Call Search
     

New to Ham Radio?
My Profile

Community
Articles
Forums
News
Reviews
Friends Remembered
Strays
Survey Question

Operating
Contesting
DX Cluster Spots
Propagation

Resources
Calendar
Classifieds
Ham Exams
Ham Links
List Archives
News Articles
Product Reviews
QSL Managers

Site Info
eHam Help (FAQ)
Support the site
The eHam Team
Advertising Info
Vision Statement
About eHam.net

donate to eham
   Home   Help Search  
Pages: Prev 1 [2] 3 4 5 6 7 ... 11 Next   Go Down
  Print  
Author Topic: Goodbye tubes.  (Read 5142 times)
M0HCN
Member

Posts: 566




Ignore
« Reply #15 on: July 12, 2019, 04:29:52 AM »

G3RZP, I know you know this, but for others, a big part of the IMD issue is that pretty much everyone runs the damn things with a swamped gate (Transformer down to a few ohms, then swamp the gate capacitance with a low value resistor).

It is easy, stable, very high gain and rather less prone to gate overdrive issues then a feedback amplifier is. You get a 36dB gain stage, taking maybe a couple of watts to get your kW, and burning power in an input attenuator to reduce the 100W from the rig to something the gates will survive. Problem is, it is pretty damn close to being open loop!

If instead you burn that 100W of drive power in the feedback resistors and wind up with say 13dB of stage gain, you have substantial feedback around the device and thus very much better IMD levels. Feedback always makes the products more complex, so you need to use enough of it, but it does substantially linearise the stage.  

The problem with a high feedback design is that the thing becomes very sensitive to overdrive because once the FET drains are banging into hard saturation, the feedback goes away and gate voltages shoot up, the same thing happens if you have a shorted output for example where there is NO RF voltage at the drains irrespective of the gate voltage (In a open loop design the PSU current trip comes to the rescue, but that does not help here or prevent the gate being blown out).

You MUST provide gate overdrive protection if doing this (PIN switch or similar and clamps), and it must sense the gate voltage, not the input power, because in the event of a load problem the two are not the same.

You should also do the load pull studies and plot the stability circles, as stability can be rather more of an issue.

On the bench I can hit -60dBc (Not PEP) IMD3 with the higher order products off my SA screen by 9th (So better then about -90dBc), turning that into a publishable design is of course the 'ask me for anything but time' activity.  


73 Dan.
Logged
G3RZP
Member

Posts: 1217




Ignore
« Reply #16 on: July 12, 2019, 10:21:22 AM »

Dan

Quote
The problem with a high feedback design is that the thing becomes very sensitive to overdrive because once the FET drains are banging into hard saturation, the feedback goes away and gate voltages shoot up, the same thing happens if you have a shorted output for example where there is NO RF voltage at the drains irrespective of the gate voltage

That was basically known many years ago (more than I've been alive!) and has since often been forgotten!! (Does any university teach analogue and RF design anymore?) See ‘Second thoughts on radio theory’ by ‘Cathode Ray’, published for Wireless World, Iliffe & Sons Ltd., London, 1955, or in Wireless World, ‘Negative Feedback’ by ‘Cathode Ray’, February & March 1946, (available at https://www.americanradiohistory.com/)

The other 'little' thing with a high feedback design is ensuring that the Nyquist plot stays stable at all power levels - and it is of course dependent on things like Miller capacity (variable with signal level!) and gain compression. (Personally, I always prefer a Bode plot, because it doesn't need normalising). And of course, stays stable with the impedance circles that it sees at the harmonic frequencies, which can be a different kettle of fish....

It is somewhat frightening to see the extent to which - especially the higher order - IMD Products have got worse over the years since tube PAs  stopped being made....


Logged
K4EMF
Member

Posts: 157




Ignore
« Reply #17 on: July 12, 2019, 10:54:17 AM »

I have to say I find reading you guys that know electronics very interesting.  Even if I don't always know exactly what you're talking about.
Logged
M0HCN
Member

Posts: 566




Ignore
« Reply #18 on: July 12, 2019, 10:57:39 AM »

Yep, fairly basic feedback system theory.

