I happened to see the General Electric BA-5A limiting amplifier and that it uses an analog delay line to get the control voltage ahead of the signal, so nothing above a set level can get through. Reading the material on the web and in a book here, I started playing with LTspice to come up with the delay line using the same capacitance and induxtance, and tweak it for flat frequency response and also determine the time delay at various frequencies. It's by no means a complete treatment but there are three transformerless models and two with transformers. The transformers are not well-modeled, but I hope it will be of interest. Note that the book mentioned has just about the full theory section from the manual for the instrument as well as a great schematic. references are included in the document attached.
Because I prefer tube and other analog gear, and note that just about all compressors and limitiers have a finite attack time that can allow some excessive peaks through, this one is very interesting. I'm probably wrong about a lot of this stuff and just trying to understand it.
Because I prefer tube and other analog gear, and note that just about all compressors and limitiers have a finite attack time that can allow some excessive peaks through, this one is very interesting. I'm probably wrong about a lot of this stuff and just trying to understand it.
Attachments
Last edited:
The thing that is getting me is that GE built the limiter for what would seem to be broadcast use and must have included music. They sell for several thousand dollars today and I hope people are not just buying them to put in a rack and look at. So what about the varying delay (group delay) over the audio bandwidth? I could never get the network to have a constant delay over the whole spectrum. One missing bit of information is the dc resistance of the inductors.