WikiPatents - Community Patent Review
Create Free Account  |  License or Sell Your Patent  |  WikiPatents Marketplace  |  WikiPatents Blog
Username:  Password:  
    
Advanced Search
Scan-bypass architecture without additional external latches    
United States Patent5925143   
Link to this pagehttp://www.wikipatents.com/5925143.html
Inventor(s)Gillis; Pamela Sue (Jericho, VT), Kolagotla; Ravi Kumar (Breinigsville, PA), Miller; Dennis A. (Jericho, VT), Noack; Maria (Austin, TX), Oakland; Steven Frederick (Colchester, VT), Rebeor; Chris Joseph (Cambridge, VT), Sopchak; Thomas Gregory (Cambridge, VT), Trinko-Mechler; Jeanne (Essex Junction, VT)
AbstractA scan architecture for testing integrated circuit chips containing scannable memory devices, such as register arrays, is operable in a bypass mode during which only a small portion of the memory cells in each device or array is inserted in the scan path to substantially reduce scan path length, test time and test data volume during testing of other logic components in the circuit chip. Additional decoder logic is employed to select a small number of words in the device or array during the scan-bypass mode, and multiplexor circuitry removes the bypassed words from the scan path. By leaving the small number of the register array words in the scan path, observability of logic upstream of the array, and controllability of logic downstream of the array, is preserved during the bypass mode without the need for additional shift register latches and other external logic components.
   














 Title Information Submit all comments and votes
 
Patent Text Patent PDF Print Page Summary File History
Plain text PDF images Print Summary File History
Drawing from US Patent 5925143
Scan-bypass architecture without additional external latches - US Patent 5925143 Drawing
Scan-bypass architecture without additional external latches
Inventor     Gillis; Pamela Sue (Jericho, VT) , Kolagotla; Ravi Kumar (Breinigsville, PA) , Miller; Dennis A. (Jericho, VT) , Noack; Maria (Austin, TX) , Oakland; Steven Frederick (Colchester, VT) , Rebeor; Chris Joseph (Cambridge, VT) , Sopchak; Thomas Gregory (Cambridge, VT) , Trinko-Mechler; Jeanne (Essex Junction, VT)
Owner/Assignee     International Business Machines Corporation (Armonk, NY)
Patent assignment
All assignments
Publication Date     July 20, 1999
Application Number     08/857,974
PAIR File History     Application Data   Transaction History
Image File Wrapper   Patent Term   Fees
Litigation
Filing Date     May 16, 1997
US Classification     714/726 365/201
Int'l Classification    
Examiner     Beausous Jr.; Robert W.
Assistant Examiner     Iqbal; Nadeem
Attorney/Law Firm     Townsend; Tiffany L.
Address
Parent Case     This application is a continuation of application Ser. No. 08/577,676, filed Dec. 21, 1995, now U.S. Pat. No. 5,719,879.
Priority Data    
USPTO Field of Search     371/22.31 371/22.32 371/22.34 371/23.5 365/201 395/183.06 364/490 324/765 714/30 714/43 714/47 714/726 714/727 714/729 714/730 714/733 714/742
Patent Tags     scan-bypass architecture without additional external latches
   
Enter a comma (,) or semicolon (;) between multiple tag words/phrases.
Describe this patent:
 Amusing   
 Clever   
 Complex   
 Efficient   
 Historic   
 Important   
 Innovative   
 Interesting   
 Practical   
 Simple   
[no votes]
Patent WIKI

Share information and news about this patent, including information and news about the technology, inventors, company, ligation and licensing.

 References Submit all comments and votes
 
*references marked with an asterisk below are user-added references
 U.S. References
 
Add a new US reference:  
ReferenceRelevancyCommentsReferenceRelevancyComments
5677915
Whetsel

Oct,1997

[0 after 0 votes]
5673276
Jarwala et al.

Sep,1997

[0 after 0 votes]
5646422
Hashizume

Jul,1997

[0 after 0 votes]
5631911
Whetsel, Jr.

May,1997

[0 after 0 votes]
5627840
Hundertmark et al.

May,1997

[0 after 0 votes]
5615217
Horne et al.

Mar,1997

[0 after 0 votes]
5497378
Amini et al.

