Probe assembly with controlled impedance spring pin or resistor tip spring pin contacts

[0001] The subject matter of the present Application is related to a manner of using a spring pin contact, having a resistor in its tip for the probing of electrical signals. That spring pin resistor tip contact is shown and described in the pending US Patent Application entitled SIGNAL PROBE AND PROBE ASSEMBLY bearing Ser. NO. 11/141,500, which was filed on 31 May 2005 by Joseph Groshong, Brock J. LaMeres and Brent A. Holcombe, and which is assigned to Agilent Technologies, Inc. For the sake of completeness and for brevity, the Application ‘SIGNAL PROBE AND PROBE ASSEMBLY’ is hereby incorporated herein by reference.

[0002] The probing of high frequency signals, whether for an Oscilloscope or a Logic Analyzer, presents certain technical problems that increase in difficulty as the frequency of the signal gets higher. Quite aside from attaining the requisite bandwidth in the ‘Scope or Logic Analyzer, careful attention has to be paid to the way the signals of interest (the ‘target signals’) are acquired through suitable probes. At the highest frequencies, say, above ten GHz for ‘Scopes and above about a one GHz clock rate for Logic Analyzers, probe related issues that could be often be ignored at lower frequencies emerge as factors that limit overall system performance.

[0003] The incorporated ‘SIGNAL PROBE AND PROBE ASSEMBLY’ is a way to deal with at least one of these issues: getting a necessary resistor in the probe's tip as close as possible to the target signal. Probes for different kinds of test equipment (e.g., ‘Scopes, Logic Analyzers, Spectrum Analyzers, etc.) have such tip resistors for different reasons, such as damping, minimizing loading, attenuation, or, perhaps impedance matching. What follows the probe tip is generally a length of coaxial cable having a particular characteristic impedance (Z₀) that should not be thought of as merely a shielded conductor, but as a transmission line. Generally there is an active circuit of some sort at the other end of the coax, such as an amplifier with a particular frequency response, or, a threshold detector. The active circuit's output is then sent by a robust buffer amplifier over suitable transmission lines to the particular measurement circuitry in use. Probe architectures vary, but in each case it remains desirable to get the tip resistor as close as possible to the target signal.

[0004] There are two general reasons for this. First, any extra distance represents length along a conductor that will exhibit both a parasitic series inductance and a parasitic shunt capacitance to any nearby AC ground. These parasitic reactances represent target signal loading over an above what the probe's coupling network proper is intended to present. These parasitic reactances can be the cause of reflections and bandwidth limitations. Second, even if those parasitic reactances do not noticeably cause reflections and reduced bandwidth, they still amount to an unwanted intervening impedance that means that the signal presented to the entrance of the probe tip's resistor is not the identically same as the one contacted by the very end of the probe's tip. This is an issue of signal fidelity within the probe, quite apart from, and even in the absence of, significant loading.

[0005] The spring pin disclosed in ‘SIGNAL PROBE AND PROBE ASSEMBLY’ reduces the length of the probe tip by placing the probe tip resistor very close to the mechanical location of electrical contact with the target signal. It is almost as if one end of that resistor were the actual mechanical point of contact. We can't really achieve exactly that, because the resistor is an SMT part with rather flat surfaces for leads, while the contacting surface for the probe needs be a durable needle point or some non-slip variation on that, such as a ‘crown point’ having a plurality of points around a central depression. The next best thing is for the target side of the tip resistor to carry the durable point, while the other end of the resistor is mounted in a socket that in turn is carried by the larger contact structure and is mechanically biased against axial movement by a spring. ‘SIGNAL PROBE AND PROBE ASSEMBLY’ goes on to show an inline array of such spring pins carried by a PCB (Printed Circuit Board) that couples each of those spring pins to an associated coaxial cable that is a transmission line of characteristic impedance Z₀ and that leads to some type of test equipment. We shall hereinafter call such a spring pin contact a ‘resistor tip spring pin’ contact.

[0006] As significant an improvement as ‘SIGNAL PROBE AND PROBE ASSEMBLY’ represents, there is still an aspect of its operation that can be improved. In particular, and as will become apparent when the Drawings are considered, the region of the spring pin: (a) which telescopes in and out of the housing containing the spring; and (b) which bears the socket that receives the tip resistor; (c) represents a variable length (depending upon how far down the entire probe assembly is moved to ensure good contact for all the probe tips carried by the assembly) that also has an arbitrary characteristic impedance (Z_x) that in general is not equal to that (Z₀) of the interconnecting cables. This mischief arises owing to the placement of the aforementioned housing right at the edge of the PCB (in the belief that this would maximize the variation in the amounts that different spring pins may move when contact is established). That is, the pin driven by the spring and carrying the tip resistor is sticking out into space by some unregulated amount, amounting to a transmission line of arbitrary Z_x. Once the pin enters the housing, however, the characteristic impedance is controlled (Z₀), and remains so from there on as the signal proceeds further into the probe assembly. At high frequencies, the transition from the short length of uncontrolled Z_x into the controlled Z₀ is a discontinuity that manifests itself as reflections in the time domain and as frequency response anomalies in the frequency domain. It would be advantageous to adjust Z_x to be the same as Z₀. What to do?

