RF Output matching using attenuators

High frequency, antennas

RF Output matching using attenuators

Postby johnf » Fri Aug 13, 2010 4:56 am

sort of counter intuitive
from my web site www.iel-rf.com and expanded


Using Attenuators as Loads


Copyright Instrumentation Engineering Limited May 2001

RF loads are usually required when testing RF transmitters. The reason for this is that the transmitter has been designed to work into an impedance that is usually equivalent to 50 ohms resistive. Modern transmitters i.e. solid state Bipolar or Mosfet, will not produce their rated output power unless they see a reasonably good load impedance that approximates the above. A good rule of thumb for estimating forward power from solid state PA’s is-
Forward power is bounded by 1/S and 1/S^2 where S is the SWR number (S:1). Typically the rated output power is held constant by the internal control and bias circuits for SWR values up to 1.3:1 or sometimes as high as 1.5:1 (this problem is caused by the load lines for the active devices in the output shifting). At his point the forward power starts to turn down as SWR increases. Typical antennas present load impedances with SWR values as high as 3.0:1. In such cases the forward power could be as low as 1/9 of the value presented into a good dummy load. The true power is 25% less than the 1/9 value due to the reflected power of the 3.0:1 SWR. (25% of forward power is reflected to the transmitter by a 3.0:1 SWR. Therefore the worst case true power delivered to the load is approximately 11dB below the power delivered to our good dummy load by our transmitter.

Most transmitters will work into a load SWR of up to 3.0:1 without damage. An SWR of 3.0:1 is equivalent to a load return loss of 6dB. A 3dB attenuator also presents a return loss of 6dB if the end away from the transmitter is open or short circuited or shunted by a pure reactance.
Dummy loads unless expensive precision types have typical measured return losses of between 15 and 20dB (SWRs of 1.15:1 and 1.05:1) with cheaper units specified at SWR of 1.2:1 ( a return loss of approx 13.5dB).

From the above it can be seen that an attenuator of greater than 3dB can be used as a load that will not damage the transmitter. However if transmitter output power is to be measured with reasonable accuracy then attenuation values of greater than 8dB should be used which will give a result accurate to  0.12dB.

Examples of attenuators as a load use.

1./ From above. If we connected an attenuator of 4dB in series with the antenna the transmitter would see a return loss of 14dB (made up of 6dB from the antenna and 8dB from the attenuator. The transmitter would then supply full power minus 4dB to the antenna, a theoretical gain of 6dB. However the antenna because of its SWR has a mismatch loss of a further 1.25 dB. We are still ahead by 4.75dB in transmitted power than without the attenuator.

2./ By putting an attenuator onto an isolator (circulator with dump load)
Instead of a load, access to the transmission system can be had to check deviation, carrier frequency, and condition of the antenna / feeder. If the attenuator were 10dB in value this would provide 20dB of return loss to the circulator and reflected energy from the circulator would be measured at one tenth actual value. Advantage of this is that all this can be done with out taking the transmission system out of service.

3./ Checking antenna SWR
When checking antenna SWR (return loss) the antenna should be isolated from the output impedance of the transmitter. This is due to the fact that most commercial land mobile transmitters do not have a 50 ohm output impedance. By putting a 10dB attenuator on the transmitter output the measuring bridge (which is designed for 50 ohm source and load) will measure the actual SWR of the antenna with reasonable accuracy. Without the attenuator it is possible that a good SWR indication is had when in fact the true SWR indicates the antenna or feeder has failed.

4./ Placing an attenuator between RF FM Broadcast exciter and RF power amplifier.
If a low value attenuator ie 1dB – 3dB is placed between exciter and power amplifier both the Exciter output and the power amplifier input see increased return loss helping stabilize power control circuits and adding significantly to RF stability of the power amplifier especially if the power amplifier is working into a high Q output load such as cavities or a multi coupler. Harmonic and spurious energy that is reflected by the output cavities is reflected through the power amplifier to its input and onto the exciter output where again it is reflected back to the power amplifier input giving rise to instability of the power amplifier. This effect is most pronounced with modern wideband solid state power amplifiers operating in anything other than pure class A. Class B, C, D, all exhibit this behaviour while class E,F,G also suffer they also tend to modify their input impedance with drive level.

