Testing a M5L 2101AP Static RAM Chip

[NivagSwerdna]

I'm trying to test a RAM chip I took out of a board which I will eventually get around to repairing...

The device is a M5L 2101AP, probably better know as 2101A which is a 256x4 Static RAM with separate input and output buses for the data.

I have attached +5V to p22, GND to p8, Driven p18 and p19 LO, p17 and p20 HI... driven a sensible address to A0..A7 and driven DI1..DI4 to sensible values...

The output is going into a uP with high impedance inputs (it's actually a PIC dsPIC33EV256GM006-I_PT) but via a 220 ohm resistor (all my digital pins have a series resistor)

What I am not seeing is nice logic levels on the Data Output Pins... instead I see values like...

0.2V, 1.8V, 0.5V, 2.0V... 

it's like it is trying to represent 0s and 1s but just at a much lower level than I expected.  I was hoping for an unambiguous output but 1.8V seems too low for a TTL '1'

Could it be that the device struggles to cope with my high impedance inputs?

I tried adding weak pull-ups or pull-downs at the PIC side but that just pulls up to 5V or down to GND respectively and I see no output.

Is there something really simple I am missing here?

Thanks in advance

PS
I have two devices... both seem to do the same

[edavid]

Pin 17 should be tied high also.

[NivagSwerdna]

Pin 17 should be tied high also.

Oops.. yes it is... corrected above.  Thanks

Additionally I am able to use the uP ADC to measure all the voltages of the chip under test... and I see

Code: [Select]

P01 0v0, P02 4v9, P03 0v0, P04 5v0, P05 0v0, P06 5v0, P07 0v0, P08 0v0, P09 0v0, P10 0v2, P11 5v0,
P12 1v7, P13 5v0, P14 3v7, P15 5v0, P16 3v6, P17 5v0, P18 0v0, P19 0v0, P20 5v0, P21 5v0, P22 5v0
You will notice anaemic voltages at the Data Out pins...
P10 0v2, P12 1v7, P14 3v7, P16 3v6

1.7V is definitely ambiguous in logic terms

[edavid]

1.7V is definitely ambiguous in logic terms

Yes, but the other voltages are fine for NMOS.  Looks like you have one bad output.

[NivagSwerdna]

Do you think if I made it drive a load it would help?  What sort of load?



If I drive the Inputs with a value between 0 and F and observe the outputs I get good values on DO4... totally bogus values on DO3... not much output on DO2 and small voltage swings on DO1...

... there's definitely something going on

[NivagSwerdna]

Time to test some more 2101 RAM chips and I still cannot get my tester to work...

Code: [Select]

          ---------------
0.0  >A3  |  1       22 |  VCC  5.0
0.0  >A2  |  2       21 |  >A4  0.0
0.0  >A1  |  3       20 | >RnW  0.0
0.0  >A0  |  4       19 | >nCS1 0.0
0.0  >A5  |  5       18 |  >OD  0.0
0.0  >A6  |  6       17 | >CS2  5.0
0.0  >A7  |  7       16 |  D3   2.9
0.0  VSS  |  8       15 |  >D3  5.0
5.0  >D0  |  9       14 |  D2   0.0
2.9  D0   | 10       13 |  >D2  5.0
5.0  >D1  | 11       12 |  D1   2.9
          ---------------
The board these come from has pins 18 & 19 to GND and pin 17 pulled up

So the only real inputs are Address, Data and pin 20... with Pin 20 low I would expect the inputs and outputs to match.

The chip is believed GOOD; certainly seems to mostly work in the original board.

I must be missing something obvious

[NivagSwerdna]

Interestingly... on the original board this directly drives a 74367...

And the Datasheet of a 74367 has...

VIH Input HIGH Voltage MAX 2.0 V Guaranteed Input HIGH Voltage
VIL Input LOW Voltage MIN 0.8 V Guaranteed Input LOW Voltage

So if I consider >2V as a 1?

[NivagSwerdna]

Write 00 to 0000

Code: [Select]

          ---------------
0.0  >A3  |  1       22 |  VCC  5.0
0.0  >A2  |  2       21 |  >A4  0.0
0.0  >A1  |  3       20 | >RnW  5.0
0.0  >A0  |  4       19 | >nCS1 0.0
0.0  >A5  |  5       18 |  >OD  0.0
0.0  >A6  |  6       17 | >CS2  5.0
0.0  >A7  |  7       16 |  D3   0.0
0.0  VSS  |  8       15 |  >I3  0.0
0.0  >I0  |  9       14 |  D2   0.0
0.0  D0   | 10       13 |  >I2  0.0
0.0  >I1  | 11       12 |  D1   0.0
          ---------------
Write 0F to 0001

Code: [Select]