Mar,1996

[0 after 0 votes]
5491666
Sturges

Feb,1996

[0 after 0 votes]
5477545
Huang

Dec,1995

[0 after 0 votes]
5473617
Farwell

Dec,1995

[0 after 0 votes]
5448575
Hashizume

Sep,1995

[0 after 0 votes]
5448576
Russell

Sep,1995

[0 after 0 votes]
5430735
Sauerwald et al.

Jul,1995

[0 after 0 votes]
5331643
Smith

Jul,1994

[0 after 0 votes]
5329533
Lin

Jul,1994

[0 after 0 votes]
5260949
Hashizume et al.

Nov,1993

[0 after 0 votes]
5257267
Ishizaka

Oct,1993

[0 after 0 votes]
5150044
Hashizume et al.

Sep,1992

[0 after 0 votes]
5084847
Whestel, Jr.

Jan,1992

[0 after 0 votes]
5056093
Whestel

Oct,1991

[0 after 0 votes]
5048021
Jarwala et al.

Sep,1991

[0 after 0 votes]
4872169
Whetsel, Jr.

Oct,1989

[0 after 0 votes]
4009348
Tate

Feb,1977

[0 after 0 votes]
3961254
Cavaliere et al.

Jun,1976

[0 after 0 votes]
 Foreign References
 Other References
 Market Review Submit all comments and votes
   
Market Size
Estimate the gross annual revenues of the relevant market sector:
> $10B
$5B - $10B
$2B - $5B
$500M - $2B
$100M - $500M
$10M - $100M
$1M - $10M
$500K - $1M
$100K - $500K
< $100K
[No votes]
$0
 
$0   $2.5B   $5B   $7.5B   $10B
Market Share
Estimate the percentage of the relevant market sector this invention will capture:
75% - 100%
50% - 74.99%
25% - 49.99%
10 - 24.99%
5 - 9.99%
2 - 4.99%
1 - 1.99%
< 1%
[No votes]
0.0%
 
0%   25%   50%   75%   100%
Reasonable Royalty
What percentage of gross sales should the inventor or assignee be paid?
75% - 100%
50% - 74.99%
25% - 49.99%
10 - 24.99%
5 - 9.99%
2 - 4.99%
1 - 1.99%
< 1%
[No votes]
0.0%
 
0%   25%   50%   75%   100%
Public's "Guesstimation" of Royalty Value
Market SizeN/A[No votes]
xMarket ShareN/A[No votes]
xReasonable RoyaltyN/A[No votes]

N/A

License Availablity
If you are NOT the owner or assignee, answer here:
Yes, license is available for purchase

No, license is not currently available



[No votes]
License Availablity
If you ARE the owner or assignee, answer here:
Yes, license is available for purchase

No, license is not currently available



[No votes]
Competitive Advantage
Does this invention have a significant competitive advantage over similar technologies?
Yes

No



[No votes]
Most helpful competitive advantage comment
[No comments]

Commercial Alternatives
Are there viable commercial alternatives for this invention?
Yes

No



[No votes]
Most helpful commercial alternative comment
[No comments]

 Technical Review Submit all comments and votes
 Claims Submit all comments and votes
 


What is claimed is:

1. A method of propagating data through a single shift register having a first group of non-bypassable data words and a second group of bypassable data words, the method comprising the steps of:

(a) providing a bypass mode for addressing and scanning the first group of non-bypassable data words of the shift register, the providing of the bypass mode including multiplexing a scan input to a scan output of the second group of bypassable data words, disabling selection of an output of the second group of bypassable data words, and enabling selection of an output of the first group of non-bypassable words, to insert only the first group of non-bypassable words into a scan path; and

(b) providing a non-bypass mode for addressing and scanning the first group of non-bypassable data words and a second group of bypassable data words of the shift register;

(c) selecting the bypass mode and applying data to an integrated circuit device; and

(d) selecting the non-bypass mode and applying data to the integrated circuit device.

2. The method of claim 1 wherein the data words comprise scannable latches.

3. The method of claim 1 further comprising the step of applying a signal to switch between the bypass mode and non-bypass mode.

