The Designer's Guide Community Forum https://designers-guide.org/forum/YaBB.pl Design >> RF Design >> Input Matching https://designers-guide.org/forum/YaBB.pl?num=1195659245 Message started by aaron_do on Nov 21st, 2007, 7:34am

Title: Input Matching Post by aaron_do on Nov 21st, 2007, 7:34am Hi all, just thought i'd get some opinions here. The question is do we really need to match the 50 ohm antenna to receiver input? The only reason I can think of is if you have an off-chip band-select filter, you need to match it to 50 ohm to ensure the proper frequency response. Any response is welcome. thanks, Aaron

Title: Re: Input Matching Post by pancho_hideboo on Nov 21st, 2007, 8:12am Radiation resistance of Lamda/2 dipole anttena is 73ohm. Holded dipoe is about 300ohm. Radiation resistance of monopole anttena and inverted F anttena is lesser than 50ohm. Slot anttena impedance is very high. To some extent, you can tune of anttena impedance, for example close to 50ohm by offset feeding or adjusting short post of inverted F. But you can't control impedance of antenna freely. If antenna resistance is small, gain of anttena is also small. I don't think 50ohm is necessary as matching reference impedance for anttena with not 50ohm impedance. So if you can connect LNA to anttena directly without feeder line, you should matching like following. http://www.designers-guide.org/Forum/YaBB.pl?num=1065493598

Title: Re: Input Matching Post by aaron_do on Nov 21st, 2007, 6:24pm Hi, thanks for the reply. So basically antennas don't really get 50 ohm impedance at all. So we should consider the antenna we are using first... Actually after some discussion with colleagues it seems the 50 ohm antenna impedance is more a concern of the transmitter side. I.E we need to have a well defined antenna impedance to get the maximum power transfer from the PA to the antenna. If we have a TR switch, however, it seems like we should be able to do away with 50 ohm matching on the receiver side as long as there is some kind of out-of-band filtering... thanks, Aaron

Title: Re: Input Matching Post by pancho_hideboo on Nov 21st, 2007, 9:45pm Antenna is reciprocal in TX and RX. So impedance is same for both TX and RX. But if you connect BPF of 50ohm, you have to match with 50ohm. Commercial product's RF BPF is almost 50ohm.

Title: Re: Input Matching Post by Mak on Nov 27th, 2007, 3:52am I believe 50ohm input impedance is very important for RX even there is no RF filter in between, because when you are connecting the chip to the antenna, you should design a transmission line (TL). If the input impedance of the RX and the (TL) are both 50ohm, the length of the TL can be choose freely, i.e., it matches the practical situation that the chip-design company won't know how their customer in selecting the length of the TL in connecting their antenna. Just my small opinion. :)

Title: Re: Input Matching Post by pancho_hideboo on Nov 27th, 2007, 4:51am 50ohm is just for easy design purpose. This is because actual RF instruments are designed as 50ohm impedance. Design and measurement are very easy if you set all interface impedances are 50ohm. So if IC customer doesn't have enough design ability, 50ohm is very preferable for such customer. But this is very classical or traditional method. Of course, loss of rf coaxial cabel is minimum around 50ohm characteristics impedance. But generally 50ohm is not best impedance for both LNA and PA if we consider gain, NF, distortion(ACPR), power efficiency, etc. Assume integrated design of IC and Antenna, such as RF module, we don't need 50ohm at all. Here 50ohm feeder line is not required. So Integrated anttena and  miliwave IC with on chip antenna don't need 50ohm. Even in relative low frequency, folded dipole and loop anttena is not feeded by 50ohm line when balanced feeding.

