TL081 on Gnd/+9V, PIC on Gnd/+5V ?

Hello, I've looked all the datasheets of the TL081/TL082, and Googled around with keywords like "single supply TL081" and others, but didn't find a schematic or an answer to my question, which in its substance is:

Having gnd and +5V which feed my PIC, if I add a +9V line, can I place a TL081 between gnd and +9V (it would be out of spec of 1V, but it's not the problem I'm asking about) and expect it to work well? I'm using it as a simple voltage follower. Do I need to make a virtual ground (e.g. via two 470 ohm resistors) and make some current flow between it and the TL081 output? (from the datasheet the TL081 doesn't seem to require any current flowing between its output and ground, but there's no mention to a circuit like this anyway). The opamp + input would receive voltages between +2V and +5V, i.e. (relative to the midpoint of its gnd/9V power) -2.5V to +0.5V. I need to attach its output to the PIC's ADC, where I'll read from 2V to 5V. Of course I can't use a TL081 on the gnd/+5V power bus, and a dual -5V/gnd/+5V would force me to modify my already built PSU and, more than that, wouldn't give me the

The circuit I'm thinking of is this:

Gnd | \ / 470 +9V +5V \ | | | 470 | |

--------------*---/\/\/----* | - | | input | | + | |

---------------------- + | | TL081 --*--------- ADC PIC .-- - | | | | | | | | Gnd | Gnd | | '-------------'

input ranges from -0.5V to +2.5V, I expect to get on the PIC ADC +5V to +2V in response to this input range.

Greets and Thanks a lot in advance for your help! TPM

This should just make it: View in a fixed-width font such as Courier.

I don't think the TL081 will swing within 2 diodes of its (-) rail. The

You could use a Schottky.

I don't know enough about a PIC to be able to say for sure but on many ADCs a high frequency high impedance input is a very bad idea.

It may be simpler to get a better op-amp but an RC between the op-amp and the ADC may be needed.

The OA need only achieve one diode drop below 2V or 1.3V to pull the output to 2V. The "typical" characteristics show the OA achieving rails within 0.625V with 10K load to GND and +/-5V supplies. Maybe he should drop a 10K from OA output to GND to improve things:

View in a fixed-width font such as Courier.

Not at all- the PIC SA A/D allows the user to program the charge up time of the S/H capacitor. So unless he there is a minimum sample rate issue,

Better to put the anti-alias RC before the OA IN(+)- no RC needed at A/D input-there is no output impedance issue.

I didn't see anything about high frequency anywhere in the thread. What are you talking about?

Bloggs doesn't mention that the reason for the diode and pullup is that you can't drive signals higher than the PIC rail into the input switcher, or it will leak current into the other analog inputs and you won't be able to read them accurately.

However, I think the pullup should be 2.5K or lower, else it will not be able to pull within half a bit of +5V PIC rail. PIC ADC's are rather leaky and need to be driven from low impedance source. And lower pullup resistance will make it harder for the TL081 to pull down to 2 Volts. If it won't make it, you could boost current with a cheap PNP transistor.

And there is another potential problem with Bloggs' circuit. With signals near +5V, the diode will be conducting very little and the OA approaches open-loop conditions. You are likely to see a lot of "fuzz" on the OA output, which will be rectified by the diode and drive the PIC input negative. You may need to add a few hundred pF capacitor from OA output to inverting input to settle the thing down.

The ADC has hardware inside it that runs from the PIC's clock. The PIC has a clock at many MHz.

That capacitor will be across the diode.

I doubt that the LT081 op-amp will create enough fuzz to keep the ADC input below the rail. As soon as it did, it would swing its output up to some more positive voltage that would back bias the diode.

Right- if Vin is exactly 5V then any noise will cause the amplifier to break into open loop- although the feedthrough to the output is greatly attenuated, there is no point in having it there. A small 10uA injected into the output node will take care of the PIC leakage with no penalty on pull-down: View in a fixed-width font such as Courier.

TLO8x +9V +5V | | | 10K | | +---390K---+ | | (+) | | Vin-------------|\ 1n4148| | >--+--| TO PIC A/D 2V

"Fred Bloggs" a écrit dans le message de news: snipped-for-privacy@nospam.com...

This circuit will keep the TL81 output happily in its allowed range while allowing perfect buffering from 0 to 5V.

|\ 2K2 IN >--------|+\ ___ | >+-|___|-. .-|-/ | | | |/ --- | | --- 10n | | | | +-----' | | ___ | TO ADC

I suppose this will give a true buffer over the full 0 to 5V range on both input and output:

View in a fixed-width font such as Courier.

In article , Fred Bartoli wrote: [...]