Methods and circuits for performing margining tests in the presence of a decision feedback equalizer

• US 7,596,175 B2
• Filed: 06/06/2008
• Issued: 09/29/2009
• Est. Priority Date: 05/20/2003
• Status: Active Grant

• Current Assignee: Rambus Inc.
• Original Assignee: Rambus Inc.
• Inventors: Chen, Fred F.
• Primary Examiner(s): Perilla; Jason M.
• Application Number:
  US12/134,691
• Publication Number:
  US 20080240219A1
• Time in Patent Office: 480 Days
• Field of Search: 375229-236, 375/295, 375/316, 327 50- 90
• US Class Current: 375/229
• CPC Class Codes: H04L 1/20 using signal quality detector
• CPC Class Codes: H04L 1/241 using pseudo-errors
• CPC Class Codes: H04L 1/242 by comparing a transmitted test signal with a locally generated replica
• CPC Class Codes: H04L 1/244 test sequence generators

Abstract

Described are methods and circuits for margin testing receivers equipped with Decision Feedback Equalization (DFE) or other forms of feedback that employ historical data to reduce intersymbol interference (ISI). In one example, a high-speed serial receiver with DFE injects the correct received data (i.e., the “expected data”) into the feedback path irrespective of whether the receiver produces the correct output data. The margins are therefore maintained in the presence of receiver errors, allowing in-system margin tests to probe the margin boundaries without collapsing the margin limits. Some receivers include local expected-data sources that either store or generate expected data for margin tests. Other embodiments derive the expected data from test data applied to the receiver input terminals.

Claims

1. A receiver comprising:

a sampler having a plurality of sampler input terminals, including at least one sampler data terminal to receive an input data stream, a sampler clock terminal to receive a clock signal, and a sampler voltage reference terminal to receive a reference voltage; and

at least one sampler output terminal;

wherein the sampler is adapted to sample the input data stream according to the clock signal and with respect to the reference voltage to produce a sampled data stream on the at least one sampler output terminal;

a multiplexer having a first multiplexer input terminal connected to the at least one sampler output terminal;

a second multiplexer input terminal connected to a source of expected data; and

a multiplexer output terminal adapted to provide alternatively one of the sampled data stream and the expected data; and

a feedback circuit having a plurality of delay elements operatively connected to the multiplexer output terminal, each delay element to provide at least one historical bit from the sampled data stream or from the expected data source from the multiplexer; and

a plurality of data-weighting circuits, each data-weighting circuit connected between one of the plurality of delay elements and at least one of the plurality of sampler input terminals and adapted to provide a weighted feedback based on the at least one historical bit from the corresponding delay element.

2. The receiver of claim 1, further comprising a comparison circuit to compare the expected data with the sampled data stream.

3. The receiver of claim 2, wherein the comparison circuit issues error signals responsive to mismatches between the expected data and the sampled data stream.

4. The receiver of claim 1, wherein the expected data comprises an expected-data stream timed to a second clock signal, the receiver further comprising a timing control node to receive a timing control signal to phase adjust the first clock signal relative to the second clock signal.

5. The receiver of claim 4, further comprising clock recovery circuitry to recover the second clock signal from the sampled data stream.

6. The receiver of claim 4, further comprising clock recovery circuitry to recover the first clock signal from the sampled data stream.

7. The receiver of claim 6, wherein the clock recovery circuitry adjusts the phase of the first clock signal responsive to the timing control signal.

8. The receiver of claim 1, wherein the sampler samples the input data stream on an edge of the clock signal.

9. The receiver of claim 1, further comprising a second sampler.

10. The receiver of claim 1, further comprising a clock recovery circuit coupled to the sampler output terminal to derive the clock signal from the sampled data stream.

11. The receiver of claim 10, wherein the clock recovery circuit times the expected data.

12. The receiver of claim 11, wherein the clock recovery circuit phase adjusts the expected data relative to the sampled data stream.

13. The receiver of claim 1, wherein the expected data source comprises a second sampler coupled to the sampler data terminal to sample the input data stream with respect to a second reference voltage to produce the expected data.

