Distributed Amplifiers

[T3sl4co1l]

I've been playing with (or at least hypothesizing about) distributed amplifiers lately.  Mainly because, for the hardware I have on hand, it's the easiest way to get a fairly large amount of RF power (for wideband lab applications)

Which, because common power BJTs and MOSFETs have internal frequency limitations (distributed resistance, excessive junction capacitances, and recombination in the case of BJTs), means vacuum tubes (which tend to have a higher internal frequency limit, but have relatively high capacitance and very high output impedance in comparison).

So I'm wondering a few things:
1. Is it practical to build one with triodes (grounded cathode)?
2. Is it practical to build one with grounded grid (cathode input) design?
3. Can a push-pull amplifier be built?  I don't mean a balanced amp (such as Tektronix used extensively, for obvious reasons), I mean class AB, lots of harmonics each side.

Partial answers:
1. The general advice seems to be, avoid output-input coupling at all costs, because you'll screw it up.  Hence, almost unanimously, pentodes or cascodes have been preferred.  This persists today with lots of articles on monolithic FET amps using cascodes.  (Though with the C(V) variation of semiconductors, that's probably even more important than with tubes.)  I've read a patent: http://www.google.com/patents/US2778886 that says it's perfectly fine, as long as the triangle of transmission lines (GND-grid, grid-plate, and plate-GND) has equal propagation velocities.  But anyway, adjusting such a circuit seems... ridiculous.  So I'd probably want to avoid that, anyway.

2. Most sources say that, if you have a lossy input line (some FETs, most BJTs), you're limited in the number of stages.  Which implies that a lossy input device simply can't provide additional GBW, no matter what.  But this looks stupid to me.

You can easily build a tapered transmission line structure, that effectively connects all those loads in parallel.  At each node, you have three connections: the input-side filter section, an amplifier load, and the filter section to the next amplifier stages.  The input side has an impedance of the parallel combination of the amplifier and the next filter section.  The last section is simply matched into the last amplifier load, no termination resistor. (Or you could add a termination resistor, but this would only have the effect of reducing the entire network impedance, while wasting power.)  Who needs termination when you can get distributed termination for free, right?

The impedances work out, so it should work correctly.  But I don't want to go and build one, and spend weeks trying to tune it, only to discover it's impossible.

3. I don't know, but suspect, that, if you take two matched but otherwise independent dist amps, and split and merge them with a 180 degree hybrid on the input and output, you won't get a PP system, because the energy from one must go down and reflect off (or be absorbed by) the other, as well as go to the output.  Which means the pushing and pulling has to be done on a more fundamental level, such as using tightly coupled inductors in a stage-by-stage complementary circuit.  Is this right?

Interesting, I see a ref that says it really can be that simple, https://www.google.com/patents/US4446445

On the other hand, this one uses hybrids on each tube, http://www.google.com.na/patents/US3571742

Perhaps it's not exclusive?

Tim

[rfeecs]

I can't believe this is the easiest way to make a power amplifier.   But if all you have on hand is stone knives and bear skins...

I believe one issue is you want a high output impedance for the devices because lower output impedance contributes to losses along the output transmission line.  That is one reason people often use cascode devices.  Pentodes would have the same advantage.  No idea if common cathode triodes would work, but if they have a low output impedance that would be a problem.

Similarly a lossy input line is a problem.  You need the signals along the input transmission line and the output transmission line to be properly phased so that the signal on the output travels toward the load and add up in phase.  A lossy input line will cause the signal to attenuate as it goes from device to device which is especially bad for a power amplifier.  People have tried to compensate for this by using series caps that scale along the input line to try to equalize the power to each device.

A push pull can be made just like any other amplifier type.  Just think of the two amplifiers as black boxes.  It doesn't matter what goes on inside.  Class AB probably would work as well.

One notorious problem with distributed amplifiers is above the cut-off frequency the phasing falls apart.  This combined with the huge number of possible feedback loops make oscillation likely if it is not accounted for in the design.