          ---------------
0.0  >A3  |  1       22 |  VCC  5.0
0.0  >A2  |  2       21 |  >A4  0.0
0.0  >A1  |  3       20 | >RnW  5.0
5.0  >A0  |  4       19 | >nCS1 0.0
0.0  >A5  |  5       18 |  >OD  0.0
0.0  >A6  |  6       17 | >CS2  5.0
0.0  >A7  |  7       16 |  D3   2.9
0.0  VSS  |  8       15 |  >I3  5.0
5.0  >I0  |  9       14 |  D2   2.9
2.9  D0   | 10       13 |  >I2  5.0
5.0  >I1  | 11       12 |  D1   2.9
          ---------------Read 0000

Code: [Select]

          ---------------
0.0  >A3  |  1       22 |  VCC  5.0
0.0  >A2  |  2       21 |  >A4  0.0
0.0  >A1  |  3       20 | >RnW  5.0
0.0  >A0  |  4       19 | >nCS1 0.0
0.0  >A5  |  5       18 |  >OD  0.0
0.0  >A6  |  6       17 | >CS2  5.0
0.0  >A7  |  7       16 |  D3   0.0
0.0  VSS  |  8       15 |  >I3  0.0
0.0  >I0  |  9       14 |  D2   0.0
0.0  D0   | 10       13 |  >I2  0.0
0.0  >I1  | 11       12 |  D1   0.0
          ---------------Read 0001

Code: [Select]

          ---------------
0.0  >A3  |  1       22 |  VCC  5.0
0.0  >A2  |  2       21 |  >A4  0.0
0.0  >A1  |  3       20 | >RnW  5.0
5.0  >A0  |  4       19 | >nCS1 0.0
0.0  >A5  |  5       18 |  >OD  0.0
0.0  >A6  |  6       17 | >CS2  5.0
0.0  >A7  |  7       16 |  D3   2.9
0.0  VSS  |  8       15 |  >I3  0.0
0.0  >I0  |  9       14 |  D2   2.9
2.9  D0   | 10       13 |  >I2  0.0
0.0  >I1  | 11       12 |  D1   2.9
          ---------------Read 0000

Code: [Select]

          ---------------
0.0  >A3  |  1       22 |  VCC  5.0
0.0  >A2  |  2       21 |  >A4  0.0
0.0  >A1  |  3       20 | >RnW  5.0
0.0  >A0  |  4       19 | >nCS1 0.0
0.0  >A5  |  5       18 |  >OD  0.0
0.0  >A6  |  6       17 | >CS2  5.0
0.0  >A7  |  7       16 |  D3   0.0
0.0  VSS  |  8       15 |  >I3  0.0
0.0  >I0  |  9       14 |  D2   0.0
0.0  D0   | 10       13 |  >I2  0.0
0.0  >I1  | 11       12 |  D1   0.0
          ---------------
So functionally correct but anaemic?

[edavid]

So functionally correct but anaemic?

Why do you say it's an[a]emic?

[NivagSwerdna]

I have a small collection of these devices, from old boards, and purchased from various sources.

I have been trying to determine which ones work and which don't.  For "identical" devices the driven data output voltages seem to vary considerably by device so I am trying to determine which ones are within spec.

Now... My definition of good is clouded by the fact that my homebrew chiptester is based on a dsPIC33EV256GM006 which has GPIO where a Input High is considered >= 0.75 VDD i.e. 3.75V which may be unrealistic for an aged TTL output... I'm surprised I haven't seen this more often.

Anyway... given TTL is a High of >= 2V (and I can measure pin voltages) I can workaround the level issues.  For some chips it seems they lack oompf and can only drive up to 1.7V which is a fail.

Code: [Select]

          ---------------
0.0  >A3  |  1       22 |  VCC  5.0
0.0  >A2  |  2       21 |  >A4  0.0
0.0  >A1  |  3       20 | >RnW  5.0
0.0  >A0  |  4       19 | >nCS1 0.0
0.0  >A5  |  5       18 |  >OD  0.0
0.0  >A6  |  6       17 | >CS2  5.0
0.0  >A7  |  7       16 |  D3   3.5
0.0  VSS  |  8       15 |  >I3  5.0
5.0  >I0  |  9       14 |  D2   3.4
3.7  D0   | 10       13 |  >I2  5.0
5.0  >I1  | 11       12 |  D1   3.4
          --------------- Good ones seem to easily exceed the 2V threshold.

Makes me wonder how the ATMega based testers cope with TTL Levels...

[edavid]

To be pedantic, they are not TTL outputs, they are TTL compatible NMOS.

V_OH on the Mitsubishi datasheet is 2.4V.  I think National is only 2.2V.  Of course it should be a bit higher with light loading, but 2.9V is pretty good.

What's your supply voltage?

[NivagSwerdna]

What's your supply voltage?

5.0V

Feel free to be pedantic. 

This is something I naively overlooked.  I guess a 74HCT buffer of some kind could perform the necessary translation but I think this is OK for now in that I have met my objective of sorting my parts.  I'm surprised it didn't bite me earlier.