Anadyne User note. 5/15/2009.

This note addresses two issues:

1. A new version of the L-17D
2. Using the L-17D at higher rail voltages.

Anadyne has made a second version of the L-17D. The changes are transparent to the user; i.e. the version 2 parts will work in existing applications without any changes to the circuit boards. However, new applications can take advantage of a new feature that improves performance. Version two parts can be identified as they do not have a 'J' on the bottom line of text on the lid. Only version 2 parts are now generally available.

The changes are discussed below.

1. A 1 pf capacitor that was in parallel with the A3 feedback resistor has been removed to speed up A3. Users who wish to slow down A3 to lower the noise can put a capacitor from pin 24 to pin 26 to replace, or increase, the one we have removed. Several users were running with a speed up capacitor from pin 24 to ground to counteract the capacitor we removed.


2. A functional schematic of A1 is included with this note to clarify the second change. A1 has a negative offset. This is caused by a mismatch in R1 and R2 in the schematic. R1 is smaller than R2 due to mask imperfections, and this causes a negative offset in the output of A1.

Pin 3, which was the old A1 DC adjust on the L-17C, was labeled “Do not use” in the L-17D application notes. The reason for this was that it was on the R1 side of the differential pair formed by Q1 and Q2. Thus the offset had to be corrected by running a resistor to ground from pin 3, and this caused A1’s output to drift in a non linear manner over temperature. That is why the alternative method was needed to zero A1's output by offsetting the positive input by attaching it to VCC1 through a large resistor. The PSSR of the amplifier was seriously degraded by this practice. In version 2 the pad connected to R1 was shifted to attach to the R2 side so that a resistor can be put in parallel with R2 to equalize the resistors in order to bring the output of A1 to zero. This improves the PSSR dramatically.

New designs should utilize this method of zeroing A1 using a resistor from pin 3 to VCC1.

Running the L-17D at higher rail voltages.

Some users have experienced difficulties when trying to run the L-17D at rail voltages above 6 volts. We have observed that the behavior at these voltages depends strongly on the layout and we have investigated the phenomenon very carefully. The problem originates in the logging section and a functional schematic of the relevant parts of that section is included with this note.

The problem manifests itself most strongly at large pulse heights where a very high frequency oscillation, 500MHz or higher is observed. We have established that the severity of the oscillation is strongly dependent on the capacitance on the pins S+ and S-. It is minimized when the capacitance on S+ is very low and the capacitance on S- is several pfs. The problem is caused by too much current running through Q1 and Q6. This induces a high frequency oscillation of the L+ output which goes directly into the output amp, (pin 21). See fig 2.

We offer several solutions to the problem below.

Please note the following:

1. VCC1 And VCC2 do not need to be run at the same voltage.
   However VEE1 and VEE2 must be run at the same voltage as they are connected through the substrate.
2. There is a pad that is currently not pinned out, as the original pinout was as close to that of the L-17C as possible to accommodate existing customers. This pad is shown in the application notes on page 10, where it is named "Log DC out adjust". This pin moves both inputs to the output simultaneously, so it is not necessary to change both S+ and S- in order to bring the log outputs to ground. Currently this pad is only accessible to projects that utilize die. However, Anadyne intends to offer an alternative pinout, (called pinout H in this note) for the L-17D, where this pad will be accessible.

In general the only reason to run the rails at voltages that are higher than the 6 Volts recommended, is to increase the maximum output from A1 to extend the range as far as possible. It is also possible that some users may want to increase the maximum output from the output amplifier. If raising the maximum output from A1 is the motivation for the increased rail voltages, we suggest solution 1, but the others may be used instead.