5./ At frequencies in the HF to VHF range for a single spot frequency rolls of 50 ohm coax can be used as attenuators by using the coax makers attenuation vs frequency charts. High values of attenuation are unlikely at low HF frequencies but useful Pa output protection can be had with RG58 or RG174 the later will have to be water cooled for powers above 100watts @ HF
johnf
 
Posts: 376
Joined: Sun Aug 08, 2010 4:51 pm
Location: Wellington New Zealand

Re: RF Output matching using attenuators

Postby Doug Coulter » Fri Aug 13, 2010 11:37 am

I like the ideas here. I'm thinking they might be handy when trying something that may have huge SWR possibilities, like driving a plasma with an arbitrary waveform, in some mode where it "goes out" for part of a cycle, something I happen to have an interest in here. Pretty hard to tune out SWR on a wideband signal with a time-varying load by the usual methods.

For sputtering, I'd think the usual automatch does OK, if you have one, and I'd further guess that the RF sources for that tend to be a little more robust than a normal transmitter as well, to handle high SWR until the plasma fires off. I know the one Jerry once sent me a schematic for looked a whole lot stouter than your usual radio amateur transmitter that is usually running right on the edge of survival.
Posting as just me, not as the forum owner. Everything I say is "in my opinion" and YMMV -- which should go for everyone without saying.
User avatar
Doug Coulter
 
Posts: 2968
Joined: Wed Jul 14, 2010 8:05 pm
Location: Floyd county, VA, USA

Re: RF Output matching using attenuators

Postby Crispin Metzler » Thu Oct 13, 2011 2:07 pm

Related......

For tube FM transmitters I was fond of loading input tanks with power resistors to use up surplus drive power. Directly reducing the Q at the grids would improve fidelity and stability. No downside really. I operated a pair of 4-400A's in push-pull at 100MHz (data sheet says 60) with NO neutralizing circuit of any kind this way to prove the concept. Even biased them Class A as a worst case test of the concept. (OK, there was also a fancy Z=5 Ohm screen bypass cap made of .031 PC Board.)

As a kid putting up TV antennas for a living we would put 3 - 6 db attenuators at each end of the line when very near the towers. It cut ghosting from line imbalance pick up etc. and prevented overloading distribution amplifiers.
Crispin Metzler
 
Posts: 5
Joined: Thu Oct 13, 2011 11:36 am

Re: RF Output matching using attenuators

Postby Doug Coulter » Thu Oct 13, 2011 11:21 pm

Right, there are plenty of times when a strategic loss of a few dB wins you more than you lose by it. I do often work the other side of the street -- how long can this battery last, how deep into the noise can I get this signal back from, and so on, and so do many others. It can blind you to the fact that it's just not always true that perfect efficiency is always the best and main goal. Efficacy might be a better term.

Now, a real master would just redesign the tank circuit that needed de-Q ing so that it didn't have too much Q in the first place, more elegant, fewer (and probably cheaper) parts, but there I go again...

I've already gotten into trouble with that here, with a couple designs that only worked because I chose components so carefully that various sorts of compensation and neutralization weren't required (for examp0le, gain fell off before any phase comp was needed to close a loop). Then someone came along and wanted to "improve" on them and couldn't get them working with the "better" parts...So I guess if you go down that road, you ought to give some sort of fair warning. ;)

////////

We have to be careful with a couple words here - most of the time when a physicist uses the word Q, it has nothing to do with what an RF guy is using that designator for. Ditto resonance...
On this thread, we're fine and all know what we're talking about, but to a nuclear physicist, Q is net reaction energy, and resonance, gosh...has several overloaded (in the software sense) meanings in just that one field. Few of them look like what we think of as tuned circuit behavior, ...mostly. Sometimes you get a "resonance" plot that really does look like a tuned circuit - one hump for one thing and another plot with a zero crossing where the hump was on the first (eg a phase plot, but for a physicist, phase means...oh crap -- there's no end). But sometimes it means something got effectively bigger in physical size (reaction cross section resonance)... or simply more likely...or...only happens in a narrow area (almost the EE definition).

You know, this invention of jargon, which in good faith probably started out to make things simpler, is now a way to keep the unwashed out of your field it seems (read any modern paper), and boy, I think it holds us back. I'm not going to propose new terminology (I'm not smart enough to make it up myself) but...something to be aware of.
Posting as just me, not as the forum owner. Everything I say is "in my opinion" and YMMV -- which should go for everyone without saying.
User avatar
Doug Coulter
 
Posts: 2968
Joined: Wed Jul 14, 2010 8:05 pm
Location: Floyd county, VA, USA

Re: RF Output matching using attenuators

Postby Starfire » Fri Oct 14, 2011 7:53 am

Nothing teaches more than melting a feeder with a miss-match :cry: - great work John, thanks
Starfire
 
Posts: 143
Joined: Thu Aug 05, 2010 5:26 pm
Location: North Ireland


Return to RF

Who is online

Users browsing this forum: No registered users and 1 guest