The AM Forum

I just finished my 75 meter AM transmitter and I'm posting the modulation scheme for some final input before I test it on Tuesday.  The RF section was already tested and I'm getting a clean 15 watt sine wave on the output.  Someone advised that I bias the driver stage to prevent thermal runaway.  Thanks, done.

The audio modulation is modified heising and I've borrowed the idea from elsewhere on the web.  It just consists of an 8200uF electrolytic in series with an audio modulated NPN transistor.  The bias on the transistor is adjusted via a 10k pot. 

The modulated audio is run through a 25mH choke.

Jon
KA1TDQ

Jon

1. 25mH corresponds to an impedance of around j8 ohms at 50Hz. Just to be sure, is the modulation resistance of the RF stage 8 ohms or less (to obtain a flat modulation characteristic down to 50 Hz)? For a 15W (at carrier) transistor transmitter... I would be surprised if the modulation resistance is 8 ohms or less.

2. Just to verify... is the 25mH inductor capable of handling the current that will flow through it without saturating?

3. Some people like to make the Heising capacitor as large as possible... but I believe (as also derived in the literature) that the flattest low frequency response (down to the point of rolloff) will result if the impedance of the Heising capacitor is equal to the impedance of the Heising inductor at the audio frequency, f,  at which the impedance of the Heising inductor is equal to the modulation resistance. I always pick the value of the Heising capacitor using that approach. For example, when I externally modulate my Ranger using a modified Heising configuration, I use the following values:

Ranger modulation resistance = 600V/0.120A = 5000 ohms
Heising choke: 20H => j5000 ohms at f= 40Hz
Heising capacitor = 0.8uF => -j5000 ohms at 40Hz

Looking forward to hearing more about your results.

Stu

Oh, I see.

The choke is capable of 4 amps, so it can definitely handle the current.

I see by your numbers how to get the L and C values. My RF resistance is 50 ohms, so my low range roll-off is 318 hz for the inductor and 400 hz for the capacitor. It'll work, but the low range will suffer.

Jon,

  Usually Heising has both a class A audio stage and the RF stage getting DC current through a common Heising choke. Adding the capacitor in series with the collector of the power transistor does not allow any DC operating point to be established.

  To run "Modified Heising" the modulator transistor needs to be able to provide power to the load through the Heising capacitor. Your modulator cannot do that unless you supply it power via another means. Perhaps an NPN-PNP complimentary output audio amplifier would do where the audio runs off the same 33v supply, and the audio output feeds the Heising capacitor.

Maybe something like this, but with a single supply:
http://diyaudioprojects.com/Solid/Jean-Hiraga-Class-A-Amplifier/

Jim
WD5JKO

Yeah, my idea seemed too simple.

Another way would be to skip the on-board modulator and run an external PA amp. Just connect the positive speaker output lead to the electrolytic (-) lead and the speaker (-) lead to ground. That is also 8 ohms.

Also, rather than wasting my 6 volt regulator that's on-board, I can use it trick out the transmitter with bright blue LED's like the computer guys do with their home built computers.

Actually, your idea is on the right track, but just needs a few slight changes. If the audio transistor can take the 33 volts of the power supply, then just remove the capacitor. If it can't, then keep the capacitor but add a dropping resistor from the power supply side of the choke to the audio transistor side of the capacitor, then add a second choke after the dropping resistor. The varying current flowing throught the choke developes an AC voltage across the choke, which feeds through the capacitor (basically the same concept as an impedance coupled audio output circuit).

Jon

You need to verify that the RF stage that you plan to modulate can, in fact, be modulated by modulating the B+ applied to that stage. I.e. if you double the B+, then the current drawn by the RF stage will double, and the RF power output will be 4x bigger. You don't have to run the stage at 15W carrier.... but whatever the carrier output level is, the power should be 4x larger when you double the B+. Similarly, you need to see how low you can drive the B+ on downward portions of the modulation cycle before the RF stage enters cutoff. You can test all of this by just varying (by hand) the B+, and measuring the associated current and RF output power.

You may need to add some type of negative peak limiting circuit to the modulated B+ to prevent the generation of distortion products when the B+ is driven down to near the cutoff level of the RF stage.

Assuming that modulating the B+ does, if fact, result in a proportional modulation of the average current drawn by the RF stage: the modulation resistance is not the same was what is (perhaps confusingly) referred to as the RF output impedance. What is often referred to as the RF output impedance is the optimum value of the RF load impedance at the fundamental frequency (with the antenna or a dummy load connected). The modulation resistance, R, is defined as: (1/R) = the change in average current drawn by the RF output stage / the change in B+ that produces that change in average current drawn by the RF output stage. It is approximately equal to the B+ at carrier / the average RF output stage current at carrier.

