THE CASE FOR PEP

By

Eric P. Nichols, KL7AJ

This article will probably be of little interest to two classes of radio amateurs: the die-hard QRP operator who sees no need for ever running high power, and the ham who thinks he must run 1.5 KW at all times, “just because he can.”

Instead, I'd like to provoke some thought about the right reasons for using high power, and the best way of going about it. This discussion will mainly on SSB, but with a few references to CW.

Let's start the discussion off with a question. What is better: a 100 watt SSB transmitter running 10dB of compression, or a 1KW SSB transmitter running super clean?

This question has been ruminated on since the earliest days of phone, sometimes with good empirical data, sometimes not. When does it help to have pep in your P.E.P.?

To make any progress in this discussion, we need to address a fundamental law of physics:

P=E/T

Power equals energy over time. It's a law of the universe to which any meaningful discussion is inextricably shackled.

All other things being equal, we can talk for ten minutes with 100 watts for the same cost as talking a minute with 1000 watts.

What are we really doing when we run speech compression? We're really compressing more USEFUL TIME into a given amount of available time. Or, perhaps better stated, we're increasing the amount of time at which a voice signal is at a certain amount of power. In reality, we can't really compress time. But we CAN eliminate useless chunks of time.

Since the human voice is highly variable over a period of time, we have a lot of time during which any “chunk” of speech (most generally a syllable) that doesn't do us a whole lot of good. The concept of speech compression then, is to bring every part of our signal (in time) up to the maximum available power. What we are doing, actually is increasing the ENERGY without increasing the power.

Assuming equal efficiency, over a substantial period of time, a 100 watt amplifier with 10 dB of compression will “cost” as much as a 1 KW amplifier with no compression. A highly compressed signal, for a given P.E.P., somehow “feels” like it is doing more. It's certainly burning up more energy.

But is it really doing us any good?

Let's look at some pros and cons. Again, let's look at two different power amplifiers, but with equal EFFICIENCY, to keep things on the same footing.

Compression definitely does a few things for us. It eliminates overshoot that could drive our amplifier into non-linearity, causing splatter, intermodulation distortion, and other ills. Paradoxically, putting up with a little intermodulation distortion (which any compression scheme creates) at a low-level, reduces a LOT of intermodulation distortion at HIGH levels. So a little bit of peak clipping is definitely useful.

It's definitely easier to build a 100 watt R.F. amplifier than a 1 KW amplifier. We can do a lot of aggressive compression, as well, with a normal 100W rig without any outboard accessories. 100W is just SMALLER than a kilowatt. A well-compressed 100W signal is more than ADEQUATE for most H.F. communications.

Aggressive speech compression just sounds “louder” at the other end. Is it any more intelligible? This is the big question. As we will investigate later, we might find that lots of compression is more noticeable than useful.

Now some pros about high power. A clean, high power linear amplifier, adjusted properly, can be run with lots of headroom. It can safely handle large voice peaks without saturation, splatter, and other untoward effects.

Very high peak powers can be achieved with very low AVERAGE power levels. You can make an effective 1 KW power amplifier, running an uncompressed signal, with no more “heavy iron” than a 100W transmitter running lots of compression. In fact, in RADAR, it's no problem to build a 1 Megawatt peak power transmitter with a 100 watt klystron! It's all a matter of our original equation. Keep the time down, and the power can go up! It's ENERGY that eats up transmitters, not power. (Within limits).

Now, an uncompressed SSB signal is not likely to have a 10,000:1 peak-to-average ratio as a radar might...but a 10:1 peak to average ratio is not unusual at all for the human voice...especially the MALE human voice. (Female voices actually have a lot higher energy density, on the average.)

Averaged over many cycles, a 1KW amplifier may be no more expensive than a 100 watt amplifier with lots of compression, ignoring the initial investment, of course.

Now, here's an interesting rub. By clipping the peaks of most SSB signals, we greatly increase the average to peak power. Theoretically, this should give us a real edge in communications effectiveness.

HOWEVER...as it turns out, the first peaks to get lopped off by clipping (the lower branches of the Christmas Tree, if you're looking at a typical monitor scope), have the greatest amount of USEFUL intelligence!

The bottom line of all this is that FREQUENCY equalization is of FAR greater importance than compression in increasing distant intelligibility. Chip Margelli of Heil Sound gave a graphic demonstration of this at the Alaska Hamfest last fall. It also means that, since you AREN'T going to be using a lot of compression, you'd best have some headroom available! Chip's optimum system used LOTS of equalization and minimal compression.

Now, someone is always quick to mention something like, “How about those infinite clipping tests that Bell Labs did in the late 1950s? Didn't they PROVE that just outputting the zero crossings of any audio signal has the MOST increase in intelligibility?”

The infinite clipping tests were done with NON-BANDWIDTH LIMITED SYSTEMS. This is the crucial point. The greater your bandwidth, the more clipping and compression will do for you. In a narrowband system, such as amateur SSB, the infinite clipping model is greatly compromised.

So, do you need that 1.5 KW?

Probably not. But it can't hurt to have some pep in your P.E.P. when you really need it.

Eric P. Nichols, KL7AJ