Harmonic stability can also be a case of making sure the exciter terminates harmonic power coming back down the link cable reasonably (Most don't), which is one of those things that makes me feel that building a transmitter usually trumps building an amplifier, much easier to the the transmitter case right when you control the whole damn chain, and it is not like small signal doings are notably expensive these days. The upside of dumping 90% of the exciter power into a pad is that a 10dB pad at least guarantees a 20dB return loss for the harmonics heading towards the rig, directly connecting puts the output part of the LPF across the gate network (phase rotated by the length of cable), this is not good.

Fortunately these days we have software that does harmonic balance and load pull calculations and that can plot the stability circles for us.....

It is not that LDMOS is bad, its just that there seem to be the same old tired designs based on prehistoric Motorola app notes from when the MRF150 was young that keep on being minimally respun for new devices (Except that some of those old app notes actually had feedback!).

Cathode Ray was awesome!

73 Dan.
Logged
G3RZP
Member

Posts: 1217




Ignore
« Reply #19 on: July 12, 2019, 03:11:01 PM »

Dan,

Back in about 1975, there was an article in 'Electronics' magazine about predicting stability in solid state PA with the impedance circles of the load impedance at the harmonic frequencies on Smith Charts....I figured back then that design of valve amplifiers was easier - although at the time, I was leading a group on HF receiver design...which (story of my life) had a reorg and never got into production afterwards...

Yes, a lot of 'Cathode Ray'(M. G. Scroggie) is still good basic knowledge today...
Logged
M0HCN
Member

Posts: 566




Ignore
« Reply #20 on: July 12, 2019, 03:59:02 PM »

The problem with stability circles on a Smith chart is that it can be tricky to tell which side of the curve is the stable one...

The basics don't change and a newbie could still do worse then reading F.Termans magnum opus (Also, Langford-Smith).

73, Dan.
Logged
G3RZP
Member

Posts: 1217




Ignore
« Reply #21 on: July 13, 2019, 05:57:36 AM »

Scroggie's  (Cathode Ray) 'Essays in Electronics' has a good introduction to Nyquist diagrams that is far, far better than I had at college!
Logged
NO9E
Member

Posts: 880




Ignore
« Reply #22 on: July 13, 2019, 07:28:41 AM »

With new transistors, the old concept of transceivers is probably obsolete. The original concept was useful to avoid replication and ensure same frequency without many cables. The new transmitter can use a digital stream or generate all modes natively, has ALC and pre-correction internally,  adjust PS voltage for optimal efficiency, etc. And the receiver is either connected via Ethernet or is a small board inside. With higher efficiency, probably what made 100W before can now make 300W with same weight and heat.

The current FCC rules limit the amp gain to 15 db. But it does not apply to a digital input.

An intermediate new design of that sort could be Expert 1.5k with kx3 fitted inside and KX3 software adapted for pre-correction etc.
21 lb 1.5 KW and low IMD. Zenki special.

Ignacy, NO9E
 
Logged
W9IQ
Member

Posts: 3211




Ignore
« Reply #23 on: July 13, 2019, 03:47:17 PM »

Quote
The current FCC rules limit the amp gain to 15 db. But it does not apply to a digital input.

How is it that the 15 dB limit does not apply to digital? The regulations don't seem to grant such an exemption:

Quote
Not be capable of amplifying the input RF power (driving signal) by more than 15 dB gain. Gain is defined as the ratio of the input RF power to the output RF power of the amplifier where both power measurements are expressed in peak envelope power or mean power. 


- Glenn W9IQ
Logged

- Glenn W9IQ

I never make a mistake. I thought I did once but I was wrong.
AC2RY
Member

Posts: 723




Ignore
« Reply #24 on: July 13, 2019, 05:39:02 PM »

Quote
The current FCC rules limit the amp gain to 15 db. But it does not apply to a digital input.