Title: Re: Input Matching Post by RFICInDaHouse on Dec 1st, 2007, 6:32pm To recapitulate, you need a well defined LNA input impedance to: 1_ Terminate a preceding SAW,BAW,... off-chip filter with 50 ohms since these filters are optimized to operate with 50ohms terminations. 2_ If no filter precedes the LNA and the antenna is on-chip, then the LNA should synthesize an impedance which is complex-conjugate of that of the antenna at the RF signal of interest for maximum power transfer. 3_ If you have any transmission line effect whether intentional or not, then you need to consider terminating the TL with its characteristic impedance to avoid reflection. That basically defines the input impedance of the LNA. 5_ It feels that I forgot something.... ;) -Ibrahim

Title: Re: Input Matching Post by aaron_do on Dec 1st, 2007, 7:58pm Hi Ibrahim, thanks for the input. However, I started this topic because I disagree with point 2. For RFIC, we are more concerned with voltage than power. I should be able to get much better gain and noise figure without matching the input to 50 ohm. For example if i have an ideal L in series with the gate capacitance of the input transistor. The gain of the input network would be wL/50 and the NF would ideally be 0 dB since an ideal LC adds no noise. I guess in future, we may see on chip band-select filtering which is not matched to 50 ohm to take advantage of this... cheers, Aaron

Title: Re: Input Matching Post by didac on Dec 2nd, 2007, 2:35am Hi, Aaron if you plot on a smith chart the reflection coefficients you will find that exists a trade-off between matching and NF, and the NF will never be 0 dB, moreover if you make a perfect noise adaptation probably most of the incoming signal will be reflected back to the antenna so I doubt you will gain anything reducing the noise to the minimum expression. Just my point of view.

Title: Re: Input Matching Post by aaron_do on Dec 2nd, 2007, 7:19am Hi didac, thanks. Your point of view is welcome. The point i'm trying to wrestle with is whether or not we care about signal reflection. We should be able to get very high voltage gain and low noise if we don't care about matching and we have some nice high Q inductors... cheers, Aaron

Title: Re: Input Matching Post by didac on Dec 2nd, 2007, 7:45am Hi, I think that it doesn't matter if we obtain a high gain if all the signal it's reflected back to the antenna, but I agree that 50 ohms it's not a useful reference impedance for a complete SOC.

Title: Re: Input Matching Post by RadioFreeq on Dec 8th, 2007, 10:09am Hi, if you have some specs for your chip, it would be a nice start point making a complex matching between the different components. if you don´t have specs, then start thinking about your goal.

Title: Re: Input Matching Post by didac on Feb 7th, 2008, 2:48am Hi, Yesterday I read a paper that puts this discussion in perspective: "From Oxymoron to Mainstream: The Evolution and Future of RF CMOS",Thomas H. Lee,RFIT2007-IEEE International Workshop on Radio-Frequency Integration Technology, Dec. 9-11, 2007, Singapore. It's the typical invited paper for a symposium that makes reflexions,history... but it has an interesting quotation that reminded me this thread(the quotation doesn't reference directly the author but google is a powerful tool): "CMOS devices take voltage, rather than power,as an input, and they have an almost purely  capacitive input, so they can't absorb any power. [G]etting … voltage gain using passive components that consume zero [power is] the most important factor in overcoming the noise of that device [6]." doing google search you end up in a interview in  Electronic News: http://www.edn.com/article/CA6339267.html. Prof.Lee in his paper mades a reasoning about why this idea is wrong, although he isn't completely neutral of course(he cited Shaeffer et al-where et al is Lee himself-), but ended up in the well know trade-off between Nfmin and input matching for power transfer(two-port noise theory). Just to put things in perspective,

Title: Re: Input Matching Post by aaron_do on Feb 7th, 2008, 8:01am Hi didac, thanks for the post. I read that section of the paper and it was quite enlightening. In fact, I don't think the Bsim3v3.2 (that's what i'm using) takes NQS gate noise into account, so the effect described wouldn't even show up. Not sure if i'm right about that. On the other hand, I think T.H.Lee is talking about optimizing NF, where the optimum NF could be as low as 1 dB. In low data rate (low-power) applications, we often can live with much higher NF. Furthermore, although the LNA has the biggest impact on receiver noise, if the rest of the system's noise is high, the LNA gain will have a much bigger impact on the overall NF than the LNA NF. Therefore, for sensor networks such as the one designed by Cook et al., I think passive voltage gain is a very good idea. I have one major issue with the Ben Cook paper, however, in that the analysis is a little deceptive. I guess that might be why T.H.Lee seemed to be attacking it in his paper. In the analysis, Cook says that the LC network has a 3 dB NF under matched conditions. However, if you read his thesis (which you can find online), you'll see that the LC network has an NF of less than 1 dB. This suggests that he severely traded off S11 in his design. Basically it is not matched, and that's why S11 was not quoted in the JSSC paper. He didn't say that the receiver had no input matching, and it made the design look a lot better. Overall though, I thought the paper was very good. cheers, Aaron P.S. Lee really seemed annoyed with the Cook paper...