[0007] A solution to the problem of uncontrolled characteristic impedance along a spring biased pin, which may have a resistor in its tip, for a probe assembly is to provide a stepped shelf of ground plane that extends outward along the pin and toward a work circuit assembly having target signals to be probed. The length of outward extension is chosen such that even when there is only (or at least) an expected minimum amount of compression of the spring while producing and maintaining contact, the entire exposed portion of the pin is over the shelf, whose depth of step has been selected to produce Z₀ for that exposed portion of the pin. A pin that is compressed too far will bring the circuit side of a tip resistor close to the ground plane on the shelf, and present an unwanted shunt capacitance. However this can be forestalled by proper test fixtures that are correctly adjusted. A completely uncompressed pin (i.e., one that is unused) may have a short length of Z_x that extends beyond the shelf, but this case is not of interest, as that pin is evidently not in contact with a signal.

[0014] Refer now to FIG. 1, wherein is shown a simplified representation 1 of a resistor tip spring pin probe assembly 1. Flat multiple coaxial cable assemblies 5 and 6 are connected to some item of test equipment (not shown) that might be a Logic Analyzer. Cable assembly 5 is attached to respective multiple traces on one (say, the top) surface of a printed circuit board 2; the other cable assembly is similarly attached to the other side of printed circuit board 2. As shown in the inset, a top row 3 of resistor tip spring pin contacts (7) is soldered to a far end of the traces on the top side of board 2, while a bottom row 4 of resistor tip spring pin contacts is similarly soldered to the traces on the underside of board 2. The traces to which the center conductors of the coaxial cables 5 and 6 are connected form transmission lines with ground planes within the printed circuit board 2. The outer shields of the coaxial cables 5 and 6 are, of course, connected to those ground planes. Coaxial cables 5 and 6 may be transmission lines.

[0015] Referring now to FIG. 2, it will be seen that a resistor tip spring pin contact 7 may have a crown tip 9 for ease in penetrating oxides or other coatings on surfaces to be probed. A single sharp pointed tip might be used as well; the crown tip is nearly as sharp and has the advantage of several points of electrical contact and a central concavity to help it stay put if applied against something not altogether flat.

[0016] Continuing, the crown tip 9 is seen to be the shaped end of an outer tip 8 that contains a socket for receiving one end of a resistor 13. Resistor 13 may be a common SMT (Surface Mount) resistor of the “0201” variety. The other end of resistor 13 is carried by a socket on the end of a plunger 10 which telescopes in and out of a housing 12 containing a spring 11, which may be a compression spring. Thus it is, as set out in full detail in the incorporated ‘SIGNAL PROBE AND PROBE ASSEMBLY,’ that the resistor 13 is as close to the point of probing as practical.

[0017] The plunger 10 is retained within housing 12, while the spring 11 provides the contact force that urges the crown tip 9 against a pad or trace (such as 16 in FIG. 3) when the resistor tip spring pin contact (7) is in use. A suitable insulating covering 14 (which may be of a suitable plastic) protects and reinforces the resistor 13 while bridging outer tip 8 and plunger 10.

[0018] Now refer to FIG. 3, and note that it shows in greater detail how the resistor tip spring pin contacts 7 are mounted to and carried by the printed circuit board 2. It will be observed that the end of the housing 12 nearest the target circuit board 15 and its pads 16 is even with the bottom edge 17 of the board 2. Target circuit board 15 bears pads or traces 16 that carry pluralities of target signals to be probed. This arrangement has certain consequences, which we now turn to in connection with FIGS. 4A–4C.

[0019] The side views in FIGS. 4A–4C show that the circuit board 2 is actually a multi-layer affair that has an associated ground plane (18, 19) underneath the conductors (20,21) to which the various resistor tip spring pin contacts 7 are soldered. The reason for this is to create strip transmission lines (22, 23) having a selected characteristic impedance Z₀ (e.g., 50 Ω) that matches that of the individual coaxial cables in the multiple cable assemblies 5 and 6. That is all well and good, and the mischief arises when the effects of the plunger's (10) extending outward by various amounts (24, 25, 26) are considered, especially as the signals of interest being probed increase in frequency. As can also be seen in the figures, that arrangement has the end 17 of printed circuit board 2 being a single undifferentiated surface. Thus, for each of the cases of: No contact (FIG. 4A, 24) between crown tip 9 and pad or trace 16 (refer again to FIG. 3); Contact producing fifty percent (or any other intermediate percentage of) spring compression (FIG. 4B, 25), and; (C) Contact producing maximum compression (FIG. 4C, 26), the characteristic impedance Z_x associated with any of lengths 24, 25 and 26 are altogether uncontrolled. That means that the transition from any of plunger portions 24, 25 and 26 is a discontinuity when formed with the controlled characteristic impedance Z₀ of the housing/signal trace combination 12/20 or 12/21 and the adjacent respective ground plane 18 or 19. That discontinuity is bad news for reasons that are well understood in and of themselves.

