How to avoid 'backwards powering' of PCCard

[Manfred]

Hello all,

I've developed a PCCard that provides a parallel interface for connecting some measuring equipment to notebooks. The problem is that when the notebook is turned off while the external equipment is still running, the PCCard gets some power via the parallel interface inputs. This power propagates via the VCC pins of the inputbuffers of my card into the notebook. Some notebook models won't power on again in this situation and I will first have to remove the PCCard (or disconnect the external device).

Does anyone know if there is a family of logic circuits which can avoid this? Since the PCCard is operating at 3.3V and the external signals are 5 V it additionally has to translate down the input voltage.

Thank you, Manfred

 

[drwho78]

There are a few of things you can do here. First, if you have control over the circuit that drives the pccard, you should use open collector/tri-state devices. You will pull-up the device using power from your pccard. That way, if you turn off the power to the notebook, no power will flow back through the inputs.

Second method is to put a blocking diode on VCC so that no power can flow from the inputs to the notebook power.

Third, you can always use transistors or opto-isolators on the inputs.

 

[GrandAlf]

Manfred, I am not quite sure what you mean. Do I understand that you have designed a parallel interface that plugs into the Laptops PC Card slot, or you have a PCB that connects to the Laptops Parallel port?.

If you could clarify this, I, and I am sure others, would try and help.

 

[sick_man]

Hello all,

I've developed a PCCard that provides a parallel interface for connecting some measuring equipment to notebooks. The problem is that when the notebook is turned off while the external equipment is still running, the PCCard gets some power via the parallel interface inputs. This power propagates via the VCC pins of the inputbuffers of my card into the notebook. Some notebook models won't power on again in this situation and I will first have to remove the PCCard (or disconnect the external device).

Does anyone know if there is a family of logic circuits which can avoid this? Since the PCCard is operating at 3.3V and the external signals are 5 V it additionally has to translate down the input voltage.

Thank you, Manfred

74ls244 and some diodes

or a bi directional buffer
and some diodes

and a diode on the vcc rail from pc and a diode to gnd with a 470R resistor attached in parallel

 

[padspcb]

I suggest u use a xilinx to do all stuff
With spartans and up series, there are no more diodes from ground and VCC pins to i/o pin, but a simple zenner of 5.6V to ground.
This means that there are no more propagation of external supply to the VCC node.
Best stuff is that u use such a chip to implement all your ports and u can limit the max current entering a pin with a simple resistor. Each pin can take up to 10ma, so u may even connect a pin to mains, to check zero crossing per example, or for hot swapping.
The nice thing about these devices is that you can reprogram them while loading the device driver, with minimum logic, in paralel mode.
So your harware will always be up to date.

 

[jazz_roa]

Manfred,

According to my experience, all you need is ABS-series buffer (for example, 74ABS245 from TI) for the parallel interface. You can get more information from TI's application notes.