November 18, 2008, 12:03:00 PM »

Try using type 836 rectifiers.  Those are high-vacuum equivalents to type 866's (not 866A's; they don't have the higher piv rating of the mv types) and have been abundant as WW2 surplus.  They should still be easy to find.  Some voltage drop will occur with high plate current, but this shouldn't be a problem when using those tubes with 100-watt class transmitters.
Logged

Don, K4KYV                                       AMI#5
Licensed since 1959 and not happy to be back on AM...    Never got off AM in the first place.

- - -
This message was typed using the DVORAK keyboard layout.
http://www.mwbrooks.com/dvorak
steve_qix
Contributing
Member
*
Offline Offline

Posts: 2599


Bap!


WWW
« Reply #8 on: November 18, 2008, 02:43:49 PM »

Hmm.. As far as diodes switching, I believe the problem arises not necessarily in the turn ON time, but the turn OFF time...   Here again, we are dealing with relatively low frequencies (under 30kHz), and I presume the diodes can at least go this fast.

Now, if they can't, there are fast recovery, high voltage diodes available (and I have used some of them in negative peak limiters), and these definitely work.

Anyway, at audio I have to say I haven't seen any switching transients or other anomalies.

Regards,

Steve
Logged

High Power, Broadcast Audio and Low Cost?  Check out the class E web site at: http://www.classeradio.org
kc2ifr
Guest
« Reply #9 on: November 18, 2008, 05:58:31 PM »

Quote
I agree with Bill on processing at low level if If IF (and only if Cheesy  ) the transmitter is capable of passing a square wave (or close to one), and properly.

Steve,
Could u explain why this is? Im not talking about hard clipping.
Bill
Logged
steve_qix
Contributing
Member
*
Offline Offline

Posts: 2599


Bap!


WWW
« Reply #10 on: November 19, 2008, 06:32:21 AM »

Quote
I agree with Bill on processing at low level if If IF (and only if ) the transmitter is capable of passing a square wave (or close to one), and properly.

Steve,
Could u explain why this is? Im not talking about hard clipping.
Bill

Hi Bill,

Hard clipping is the discussion at hand and that's why I mentioned the square waves, etc.  Even soft clipping could be an issue - depending.  For normal processing, in theory, there shouldn't be any problems  Cool

Regards,

Steve
Logged

High Power, Broadcast Audio and Low Cost?  Check out the class E web site at: http://www.classeradio.org
w1vtp
Member

Offline Offline

Posts: 2638



« Reply #11 on: November 19, 2008, 08:39:59 AM »

Bill, Steve

It might be educational to show a spectrum display of a pure sine wave versus one that has been clipped.  The problem as I see is it that once there is clipping in an AF mudulation signal there are a LOT more unwanted high frequency components to deal with IE:  splatter.  If we try to filter the high freq comp... that solves the broad band stuff but now we have a lot of junk cluttering the spectrum we WANT to hear.

First cup of "bean" hasn't kicked in yet but that's the way I see it right now.  Interesting thread.  I love this forum

Al VTP (Very Tired Person this morning) Smiley
Logged
K3ZS
Member

Offline Offline

Posts: 1037



« Reply #12 on: November 19, 2008, 09:09:45 AM »

I use diode negative peak clipping only to protect the mod transformer.   I try to keep the mod level so that it only
clips rarely.    Any sharp clipping creates harmonics and splatter.   There was an article a few years ago that
"softened" the diode clipper with an added resistor in series with the zener string (this was the clipper using zener
diodes).
Logged
kc2ifr
Guest
« Reply #13 on: November 19, 2008, 04:52:43 PM »

OK Steve,
I see.........never used clipping here......just heavy on the compression using this little box.
http://www.inovon.com/?a=2&s=1&i=1

So far its working GREAT. I use it with my 500.
I use the Urei modulimiter with the 4x1 and I love it also.
Scroll down on this page to see the BL40 Modulimiter....
http://www.uaudio.com/webzine/2005/august/index4.html
Both units are designed to be used with broadcast AM transmitters.

Bill
Logged
Steve - WB3HUZ
Guest
« Reply #14 on: November 19, 2008, 05:09:50 PM »

Consider that the clipping is occuring at negative 95+ percent moduation. What is the power of any harmonics/distortion products at that point?


Bill, Steve

It might be educational to show a spectrum display of a pure sine wave versus one that has been clipped.  The problem as I see is it that once there is clipping in an AF mudulation signal there are a LOT more unwanted high frequency components to deal with IE:  splatter.  If we try to filter the high freq comp... that solves the broad band stuff but now we have a lot of junk cluttering the spectrum we WANT to hear.

First cup of "bean" hasn't kicked in yet but that's the way I see it right now.  Interesting thread.  I love this forum

Al VTP (Very Tired Person this morning) Smiley
Logged
steve_qix
Contributing
Member
*
Offline Offline

Posts: 2599


Bap!