The modulation resistance is usually greater than what is referred to as the RF output impedance. In a typical class C RF amplifier, the modulation resistance is a little less than 2x the RF output impedance. In a class E amplifier, the modulation resistance is also a little less than 2x the RF output impedance. In a class B RF amplifier, the modulation resistance is around the pi/2 = 1.57  x  the RF output impedance

Stu

Great explanation! Thanks, I didn't know lots of that.

I'm getting things set up tonight for a bench test tomorrow. Hopefully things go well. If they do, I will get some good measurements to incorporate a negative peak limiter to clamp the voltage before it gets too negative.

Jon

Since you're running such low power, you could get DC to (whatever your high end is) audio with near 0 distortion by using a class A series modulator.  That would certainly be the easiest and cleanest modulator you could use.

Place the modulator in series with the positive voltage going to the RF amplifier.  If your RF amplifier can handle, say 6x the power you're running at carrier, use a power supply that's 2.5 times higher than the carrier DC, and set the carrier DC appropriately.  This will give you up to 150% positive modulation.  A power supply of 3x the DC will allow up to 200% positive modulation.

Stu's comments about whether the RF amplifier can handle the power (and voltage) created by adding a modulator cannot be over stated.

I haven't had a chance to test it yet, but probably will late this afternoon. My guess is that 33 volts will be too much carrier voltage considering the added modulation voltage. I can easily rearrange the power supply configurations for say 12 volts carrier and 33 volts modulation.

I may do 12 anyway to be safe. I tried testing 66 volts on the final during the RF testing and immediately popped the transistor.

When I make the negative peak limiter, what negative voltage am I limiting? I'm sure it's more than the 12 volt supply and I wouldn't simply just clip it at say .5 volts.

Also, the circuit would be between the output of the choke and ground... correct?

Oh no... I just thought about it... its in parallel with the choke to limit negative excursions the other way.

Keep the supply at 33 volts (or whatever it is), and use a series MOSFET or transistor and do the modulation that way and set the carrier DC at 12V using the modulator. Heising is not a great modulation system - there are many limitations, and it is ultimately class A, so it is not more efficient than a class A series modulator, but has many more problems and limitations.

Since he is planning on not running much in the way of carrier power, it might be better to run heising and modulate only up to 95% or so. Running 200% modulation on a 15 watt or so carrier is just asking for trouble. It might be ok when you have an abundant amount of carrier power, but when you're running low power like that, you're not gonna be be heard. The trick to getting 100% or more from heising modulation is to run more voltage on the modulator than on the RF amplifier. Another way of doing it would be to reduce the voltage down to say maybe 24 volts. Hook up the modulator and final like I said, cap between the final amp and the modulator, modulator fed voltage through a second choke. Butinstead putting the dropping resistor between the modulator and power supply, put it on the RF amplifier side, to reduce the RF amp voltage down to 12 volts or so. Reducing the voltage on the RF amp makes it easier for the modulator to go up to 100%, since there's a little less carrier power.

I like the idea of a series mosfet for the frequency response. I plan on using a linear anyway. I'm on the road right now (parked) so any good links to a good mosfet audio circuit? I already have the preamp driver stage, so I just need the biasing setup.

http://files.myopera.com/Frister/projects/modulator.jpg

Here's something to start with. I think this circuit was for an IRF510, but could probably be adapted for some more power if you need it.

Thanks! I'm going to build it. I will keep the big choke and capacitor on the transmitter to help bring dow the DC closer to 30 volts and aid in regulation.

So I'm guessing that a negative peak limiter would go in parallel with the output NPN and the 10uH choke. After getting a good carrier level then build the limiter to suit.

What components do you have lying around the shack?  I could design a circuit around these, depending on what you have.  I need to know about MOSFETs, transistors, op-amps, etc.

I think I have a standard 741 opamp, some 2N3866s, a high voltage N channel mosfet, a 250k variable resistor, 5 volt regulator, and a few 7984 tubes.

Jon

Can you post the schematic of the Class C transistor RF stage that you will be modulating?

Stu

This is fairly accurate.  The output network is for 50 ohms.

