Dynamic random access memory with low-power refresh
View Patent ↗A technique to reduce refresh power in a DRAM. In one embodiment, all of the DRAM memory cells are refreshed at a first rate and a subset of the memory cells are refreshed a second rate greater than the first rate. In another embodiment, the DRAM has a refresh controller that generates a refresh address and controls the refresh of the memory cells addressed by the refresh address. A marker memory is used by the refresh controller to determine which of the memory cells requires refreshing at a rate faster than the refresh rate of the remaining memory cells. Testing the DRAM uses a method to determine which of the memory cells are to be refreshed at the faster rate and to store the results in the marker memory.
1. A dynamic memory comprising:
a plurality of memory cells;
a refresh controller adapted to refresh all of the memory cells at a first rate and refresh a subset of the memory cells at a second rate;
wherein the second rate is greater than the first rate;
wherein the refresh controller is further adapted, for all the memory cells in the dynamic memory and at the second rate, to: 1) generate a refresh address, 2) read a value in a marker memory addressed using the refresh address, and 3) selectively refresh, depending on the read value, memory cells addressed using the refresh address; and
wherein the memory cells are organized into a number of rows and a number of columns addressed by corresponding row and column addresses, the memory cells in a row are refreshed at substantially the same time, the refresh address is a row address, and the number of values in the marker memory is equal to the number of rows.
2. The dynamic memory of claim 1 , wherein the marker memory is a non-volatile memory.
3. The dynamic memory of claim 1 , wherein the second rate is an integral multiple of the first rate.
4. The dynamic memory of claim 1 , wherein the subset of memory cells are those memory cells unable to retain data when refreshed at the first rate over all operating temperatures and power supply voltages of the dynamic memory.
5. The dynamic memory of claim 4 , wherein the second rate is sufficient for the subset of memory cells to retain data stored therein over all the operating temperatures and the power supply voltages of the dynamic memory.
6. The dynamic memory of claim 5 , further comprising a test apparatus, the test apparatus adapted to: 1) test all the memory cells at a refresh test rate to determine which of the memory cells are unable to retain data at the refresh test rate, and 2) enter at least one value in a marking memory identifying the memory cells unable to retain data at the refresh test rate;
wherein the memory cells unable to retain data at the refresh test rate are the subset of memory cells.
7. A method of refreshing a dynamic memory having a plurality of memory cells organized into a number of rows and a number of columns addressed by corresponding row and column addresses, the memory cells in a row are refreshed at substantially the same time, the method comprising the steps of:
a) refreshing all of the memory cells at a first rate;
b) repeating, at a second rate greater than the first rate and for all the rows in the dynamic memory, the steps of:
b1) generating a refresh row address;
b2) reading a value in a marker memory addressed using the refresh row address; and
b3) selectively refreshing, depending on the read value, memory cells in a row addressed using the refresh row address;
wherein the number of values in the marker memory is equal to the number of rows of the dynamic memory.
8. The method of claim 7 , further comprising the steps of:
testing all the memory cells at a refresh test rate to determine which, if any, of the memory cells are unable to retain data at the refresh test rate; and
entering at least one value in a marking memory identifying those memory cells unable to retain data at the refresh test rate.
9. The method of claim 7 , wherein at least one of the memory cells in the rows of memory cells is refreshed in step b3 is unable to retain data when refreshed at the first rate over all operating temperatures and power supply voltages of the dynamic memory.
10. The method of claim 9 , wherein the second rate is sufficient to assure that data stored in the at least one memory cell is maintained over all the operating temperatures and the power supply voltages of the dynamic memory.
11. The method of claim 7 , further comprising the steps of:
b4) incrementing the refresh row address; and
b5) repeating steps b1) though b4) until the refresh row address equals or exceeds a specified value.
12. The method of claim 7 , wherein the marker memory is a non-volatile memory.
13. The method of claim 7 , wherein each value in the marker memory is a multi-bit value.
14. The method of claim 7 , wherein the marker memory is loaded with test result data from a data retention test of the dynamic memory using a refresh test rate.
15. The method of claim 7 , wherein the second rate is an integral multiple of the first rate.