NEGATIVE PEAK LIMITING/APACHE..?

[ve6pg]

..HI AGN FROM TIM...GOT MY APACHE RUNNING,LOTS OF RF,AUDIO SOUNDS FB...BUT WATCHING THE 'SCOPE,I SEE ALOT OF NEGATIVE PEAKS...I'VE REMOVE THE 6AL5 CLIPPER TUBE,BUT STILL SEE ALOT OF NEGATIVE PEAKS..OK..I'VE READ THE ARTICLES ON THE AM WINDOW,ETC ABOUT USING DIODES TO REDUCE THE NEGATIVE PEAKS,BUT ISNT THIS THE SAME AS USING THE 6AL5 CLIPPER CIRCUIT?..
...SK..

[WBear2GCR]

No, the clipper is a different animal.

Try reversing the polarity of your microphone?
If that is a problem for some reason you can reverse either the primary or secondary driver xfmr's leads.

Make sure ur properly loaded up too... you could be seeing "downward" modulation??

As far as "negative peak limiting" the purpose is to prevent the xmitter's finals from being driven into the OFF zone - ie. no output. The negative swing of the modulation signal exceeding the B+ voltage...

The modulator itself won't have this problem since the mod tubes are push-pull. The input and predriver tubes probably can't be driven into an "off" state as the transmitter is currently set up...

The phase may have flipped when the "clipper tubes" were bypassed, since each pass through a grid, out to a plate flips the phase 180 degrees...

I hope I got this right...

 

Btw, online the use of CAPS LIKE THIS IS CONSIDERED the same as "shouting..." fwiw. 

      _-_-WBear2GCR

[W2PFY]

WHAT??  I CAN'T HEAR YOU

[K1JJ]

Something to be aware of.....

If you flip the leads [phase] of a high level modulator chain that uses a negative feedback loop, the chain will probably take off and oscillate. I doubt the STOCK Apache has audio negative feedback around the modulator tubes back to the earlier driver stages, so no worry -  but some may have, if modified.

As mentioned in a previous post, a good way to switch phase would be in a stage BEFORE the negative feedback loop, like in the mic, first preamp, etc.  Over here, I simply put a DPDT toggle switch into a shielded, grounded  box that flips my RE-20 leads around. This way I can easily low level change audio phase for any rig using the same mic  - whether if be a ricebox, plate modulated, PDM, or whatever.

73,

T

[flintstone mop]

Hi Tim
Are you using any audio processing into the Apache?? I forget what mic processor guys were using on this board, the VX something or other. One of those 1 rack unit high marvels for compressing and limiting your mic audio. Excessive "base-lining" that bright line you see closing the carrier,as youare seeing on your' scope, is not good for the mod transformer and for your neighbors on the band. The modulator is actually seeing an open load when the PA tubes are over modulated and it causes splatter. Do not use the clipper in the Apache.    That was one of the circuits that make it Scratchy Apache. You are discovering the slightly painful way to make a nice sounding loud AM signal. You can't throw hi-fi audio into an AM transmitter and it is loud and full. This is the real secret of AM sound.
I did not care for negative cycle loading in my Apache. It worked, but the audio reports were not too good and I don't think the modulator liked the added components. The magic seemd to happen outside of the box, in my case.
Fred

[Bacon, WA3WDR]

Hi Tim,

Unfortunately, a good modulator will tend to overmodulate a transmitter very easily.  Outboard processing, preferably good compression, is the way to control overmodulation without distortion.  This works best when the transmitter frequency response is as wide as possible.  Proper equalization is necessary too.  Muffled audio will not be clear at the far end, and it will tend to sound more distorted as well.

Like most amateur AM transmitters that use tetrodes, the Apache overmodulates the screen grid of the final, which tends to cause sudden negative peaks going down to cutoff.  You can get slightly better results by arranging a resisive divider for the screen grids, instead of just a single resistor from the screens to modulated B+.  Check these articles:

Screen Modulation Optimization for the Heathkit DX-100
Dino, WA1KNX
(Note: this is not about screen modulation by itself, it is about screen modulation along with plate modulation.)
http://www.amwindow.org/tech/htm/scrnmod.htm

From the original article:
Class C Optimization for Ultra Low Distortion
Dino, WA1KNX
http://www.amfone.net/AMPX/71.htm
(Note: there was a wiring error in the original schematic from AM P/X, that is already corrected in the online schematic.)