4. The method of claim 1 wherein the shift register is incorporated into an integrated circuit device.

5. A method of testing an integrated circuit device comprising the steps of:

(a) providing a scannable memory array device having a first shift register chain, the first shift register chain having a first group of non-bypassable data words and a second group of bypassable data words;

(b) providing a bypass mode for addressing and scanning the first group of non-bypassable data words, the providing of the bypass mode including multiplexing a scan input to a scan output of the second group of bypassable data words, disabling selection of an output of the second group of bypassable data words, and enabling selection of an output of the first group of non-bypassable words to insert only the first group of non-bypassable words into a scan path; and

(c) providing a non-bypass mode for addressing and scanning the first group of non-bypassable data words and a second group of bypassable data words;

(d) selecting the bypass mode and applying test data to the integrated circuit device; and

(e) selecting the non-bypass mode and applying test data to the integrated circuit device.

6. The method of claim 5 wherein the data words comprise scannable LSSD L1/L2 latches.

7. The method of claim 5 further comprising the step of applying a control signal to switch between the bypass mode and the non-bypass mode.
 Description Submit all comments and votes
 


BACKGROUND OF THE INVENTION

The present invention relates in general to a method and apparatus for bypassing scannable memory devices, such as scannable register arrays, which eliminates the need for external shift register latches and other logic components.

As the packing density of integrated circuit chip devices has increased greatly in recent years, various schemes have been devised for providing increased testability of the devices. One particularly effective testing scheme employs level sensitive scan design (LSSD) in which chains of shift register latches (SRLs) are integrally formed in the device circuit to facilitate testing. The SRLs provide internal control points and observation points in the circuit to permit application of control test inputs to the circuit's various logic devices, and observation of the outputs which result from the inputs. The SRLs form a shift or scan path through which test data can be propagated.

In the operation of an LSSD testing scheme, a serial test data stream is first propagated through the scan path. A control pulse is then applied to the logic circuitry in the device to cause it to operate through one cycle using as input, the data in the SRLs at the various control points. The resulting serial data stream is then propagated out of the scan path to an external buffer where it is analyzed to determine if the digital values at the various observation points indicate that the logic circuitry is functioning properly. This procedure is repeated numerous times to ensure that all logic circuitry is tested.

Although the scan design testing scheme is very effective, the maximum scan path length, and therefore the maximum number of SRLs in the scan path, is limited by time constraints since the test data is propagated serially, and each SRL adds an additional scan cycle delay to the total time required for propagation of the test data through the scan path. Thus, the more SRLs that are added to the scan path, the longer the length of time required for the testing procedures. This becomes more and more of a problem as integrated circuit fabrication technology continues to improve and permit greater and greater packing density.

The current design of many memory device circuits also adds to the foregoing problem. More particularly, many memory devices, such as register arrays, are implemented with memory cells or bits that are each comprised of an LSSD L1/L2 latch, which facilitates connection of the array bits in the LSSD scan path. This permits the register array to be tested in the same manner as the logic circuitry, thereby eliminating the requirement for array test techniques, such as those using Array Built-In Self Test (ABIST), that require additional area for supplying test patterns from the test logic through the surrounding chip logic to the register array. However, the test time, volume of test data required behind each scan input, and number of I/Os that are consumed for scanning become prohibitively large if multiple scannable register arrays are provided on the same circuit chip, since each bit in the arrays forms part of the scan path length. This necessitates that a plurality of scan paths and associated I/O pins be provided on each chip to avoid exceeding a maximum allowable number of SRLs in each scan path. Unfortunately, each additional scan path requires the use of two additional I/O pins for applying and receiving signal data to and from the scan path. This can present a serious problem for I/O constrained designs, especially since the number of economic test function I/Os is limited.

One proposed solution to this specific problem presented by scannable memory devices, such as scannable register arrays, is to employ a bypass scheme wherein the bits of the register arrays can be selectively bypassed during testing to reduce the scan path length substantially. With the bypass scheme, circuit testing is performed in two modes. In the first, bypass mode, all logic surrounding the register arrays is tested using the reduced scan path length. Then, the circuit is placed in the non-bypass mode for the remainder of the test, during which only the register ar rays are tested. The total time required for testing the circuit is thus substantially reduced since the scan path length is substantially shorter during the logic component testing, and only the array bits are tested during the non-bypass mode.

To bypass the register array, extra latches and multiplexors must be inserted to observe upstream logic and control downstream logic. A set of SRLs known as "listening latches" is thus introduced with one latch at each input of the register array. These listening latches are multiplexed with the register array outputs to preserve the observability of upstream logic and controllability of downstream logic during scan-bypass mode. Although this arrangement reduces the test time significantly, the number of listening latches and multiplexors needed by this scheme can be large if the number of inputs to and outputs from the register array is large. This increases the overall chip area necessary to implement the circuit. Also, the multiplexors introduce undesirable additional delay in the functional path from the register array.