Title: Re: Input Matching Post by didac on Feb 7th, 2008, 12:55pm Hi aaron, As I see you have readed in depth the paper from Cook and his thesis(could you post the link, I'm unable to find online-at the end of the day I lost my google powers-). First I have to say that it's indeed an interesting paper(I'm just in the second reading of it, and just now with pen and paper but it's quite innovative in the solutions they propose), I think that most of the "angry answer" of prof.Lee comes from the interview and not from the paper itself(the part of CMOS doesn't consume power only voltage) and as I said he is not entirely neutral(his work with Shaeffer and his own book for example). I had to read and analyze more in depth the paper(and the thesis if I have the change) but at first some things impresse me a little bit: -as you said lack of s11 measurement(at 50 ohms or whatever impedance they use in the antenna-which antenna?-) -at first glance I didn't find the data about sensitivity,data rate(250Kbps?),BER and intended distance for the receiver(I suppose these are carried out in thesis),now I'm making some educate assumptions to pick up numbers. I agree with you that if you have a big gain at the LNA we can consider(we are engineers) that all the noise is from LNA and if as you said you can survive with big NF it won't be the problem, the problem comes from how much power from the antenna is effectively amplified(enters in the receiver) so the overall SNR at the demodulator input allows the detection. PS:I'm not sure about NQS but I think that are included in BSIM4 models and not in BSIM3-this means when we switch to BSIM4 we cannot blame anymore the modelling guys for the NF excess... PS2:I will read again the paper and if I could the thesis-I like classical theory-as you can deduce from my previous posts- but I'm not closed to learn new things.

Title: Re: Input Matching Post by aaron_do on Feb 7th, 2008, 5:50pm Hi didac, here's the link for the thesis. http://www.eecs.berkeley.edu/Pubs/TechRpts/2007/EECS-2007-57.html As far as I know, there was no antenna in the design. Furthermore, there was no input band-select filter. However, if you read the section on mixer input impedance, he talks about how the input impedance forms a bandpass response which can be used to effectively cancel out of band interferrers. That was his justification for having no band-select filter. However, I don't think this effect was thoroughly researched. One more point to note is that antennas themselves have their own S11, and i believe that the actual antenna input impedance can be vastly different from 50 ohm. Since the Cook matching network relies on a 50 ohm source, I think it is a bit risky. I should also point out that his analyses do not follow his implementation. For example, i don't have the paper on hand, but i think he showed simulations of the mixer input impedance with RSW=100 ohm, where the input impedance was around 5 kohm. However, in his implementation, RSW=5 ohm which leads to much lower input impedance (less than 1 kohm). The mixer input impedance is the basis for the high voltage gain of the input matching network. As a matter of fact, I don't think i agree with his calculations of mixer input impedance either. Perhaps you can verify them. It seems very strange to me that the input resistance depends on the source resistance. Anyway I've read the paper several times now so I guess i could go on forever. Hope you recover your google powers :-) cheers, Aaron