[0020] The remedy for that unfortunate situation is shown in FIGS. 5A–5C. In FIG. 5A we see a partial side view of an improved probe assembly 37 that has two rows of resistor tip spring pins 7. One row is on a first (as seen in the figure, a top) side of a multilayer printed circuit assembly 36, while the other row is on the second (bottom) side. FIG. 5A corresponds to FIG. 4A in that each depicts a case where the plunger portion 10 of the resistor tip spring pins 7 is fully extended (and thus probably not in contact with any target signals). If they are in such contact, then instances of Z_x will be present, and further compression (approach of the probe assembly 37 toward the target circuits upon a work circuit assembly) is indicated. This probably amounts to adjustment of a test fixture that holds things in position relative to one another.

[0021] As in FIG. 4A, Z_x will be present because most or all of the entire length 27 of the plungers 10 extend beyond the ends 64 and 65 of the housings 12, causing portions 28 to extend beyond the edge 62 of ground planes 31 and 33. Those ground planes form transmission lines 34 and 35 with the signal traces 30 and 32, to which are soldered the respective top or bottom instance of the resistor tip spring pins 7. Dielectric layer 69 separates signal trace 30 from ground plane 31, while dielectric layer 70 separates signal trace 32 from ground plane 33. The two ground planes 31 and 33 are separated by dielectric layer 71 that shares with them the end or edge 62. Dielectric layers 96, 70 and 71 may be of FR4, or other suitable substrate material. It will be appreciated that edge 62 cannot be moved (leftward in the drawing) all the way toward the work circuit assembly (not shown) having the target signals, as to do so would likely create mechanical interference with the work circuit assembly and prevent the movement of the plungers 10 into the housings 12. That, of course, would force at least some Z_x to remain, which is most undesirable. Thus it is that the start of the bulge 72 for the covering 73 of the resistor 13 is at a location that will be proximate (i.e., above or below, according to which is the row of contacts) edge 62 while additional compression of plungers 10 into housings 12 is still possible. In this way a variety of different compressions will eliminate Z_x and allow just a remaining Z₀ of some length.

[0022] Things will be arranged, then, so that when all the resistor tip spring pins 7 are in contact with their respective target signals, all of the plungers 10 will be pushed into their respective housings 12 to respective amounts so that no Z_x is present for any plunger. And as will be explained shortly, the characteristic impedance Z₀ of the length 29 is the same Z₀ as of the transmission lines 34 and 35, (which may be strip lines), which is of course selected to be the same Z₀ for the characteristic impedance of the interconnecting coaxial cable assemblies 5 and 6.

[0023] We turn now to FIG. 5B, which is a partial side view of an arrangement 38 similar to that 37 of FIG. 5A. One difference is that length 39 is equal to the distance 29 by which the ground plane 31 extends beyond edge 61 in a direction toward the work circuit assembly and the target signals (which for brevity, neither of which is shown). The pesky Z_x has vanished, leaving only Z₀. The plunger 10 will preferably not yet be at its maximum penetration into the housing 12. The other difference is that FIG. 5B may also be taken as representative of a situation 38 where there is but one row of pins 7 disposed upon only one side of the printed circuit assembly 36. (As if in FIG. 1 row 4 of contacts 7 and its associated cable assembly 6 were never present.)

[0024] FIG. 5C shows a partial side view very much like the one of FIG. 5B, save that now length 40 is less than distance 29. As before, the original Z_x is gone, leaving a region of length 40 of Z₀, and some new region 74 of length that is approximately length 29 diminished by length 40 (for brevity we are neglecting fringe effects) and whose characteristic impedance is some unknown Z_y. We can do two things to mitigate the possibility that Z_y will re-introduce the mischief of Z_x that we have worked to hard to eliminate. First, it might be that Z_y is sufficiently close to Z₀ (say, within ten percent) that we needn't be concerned, or, that if it is not then we can make it so. A particular dielectric constant for the covering of the resistor 13 might help in this. Second, we can adjust the fixtures in use (not shown) so that the situation of FIG. 5C does not, or seldom, occurs, and that if it does, then the length 74 is sufficiently small that Z_y may simply be ignored. If length 74 is sufficiently small then the frequencies needed to ‘see’ it as a significant quantity may well be higher than either or both of the upper frequencies produced by the SUT or the test equipment in use.