SUMMARY OF THE INVENTION

To address the foregoing problem, the present invention provides an improved scan-bypass scheme that substantially reduces scan path length and circuit test time for circuits containing scannable memory devices, such as scannable register arrays, and does so without the requirement of any additional external logic devices, such as listening latches or multiplexors. This is achieved through use of circuitry which substantially reduces the apparent logical size of each memory device during the scan-bypass mode. More particularly, circuitry is provided which removes all but a selected small number of the data words in each memory device from the scan path during the scan bypass mode. The selected words are kept in the functional path during the scan, while the scan chain connecting the remaining words of the array is bypassed. Leaving the small number of selected words in the scan path during bypass mode serves the same purpose as the listening latches in the previous scheme by allowing the observability of upstream logic, and the controllability of downstream logic, however, it does so without the requirement for additional external SRLs or multiplexors. As in the previous scan-bypass scheme, all of the latches in each memory device are tested during a non-bypass mode of operation in which all of the words in each memory device are inserted in the scan path. The testing time and volume of data going to the tester are still substantially reduced during the non-bypass mode of operation since only the memory cells are being tested.

In a preferred embodiment of the present invention, additional decoder logic circuitry is employed to select only the small number of words in the memory device during the scan bypass mode. At the same time, the additional decoder logic prevents the bypassed words from being selected or addressed. Preferably, the additional decoder logic circuitry comprises a plurality of AND gates and an OR gate which logically combine the decoder select inputs with a bypass control signal. Switching devices, preferably multiplexors, are also provided which are responsive to the scan bypass control signal to implement the actual bypass of the latches in the scan path which correspond to the bits or cells of the bypassed data words in the memory device.

BRIEF DESCRIPTION OF THE DRAWINGS

The advantages and features of the present invention will become apparent from the following detailed description of a preferred embodiment thereof, taken in conjunction with the accompanying drawings, in which:

FIG. 1 is a general block diagram illustrating a generic implementation of the present invention;

FIG. 2 is a diagrammatic illustration of a conventional LSSD logic testing scheme; and

FIG. 3 is a logic block diagram illustrating a specific memory device design which is constructed in accordance with a preferred embodiment of the present invention.

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENT

FIG. 1 illustrates a generic scannable memory array device 10 which incorporates the scan-bypass architecture of the present invention. As is conventional, the memory device 10 comprises a plurality (n) of multiple bit data words 12 which receive data from a data in bus 14, and transmit data over a data out bus 15. Each of the data words 12 is comprised of a plurality of scannable shift register latches (SRLs), one for each bit, which enable the integrity of the memory array device 10 to be tested using a scan design testing method that treats the memory device 10 as if it were a random logic block circuit. For this purpose, a scan input line 16 from upstream SRLs (not shown) in a scan path is connected to the first of the data words 12 labelled Word 0 so that a test data stream can be shifted into and out of the data words 12. A plurality of scan out interconnection lines 18 is provided for interconnecting adjacent ones of the data words 12 to permit serial propagation of the test data stream through each bit of the memory device 10, and through a scan out line 19 to downstream SRLs in the scan path. As is also conventional, an address decoder circuit 20 is connected between an address bus 22 and a plurality of select lines 24, one for each of the data words 12, for selecting which of the data words 12 are to receive data through the data in bus 14, or transmit data through the data out bus 15.

The heart of the invention lies in the implementation of a bypass scheme which removes the majority, but not all, of the data words 12 from the scan path during testing of logic components upstream and/or downstream of the memory device 10 so that the scan path length, test time and test data volume are substantially reduced. To implement this scheme, the decoder circuit 20 incorporates additional logic circuitry. When the bypass mode is active, a bypass control signal is sent through a bypass control line 26 to the decoder circuit 20 which causes the logic circuitry therein to select or address a small number of the data words 12. In the example illustrated in FIG. 1, only the first two data words labelled Word 0 and Word 1 are selected during the bypass mode. The bypass control signal also causes a switching device 28, such as a multiplexor, to connect the scan output line 17 to a bypass line 30 which is connected directly to the scan output interconnection line 18 from Word 1. In this manner, the serial test data stream on the scan input line 16 propagates only through the plurality of latches forming Word 0 and Word 1, and is then directed through the scan bypass line 30 to the scan output line 19, and onto the downstream SRLs in the scan path. It is important that the decoder circuit 20 not select any of the bypassed data words during the bypass mode so that test data on the data bus 14 is not written into the bypassed words, and the data in the bypassed words is not transmitted on the data out bus 15, during testing. In addition, it is equally important that at least some of the data words be selected to provide the necessary observability of upstream logic components connected to the data in bus 14, as well as controllability of downstream logic components connected to the data out bus 15.

