Test coupon for measuring a dielectric constant of a memory...

Details Test coupon for measuring a dielectric constant of a memory... Test coupon for measuring a dielectric constant of a memory...

2001-12-28

2003-12-30

Deb, Anjan K. (Department: 2858)

Electricity: measuring and testing

Fault detecting in electric circuits and of electric components

For fault location

C324S639000, C324S663000, C174S250000, C174S262000

Reexamination Certificate

active

06670816

ABSTRACT:

BACKGROUND OF THE INVENTION
1. Technical Field of the Invention
This invention relates generally to technology for measuring a dielectric constant of a memory module. More particularly, this invention relates to a test coupon used to measure the dielectric constant of a Rambus Interface Memory Module (RIMM), as well as a method of using the test coupon.
2. Description of Related Art
In general, Rambus Dynamic Random Access Memory (RDRAM) is supplied in a Rambus Interface Memory Module (RIMM) that can include four to sixteen individual RDRAM devices. Using controlled-impedance transmission lines, Rambus memory systems are able to offer high bandwidth. To control the electrical and logical characteristics of these memory systems, the parameters of each system component (RIMM, motherboard, etc.) must be kept within a specified range. These parameters include impedance, propagation delay, propagation delay skew, and others. In particular, since the Rambus channel operates at high data rates of up to 800 million transfers/second, controlling high frequency properties, such as reflection and crosstalk is critical.
Propagation delay is an important parameter since the flight-time of the electrical signals must meet the maximum and minimum specifications for the system to function logically. The propagation delay of a RIMM is depends mainly on the wave velocity of the RIMM Printed Circuit Board (PCB) and on the capacitive loading of the surface-mounted RDRAM devices. In typical RIMMs, the bare-board PCB delay accounts for 60-90% of the total delay.
The PCB propagation delay is primarily a function of the dielectric laminate materials (Prepreg, core, etc.). Although RIMMs are manufactured with standard FR-4 materials, careful material selection and stack-up optimization is necessary in order to achieve the specified electrical impedance and propagation delay properties. The RDRAMs load the bare board PCB capacitively, increasing the total propagation time of the trace signal. Delay from the RDRAM devices increases as the mounting density of the RDRAM components on the PCB increases. Further, RIMMs will have higher propagation delay and lower impedance when built with RDRAMs with higher input capacitance. Conversely, RIMMs will have lower propagation delay and higher impedance when built with RDRAMs of lower input capacitance. Therefore, to maximize high volume production yield, the RDRAMs and PCBs should be optimized to meet both the impedance and propagation delay specifications when assembled as a module.
One method for measuring propagation delay is to measure propagation velocity. To determine the propagation velocity, and hence the propagation delay, the structure delay is determined by measuring the difference in time it takes for the pulse to propagate through the structure. Parameters related to the transmission time in RIMMs include the propagation delay Tpd and delay skew &Dgr;Tpd between each Rambus Signal Level (RSL).
FIGS. 1 and 2
illustrate a delay skew &Dgr;Tpd measurement method according to the prior art. Referring to
FIG. 1
, propagation velocity is measured using Time Domain Reflectometry (TDR) in a Time Domain Transmission (TDT) mode. A TDT mode test is performed by placing low capacitance, high impedance probes
6
and
8
at both launching and reflection points. A pulse
4
is applied to one end (the launching point) of a test coupon
2
with a 50-ohm probe
6
and a signal is captured at the other end (the reflection point) using a second probe
8
. The advantage of the TDT mode over other TDR techniques is that the captured signal has propagated only once down the coupon, yielding an improved rise-time response. In the graph shown in
FIG. 2
, the x-axis represents time and the y-axis denotes a voltage level. By measuring the time difference &Dgr;T between the signal
10
input at the launching point and the signal
12
captured at the reflection point, the propagation delay and dielectric constant can be calculated.
FIG. 3
illustrates another conventional method for measuring propagation delay and velocity. Referring to
FIG. 3
, this method uses a dual-ported continuous impedance coupon. Using a vector network analyzer, the propagation delay is measured as a phase delay of a 400-MHz sine wave between two ports. This method yields the most accurate measurements in absolute time since the electrical length is fixed and reflections are minimal due to the continuous impedance of the trace. This diagnostic coupon may be placed on the side-panels of a typical RIMM module panel.
Unfortunately, however, due to limitations on the main time scale of the currently available TDR measuring instruments, the prior art test methods shown in
FIGS. 1-3
cannot ensure a reliable and precise measurement. It is therefore impossible to determine either the exact launch edge or the exact reflected wave edge. The uncertainty in identifying these edges is a major obstacle to accurately measuring propagation delay by TDR. For example, as shown in
FIG. 2
, displaying the launching and reflection points on one screen of the measuring instrument makes it difficult to locate the true launching and reflection points, which are the references used to measure the propagation delay. The measured propagation delay Tpd value may therefore be erroneous because it is based on inaccurate launching and reflection points. As a result, in conventional propagation delay measurement methods, the precise dielectric constant of a RIMM PCB, and thus the properties of the PCB, are difficult to know, and controlling variations in properties of the board itself is difficult.
Moreover, since the device specification requirement is more stringent than that for a clock, and requires a delay skew &Dgr;Tpd within the range of ±10 ps, it is difficult to precisely estimate and manage the delay skew &Dgr;Tpd. Some reasons for this difficulty include the fact that the parameters affecting the delay skew &Dgr;Tpd of a RIMM vary depending on properties of the module PCB and mounted RDRAMs, on the interoperation when combining the RDRAMs with the PCB, and on errors in the manufacturing process.
In order to make delay skew &Dgr;Tpd more consistent, RSL traces each having the same length can be used to make the propagation delays of the traces more similar. This is not sufficient, however, because even if the total lengths of the traces are identical, each of the RSL traces may still exhibit different properties due to asymmetrical structures used in forming the traces. The individual RDRAMs mounted on the PCB may also have slightly different properties depending on their pins.
Furthermore, because of manufacturing inconsistencies, the impedance and the propagation delay of the PCB trace may vary among products and may even vary among traces in the same product. This can lead to variation in the operational characteristics within a given range. All of these factors can change the channel delay. Although the amount of variance due to these factors is relatively small and may not be a serious problem in light of the overall propagation delay, this variance can cause a significant problem with respect to the accuracy of the delay skew &Dgr;Tpd measurements. The properties of each RDRAM component and PCB should therefore be assured in order to control the properties of the mounted RIMMs. The existing RIMM PCB design, however, cannot provide precise measurement because of a via effect and a bending effect of each signal.
SUMMARY OF THE INVENTION
An object of the present invention is to provide a method of measuring a dielectric constant of a PCB that can precisely measure the properties of a produced memory module, and control, based on the measured data, the properties of a PCB.
Another object of the present invention is to provide a test coupon for use in the method of measuring a dielectric constant.
Another object of this invention is to provide a cost-effective way to measure the properties of a memory module PCB using existing measurement instruments.
A still further object of this invention is to assure cons

Affiliated with

Also associated with

Rating

Test coupon for measuring a dielectric constant of a memory... does not yet have a rating. At this time, there are no reviews or comments for this patent.

If you have personal experience with Test coupon for measuring a dielectric constant of a memory..., we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Test coupon for measuring a dielectric constant of a memory... will most certainly appreciate the feedback.

Rate now

Comments { 0 }

Profile ID: LFUS-PAI-O-3122196