# Using toroidal power transformers as a modulation transformer

Discussion in 'Amplitude Modulation' started by W2BTK, Jan 11, 2019.

Tags:
1. ### W2BTKHam MemberQRZ Page

Well Hello.

So I want to talk about how to use a modern 4/8 ohm low THD audio amplifier as a modulator for plate modulation with AnTek toroidal power transformers as a modulation transformer. I figured I would share this information because quite a few people ask me about my externally modulated Johnson Viking II. I did not come up with this method, but heard about it from a friend and did not see this info anywhere else on the net about how to outsource a modulator for a boat anchor. (By the way my 807s were donated to a worthy individual.)

What I am doing is using a 300w Class H Crown Audio amplifier as a modulator for the Viking II. Any low THD audio amplifier can be used with similar results on any boat anchor or similar plate modulated rig. To do this, a custom modulation transformer is needed. Don't worry though, you can make one with Antek power toroids for CHEAP! Let's go over how.

For this example we will use a mod impedance of 3000 ohms, which is what you would get with a Viking II at .250 ma plate
current and 750v. So as we know, you need a mod transformer to make the 4 ohm output of the audio
amplifier play into the 3000 ohm impedance of the finals. In order to do this, we will have to calculate the turns ratio needed for that impedance transformation. Take the impedance ratio of 4:3000, simplify it to 1:750 and then take the square root of both sides.
You end up with 1:27.3... this is the exact turns ratio you need for your modulation transformer. Something close to that will work splendidly, don't be too concerned with being 100% exact. This math can be applied to many different rigs to get a turns ratio, as long as you know your plate voltage and current.

Now that you know the turns ratio you need, you can start to look at different power transformers to use. The AnTek toroidal transformers are exceptionally well built and can take serious abuse. All of their offerings are 115v primary and they offer many secondary voltages in a variety of VA ratings. However, you will not find a single transformer that will do the job by itself. BUT, if you take two of them and tie the 115v primary windings of the two transformers together, you are left with the just the secondary windings
that can be treated as a primary and secondary of your choosing. Now we can make many different turns ratios. For example, my Viking needs a 1:27.3 turns ratio so I used a 115:55 transformer and a 115:800 transformer together. Since these toroids have TWO sets of secondary windings, I was able to put the two sets of 55v windings in parallel, and the two sets of 800v windings in series to create a 55:1600v transformer, which gives you a 1:29 turns ratio. It is damn close and works awesome, especially since at 200ma plate current and 750v you need a 1:30.6 turns ratio to be 'perfect'. Again, anything close will work marvelously. Two 800VA Anteks in series will work up to 3000v plate. Very good for any boat anchor or even a bigger final. A friend of mine has run these toroids in his modulator setup with 3300v and it was fine, but I wouldn't recommend that. Just know it has been done.

Here are schematics of how I wired my mod transformer and the modulator itself in my Viking II. I use a 300w modulator, but it is only opened up about 1/4 of the way. 300w of audio is enough to blow the finals out of the top of the Viking II, but It's nice to have duty cycle and headroom if you don't abuse it. You don't have to use the three diode ultra-modulator in your modulator if you don't want to, but I think it's cool because it protects the transmitter and it gives people something to moan about. I got the design for the three-diode circuit from the 1956 QST article on AM ultra-modulation by Ollie Allen. Let's not let that be the primary focus of discussion.

Anyways I hope the AM community finds this interesting. Give it a try, it works!
Happy Hamming.
Bill W2BTK

Last edited: Jan 11, 2019
W9WQA, N1BCG and AG5CK like this.
2. ### K4KYVSubscriberQRZ Page

My immediate concern when looking at the top diagram was that the unbalanced DC passing through the secondary would cause core saturation in the transformer. Looks like that issue is addressed in the second diagram, but the 0.1 mfd blocking capacitor seems a little low. Many BC transmitters have a higher modulating impedance than 3K, ≈ 5000 ohms or more, but use a 1 or 2 mfd blocking cap. My homebrew transmitter operates at 4000-5000 ohms modulating impedance (depending on how heavily I load the PA), and my blocking cap is 4 mfd. Higher capacitance = less reactance = more robust low frequency response. I have seen transmitters that used as much as 8 mfd of blocking capacitance.

The concept of using the toroidal power transformers is an excellent idea, if they end up with a good audio frequency response plus acceptable phase-shift characteristics and linearity. 60~ a.c. power transformers have been used with good results in many audio applications, but it depends on the transformers used. The only way to find out if it works is to try it.

Here's my 2¢ worth, FWIW, regarding Ultramodulation:
I tried the ultramodulation circuit many years ago. It worked precisely like the 1956 QST article said, with the plate current meter kicking up during modulation, higher positive peaks on the scope, and the rf ammeter in the transmission line kicking up higher than usual on voice syllables, but reports consistently told me that the audio had a noticeably raspy/fuzzy distortion but that the modulation did not sound any louder than normal. I used it for a while but finally removed it from the circuit, since it works by inherently distorting the sine wave, plus I didn't like wasting expensive and hard-to-generate audio power to heat up a resistor. Although the average plate voltage and plate current kick up during modulation leaving an impression of more modulation on the carrier, most of the increase in RF antenna current is due to increased DC input to the final during voice syllables, from adding the rectified audio voltage produced by the circuit to the +HV from the power supply.

This circuit may reduce the likelihood of buck-shotting from overmodulation peaks in the negative direction, but the distorted waveform is likely to produce more subtle wider sidebands that spill over into adjacent channels, unless a low-pass filter is used between the modulation transformer and final. The high-level low-pass filter using a "splatter choke" brings on its own problems, for example the frequent failure of the modulation transformer in the Collins KW-1.