Title: Re: Input Matching Post by didac on Feb 9th, 2008, 8:53am Hi aaron Thanks for the link, I downloaded the thesis yesterday evening and I readed it before closing the day. At this moment I didn't go through replicating the analysis of mixer input impedance. I spended some time playing with the matching network and duplicated their results-+/-0.5dB here or there,I used the QL they indicated and assumed ideal capacitors- for figure 25 page 70 and with the NF they indicate for passive network I can say for sure that they are bad matched(s11 around -5dB) at 2.4G, then I did a little experiment I consulted the ITRS for the predicted Quality factor of inductors and I see that it's around 50 for 2020-and it's marked as red so at this moment they don't have a clue how to manufacture it-(it's not referred to the same inductance or band that we are speaking here) but just to made the experiment more interesting I assumed a QL=100 and repeated the simulations, as expected NF was at all efects null,voltage gain boosted but mismatching is worsened(in fact what I think that limits the voltage gain here is QC-ideal capacitors- that depends on Rs w and the values of the capacitors themselves, so maybe I could optimize a little bit this thing). My first impression is that the overall SNR is degraded compared with a typical matched amplifier due to the loss of power delivered by the antenna, but I must run more numbers to be sure. About other parts of the thesis:I immediately discarded the low-IF architecture since they indicated only 20dB of IRR and I considered it too low and focused in the Direct Conversion architecture(that gives them a better NF after all and the power consumption is still good). About the antennas is true that their impedance could be different of 50 ohms(I don't even remember an antenna right now that gives a real part of 50 ohms), but it's very common if you are an antenna manufacturer to match them to 50 or 75(for TV) this way one company usually designs the chip,another the antenna and another the final product(a friend of mine works for a company that designs bluetooth headseats and they do this way). If you are in a company that manufactures the whole module you have more freedom on the choice of impedances(but better check with the PCB designer because e ohms of characteristic impedance could be very funny). I will try to scratch my head a little bit more this evening, PS:finally I recovered my google powers I think that was a collateral effect of no drinking coffee for several hours... In fact is a very interesting thesis to read and discuss, but one thing is clear:they traded off all(BW efficiency,hardware implementation-it reminds me a little bit to an old AM radio-) to make a good number in energy/per bit.

Title: Re: Input Matching Post by didac on Feb 9th, 2008, 12:38pm Hi again, Another thinking that I had is that the increasing Q of the overall network will lower the BW so here we have another trade-off(I think that it helds the Bode-Fano criteria of the achievable matching over some bandwidth) about how much ISM band is needed for the system to work.

Title: Re: Input Matching Post by aaron_do on Feb 10th, 2008, 1:23am Hi didac. your conclusion seems correct. Energy/Bit is the focus... BTW when i say that the antenna impedance is different from 50 ohms, I mean that the antenna is designed to be 50 ohms, but in reality, the S11 of the antenna is better than around -10 dB.  I don't know this for sure because I don't know much about antenna design, but I remember looking at the data sheet for an antenna used for ZigBee, and the S11 was not great. This suggests that LC network used in this paper could have huge variations in the center frequency, bandwidth and gain. A normal active LNA doesn't have this problem because the matching network is usually much lower Q, and most of the gain comes from the active devices (i suspect in most cases this is true). Now i'm going to read up on the Bode-Fan criteria :D cheers, Aaron

Title: Re: Input Matching Post by didac on Feb 11th, 2008, 10:04am Hi aaron, You are right antennas are nominally 50 ohms maybe at one frequency across the band so this spread of values should be taken into account in this matching network...One think that comes to my mind is to study the feasibility(following the AM radio anology) of using one of the capacitors like a varactor, to tune in the desired channel... this way I suppose we can have good matching in the channel of interest without trying to match the whole band, what I don't know right now is the effect on gain or NF...also a control loop will be needed to tune it correctly so it's hardware overhead. Lets see if I find a couple of hours to work more in this. keep in touch