WWW
« Reply #15 on: November 19, 2008, 07:42:52 PM »

Consider that the clipping is occuring at negative 95+ percent moduation. What is the power of any harmonics/distortion products at that point?

Ahhhh... the question of the hour  Wink   That depends on how MUCH clipping is happening.  I suppose if the unclipped waveform is symmetrical, and the clipped waveform is 200% positive (and clipped off at 95% negative), that would produce a LOT of harmonics.  I've run a few experiments in this area, and the [unfiltered] bandwidth will expand.  But, that is an EXCESSIVE amount of clipping.

If you just clip off the occasional peak, I have not found any significant bandwidth issues, and in fact, it appears to be quite tollerable as compared to blatent overmodulation (in the negative direction), which can produce splatter like nobody's business (remember Irb?).

Regards,

Steve
Logged

High Power, Broadcast Audio and Low Cost?  Check out the class E web site at: http://www.classeradio.org
W8IXY
Contributing
Member
*
Offline Offline

Posts: 124


« Reply #16 on: November 20, 2008, 01:26:44 PM »

Hello AMers!

I have been involved with AM broadcast and amateur radio AM for almost 50 years now....

In AM broadcast, ALL of the audio processing is done prior to applying the audio to the transmitter.  The transmitter is (hopefully) capable of reproducing a modulated wave form identical to the one supplied to its input.

Any manipulation of negative peaks, equalization, increasing average density, and so on is done and filtered before the audio even gets to the transmitter input stage.  Any non-linearity in the transmitter (including negative peak limiting) is a form of DISTORTION and actually works against you.

The ideal would be to build a transmitter that can handle any waveform you feed it, such as an asymmetrical waveform that is processed to have positive peaks that exceed 200% positive modulation, and almost 100% negative modulation.  Phase shift would need to be very low, and frequency response would need to be ruler flat.  Then all "manipulation" of the audio would be done at low level, and the transmitter would faithfully follow that audio wave form.

I designed all the AM processing parameters for the OMNIA One audio processor here at Telos systems, and have some "very highly processed" audio modulating the 50,000 watts here on WKNR/AM850 in Cleveland.  If you are within range, tune it in.  Your transmitter needs to be as "transparent" as possible, and the processed audio will fly through it and result in UNREAL good sounding audio on the air.

73
Ted  W8IXY
Logged
steve_qix
Contributing
Member
*
Offline Offline

Posts: 2599


Bap!


WWW
« Reply #17 on: November 20, 2008, 02:36:49 PM »

Hello AMers!

  Your transmitter needs to be as "transparent" as possible, and the processed audio will fly through it and result in UNREAL good sounding audio on the air.


Hi Ted, good to hear from you!!!

We have a few transmitters on the air with transparent audio.  These are generally the low level modulated transmitters, or the solid state class E, high level modulated transmitters. 

However, we do have a LOT of transmitters that won't do it - so the compromise is to do a bit of negative peak clipping at high level  Wink   It works rather well if not carried to an extreme.  I've done it with broadcast rigs (like the Collins 21E, RCA BTA1-R, etc.) that, while pretty good, would not handle a clipped, asymmetrical waveform very well.

Good stop-gap measure 'till we got an MW5 (this was back in the early '70s).

Regards,

Steve WA1QIX


Logged

High Power, Broadcast Audio and Low Cost?  Check out the class E web site at: http://www.classeradio.org
w1vtp
Member

Offline Offline

Posts: 2638



« Reply #18 on: November 21, 2008, 01:24:52 AM »

<copied from Steve QIX>  Ahhhh... the question of the hour 

My comments.  The question is if the transmitter is at ~95% negative modulation peak (null) and why wouldn't splatter be a non-issue as there is little power at that point coming out of the transmitter.  Harmonic distortion is distortion no matter if it occurs on negative or positive.  Some folks have touched on the key in this thread -- the diode with keep alive is a great way of reducing the HARD square wave corners in negative clipping.  But I think that excessive positive modulation can produce similar results IF the audio peaks are accompanied with some clipping at the positive peaks too.

It always comes down to good design and proper operation of our AM transmitters.  This thread is an excellent one for me.  I would still like to see some spectrum analysis under controlled testing conditions and compare a good negative modulation protection with an honest to goodness negative over modulation.  Could even throw in some positive clipping spectrum displays.

This is fun -- I think we as AMers are much more conscious of proper operation on both the negative peaks and possible flat topping.  We really do not want yellowy distorted sounding modulation and that usually results in good BW performance.

Cheers, Al VTP (who couldn't persuade my bod to shut down for the evening).

Logged
steve_qix
Contributing
Member
*
Offline Offline

Posts: 2599


Bap!


WWW
« Reply #19 on: November 21, 2008, 04:20:23 AM »

Of course Al has a great point.  If there is clipping of the positive peaks too, then obviously there will be subsequent harmonics (aka splatter) produced.