Hmmm... The schematic is a bit hard to read.  Looking at the transistor number, the NTE1127 is an IC.  The NTE112 is a diode.  I'm unable to discern whether the last character in the part number is an F or a 7 (looks like a 7), but either way I couldn't figure it out  ;)

Yeah, I made a mistake on that.  The final is an NTE291.  I've also attached a photo of my semiconductor stash.  It includes:

(5) TC4420 mosfet drivers
(2) 7805 voltage regulators
(1) 2N3907 power transistor (vintage 1960's I think)
(1) NTE4011B quad NAND gates (I think)
(1) SA602 for your oscillator/multiplexer needs
(2) good ole 2N3866's
(2) NTE291's
(1) NTE969
(1) NTE51
(5) IRF740's (insert "Halellujia" song here)
(1) 16006 audio NPN transistor I was originally using as a mic preamp
(2) SN74AC74DR dual positive edge triggered flip-flops
(1) LT6231CS8 dua opamp
(2) FST3253MX multiplexer/demultiplexer TTL

Also, there is more voltage available from the transmitter (around 66 vdc). So I can still maintain around a 10 watt carrier with a class A modulator. 

Ok, try this one.  It uses the MOSFETs you have.  Some of the values are very non-critical. If there is a part you have, and it's close but not the exact value, ask about it and chances are it will work.

(http://www.classeradio.com/lp_modulator.jpg)

The 50VDC power supply input value may be a bit too high for your application.  In ideal class C, the RF peaks will get up to about twice the applied DC, but you do want a safety factor and of course if the tuning is wrong or if the load disappears or changes, things happen!

The transistors in the input stages (2n3904) can be substituted with similar devices - 2n2222, etc.   The B (beta) is the important parameter.  You may have to adjust R19 if the transistors are different.  You want about 5 volts at the collector of Q5.

Adjust the carrier for about .3 times the applied DC.   This will give you good headroom for positive peaks.  Keep in mind, with this design, the output will not make the "rail" (full power supply voltage) due to the gate threshold value of the series modulator MOSFET.  It will be about 4 or 5 volts lower than the DC voltage.  So, for the values shown, you can expect about 45VDC max out of the modulator under the highest positive peak conditions.

Thanks a bunch Steve! I feel honored... I have a personalized and customized schematic. 

I will heat sink the fet with no problem.  I have a metal chassis mount already with fins (electrically isolated) with a cooling fan as well.  I like the fact that the negative peak limiter is included too.

I'm going to build it as-is rather than modifying anything.  It'll be maybe a $10 trip to Fry's to get the little pieces. 

I've got lots going on right now until the weekend, but I will put it together and let you know how it turns out.

Thanks again!

Jon

Ok, sounds good!  Hopefully, it will work  ;)  (it should!).

Good idea to get the parts - I would stock up on some very common components such as 2n3904 and 2n3906 transistors, assorted resistors containing multiple values, capacitors (1uF, 3.3uF, 10uF, 47uF, 100uF), etc.  The diode in the negative peak limiter is a 1n4007.  These are good diodes to have around.  Also, I'd get a bunch of 7805s, 7812s and 7912s along with some TL074s and TL072s.  Maybe a couple of 25VCT power transformers if there are any to be had - very useful.

Armed with this, you will have flexibility in building analog circuits quickly.

When I first looked at your W1 call, I thought - heck, this guy is local.  NOT!  I checked QRZ.... :D

Yeah, technically I'm a Maine-iac from back in the day. 

If I do .3 volts carrier from 33 volts, that will put me at 10 volts carrier.  Using Vcc^2/(2Power)=Zout, I get 1-ish watts for 50 ohms. 

If I run the numbers for 18.4 ohms, I will get 2.7 watts carrier.  18.4 ohms gives me a .01 capacitor for the first part of the Pi output network (a good even number for a capacitor).

2.7 watts will probably not strap anything, even in Arizona.

Yeah, voltage is really king in this kind of amplifier if you want to keep the impedance up.  You could either try to run more current (I figured you'd be running about 12 volts at about 1 amp), or you could substitute one of your IRF740s for the transistor (you may have to change the RF amplifier circuit a bit), and run a whole lot more voltage.  With this, you should be able to increase the power supply voltage to at least 60 volts, and run 20 volts at an amp or so at carrier.  That would be good for driving the linear (which is what I think this thing is for anyway!).

I'm almost finished.  I just need to run in town later today and get two 10uF electrolytics.  From the picture, the smaller blue trimmer is the negative peak limit control (used a 5k instead because that's what I had), and the larger blue trimmer is the carrier adjustment.  I've set both to mid-scale.