Here's a classic tube distributed amplifier, HP Journal Volume 1, Number 1:
http://www.hparchive.com/Journals/HPJ-1949-09.pdf

[rfeecs]

I am endeavoring, ma'am, to construct a distributed amplifier circuit using stone knives and bearskins.

[xygor]

Tektronix did a common cathode triode one.
http://w140.com/tekwiki/wiki/File:Tek_581_vertical_output_amp.png

[T3sl4co1l]

I can't believe this is the easiest way to make a power amplifier.   But if all you have on hand is stone knives and bear skins...

That, and the network theory experience.  But people seem to place little value on those experiences so I put in a more concrete explanation.

I believe one issue is you want a high output impedance for the devices because lower output impedance contributes to losses along the output transmission line.  That is one reason people often use cascode devices.  Pentodes would have the same advantage.  No idea if common cathode triodes would work, but if they have a low output impedance that would be a problem.

Yeah, but that's just copying the same "rule of thumb" I mentioned, without providing a proof that that is the ultimate limit.  And it doesn't refute my proposal for a tapered line, which should develop equal voltages and phase shifts at all nodes, which would quite handily disprove the common advice.  So I'm looking for theory on that.

A push pull can be made just like any other amplifier type.  Just think of the two amplifiers as black boxes.  It doesn't matter what goes on inside.  Class AB probably would work as well.

"Black boxes" doesn't make sense to me.  A 'box' is only a 'box' in so much as your external description is sufficient to allow you to simplify it.  But I don't think that's possible, in that, to have a complete description of the amplifier's phase and distortion (in order to cancel out even harmonics, in this case), you have to look at all the devices and networks inside the thing to begin with.  Not to say that you can't try and abstract it, but that when you do so, you've abstracted away no superfluous information!

Or to put it another way -- can you prove that it is sufficient to use splitter/combiners to obtain true class AB, low distortion performance?

I expect that's the hardest part, and the root of what I'm asking about -- it seems theoretical knowledge of these sorts of complex networks was always in the rarefied elite, and has now lapsed into the domain of the arcane and ancient!

One notorious problem with distributed amplifiers is above the cut-off frequency the phasing falls apart.  This combined with the huge number of possible feedback loops make oscillation likely if it is not accounted for in the design.

Here's a classic tube distributed amplifier, HP Journal Volume 1, Number 1:
http://www.hparchive.com/Journals/HPJ-1949-09.pdf

That's a strange statement: their schematic (I looked up the manual) doesn't show the 'speed-up' caps of a bridged-T network, only the traditional cascaded inductor structure.  Of course, parasitics and mutual inductance are not shown, either, but it still seems disingenuous.

Also in either case, the problem with that example is, nice though it is, it runs an incredibly steep load line: 1400 ohms per tube, well under the maximum power output load for that device.  If you need maximum power (which is the joint goal I'm investigating), things get a lot harder.

Tim

[T3sl4co1l]

Tektronix did a common cathode triode one.
http://w140.com/tekwiki/wiki/File:Tek_581_vertical_output_amp.png

Sort of -- that actually shows they didn't want to try, in a sense: because they use a balanced structure (matched pairs of TLs and triodes), they have complementary voltages available, so can neutralize each stage.  This type of neutralization cancels Miller effect but doubles Cgp's effect on loading, thus reducing their total bandwidth.

Tektronix was probably the most visible user of distributed amplifiers, but they also did exclusively small signal applications.  At most, a watt or two was needed to drive the CRT, and their amplifiers (plus a lot of triggering and power regulator circuitry, of course) notoriously consumed hundreds of watts!  Their driving constraint was absolutely pristine signal quality, damn the expense.

I think CATV used a lot of dist amps, even into the SS era (i.e., using transistors), for obvious reasons -- but again, fairly low power levels, even for distribution head amplifiers.  Efficiency might've been a bigger concern for them (and reliability even more so).