I think for the purposes of this particular discussion, we are "assuming" the transmitter is not clipping the positive peaks at all, at the actual level of modulation in use at the time.

I wouldn't apply negative peak limiting (at least not much!) to a transmitter with limited positive peak capabilities.  Heck, my voice alone WITHOUT negative peak limiting has a natural 30% to 40% asymmetry - and I have actually observed more than this at times.

I can speak for myself - the negative peak limiting I use is pretty minimal.  It is there only to deal with the occasion high peaks that would otherwise cause overmodulation in the negative direction.  If carried too far, the point of a good, high fidelity transmitter is diminished because there WILL be distortion.

On the other hand, it sure is nice to be able to really pour on the audio when conditions are rough.  At this point, pure fidelity gives way to copyability  Cheesy  But, never to the point of positive peak clipping, too!

Regards,

Steve
Logged

High Power, Broadcast Audio and Low Cost?  Check out the class E web site at: http://www.classeradio.org
w4bfs
W4 Beans For Supper
Contributing
Member
*
Offline Offline

Posts: 1433


more inpoot often yields more outpoot


« Reply #20 on: November 21, 2008, 06:54:29 AM »

this is a good discussion ... rather than hijack this one believe I'll start a new one about transmitt audio...73 ....John

10:38 am  ...just got off the air in a qso with Bill, W8VYZ in snowy Astabula ... we discussed his 3 diode circuit ... he verified that he was using an 836 in series with a load resistor of 4 - 200 Watt units that Bill said got 'very warm' during use ... he didn't specify their value ... Bill mentioned a 80% negative mod limit ... also said he was using 10 - 1kV piv diodes in a string for each of the other diode functions ... Bill kept referencing the CQ and QST magazine articles in the 1950's and reported that the QST article was "absoutely correct' .... this was an interesting qso and Bill is wanting to assert am operation on 7290 kHz ...73 ... John
Logged

Beefus

O would some power the gift give us
to see ourselves as others see us.
It would from many blunders free us.         Robert Burns
steve_qix
Contributing
Member
*
Offline Offline

Posts: 2599


Bap!


WWW
« Reply #21 on: November 22, 2008, 05:23:49 PM »

It's really ok to overmodulate in the positive direction and assuming the waveform is not modified in any way (distorted/clipped/etc.), doing so will produce no extra harmonics or splatter than if the transmitter were modulated
less than 100% (positive) - again, assuming no distortion positive or negative and all other things being equal.

Most naturally produced waveforms (musical instruments, human voices, etc.) are naturally asymmetrical.  My voice waveform certainly is - by quite a large amount.

As long as nothing distorts the waveform  AND if the waveform is fed to the transmitter in such a way as to allow higher voltage peaks to modulate the transmitter in the positive direction (correct phase), all will be OK, and since there is no distortion, no splatter will be produced.  Only overmodulation in the negative direction produces splatter (assuming no other distortion or limitations in the transmitter).

Regards,

Steve


Logged

High Power, Broadcast Audio and Low Cost?  Check out the class E web site at: http://www.classeradio.org
Steve - WB3HUZ
Guest
« Reply #22 on: November 22, 2008, 06:02:55 PM »

Quote
Most naturally produced waveforms (musical instruments, human voices, etc.) are naturally asymmetrical.  My voice waveform certainly is - by quite a large amount.


Indeed. The first attachment is a scope shot of my voice while I was saying the universal AM test phrase (uuuureeeiiiii). You can see it is asymmetrical at almost a 2:1 ratio.





* pos_audio.jpg (92.89 KB, 800x600 - viewed 489 times.)
Logged
Steve - WB3HUZ
Guest
« Reply #23 on: November 22, 2008, 10:00:36 PM »

If that happens, just flip the phase/polarity. Simple really.

See below.




* neg_audio.jpg (97.18 KB, 800x600 - viewed 518 times.)

* negative_mod.jpg (105.97 KB, 800x600 - viewed 464 times.)

* positive_mod.jpg (106.66 KB, 800x600 - viewed 475 times.)
Logged
WB6VHE
Guest
« Reply #24 on: November 24, 2008, 01:46:48 PM »

Since nobody appears to have mentioned it, I refer those interested to
an article in Electric Radio, No. 154, March 2002, by Tom Bonomo, K6AD.
Title of article: Modulation Transformer Protection for AM Transmitters.
I have been using this circuit for over a year and it works great!  Prior to this
I have blown two (expensive!) mod transformers when the RF amp load went
away (tank cap flashover).  While the idea of the article is to protect your mod iron,
the  circuit also functions as a neg peak limiter.  I can shout into the mike (not that I want to)
now, and it is not possible to overmodulate on the negative peaks.
The audio both looks (on a scope and looking at the sig on a spec analyzer)
and sounds great.  The resistors hardly get warm.  I use this with my
homebrew 813 rig, 811A's for modulators, roughly 400 Watts input to the rf amp.
Logged
Pages: [1] 2   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 0.082 seconds with 18 queries.