Linear RFPA: push-pull vs. single-ended - RF Cafe Forums
Because of the high maintenance needed to monitor and filter spammers from the RF Cafe Forums, I decided that it would be best to just archive the pages to make all the good information posted in the past available for review. It is unfortunate that the scumbags of the world ruin an otherwise useful venue for people wanting to exchanged useful ideas and views. It seems that the more formal social media like Facebook pretty much dominate this kind of venue anymore anyway, so if you would like to post something on RF Cafe's Facebook page, please do.
Below are all of the forum threads, including all the responses to the original posts.
Post subject: Linear RFPA: push-pull vs. single-ended
Unread postPosted: Thu Jun 03, 2004 3:05 am
Does anyone have actual experience in push-pull PA design? What is the expected improvement in linearity compare to AB class single ended?
Unread postPosted: Thu Jun 03, 2004 3:24 pm
Your question is a bit vague.
When you design PA, you have 2 requirements which are contradicted, linearity vs. efficiency. One comes on the account of the other. Most often there is no PA as a stand alone unit, unless your linearity requirements are very mild. Your PA will probably be integrated with some kind of linearizer, of some topology depends on the bandwidth you have to deliver.
Most push-pull transistors are intended for high power levels (Mostly above 100W PEP) like Motorolas' MRF21180. For low and medium power levels most transistors available are single-ended. The approach for designing a PA is to get the best efficency by an operating point towards saturation, and improving the linearity to meet the requirements.
You will need to provide more details about your application: Power levels, frequency of operation etc; please do so and I will be keen to help you.
Unread postPosted: Fri Jun 04, 2004 2:21 am
Thank you for your post Itay, yet it does not answer my question, which is very straightforward, is a push-pull amplifier more linear compare to single-ended one? Provided, that both are AB class driven with equal conducting angles. Everything you have mentioned including Gemini packaging and necessity of linearization around amplifier is of secondary concern, my question is about amplifier itself.
There are tones of publications available about push-pull amplifiers, but I found only one (http://www.fcsi.fujitsu.com/products/MWpdf/an014.pdf) scrutinizing the issue with the final conclusion: No, the push-pull is no more linear than single-ended. This puzzles me.
Unread postPosted: Mon Jun 07, 2004 9:24 am
If by push/pull you mean a balanced amplifier, then yes, a balanced amplifier can be more linear. A balanced amplifier has much, much less 2nd harmonic distortion than a single ended class AB. So, if your RF band of interest is wider than an octave, there is a distinct advantage to using a balanced amplifier. But, if you're only concerned with a narrow band of frequencies, and the 2nd harmonic will be knocked down by filtering, then there's really no difference in linearity.
One other nice thing about a balanced amplifier is that if you lose one transistor, you still have output. This is good in some high reliability applications where diminished output is better than no output.
Hopefully this was what you're looking for?
Unread postPosted: Mon Jun 07, 2004 10:50 am
Thank you for your input, although I do not understand which topology you are describing. Could you clarify please what exactly balanced amplifier means within your definition?
Unread postPosted: Mon Jun 07, 2004 1:01 pm
Balanced amplifier means 2 different amplifiers connected in the input and output with hybrid quadrature couplers. The advantage of this amplifier is that it gets half the input power level (-3dB) at each input, thus allowing it to work in a greater backoff and become more linear. At the output, the powers are being combined and you get twice the power as you would get with a single amplifier, this gives you a higher P1dB (theoreticaly 3dB higher, practically less due to the hybrid coupler insertion loss). The disadvantage of this topology is of course the power dissipation, which is twice than single ended amplifier.
To your question, absolutely Push-Pull amplifier is more linear than single ended one because you utilize the entire cycle of the input signal, namely the amplifier delivers the entire cycle without cutting it and by this you produce a linear amplifier with less distortions. You use class A-B or class B transistor amplifiers and combines them in Push-pull so at any given time one of them conducts, this can be done in a balanced amplifier topology or in the push-pull topology.
Unread postPosted: Mon Jun 07, 2004 5:28 pm
My understanding of topology for balanced amplifier coincides with yours, but I am afraid not with Darrin’s. Yet since you are on a position to comment Darrin’s posting, would you explain please the issue of second harmonic suppression in balanced amplifier? I am accustomed to idea that this feature does not belong to balanced amplifier inherently.
Furthermore, I quote your statement “You use class A-B or class B transistor amplifiers and combines them in Push-pull so at any given time one of them conducts, this can be done in a balanced amplifier topology or in the push-pull topology.” How this can possibly be done in a BALANCED topology?
I’d vote with both my hands for the idea of enhanced linearity of RF push-pull amplifier due to conducting angle expanded to 360deg. This is well-known fact at audio field. The problem is that for some reason this seems not to work at RF, presumed that linearity concerns spectrum re-growth close to carrier rather than second harmonic suppression (http://www.fcsi.fujitsu.com/products/MWpdf/an014.pdf). Not even a single source I found mentions that RF AB class push-pull amplifier is more linear than single ended counterpart. Can anyone explain why, or prove opposite?
Unread postPosted: Thu Jul 08, 2004 5:41 pm
The balanced amplifier is different from differential amplifer, the later can be constructed by 180 degree balun or hybrid.
In most cases the 2fo can not be eliminated as Darin said since they are out of band. However 2fo can used rather than cancelled,and it is still very interesting to have push-pull for high power delivery and ease to match(higher impedance). Another insteresting developement is for Differential LNA design above 2 GHz, since this structure absorbs some noise sources which exist in common mode and not need GOOD rf ground.
If talking same power level, it is clear there is 3 dB power back off, good? If not, at least the linearity is not worse then single-ended, OK?
The differential amplifier is not limited by the concept of push-pull anymore. It can be wokring at calss A, AB, C, E...
Old Tele man
Post subject: re: RFPA linearity
Unread postPosted: Sun Oct 17, 2004 1:28 am
Joined: Sun Oct 17, 2004 1:07 am
Location: Tucson, AZ
a very good book on this subject is:
Steve C. Cripps, Ph.D.
RF Power Amplifiers for Wireless Communications, 1999;
Artech House (microwave library), Inc.; Norwood, MA;
Devil sez: "...Yes, but it's a DRY heat!"