LC vs RC filters for high/low pass duties...

[mutterd]

Ok, so i just did a session at a studio here in town that has the most amazing collection of crazy old outboard stuff - handmade Manley veri-mu's, blue stripes, 1178's, racks of McCurdy tube pre's, Tele tube pre's, a neve 8056, a whole mess of pultec's and mccurdy pultec style eq's…

one of the things i was most shocked by was the pultec HLF's - they are just simple, variable high and low pass filters that just sound amazing…

so i did some digging, and got to thinking - it seems most of what you read now a days, and around here suggest using RC circuits for high and low pass, but in both the Tremaine's Audio Cyclopedia and the RCA Radiotron book they talk almost exclusively of constant K LC circuits…

in perusing thru some old schematics i found that a lot of the older designed utilized LC filters - neve did a variable high pass with a tapped inductor… Helios used some LC filters - AND - these pultecs are LC as well:

http://www.gyraf.dk/schematics/Pultec_HLF3.pdf

so for the channel strips I'm working on I had settled on simple, single stage RC filters for both high and low, but now i am thinking:

is there an advantage to RC circuits other than the low cost and small size?
Is there a reason that LC circuits have kind of fell out of favour?
Or have they?

I'd love to get you guy's insight on this - I did the math and have started sourcing the stuff - figure i'll give em a shot in my prototype box…

thanks guys,
Timothy

 

[Monte McGuire]

Inductors are expensive, heavy, large, and they can create distortion from hysteresis or saturation. Still, they can be designed to work very very well, as you have discovered.

I think that the advent of cheap and serviceable IC op amps has allowed designers to build filters that do not _require_ inductors, whereas a long time ago, a single amplifier was a complex and large device, something that few would consider as a way to allow a capacitor to take the place of an inductor.

 

[mutterd]

Thanks Monte,

yeah - thats kind of what i was thinking…

but on a small scale i wouldn't mind dropping a few $$$ into each filter - this might even inspire me to make the filters variable...

 

[Monte McGuire]

One advantage of LC filters is that they don't use an amplifier's feedback / gain to achieve a specific frequency response, so they don't tend to magnify the distortion of the amplifier used. A simple LC filter, even if it has a makeup gain amp following it, will be able to use the makeup gain amp as a simple, flat, dozen or so dB makeup gain stage, where it will behave pretty nicely.

Subjectively, this problem can be found in early IC op amp console peaking EQ circuits where the amps available weren't all that nice to begin with. When pressed into providing extra gain in the upper midrange, their marginally acceptable distortion can get magnified with certain topologies, and end up creating some positively annoying distortion. That same amp, asked to provide a flat 10-15dB of gain, following a lossy LC filter, would be a lot less offensive. So, there is some merit to using LC filters, especially if you don't have clean amplifiers to use for an active filter.

These days though, one can design active, capacitor-only EQs and filters that perform incredibly well, and one can also design LC, mostly passive EQs and filters with modern makeup gain amps that also perform extremely well. It's a choice...

 

[mutterd]

Hey thanks - thats really great info for sure…

this channel strip would be between a 600:600 input transformer and a spectra sonics 101 amplifier card…

i'm not at all worried about distortion in that particular amplifier, but it would in fact be a passive filter stage followed by a make up gain amplifier.

at this point i planned the filter as passive circuits, and unless i run into some problems, I'd like to keep it that way…

so do you have any experience with LC filters - seems like the T (or "full section") or the "Pi" filters give the best attenuation but also have more components - is that the way to go, or can I get away with an L (or "half section") filter…


thanks again!
Timothy

 

[emrr]

http://groupdiy.com/index.php?topic=50848.0

http://groupdiy.com/index.php?topic=52435.msg668501#msg668501

http://groupdiy.com/index.php?topic=28452.msg343906#msg343906

http://groupdiy.com/index.php?topic=51217.msg681567#msg681567

Maybe the Passives?
http://groupdiy.com/index.php?topic=56845.msg724710#msg724710

 

[mutterd]

Thanks so much Doug -

I saw your posts on the filter loading -  but that full pultec thread is a new one - gonna dig thru that right now…

so tell me - do you have any other info on those rotary Langevins? They look like inflated Langevin RMX's!!! the plots look pretty amazing…


Timothy

 

[emrr]

Added some more links above.  The langevin and Altec are pretty much self explanatory from a math perspective, and (as you said) I think the audio cyclopedia has some design charts that start to make sense as you stare at them. 

 

[mutterd]

Thanks Doug,

so tell me - are your issues with these LC filters just that the corner frequencies are a little weird?

how about if this same kind of idea was built into a channel strip with say 50/75/100/125 as frequency points...

someone else mentioned it in one of the threads you posted, but on some bass heavy sources the filter actually bumps the frequency around the corner frequency slightly, so if you say high pass a bass drum at 50Hz you get a nice little bump before it starts to fall off - so it actually adds a lot of excitement to the source…

thanks again Doug.
Timothy

 

[emrr]

I have no problem with the corners, this you can tailor with loading.  The problem is all the vintage and clone types are too limited in range for modern use. 

I have an RCA film dialog filter that is great, only HP steps between 70 and 150, and 24dB/oct passive.  It's also 5RU and about 40 lbs. 

I have another set that were for lab use, which have extreme bumps no matter the loading.  They are cool as special effect filters but kinda useless 95% of the time to me.  I tend to use them in parallel on guitars.

 

[Matador]

Ian's thread is a treasure trove on information about how the inductors behaved in the Pultec's.

http://groupdiy.com/index.php?topic=44213.0

 

[mutterd]

thanks for that…

I figured Ian would have a few golden nuggets to drop on this subject..

just found this too - not sure if it ever got fully realized, but there is some good reading here:
http://groupdiy.com/index.php?topic=24376.0

 

[leigh]

is there an advantage to RC circuits other than the low cost and small size?
Is there a reason that LC circuits have kind of fell out of favour?
Or have they?

I had been wondering this same thing recently, and found one perspective on the matter in Douglas Self's Small Signal Audio Design book. He puts it as:

For several reasons, inductors are unpopular with circuit designers. They are relatively expensive, often because they need to be custom made. Unless they are air-cored (which limits their inductance to low values) the core material is a likely source of non-linearity. Some types produce substantial external magnetic fields, which can cause crosstalk if they are placed close together, and similarly they can be subject to the induction of interference from other external fields. In general they deviate from being an ideal circuit element much more than resistors or capacitors.

And elsewhere adds:

I can testify that by 1975 gyrators were the standard approach [for tone controls], and the use of inductors would have been thought risible.

But it was actually while reading about Gyraf's amazing-sounding Gyratec XXI Magneto-Dynamic Infundibulum that I started wondering about the possibilities of all-passive tone controls again. (The Gyratec has a lot more going on that just a tone control, though.)