Title: Re: Input Matching Post by RFICDUDE on Feb 12th, 2008, 8:30pm I think Lee’s comment is directly entirely to the comment in the interview article. I don’t think you can infer anything else from what he said in the paper. Lee is making the point that here is a federally funded project on transceivers from a major research institution and we have comments obviously dismissing the well established fundamentals of MOSFET gate noise and impedance at high frequencies. Coming from a Ph.D. on transceiver design, it is an open mouth insert foot type of comment. Cook says the gate is essentially a capacitor and does not require power matching, but what is the backplate of the cap? The channel! Cook’s comment would be correct if the backplate of the gate is simply connected to ground, but instead it is coupled (hopelessly) to the channel. Because of the channel coupling directly to the input, the output gain doesn’t diminish this noise source. In addition, the input impedance is not just a capacitance. There is a substantial real part at high frequencies and it is related to the channel conductance. At low frequencies the passive gain argument holds, but higher frequencies it breaks down. Cook doesn’t bother to analyze the problem to tell us where this approximation breaks. Having said all this, Cook does make a point and raises some food for thought, but more rigorous analysis would really enlighten us as to what is really being traded off from his approach.

Title: Re: Input Matching Post by aaron_do on Feb 13th, 2008, 4:41am Hi, I agree entirely with what you say. Cook seemed to leave out some of the drawbacks of his work. The only further comment I have to make is that I felt that the wording by Lee was a little unkind... "an understanding... is still not as widespread as it should be" "the following recent quote packages several tragic misapprehensions in two sentences" "In light of so recent a disappointing example" it seems a bit unfair to assume that Cook did not understand the problem. If you are not trying to squeeze the last 0.1 dB out of the LNA NF, then I think that Cook's arguement has merit. Obviously there is a limit to how much passive voltage gain you can get, but I think that Cook understood that most people don't even nearly approach that limit.

Title: Re: Input Matching Post by vivkr on Feb 14th, 2008, 11:19pm Hi Aaron, Just an aside. I believe that there is a simple way of extending the capabilities of a BSIM3v3 model to include NQS effects by just adding a resistance in series with the gate. I don't recall all the details, but you can find it here: "A-Simple Subcircuit Extension of the BSIM3v3 Model for CMOS RF Design" Suet Fong Tin,.... IEEE Journal of solit-state circuits, Vol. 35, No. 4, April 2000 At the end, your fab still has to do this modelling for you of course. Vivek

Title: Re: Input Matching Post by RFICDUDE on Feb 17th, 2008, 1:35pm Yes, I agree that there are many applications where achieving NFmin is not required. And yes, Lee was a bit harsh. Cook could have easily saved himself some criticism by doing a better job of carefully qualifying under what conditions his approximation is valid and outlining other applications where it is applicable. Instead, his comments come off as general, as if he did not full understand the potential consequences of generalizing his approach. I think a very good question is … how good must a wireless transceiver system be to work reliably? Cook offers a novel perspective, but not a definitive answer to that question. And honestly, I don’t have an answer either. I am a big fan of people stirring the pot for the sake of getting others to think differently about common problems. Cook certainly stirred things up a bit. If you are interested, there is a very good and detailed paper on CMOS LNA design (matching for NFmin and S11) Trung-Kien Nguyen; Chung-Hwan Kim; Gook-Ju Ihm; Moon-Su Yang; Sang-Gug Lee, "CMOS low-noise amplifier design optimization techniques," Microwave Theory and Techniques, IEEE Transactions on , vol.52, no.5, pp. 1433-1442, May 2004.

Title: Re: Input Matching Post by didac on Feb 18th, 2008, 9:35am Hi, First of all this a pretty interesting discussion. I haven't worked with this issue for a while but I think that if we can forget about the doomsday scenario of induced gate noise I see a big drawback from the point of view of matching, I think that Cook in his thesis is trading off directly SNR due to mismatching so I don't see any advantatge right now-as I said I must dig deeper in this issue-, but assuming that we can find passive gain structures with good matching I'm pretty sure they will be very narrowband so then we will face a architectural problem to cover the whole band. PS:vivek I think that the BSIM3 extension that you reference it's already in use in RF transistors models, they add the gate resistance and the substrate network to model more accurately the high frequency behaviour, this is the reason that RF transistors are not very friendly in the sizing procedure-fixed width,length and variable number of fingers-.From what I know the gate resistance models correctly gate noise due to poly material but doesn't do anything with induced gate noise.

The Designer's Guide Community Forum » Powered by YaBB 2.2.2! YaBB © 2000-2008. All Rights Reserved.