Shifting logic band 3V/5V to TTL

[nilium]

Hi all:

I have to adapt some LED driver outputs (8 channels) that are either 3V (on) or 5V (off) to drive the inputs of a DAC (AD557), all using a single suppy voltage (either 5 or 12VDC). As it is, both 3V and 5V are interpreted as high by the DAC. I was thinking of using voltage threshold detectors (one per channel) but I was wondering if someone out there has a better idea. Thanks!

NV

[kurt]

Could it be that the LED driver outputs stop at 3V because that's when the LED's stop conducting? Assuming they are still attached that is. If so you might only need a pull down resistor to pull it all the way.

[T3sl4co1l]

Can you do anything with the LED driver?

AD557, how freaking old is this thing?

PNP arrays might be the quickest.

Tim

[nilium]

@kurt: The LEDs are common anode (5V). Each output of the LED driver is connected to each LED's cathode. When the driver drops the potential from 5V to 3V current flows through the LED and voilà. Even if I remove the LEDs, the output of the driver will be 3V or 5V. Not sure how a pull-down resistor would help. Would you please elaborate?

@Tim: I don't have the liberty to wack off the driver, unfortunately. The main board is already made and good. Now I'm trying to use the LED outputs as GPO to feed the DAC. Not sure how old the AD557 is but I needed an 8-bit DAC and found this one. In regards to using PNP transistors, are you suggesting to hold the emitter at 5V, grounding the collector via a pull-down resistor, connecting the output of the LED driver to the base, and getting an inverted output at the collector? I think it may work: When the base is at 5V, the PNP is off hence the collector is at 0V; when the base is at 3V the PNP is on, therefore the collector output 5V - V_EC, 4.4V approx.

[T3sl4co1l]

Yes, that... except with a base resistor to limit current (and not hog the current going to the LEDs).

I don't know if you'll need a subsequent logic inversion or not.  If so, you can operate the transistors as common-base instead, in which case you will need a pull-up resistor across the LEDs.

You could also add a resistor in series between driver and LED, which drops some voltage without compromising the driver's compliance voltage range.  This probably won't be enough to use the AD557 directly (0.8V maximum V_IL doesn't leave much room for compliance, whatever the driver is capable of).  A pull-up (across the LEDs) should still be used, because LEDs don't pull strongly when left open circuit.

A CMOS DAC would be more appropriate, e.g.
http://www.digikey.com/product-detail/en/MAX7624CSE%2B/MAX7624CSE%2B-ND/3517281
though it's still TTL threshold.

Tim

[nilium]

That worked like a charm! Vcc at the emitter, pull-down resistor at the collector, signal at the base, and output from the collector... The main board has resistors between the driver and the LEDs so I didn't even have to worry about adding another resistor at the base. I'll do the logic inversion in the firmware, that's no problem! BTW, I am removing the LEDs since they are on a separate board connected by a ribbon cable.

Thanks a lot for your help!

NV

[T3sl4co1l]

So, the driver is probably saturating anyway...?  And you *don't* need the LEDs?

Hell, just remove the damn LEDs, it probably pulls near enough to GND on its own!  Use a pull-up that's big enough such that it doesn't violate V_IL(max).

Tim

[nilium]

Well that's the thing, having removed the LEDs the driver outputs either 3V or 5V depending on whether the channel is "on" or "off". To use that as inputs to the DAC (TTL), one and only one of them has to be below 0.8V. The PNP deal worked great to discriminate between those two levels, outputting 4.4V or ground respectively.