Behringer 923 Filters & Noise

923 Filters & Noise

Introduction In August 2020 I bought a set of Behringer System 55 modules, as my first journey into Moog modular. My plan is to recreate the Klaus Schulze IIIP synthesizer and Dual Sequencer setup that Klaus bought on 22nd December 1975, and used successfully for many years.

Rather than rack up a set of Behringer modules and ride away making music, I thought it would be useful to explain how modules work, how to calibrate them, and how I have set them up and augmented them. I will also check their specifications, to see if they replicate the originals accurately and so I can understand them better.

The 923 Filters This module initially looks rather boring as it has just two fixed filters, with no voltage control. However the addition of white and pink noise sources makes it more useful as a source of filtered noise, ranging from low red noise to band limited high frequency noise. The original single pole 6dB filters used a couple of early Op Amps (741, 308) and a resistor and capacitor network with the capacitor value at 330nF.

They have a bandwidth of 10 Hz to 10kHz which correlates with the panel markings, and the 6dB cut provides a gentle response. The filters can be patched in parallel to give a band pass filter, and feedback can be added by using a CP3A-M Mixer module to patch the output and mix it into the input.

923 Filter PCB

The 923 Noise Generator The original Moog design uses a noise diode rather than the usual transistor in avalanche design, and it generates both white and pink noise sources at around +5V. It can be tricky to get a modern SMD transistor to produce the correct noise spectrum, but the Moog design uses discrete transistors rather than high gain Op Amps like ARP.

Setup Here is a link to the Behringer user manual, 923 Quick Start Manual, in case you did not get a hard copy manual in the box. The original Moog filters needed offset trimmers because of the high offset from the old Op Amps, however modern Op Amps have much better offset at 0.5 mV and therefore they are not needed. Behringer have used their standard JRC4580 Op Amps, and 330nF polyester capacitors.

The Moog noise circuit has a gain trimmer, which is also on the Behringer module and easily accessible by the customer. It should be set so that the white noise has an output level of +10V peak to peak. The pink noise level is slightly higher. The module was aligned properly, so no need to change anything.

Reverse log

Pot Issue The reverse log 50k potentiometers are the correct value but users have mentioned a stepping effect as the pots are turned to maximum, and this is indeed the case. Its very hard to get a smooth response as the filter is cut to maximum. Swapping the pots out for linear or audio taper will make the situation worse, so I measured the Behringer pot taper to see if a replacement reverse log pot would be any better.

In the last part of the pot travel from 7k5 to 10 kHz the resistance is going from 200R to zero within a few mm of travel, and this causes the problem. The original Moog pots were larger and had larger knobs, which all contributes to being able to dial in the settings at the end of the travel. The accuracy of the taper is important and the Behringer pot does measures closely to the reverse log taper – see picture.

An alternative and more expensive pot may be better at this travel limit, and Alpha do make a C50K but with a solid shaft. I will try this out.

Summary A simple module good at creating filtered noise or subtle sound shaping. The 923 arrived calibrated and went straight into the rack. I don’t think you need to get the 903A Noise module as well, unless you need two noise sources. It is exactly the same circuit inside, and I have put a 923 into my modular. The filter potentiometer performance is disappointing and makes the filters hard to use. I have replaced the two 330nF filter capacitors with Wima polypropylene versions, they mount on the rear of the PCB.

Copyright AMSynths 2019