[0025] The fabrication of strip and coplanar transmission lines on printed substrates is a conventional technique that is generally well understood. It involves the relation
Z₀=√{square root over ((L/C))}
where amounts of distributed inductance and capacitance per unit of length. Properties that go into selecting the dimensions of such transmission lines are the size of the conductors, their distance apart and the dielectric constant that separates them. It is a routine matter to select the parameters that determine the Z₀ of the transmission lines 34 and 35, all the way up to edges 61 and 63, respectively. The presence of pins 7 soldered to signal traces 30 or 32 has only a minor effect on the characteristic impedance as a signal leaves the plunger 10 and enters the housing/signal trace combination. Of greater interest is how region 29 can be adjusted to have the same Z₀ as do the transmission lines 34 and 35.

[0026] What we need to do is keep the same ratio of L and C for the region 29 as for the transmission lines 34 and 35. We shall follow the usual convention of understanding the distributed reactive properties of a transmission line as is they were a very large number of suitable small but adjoining lumped constants. In the following discussion, statements about increasing or decreasing the values of C and L should be understood as values per unit of length. C (per unit length) is influenced by both distance and dielectric constant. C is inversely proportional to distance, which goes up some, tending to decrease C in region 29. C is proportional to the dielectric constant separating the elements of the capacitance. The air dielectric of region 29 will be one, while that of FR4, for example, is usually taken to be four. Both of these relationships will tend to decrease C (per unit length) in region 29, since the separation distance increases and the dielectric constant will likely drop, or at best, stay the same. C is also affected by the effective area of the plates, which is to say, by width per unit length; it is proportional to that effective area (width). L (in this case, self-inductance per unit length) will increase as the diameter decreases (increases proportionally to the reciprocal of the diameter). What we have to work with, then, is an increase or decrease in the diameter of the plunger 10 (varies L), the change in width (effective plate area to the ground plane 31 or 33) as the diameter of the plunger 10 changes compared to the (previous) width of the signal traces 30 and 32, and the distance, or height, of plunger 10 above or away from the extended shelf 29 of the ground planes 31 and 33 (varies C). It will be appreciated that the length of a transmission does not affect its value for Z₀.

[0027] For example, a height 76 of 0.0075″ for stepped region 29 with a thickness of 0.0020″ of FR4 for intervening dielectrics 69 and 70 will produce a Z₀=50 Ω for a signal trace width of 0.003″ and a plunger diameter of 0.0130″. A height 76 of 0.0020″ for stepped region 29 with a thickness of 0.0085″ of FR4 for intervening dielectrics 69 and 70 will produce a Z₀=75 Ω for a signal trace width of 0.0165″ and a plunger diameter of 0.0130″.

[0028] Finally, the reader is referred to FIGS. 6A and 6B. These views 41 and 59 depict probe assemblies having spring pins 42 without resistors in their tips. View 41 of FIG. 6A otherwise corresponds to FIG. 5A, while view 59 of FIG. 6B otherwise corresponds to FIG. 5B, especially in the notion that there may be one (FIG. 6B, 59) or two (FIG. 6A, 41) rows of spring pins 42.

[0029] In these figures the spring pins 42 have plungers 43 the carry crown tips 60, or perhaps some other form of pointed tip. Housings 44 carry the plungers in the same way as before (see FIG. 2), and are soldered to signal traces 45 and 56, which in turn cooperate with intervening dielectric layers 46 and 55 to produce from ground planes 49 and 54 transmission lines 50 and 75, respectively. These transmission lines have Z₀ as their characteristic impedance, which will match that of the interconnecting flexible coaxial cables (e.g., 5 and 6 of FIG. 1, but which are not shown in FIGS. 6A–B). The edges 48/67 of the housings 44 from which plungers 43 emerge is aligned with the ends 47/68 of the signal trace/dielectric layer combinations 45/46 and 56/55. These alignments are one terminus of a region 51 of characteristic impedance Z₀ the exists between the end 47, along the amount 51 that the ground plane (45, 54) extends beyond end 47 and beneath the emerging plunger 43 until the other terminus at end 66. In the event that there are to be two rows of spring pins 42 there will also be an intervening dielectric layer 53 to separate the two ground planes 49 and 54. The result is a multilayer printed circuit assembly 57.

[0030] Corresponding to the explanation of FIG. 5A, in FIG. 6A a Z_x of length 52 is formed as the plunger 43 extends beyond the edge 66 of the ground plane 49.

[0031] Corresponding to the explanation of FIG. 5B, in FIG. 6B Z_x is absent, leaving only a Z₀ of length 51.