14. The receiver of claim 13, wherein the second sampler samples the input data stream with respect to a second clock signal.

15. A method for establishing margins of a sampler adapted to sample a stream of input data symbols on edges of a clock signal and with respect to a reference voltage to produce a stream of data samples, the method comprising:

creating a stream of expected symbols matching the input data symbols;

calculating a weighted sum of a plurality of the expected symbols;

feeding back the weighted sum of the plurality of the expected symbols to the sampler to equalize the stream of input data symbols;

while feeding back the weighted sum of the plurality of expected symbols, adjusting at least one of the reference voltage or a phase of the clock signal over a plurality of values to introduce sample errors in the stream of data samples; and

correlating the sample errors with the values to identify signal boundaries.

16. The method of claim 15, further comprising calculating an optimal sample point from the signal boundaries and adjusting at least one of the reference voltage or the phase of the clock signal to the optimal sample point.

17. The method of claim 15, further comprising sampling the stream of input data symbols to create the stream of expected symbols.

18. The method of claim 17, wherein the expected symbols are sampled using a second sampler in a margin-test mode and the input data symbols are sampled using the first-mentioned sampler in an operational mode.

19. The method of claim 15, further comprising comparing the stream of data samples with the stream of expected symbols to detect the sample errors.

20. The method of claim 19, further comprising sampling the stream of input data symbols with a second sampler to create the stream of expected symbols.

21. The method of claim 20, further comprising comparing the stream of data samples with the stream of expected symbols to detect the sample errors in a margin-test mode and recovering the input data symbols from the first sampler in an operational mode.

22. The method of claim 15, wherein adjusting at least one of the reference voltage or the phase of the clock signal over the plurality of values includes adjusting the reference voltage over a first range of values and adjusting the phase of the clock signal over a second range of values.

23. An integrated circuit comprising:

a reference-voltage source to provide a reference voltage that varies over a reference-voltage range;

clock circuitry to provide a clock signal over a range of clock phases;

sampler having a sampler data terminal to receive an input data stream, a sampler clock terminal to receive the clock signal, and a sampler voltage reference terminal coupled to the reference-voltage source to receive the reference voltage, the sampler to sample the data stream with respect to the voltage reference on edges of the clock signal to produce a stream of data samples;

an expected-data source to provide an expected data stream, the expected data stream including a plurality of historical bits;

a weighting circuit coupled to the expected-data source to receive the expected data stream and to calculate a weighted sum of the plurality of historical bits;

a feedback path extending from the weighting circuit to the sampler, the feedback path to apply the weighted sum to at least one of the sampler data terminal and the voltage reference terminal; and

a comparison circuit to compare the stream of data samples with the expected data stream and issue error signals responsive to mismatches between the stream of data samples and the expected data stream;

wherein the reference-voltage source varies the reference voltage over the reference-voltage range and the clock circuitry varies the clock signal over the range of clock phases while the comparison circuit compares the data samples with the expected data stream.

24. The integrated circuit of claim 23, wherein the weighted sum is applied to the sampler data terminal.

25. The integrated circuit of claim 23, wherein the expected data stream is a version of the input data stream.

26. A receiver adapted to perform margin testing of a pulse-amplitude modulated digital input signal, comprising:

a binary sampler having a data input terminal to receive the input signal, a reference voltage terminal to receive a reference voltage against which the input signal is measured to produce binary output values, a clock terminal to receive a clock signal that triggers sampling of the input signal according to a timing established by the clock signal, and a data output terminal to output the binary output values;

an expected data source to provide a sequence of expected digital values;

a comparison circuit coupled to the data output terminal and the expected data source to compare the binary output values from the sampler with the sequence of expected digital values, the comparison circuit generating an output signal responsive to the comparison; and

a feedback circuit to feed back to the sampler a selective one of either the binary output values or the sequence of expected digital values;

where the receiver further includes digital logic to establish a margin test mode, during which at least one of the reference voltage or the timing established by the clock signal is varied as the binary output values are compared with the sequence of expected digital values, to produce errors as a function of variation of reference voltage or sampling time.

27. The receiver of claim 26, wherein the sequence of expected digital values are samples of the digital input signal.