And here is a simpler trick that works almost as well:

The Self-Modulated Screen Grid
Bacon, WA3WDR (Hey, that's me!)
http://www.amfone.net/AMPX/98.htm


If you are into construction, you can build a compressor:

ASYMMETRICAL PEAK LIMITING
Bacon, WA3WDR
http://www.amfone.net/AMPX/issue_number_79.htm

[WBear2GCR]

The HB-3 RCA Transmitting Tube Handbook has a note on the 6146 page:

"Obtained preferably from a separate source modulated with the plate supply, or from the modulated plate supply through a series resistor." Referring to the Grid No.2, of course.

So this is the source of this circuit connection found so often.

The gain & output of the tube, afaik, is increased as the screen supply voltage increases, and decreased as it is lowered. So, it seems to me that this is a form of "bootstrapping" where when the tube is modulated; on postitive going modulation, the tube is turned on harder than it would be with a fixed supply.

In effect this is a sort of "Ultra Linear" operation, which is found in audio output circuits, in which the screen is a tap at about 70% of the primary winding, a series current limiting resistor is applied and the screen gets to follow the waveform up and down.

Regardless, the voltage to the screen needs to be current limited and is spec'd by RCA at being set far lower than the B+,  at ~160vdc for 750 on the palte - typical Class C operation,  I don't think this is made clear here: http://www.amfone.net/AMPX/98.htm

Also, I'm not sure of the switch positions in my Apache schematic, but both that and the DX-100 schematic show a 6AQ5 sitting on the bottom of a resistor string fed from the modulation point on the bottom of the RFC, with a set pot on its grid... looks like a screen bias set from what I can see?

One could use that control point (the grid) to do some screen "modulation" in an asymetric way?

Did earlier DX100s not have this 6AQ5 in the screen supply??

Looks to me like for the APACHE in the CW position the 6AQ5 does the biasing, but in the phone position there is a string of VR tubes substituted?? Well, it's one of the two, and the VR tubes would make the screen voltage "fixed"...  I suspect the manual's circuit description tells us what's intended??

I don't think the screen is the right place to try to prevent cut-off due to heavy modulation... quite frankly, I'm not sure what >100% positive peaks would look like on a trapezoid scope! Would it still look trapezoidal, or would it tend to curve off, and where? (that Class E stuff has me thinking about >100% positive modulation)

    _-_-WBear2GCR

[Ian VK3KRI]

I don't think the screen is the right place to try to prevent cut-off due to heavy modulation... quite frankly, I'm not sure what >100% positive peaks would look like on a trapezoid scope! Would it still look trapezoidal, or would it tend to curve off, and where? (that Class E stuff has me thinking about >100% positive modulation)

    _-_-WBear2GCR

>100% Positve just extends the height and width of the trapeziod. If all is well it just looks like a normal trapeziod, but the no modulation resing place isn't in the centre.
                                                               Ian VK3KRI

[Steve - WB3HUZ]

The clipper in the Apache clip BOTH the positive and negative peaks. In other words, it's a distortion generator and not much else. You are better off without it, regardless of what you may see on your scope.

[Bacon, WA3WDR]

True, the screen is not the place to put negative peak clipping.  But when a screen grid is modulated more than it should be for the amount of plate modulation being used, the negative peaks get stretched beyond about 85% negative, causing distortion from that, and making flat-line overmodulation happen even more readily if the audio level increases a little bit from there.  Improving modulation linearity by reducing the amount of simultaneous screen mod helps a little bit.

If the screen is not modulated along with the plate, the plate current does not change much with plate voltage, and positive modulation peaks are flattened.  If the screen is modulated too much along with the plate, a kink starts to appear at negative modulation peaks.

You could experiment with a one-sided peak clipper at low level, but because the waveform is AC-coupled, you get some tilt, and the clipping level shifts around, so the clipping action is not stable unless you arrange some dynamic control of the clipping level to make it stable at the output.  The asymmetrical peak-limiting compressor with servo-clipping applies this kind of control.  You can set it up for any degree of asymmetry, or you can operate it half-wave, detecting and operating on only the negative peaks, by leaving out the positive diode.