Interfacing Olde passive EQs in year 2021

GroupDIY Audio Forum

Help Support GroupDIY Audio Forum:

This site may earn a commission from merchant affiliate links, including eBay, Amazon, and others.

Brian Roth

Well-known member
Joined
Aug 20, 2005
Messages
3,249
Location
Salina Kansas
Maybe I'm posting into the wrong section here....shrug.

I had a guy ask me about an Altec 9061A and a Cinema Engineering 4038b which I ASSume needs to be integrated into a modern system.  I find NO spec sheets/app notes/whatever for either EQ device.

Having no hands-on info for those gizmos, I fall back in my mind into the old world of impedance matched systems.  Yep...I know enough to be dangerous...LOL!  45 years in this biz and what I know best is "bridging in/low Z out."

My stooopid idea here is to use a THAT 1646 with 300 Ohm buildout resistors on each output leg as a driver into each EQ, a 600 Ohm R across the out of each EQ and followed by a bridging diff amp to feed the makeup gain amp.

Bri







 
Brian Roth said:
Maybe I'm posting into the wrong section here....shrug.

I had a guy ask me about an Altec 9061A and a Cinema Engineering 4038b which I ASSume needs to be integrated into a modern system.  I find NO spec sheets/app notes/whatever for either EQ device.

Having no hands-on info for those gizmos, I fall back in my mind into the old world of impedance matched systems.  Yep...I know enough to be dangerous...LOL!  45 years in this biz and what I know best is "bridging in/low Z out."

My stooopid idea here is to use a THAT 1646 with 300 Ohm buildout resistors on each output leg as a driver into each EQ, a 600 Ohm R across the out of each EQ and followed by a bridging diff amp to feed the makeup gain amp.

Bri
KISS. You can add resistors in each leg of a "modern" source, that is typically 50-200 ohms. Just add a 100r into each leg.
For the receiving side (typically 10-20k), just strap a 620r resistor across, you'll be matched with less than 0.1dB error.
Now, you'll have to raise the gain somewher, to compensate for the inherent 6dB loss.  :(
 
Thanks Abbey!  What I'm thinking is a "wrap around" for those Olde School passive devices so they Play Nicely with newer systems...and the possible patchbays.

Balanced bridging input -> 600 Ohm EQ device -> Balanced input Follower circuit which has required makeup -> output ckt with something like another 1646 which plays well in a patchbay. 

Using a 1646 with 300 Ohm buildout R's on each leg to drive the Olde EQ unit make any sense???

Bri








 
There's a lot of info about these in another recent thread, may be the Cinema version, may be the Langevin version.    They don't need particular loading, loss is -12 or -14 range. 

You can go as simple as an Edcor 600:10K on the output side for modern interfacing and makeup most of the loss.  I've done it.  It works. 
 
If you know that the source will always be a vaguely modern output like an op amp, which is usually at most 100R (and not another filter or some high Z tube out), you really could just add a series resistor to the input terminals. Maybe create a cable even if it's just a terminal strip. Meaning you could add 500R but in practice, the source resistance doesn't really need to be exactly 600. It can be quite a bit less so you could do like a 220R with the assumed 100R or maybe even a 100R. It should not really have an impact on the response. And, if the circuit is a typical filter of 600R characteristic impedance, you will have much less insertion loss. Meaning make up gain becomes less important (but it might be enough that if you want to switch the thing in and out at the bay and have the level be exactly the same, you might need a make-up device).

If you know that the destination will always be a vaguely modern intput, which is almost always 10K, you can just strap a 680R resistor across the output terminals or integrate it into a clearly marked cable if it's nto a terminal strip. This load does need to be near 600. So 10K || 639R makes 600R. But the next standard value up of 680R makes 637R which is just fine.

You should definitely break out the spectrum analyzer software for this though. Run white noise through and look for loss and peaking. If you have loss at the high end, increase the load resistor value to 1K and see what happens. If you get peaking decrease the value to load more. But try it with different sources.

The only caveat with all of this is if you want to patch one 600R passive filter into another. Then it won't work. You might think that wouldn't happen much but unfortunately I find that I sometimes do have some high impedance device of some sort that really needs to be buffered. But there are some options.

