Test & measurement
Figure 2: HS amplifier HD2 test setup
possible scenarios by reversing their connections on one or both sides of the setup. The results are shown in Figure 3.
phase imbalance found in Figures 6 and 7, from Vendor 1B and Vendor 2B, respectively. Notice that with better imbalance performance, HD2 distortion variance is reduced accordingly, as seen in Figure 7.
same traces shown in Figure 7, testing only Vendor 2B’s balun performance as the attenuators are swapped between the inputs and outputs. This generates another set of four traces, shown as dashed lines in Figure 9. The result is that we are back where we started, as this shows up as more variation in the test measurement. This further underscores that small amounts of mismatch in either side of a differential signal pair matters at high frequencies. Keep in mind to document your test conditions in detail.
Figure 3: Testing HD2 performance with Vendor 1A balun using different balun orientations
The amount of variance in HD2 distortion
curves revealed in Figure 3 proves a further look at the balun’s performance is needed, specifically for phase and magnitude imbalance. The following two figures show the phase and magnitude imbalance of several baluns from various manufacturers. A network analyser was used for the imbalance test measurements.
Figure 6: Retesting HD2 performance with Vendor 1B balun using different balun orientations
Figure 9: Testing HD2 performance with Vendor 2B balun only using different balun orientations and attenuation pad swaps
Balancing it all Out
In summary, all things matter when developing fully differential signal chains in the GHz regions; that is, attenuators pads, baluns, cables, traces on a PC board, etc. We have proven this mathematically and in the lab using a high speed differential amplifier as our test bed. So, before we start to blame the part or the vendor, please take extra special care during the PCB layout and lab testing. Lastly, you might be asking yourself, so how
Figure 7: Retesting HD2 performance with Vendor 2B balun using different balun orientations
To further illustrate how phase imbalances Figure 4: Phase imbalance of various baluns
directly affects the performance of even order distortion, Figure 8 shows HD3 distortion over the same conditions as the previous HD2 figure. Notice how all four traces are roughly the same, as expected. Therefore, as proven in the mathematical derivation example shown previously, HD3 distortion is not as sensitive to imbalances in the signal chain.
much phase imbalance can I tolerate? When I pick up a balun, for example, and it says x number of degrees of phase imbalance at x GHz, what does that mean in terms of degradation to my part or system? Can I expect a certain amount of loss or degradation in dB’s of linearity performance? This is a difficult question to answer. In the
Figure 5: Magnitude imbalance of various baluns The red traces in Figures 4 and 5 correspond to
the actual balun used to acquire HD2 distortion data in Figure 3. This particular balun, from Vendor 1A, had one of the highest bandwidths and good pass-band flatness, but worse phase imbalance as compared to the other baluns over the same 10GHz frequency test band. The next two figures represent a retest of HD2 distortion using the best balun that had the lowest
18
Figure 8: Testing HD3 performance with Vendor 2B balun using different balun orientations
Until this point it should be assumed that the
input and output connected attenuator pads, as shown connected in Figure 2, are stationary and did not change during the balun orientation measurements. The next figure represents the
ideal world, if everything in your signal chain were matched perfectly, there would be no even order distortion to contend with. Second, it would be nice to have a rule of thumb or equation that says for every x° of phase imbalance, one should expect x dBs of loss in linearity (HD2 degradation). However, this just cannot be. Why? Because every component, be it active or passive and differential in nature, has some sort of inherent phase mismatch. There is just no way to perfectly balance an IC design internally, or cut cables with absolute perfectly matched length. So, no matter how small these mismatches are, they become more pronounced as higher and higher frequencies are utilised in a system. Let us wrap this up by saying Analog Devices
will do its job as best it can by keeping those IC layout mismatches small where fully differential inputs and outputs are used. The company hopes you do the same when testing products in the lab.
Analog Devices
www.analog.com September 2018 Instrumentation Monthly
Page 1 |
Page 2 |
Page 3 |
Page 4 |
Page 5 |
Page 6 |
Page 7 |
Page 8 |
Page 9 |
Page 10 |
Page 11 |
Page 12 |
Page 13 |
Page 14 |
Page 15 |
Page 16 |
Page 17 |
Page 18 |
Page 19 |
Page 20 |
Page 21 |
Page 22 |
Page 23 |
Page 24 |
Page 25 |
Page 26 |
Page 27 |
Page 28 |
Page 29 |
Page 30 |
Page 31 |
Page 32 |
Page 33 |
Page 34 |
Page 35 |
Page 36 |
Page 37 |
Page 38 |
Page 39 |
Page 40 |
Page 41 |
Page 42 |
Page 43 |
Page 44 |
Page 45 |
Page 46 |
Page 47 |
Page 48 |
Page 49 |
Page 50 |
Page 51 |
Page 52 |
Page 53 |
Page 54 |
Page 55 |
Page 56 |
Page 57 |
Page 58 |
Page 59 |
Page 60 |
Page 61 |
Page 62 |
Page 63 |
Page 64 |
Page 65 |
Page 66 |
Page 67 |
Page 68 |
Page 69 |
Page 70 |
Page 71 |
Page 72 |
Page 73 |
Page 74 |
Page 75 |
Page 76 |
Page 77 |
Page 78 |
Page 79 |
Page 80