The end product of "Ultramodulation" is a certain degree of controlled-carrier modulation combined with even harmonic distortion, and only a small increase in actual sideband power.

Pretty much the same as with other variations of "negative cycle loading" that became briefly popular during the late 1950s and early 60s, in attempts to make AM more competitive with SSB as the latter was increasingly gaining acceptance in amateur circles.

3. ### W2BTKHam MemberQRZ Page

I actually use a 4uf oil filled DC blocking cap, because it was free. I just wrote in a .1 to show that you need some type of DC blocker there to make it work. I knew that the ultra-modulator would take over the discussion. What can ya do. I'm mostly amazed at what these transformers can do.

I found the originator of the AnTek-in-a-row idea, Stu AB2EZ and I would like to mention and thank him for applying this concept originally.

4. ### K5UJHam MemberQRZ Page

Thanks for taking the time to post your topic here. It's nice to have a new idea to discuss. Of course the AnTeks have to be mounted in such a way that they are adequately insulated for h.v. I first heard about this from Borgi K9YQQ and he got the idea from someone, but the identity of the person who first came up with this method seems to be always obscured. Another impression I have is (I find this somewhat amusing) that any time I hear of someone running a standard audio amp that is designed for speakers, into an AM rig, it always seems to be a Crown. Everyone seems to have one of these Crowns, but I never know where they get them.

Anyway, I guess the coils on the primary of the first transformer are paralleled for more current handling? I always thought to separate cores could not be used to function as a single transformer. Oh, I think I see...they're paralleled. I think I was told once it doesn't work if two separate transformers have their primary and secondary coils in series with each other to make a push pull interstage primary and secondary side by side. It's okay if they're paralleled and equal because then each core gets the full wave form. When it comes to transformers I excel at confusion.

update: Okay it was Stu AB2EZ who originated the idea. He's a smart dude.

5. ### W2BTKHam MemberQRZ Page

Yes, and I believe that's how it's possible for them to handle what they handle doing this. They also have very low leakage reactances so you can use more than one. To discuss further than that though, I'd either have to pass the torch or read a book. They are just very very well built/shielded transformers. And based in NJ to boot- support a local business!

6. ### K4KYVSubscriberQRZ Page

Not my intent; perhaps that would be a topic for a separate thread.

I wish condx had been just a little better between us on 160m last night. I would have liked to have had more discussion over the air, but QSB was the problem.

Yep.

The same goes for power transformers in a full-wave rectifier circuit. If you use two separate transformers with the HV windings in series, and the common connection between them serving as the mid-tap, what you really have is a pair of half-wave rectifier power supplies in series, with the same inefficiencies in the transformers as with any other half-wave supply. I once used a pair of identical TV power transformers in this circuit. The preferred way is wire up two separate full-wave supplies with the transformers, and then connect the HV outputs in series so that each transformer is supplying power over both halves of the a.c. cycle.

W2BTK likes this.
7. ### N2DTSHam MemberQRZ Page

Been done for 35 years, Eric (WB2CAU) did it that long ago with a solid state amp.
Nothing special about the toroidal transformers except they are hard to mount and somewhat ugly.
Power transformers have been used lots in the past, with the current off them.

I also used to use the 3 diode setup, but I used the 14kv 1 amp bricks and they must not have been fast enough because I always got
reports I was very wide.

Modulation over 120% just sounds bad in many receivers, best to compress some at low level if you want LOUD.

Your rig sounds good because you do not seem to over do it....

8. ### KC2ZFAHam MemberQRZ Page

k4kyv: “most of the increase in RF antenna current is due to increased DC input to the final during voice syllables, from adding the rectified audio voltage produced by the circuit to the +HV from the power supply”

Doesn’t that detune the PA tank ?

9. ### K4KYVSubscriberQRZ Page

How would it have anything to do with PA tank tuning? The rectified audio from the ultramodulation circuit develops a DC voltage, which appears in addition to the fixed DC voltage from the power supply, the sum total resulting in higher DC voltage at the PA and thus more (unmodulated) carrier output. No different from what happens with controlled carrier modulation or if you manually run up the variac on the HV plate supply, or with the instantaneous increase in plate voltage during positive modulation peaks. Increase the plate voltage, let us say 10%; the plate current also increases 10%, and the power level increases 21%, since P ∝ V².

A class-C final with good modulation linearity presents a load to the modulation transformer that acts identically like a power resistor. The load impedance remains constant regardless of voltage and the plate current increases in proportion to the total plate voltage, just like what would happen with a resistor. Instead of dissipating 100% as heat as would a resistor, most of the energy is converted to RF to the antenna, but some is dissipated as heat since no PA is 100% efficient. In a properly operating class-C amplifier, none of this affects plate tuning.

WZ5Q likes this.
10. ### WB2GCRHam MemberQRZ Page

Afaik, Stu tried and identified the Antek brand (local to NJ, fwiw) of power transformers as being suitable for use
in this application. The method of reverse driving a transformer, power or audio "backwards"
predates this by a long time.

As far as getting a Crown MacroTech, just buy one on ebay, or from the local professional
PA/SR company. Usually they've "downsized" the weight by going with a Class D amp for
lugging around the road, and many have retired their old ones. If ur good at repairs, you
might find them with blown finals, and other parts very inexpensively.

Even the "smaller" ones when run in mono put out substantial power.

_-_-