Wideband matching for an integrated radio receiver LNA

[Yansi]

Hello,

Working on a project, I came across an interesting problem I do not have enough experience to figure out myself.

I have an integrated radio chipset (Silabs EZRadio Si4355, but does not matter that much), that has a differential input to the LNA.

In a standard way, a 3 or 4 component impedance match is used both to match to 50 ohm and to balance the signal for the differential input. The matching procedure principles are mentioned in AN643 RX LNA Matching. Okay, understood.  But that kind of impedance match is only narrowband, working for few tens of MHz bandwidth at best, in the desired band of interest.

The question now is, is there a way, to make a wide band impedance matching, so that it will work across multiple bands, such as 315, 434 and 868 MHz?

I guess that the phase balancing for the differential LNA input in that case should be made by a suitable wide band balun transformer. But if I accept a bit of performance loss and use the LNA single-ended mode, how should one proceed to impedance match the LNA input?

The input impedance of the LNA looks for example like on the smith chart below (some hundred ohms in parallel with about 1pF of capacitance). What steps should I follow to match it it at least better than VSWR of 2? ... if that's even possible

Many thanks for any suggestions,
Y.

[T3sl4co1l]

Hrm, no axes on the plot?

Glancing at the appnote, they show an R || C equivalent, floating; presumably there is also some common or normal mode impedance as well, which we should like to know.  This is very much like an ADC's equivalent circuit, which you can also look at for ideas.

Treating it for now as a single port (just assuming that CM magically works out, so we need only concern ourselves with DM), we have a gain-bandwidth relationship:
BW = 1 / (2 pi (R || Rs) C)
Max gain: Rs = R
If we reduce Rs, we increase BW, but we must step down our signal source to a lower impedance, thus sacrificing gain -- proportionally.  (If we increase Rs, BW does not increase much -- it's hard limited by R -- while gain drops.)

We can peak the capacitance, using some combination of filter networks, extending bandwidth modestly.  The simplest case being a series inductor, having a value around X_L = R^2 C (the exact value depending on which type of filter we want; m-derived filters are very common here), and affecting a nice 30% or so extra bandwidth.  More advanced cases use tapped ("tee") coils, and can use several additional poles (the returns for which diminish rapidly, so that we aren't very much concerned about 3+ pole types), pushing 100, even 200% additional bandwidth.  Classic Tektronix scopes did this endlessly; a good reference is Starič and Margan, Wideband Amplifiers (excerpts of which the author has available).

Note we don't need to match Rs = R.  If Rs > R, we simply lose gain (as reflected power), and bandwidth stays at its minimum.  If Rs < R, bandwidth goes up, at the expense of gain.

If there's a source termination requirement (i.e., a ~constant 50 ohm or whatever input resistance), we need a termination resistor.  If Rs = R, we're done; if Rs < R, simply terminate with the balance, Rt = Rs R / (R - Rs).  (The exact location of the terminator can be anywhere along the peaking network, depending on what's being optimized: flatness of SWR, gain, phase, etc.  More generally, we put in whatever (potentially lossy) equalizing networks are needed to achieve the desired result.)

Finally, to go from something unbalanced like 50 ohm BNC or SMA or whatever, to the differential input: use a balun transformer, preferably a very small one given the high frequency range.  This probably won't extend very low, but an LF cutoff of some 10s of MHz would seem to be more than enough given the frequencies of interest, and should be no problem finding suitable parts.  If LF to DC operation is desired, very fast fully-differential op-amps could be used.

Tim

[Yansi]

Hello Tim,

unfortunately, RFSIM99 does not provide any axes masking in the graph, as far as I know. Can I display them somehow? Anyway, was just to get a sense what we are working with.

Yes, Silabs talks and uses the impedance value as purely differential.  Not sure about the common/normal mode impedance though, will ask them.  I think they may have just measured the normal mode impedance of one LNA input against ground (with the other grounded) and then multiplied by two to get the differential, assuming both inputs are the same?

I would like to avoid using balun transformer, thinking it may be cost prohibitive, although I know there are a lot of small ones available, that would suite the band of operation. Will investigate that possibility later.

Meanwhile playing in RFSIM99, trying to come up with something, I got an interesting tip from Silabs, the AN1180, which covers all sorts of matching topologies including wide enough band ones.  Most relevant seem sections 7 and 8 and in appendix 10 there is one solution applied to Si4x6x, which is almost what I need.  I may also be switching from Si4355 to Si436x radio, then I could indeed just copy it, tweak and be done with it. Certainly a win-win solution!

Anyway, I have never found any impedance charts or tables directly for the Si4x5x series of radios, so I am thinking the Si4x5x and Si4x6x LNAs are about the same.

Will report back later with the simulation.

[T3sl4co1l]

Heh, they even show multiband, neat.  Well, there you go -- if you want to spend the time optimizing all that, I mean.

Note that the "wideband" matching networks are still less than an octave -- fine for their 170MHz band example (Fig. 4.6), not suitable for your case.  That leaves the balun, or multiband.

If this is for production, it might be worthwhile, I don't know.  Or even getting a multilayer (or SAW?) balun made.

Heh, I wonder if you can get planar cores small enough to do a reasonable planar balun.

Tim

[Yansi]

I need only slightly above an octave, 315, through 434 to 868 MHz. "Just" ... I know this is no fun to design without enough experience.

So I jumped back to RFSIM99 and copied there their circuit from appendix 10, the figure 10.4 (Recomended 5-Element Multi-Band RX Matching Schematic for Si4x6x). (link to the AN1180 here again)

I do not think the result is correct, at least it does not match AN1180 result. The return loss around 434MHz is not as good, in fact, quite bad.  Please see below the sim schematic and the S11 result from 140 MHz to 960 MHz, graph is 5dB/div, zero is the center line.

Marker is at 434 MHz, the first dip is at 332 MHz, second dip peak at 856 MHz. So the result for my three bands of interest is:
-11 dB at 315 MHz (acceptable)
-5 dB at 434 MHz (meh..)
-16 dB at 868 MHz (good)
-10 dB at 915 MHz (still acceptable even here)

Question is, what to tweak and how to make it better at 434. It will be the most used band of them all.
Questions I faced Silabs with are:
1) Why didn't they consider the common/normal mode impedance (sure, considering just the diff mode impedance seems very wrong)
2) Where are impedance measurements of Si4x5x chipsets. All I can find is the damn table in AN643, which is for Si4x6x, which are much more capable radios, so the LNA may be different.

[T3sl4co1l]

I think you'll find the CM is all over the place on that one.

It's not obvious how much CMRR they actually need, YMMV...

Tim

[ConKbot]

https://www.digikey.com/en/products/detail/B0310J50100AHF/1173-1050-1-ND/3069248

Down at 300 MHz, it's straying pretty from ideal , but for $1 and 0805 size, it's not the worst.

Practically a steal compared to Marki's catalog https://www.markimicrowave.com/baluns/baluns-products.aspx