How is it that the 15 dB limit does not apply to digital? The regulations don't seem to grant such an exemption:

Quote
Not be capable of amplifying the input RF power (driving signal) by more than 15 dB gain. Gain is defined as the ratio of the input RF power to the output RF power of the amplifier where both power measurements are expressed in peak envelope power or mean power.  


- Glenn W9IQ


There is NO power at digital input - only a number usually representing a value between -1 and +1 at each sample.  Actually the best interface will be the optical one - thus not even electric connection between exciter and amplifier.
« Last Edit: July 13, 2019, 05:42:17 PM by AC2RY » Logged
W9IQ
Member

Posts: 3211




Ignore
« Reply #25 on: July 13, 2019, 05:48:22 PM »

Humorous...

- Glenn W9IQ
Logged

- Glenn W9IQ

I never make a mistake. I thought I did once but I was wrong.
SM0AOM
Member

Posts: 253




Ignore
« Reply #26 on: July 14, 2019, 01:04:11 AM »


Harmonic stability can also be a case of making sure the exciter terminates harmonic power coming back down the link cable reasonably (Most don't), which is one of those things that makes me feel that building a transmitter usually trumps building an amplifier, much easier to the the transmitter case right when you control the whole damn chain, and it is not like small signal doings are notably expensive these days. The upside of dumping 90% of the exciter power into a pad is that a 10dB pad at least guarantees a 20dB return loss for the harmonics heading towards the rig, directly connecting puts the output part of the LPF across the gate network (phase rotated by the length of cable), this is not good.

This is quite often overlooked.
It turns out that limiting stage gains using feedback is very good for IMD suppression.

When the transition between bipolar and VMOS transistor came in the late 80s many designers used the increased stage gains for simplifying the transmitters. This often caused worse IMD performance compared to the bipolar chains.

Another aspect was that the low-pass filters at the output needed to have a "harmonic dump" as otherwise the reflected harmonics into the VMOS amplifiers made IMD performance markedly worse. It appears that reflected harmonics have less influence in bipolar transistor amplifiers.

Making an ISB-rated VMOS or LDMOS power amplifier is a major undertaking, and needs careful attention to feedback and gain distribution.
Logged
W1VT
Member

Posts: 3289




Ignore
« Reply #27 on: July 14, 2019, 04:48:09 AM »

Making an ISB-rated VMOS or LDMOS power amplifier is a major undertaking, and needs careful attention to feedback and gain distribution.

With relatively affordable standalone spectrum analyzers and other cheaper alternatives  you no longer need a facility like the ARRL Lab to build and test something like that.
Logged
W1BR
Member

Posts: 4178




Ignore
« Reply #28 on: July 14, 2019, 07:33:45 AM »

A clean two or three tone signal source is the problem for most of us.  Have the spectrum analyzers, but generating 100 watts with low IMD takes some work.
Logged
M0HCN
Member

Posts: 566




Ignore
« Reply #29 on: July 14, 2019, 08:04:51 AM »

It is somewhat unclear to me that the optimum harmonic termination for linearity is actually the same as the 50R design value of the fundamental termination.

It is well known that playing with reflected harmonics can make a meaningful difference to efficiency, but I do wonder if there is another point where they can make a real difference to linearity? Anyone got a reference to the theory on this?

As to IMD test sources, If your rig is worse then the amp you are working on, fix the rig first, it will make more difference!

Note that high power single tone sources do not impose a linearity requirement, so a pair of class C 1kW stages can produce two tones that can then be combined in the usual way with 13dB of losses and still get your 100W of drive.

Regards, Dan.

Logged
Pages: Prev 1 [2] 3 4 5 6 7 ... 11 Next   Go Up
  Print  
 
Jump to:  

Powered by MySQL Powered by PHP Powered by SMF 1.1.11 | SMF © 2006-2009, Simple Machines LLC Valid XHTML 1.0! Valid CSS!