mcf52259 cs0 cs1 draw current when power off

[drummer]

My previous project used a multiplexer to select one of 8 SRAM's used for battery back up storage.

When powered off a coin cell provided power to the external SRAM and the multiplexer so that the SRAM would not be accessed.

The multiplexer enable signal was disabled by the reset.

My new project only uses two SRAM's so I eliminated the multiplexer and drove each SRAM with its own chip selects.

The chip selects have a 4.7K resistor pulling up from VSTBY to prevent memory from being accessed.

On one side of the 4.7K I have 2.77V and on the other side 1.25V.

This means that 322 uA is flowing to each chip select output causing the coin cell to drain rapidly.

Is there anything I can do to make the chip selects power down?

Should I re-spin and put diodes in series?

Any suggestions out there would be appreciated.

[TomE]

I assume you're powering off the MCF52259's Vdd supply.

From the Data Sheet:

Table 4. Absolute Maximum Ratings

Note 4: All functional non-supply pins are internally clamped to VSS and VDD.

And you're not meant to supply voltage on any pin above the level on the Vdd pin. Or at least (and this is well specified on this chip) you can, but you have to current-limit it to the "instantaneous Maximum Current" limit.

> On one side of the 4.7K I have 2.77V and on the other side 1.25V.

I think you'll find that the Vdd pin is at about 0.6V too, so that's one diode drop from the CS pins to VDD and something is drawing current from Vdd to bring it close to ground.

> Is there anything I can do to make the chip selects power down?

Don't remove Vdd, but put the CPU into Stop Mode 3 where it it documented to draw 150uA. But that's too much for your circuit.

> Should I re-spin and put diodes in series?

How's that going to help? it will still be conducting to the Vdd pin.

You need to put that multiplexer (or equivalent gates) back. You could also use an "analog switch" chip.

Tom

[drummer · Accepted solution]

Thanks Tom for sharing my pain.

My solution was to put the chip selects in series with a MOSFET that is controlled by RSTO. When the reset is low the chip selects are disconnected from the cpu. The source connects to the cpu, drain to the memory and gate to RSTO.

All is well except I have to re-spin.