A non-bypass mode of operation is employed for testing all of the data words 12. During this mode, the decoder circuit 20 selects or addresses all of the data words 12 in the memory device 10, and the switching device 28 connects the scan output 19 from the last of the data words 12 (Word n) to the scan output line 19 so that all of the data words 12 are logically connected in the scan path.

FIG. 2 is a general illustration of a conventional LSSD scheme for testing random logic circuitry which can be employed for testing memory devices, such as those illustrated in FIGS. 1 and 3, that incorporate the scan bypass architecture of the present invention. The testing scheme employs a scan path comprised of a plurality of scannable L1/L2 SRL latches 50, three of which are illustrated in FIG. 2. The total number of the latches 50 in the scan path can be selected as desired, depending upon various constraints, the most important of which is testing time. In practice, the length of the scan path is typically on the order of 1,000 or more of the latches 50.

Each of the latches 50 is comprised of a first latch circuit 52 labelled L1, and a second latch circuit 54 labelled L2 which has its input connected to the output of the L1 latch circuit 52. Each of the L1 latch circuits 52 includes two inputs for receiving data, a first scan input 56 which receives serially propagated data along the scan path from an L2 output 58 of the previous L1/L2 latch, and a data input 60 which receives data from an output from a corresponding upstream random logic block 62. A plurality of downstream random logic blocks 64 receive as input, outputs 65 from either the L1 latch circuit 52 or L2 latch circuit 54 of the corresponding L1/L2 latch 50.

Three clocks are provided to control operation of the L1/L2 latches 50 during testing of the random logic blocks 62 and 64, and the L1/L2 latches 50 themselves. A first clock known as the A clock is connected via a first clock line 66 to each of the L1 latch circuits 52, and is activated to clock the data on the scan input line 56 into each L1 latch circuit 52. A second clock known as the B clock is connected to each of the L2 latch circuits 54 via a second clock line 68, and is activated to transfer the data in the L1 latch circuit 52 to the L2 latch circuit 54. Finally, a third clock known as the C clock is connected via a third clock line 70 to each of the L1 latch circuits 52 which is activated to transfer the output data from the upstream random logic block 62 into the corresponding L1 latch circuit 52.

During testing, test data is propagated into the scan path by alternate repetitive actuation of the A and B clocks for application of control test data to the downstream random logic blocks 64. Once the test data has been loaded into the scan path, the C clock is activated once to load the outputs from the upstream random logic blocks 62 into the L1/L2 latches 50, and this provides the necessary observability of the upstream logic. The resulting data stream is then propagated out of the scan path for analysis, again through alternate repetitive activation of the A and B clocks.

The memory device 10 illustrated in FIG. 1 can be implemented with a plurality of the L1/L2 latches 50 of FIG. 2 so that the memory device 10 is interposed between the upstream random logic blocks 62 and the downstream random logic blocks 64. Each of the cells or bits in the data words 12 would therefore be implemented by one of the L1/L2 latches 50, and those in Word 0 and Word 1 would remain selected and in the scan path during the scan bypass mode to provide the necessary observability of the upstream random logic block 62 and the controllability of the downstream random logic block 64.

Turning now to FIG. 3, an eight word, four bit, decode two, single write port, compilable Register Array (RA) 100 is illustrated which is constructed in accordance with a preferred embodiment of the present invention. The RA 100 is comprised of eight four bit memory quad cells labelled 102, 104, 106, 108, 110, 112, 114 and 116. Each of the memory quad cells is formed of four L1/L2 latches, for one bit of each of four of the data words. For example, the first memory quad cell 102 forms the Row 0 bits for the first four words WD0, WD1, WD2 and WD3. Each cell is thus identified in FIG. 3 by a word number followed by a row number as illustrated.

As is conventional, each of the data words is addressed using decoder logic. In the specific case of the eigh