Document created by analog-archivist on Feb 23, 2016
Version 1Show Document

### Q

Some times ago we have discussed about an application for measuring the optical
input power with the AD8307AR. I have got some samples and I
have built a circuitry which is attached at the end of this document. Could
you or an responsible application engineer / chip designer check if this  would
work. At least I have got the impression that it is not easy to get  accurate
results. Furthermore I have also attached our measurement results. So you
can easily see that the output voltage is not a linear function over the
optical input power. I have theoretically checked the functionallity and this
should give a linear output level in the range of 0.7V up to 1.7V.
What are the main features of our application and of the circuitry? I will
- first of all we have to measure a dc voltage. This is now a fact. The last
times I was not sure what is to measure (DC bias or the overlayed AC voltage).
- the above mentioned fact means a dc-coupled application for the AD8307AR.
(You can see it in may application circuitry). - one end of the measuring
resistor is connected to +3.3V.
We have noticed that there is a big influence of the absolute voltage to the
output level if there is no optical input power (then I expect only a  noise
level which will result in a finite output voltage nearby 0.5V or lower)
.
Unfortunately this basis output level is in the near of 1V and for a measuring
resistor of 1k instead of 180 ohms it is nearly 1.5V. I  guessed that there is
some ripple on the power supply voltage (+3.3V). However I couldn't see
anything with the scope.
Here are my questions:
- Would you prefer the AD8310AR (has an additional output buffer and better
any possibility to reduce the big influence of the supply
voltage (+3.3V)? I will try ones more to feed the circuitry separately.- Is it
necessary to shield the whole circuitry? If yes, this would raise the effort
quite much.- The DC-voltages at the differential inputs is about +3V whereby
the supply voltage for the AD8307AR is +5V. Does this work properly? In the
datasheet is only mentioned that the DC level has to be min. 2V over GROUND.
- Shall I provide a DC voltage at the OFS input or is it ok. like I have done
it?

### A

useful to look at the data you sent in terms of log conformance; i.e. the
deviation from the expected logarithmic transfer function. If you look at the
analysis tab on the enclosed spreadsheet, you'll see that the maximum deviation
from the expected log characteristic is within +/-0.5dB for all cases, which is

There are, however, a couple of issues when working with DC signals which you
should be aware of. The inputs to the AD8307 are DC coupled so in theory, there
is nothing to prevent you using the AD8307 to measure DC signals. The inputs to
the AD8307 are fully differential and have a nominal common mode voltage of
Vdd/2 (2.5V for a 5V supply).  If you DC couple the inputs you need to take
this biasing arrangement into account. Using the high side current sense
circuit in your schematic, you essentially have a single ended source
referenced to the 3V supply, the common mode voltage is also 3V. Any time you
change the nominally 3V supply voltage you also change the bias conditions for
the input stage of the AD8307 so you will see a change in intercept, slope or
both. It will be important to keep the 3V line absolutely solid, we do not spec
any common mode rejection for the AD8307 as it is designed as a IF power
detector.

The second thing to bear in mind is that the automatic offset null loop does
not operate when the inputs are DC coupled. The offset null loop tries to
remove any DC offset at the inputs and the circuit cannot distinguish between
the DC input signal and the DC offset. It is therefore necessary to disable the
offset null circuit by overriding OFS with a DC voltage greater than ?400uV.
Offset adjustment is then performed manually either by applying a signal  to
OFS as described in p18 of the datasheet or you can use the AD8318 to level
shift the signal, provide single ended to differential conversion and inject an
offset compensation voltage as shown in figure 36 of the AD8130 datasheet.

I've answered your specific questions below, but I suspect many of them are

- Would you prefer the AD8310AR (has an additional output buffer and better
The AD8310 has a voltage output and is faster to respond than the AD8307. But
operating with DC signals you'll still need to take the same precautions

- Do you see any possibility to reduce the big influence of the supply voltage
(+3.3V)? I will try ones more to feed the circuitry separately.
If you want better common mode rejection, use the AD8138 to level shift the
signal and convert to differential form.

- Is it necessary to shield the whole circuitry? If yes, this would raise the
effort quite much.
There is no inherent need to shield the AD8307 or AD8310, good layout and
grounding should be enough. But if the part is expected to operate in a very
noisy environment or the signal is likely to be contaminated by many ESD
events, you might need to consider sheilding

- The DC-voltages at the differential inputs is about +3V whereby the supply
voltage for the AD8307AR is +5V. Does this work properly? In the datasheet is
only mentioned that the DC level has to be min. 2V over GROUND.
This works fine, but do not expect the AD8307 to insensitive to changes in the
common mode voltage

- Shall I provide a DC voltage at the OFS input or is it ok. Like I have done
it?
You need to disable the internal auto null circuits and manually compensate for
offset using one of the methods described above.