The only dist amps I know of, offhand, where sheer power and bandwidth are the requirement, are lab amps, such as Amplified Research's portfolio.  As far as I know, these were mostly class A, SE amps as well (using radial-beam tetrodes and planar triodes to push 200MHz and more), since efficiency wasn't much of a concern either.

And almost all broadcast has been of rather limited bandwidth, with television being the widest at about 5MHz (nearly 6 if audio is included at the same time), so there's never been much use in commercial radio from what I've seen.

Tim

[rfeecs]

Distributed amplifiers have been analyzed to death.  A search on IEEE Xplore yields more than 800 papers:
http://ieeexplore.ieee.org/search/searchresult.jsp?queryText=.QT.distributed%20amplifier.QT.&newsearch=true

They were one of the first MMIC applications.  They are widely used in fiber-optic communications systems.
As you say, not many applications for very high power broadband RF.  Test equipment is certainly one.  Military jammers is another.

But now they are used mainly for higher frequencies.  6 GHz and above.

[Wolfgang]

... one of the pests of distributed amplifiers is stability.

If the have a somewhat higher gain, they form a transmission line (with frequency dependent characteristics) that is prone to reflective excitation (i.e. standing waves or self-sustained oscillations) if the terminations are not perfect.

I value or sporty endevour but nowadays there *are* easier ways to climb a mountain.

I would recommend a lot of simulation before investing a lot in experiments.

Much luck 

[capt bullshot]

The only dist amps I know of, offhand, where sheer power and bandwidth are the requirement, are lab amps, such as Amplified Research's portfolio.  As far as I know, these were mostly class A, SE amps as well (using radial-beam tetrodes and planar triodes to push 200MHz and more), since efficiency wasn't much of a concern either.

Afaik, these (at least the modern kind of) amplifier use a bunch of identical transitorized class A stages, paralleled by power dividers and combiners. For more gain and power, a multi-stage topology is added (so imagine one stage as the first input amp, splitted to e.g. two intermediate amplifiers, these splitted to each 4 power amplifiers and then 8 amplified signals combined to one output).
Again, afaik, this is also the topology modern FM radio transmitter stations use (combining many transistor amplifiers outputs).
It's all a matter of exact phase alignment.

[T3sl4co1l]

Afaik, these (at least the modern kind of) amplifier use a bunch of identical transitorized class A stages, paralleled by power dividers and combiners. For more gain and power, a multi-stage topology is added (so imagine one stage as the first input amp, splitted to e.g. two intermediate amplifiers, these splitted to each 4 power amplifiers and then 8 amplified signals combined to one output).
Again, afaik, this is also the topology modern FM radio transmitter stations use (combining many transistor amplifiers outputs).
It's all a matter of exact phase alignment.

That works for narrowband (especially for simplifying the splitters, potentially?), but just gets more and more troublesome for wideband.  FM BCB is very narrow, the relative bandwidth (for the entire band) is only 10%, and for a single channel, 0.2%.

Also, notice the necroposting

Tim

[coppercone2]

My plan for a broadband lab amp is to take all my TWT devices and hook them up together with diplexers to form a unified 0.5-26.5GHz amplifier that goes from 50W at low end to 1W at high end out of like 7 different test equipment chassis.

I have the amps on a rackmount but I am left without a diplexer network at this moment. I thought it would be a cool thing especially if it was hooked up to a Agile DDS chip., or a few of them connected to mixers/upconverters, it would be like one of those EW simulators. I found that trying to buy budget diplexers with the same frequency bands as my amplifiers was futile. $$$
I managed to get a few 'interesting' overlap's from eBay but I want like a smooth lab-equipment amplifier not hodge podge.

If you wanna do this though I would recommend just buying the 800$ or so MMIC modules that are sold from AD, all those ancient amplifiers kind of look like a joke when I think about it. I feel like there should be a surgery table with Frankenstein down there, you would think you walked into a portal to the 1970's.