Also, I had a 250k smooth pot for the microphone audio adjust and used it instead of the 100k. 

I modified the power supply to give me 50 volts instead.  For testing, I tapped the lower electrolytic tap (around 33 volts) to supply just the RF stage.  I didn't want to tap directly on the 50 volt+ tap since that would probably blow the final NPN. Under load, the HV tap is about 54 volts.  Once I add the audio stage in, that should probably be nicely around 50 volts.

If I set the carrier to .3(50)=15 volts, then the maximum rail voltage (about 45 volts) would be my voice peak limit.  If 100% modulation is 30 volts, then 200% peak modulation would 45 volts, yes?  If that's true, insert a Borat "Very nice!" here.

Also, carrier power at 50 ohms and 15 volts would be about 2.5 watts.  Not a lot of power at all, but it opens up quite a few posibilities for linears.

Jon
KA1TDQ

Once you get this working, the next step is higher power!!

Oh, the audio input is not microphone level by any stretch of the imagination.  You probably know that, but just in case - the last post you did use the phrase "microphone audio adjust", and the modulator, as designed takes LINE LEVEL input  :)

On the modulation question - 200% modulation is 3x the carrier DC in a high level modulated system.  So, your 15VDC carrier, if the modulator supplies 45V on an audio peak would indeed be 200% positive modulation, assuming all of the RF amplifier components, and the amplifier itself as a unit follows a perfect square law (linear) characteristic.

Question:  have you actually used the RF amplifier so you know it works correctly?  I think you have, but I don't remember  :P

Yes, I tested the RF section before I started the modulator this morning. 

I've attached a copy of the RF output waveform I'm getting with the carrier adjusted to 15.0 vdc.  It looks like an 18 volt pk/pk wave.  Also, I measured the HV during keyed state, and I'm still getting 51.4 volts loaded (plenty for 200% modulation).

While looking for the screws for the cover I found a couple non-polarized 10uF/25v caps that I had pulled from a tv some time ago.  I threw them in and it should be fine.

I need to replenish my audio stuff.  All I have left is a cheap Goodwill microphone that I picked up for a couple dollars and no "good" line level audio stuff.  This will give me plenty of time to get an amplifier going.

Jon

If you get a chance, can you re-do your last post with the B+ adjusted to 7.5V (instead of 15V) to see if the peak-to-peak amplitude of the RF output changes by a corresponding factor of 2.

Stu

Hey, if you get a chance, post a scope picture of the collector waveform at a known DC (just tell us what it was/is when you take the picture).  That would be very interesting !

Regards,

Steve

I've attached a couple scope pictures, both taken with the carrier adjusted for 7.5 volts.  The one taken at the collector seems to show to waves, one small one really large.  It's not there after the filter in the other photo.

Jon

What is the peak-to-peak amplitude of the RF output signal when the B+ is 7.5V? If the oscilloscope scale has not been changed... then it appears to be more than half of what the peak-to-peak amplitude was (18V p-p) when the B+ was set to 15V.

Stu

That's pretty neat.  I wonder if it is possible to increase the sweep rate and show, say, 3 or thereabouts waveforms on the screen so we can see what is really looks like - and increase the volts-per-division so that none of the waveform is off the screen.  How many volts peak are on the transistor's collector with a known DC voltage - that's a really important number.

Very interesting!

Regards,  Steve

I always thought that there was one simple wave.  I guess there's more.

I've attached a photo of the finished product.  To get the 200% positive peak capability and set the voltage on the final at 15 vdc, I'm getting an 18 volt pk/pk output wave.  If my math is correct for a 50 ohm load, that's 6.5 watts. 

Thanks to everyone who helped me complete this.  I bought the enclosure at a local hamfest for $1 and was debating even then.  I'm glad I bought it and the project morphed into this.  I'm going to have lots of fun with this radio! And Steve, thanks especially for the modulator design.

I'm going to post a nice picture on my QRZ profile sometime this weekend to put a bow on this thing.

Jon
KA1TDQ

Jon

18V p-p  is equal to 9 volts peak.

9 volts peak across a 50 ohm load corresponds to 0.81W into the load [I.e. 0.5 x 9V x 9V / 50 ohms]

Assuming the output RF amplitude is linearly proportional to the value of the modulated B+... then:

At 100% positive peak modulation (for example) your peak envelope power will be 0.81W x 4 = 3.24W

At 150% positive modulation peaks (for example) your peak envelope power will by 0.81W x 6.25 = 5.06W

Stu