One option is to do the impedance matching at the patchbay. More specifically, and I'm just thinking out loud here - I've never done this before, instead of having one pair of jacks on the bay for the device (let's say you use upper for output and lower for input) you have two pair for the same device right next to each other but the left pair is labeled "NO DAMP" and the right pair is labeled "DAMPED". So on the right input you have a 220R in series with the + signal wire and on the right output you have a 680R load across the + and - signal wires (note that the NO DAMP output will have to be normalled to break the load of the DAMPED output). Now you can plug into any combination and get the right damping. For example, lets say the source is another filter, you can patch into the NO DAMP input. But the destination is a modern 10K so you patch from the DAMPED output. And so on.

The other option is to build a buffer device that is just a little extruded aluminum enclosure with cable gland and some wires coming out of it and you tuck that behind the rack. It two buffers each with 10K inputs and 100R impedance balanced outputs but of course with send / returns that are damped properly. Total two dual op amps. Use a 24V external SMPS and make the virtual ground with a voltage divider and a few transistors for +-12V. Something like that.
 
EmRR said:
They don't need particular loading
Ok. So you're saying that the Altec 9061A and Cinema Engineering 4038b specifically are not passive 600 ohm characteristic impedance or that the particular circuits just don't exhibit peaking if not loaded?
 
squarewave said:
Ok. So you're saying that the Altec 9061A and Cinema Engineering 4038b specifically are not passive 600 ohm characteristic impedance or that the particular circuits just don't exhibit peaking if not loaded?


Altec Engineering Notes, Technical Letter No. 192 -  'Amplifier Impedance Effects on the Transfer Characteristic of Filters and Equalizers.'  It's online. 

Bridged 'T' constant Z networks with limited action. 
Response curves are virtually unaffected by source and load Z.

The key is the limited action range of the bridged 'T' environment, it's always somewhat isolated from the outside world. 

I have corroborated this with multiple examples of Altec, Langevin, Cinema, and Hycor variants.  You really can stick a 600:10K transformer on the output to make up most of the losses, so long as load Z is 10K or greater.  There's no difference in the curves whether connected +/- and +/- to balanced inputs and outputs, or +/- to -/+, or unbalanced on one side and balanced on the other, so long as you don't leave a connection floating, as I discovered troubleshooting in the previous Cinema thread. 

https://groupdiy.com/index.php?topic=74726.msg946655#msg946655
 
EmRR said:
Altec Engineering Notes, Technical Letter No. 192 -  'Amplifier Impedance Effects on the Transfer Characteristic of Filters and Equalizers.'  It's online. 

The key is the limited action range of the bridged 'T' environment, it's always somewhat isolated from the outside world.
Ah, yes. I remember this paper. But like you say, it's only for filters with resistors limiting the Q like the 61A.

Is this true of the Cinema Engineering 4038b? The Internet is drawing a complete blank on that one.

Note that if you use a 600:10k transformer, accidentally plugging it into a 600R load would be a disaster. So that might not be ideal if it's supposed to be in the patchbay.
 
squarewave said:
Ah, yes. I remember this paper. But like you say, it's only for filters with resistors limiting the Q like the 61A.

Is this true of the Cinema Engineering 4038b? The Internet is drawing a complete blank on that one.

Note that if you use a 600:10k transformer, accidentally plugging it into a 600R load would be a disaster. So that might not be ideal if it's supposed to be in the patchbay.

They are all the same EQ, all Art Davis circuits.  Altec, Langevin, Cinema, and Hycor thingies that look like the same thingie.  Rackmount or console channel package. 

Response curves are virtually unaffected by source and load Z.  Probably not a disaster into 600, but it will certainly load.  If you know you're going to plug it into a Pultec.....then.....don't.....do something.....

Really, that’s an example of how simply it can be done. If you’ve got some other obvious way to deal with make-up gain, it doesn’t much matter how you go about it.  Tons of these EQ’s in people’s studios with no dedicated supporting electronics. 
 
EmRR said:
Tons of these EQ’s in people’s studios with no dedicated supporting electronics.

^ This ^

I wired up three Altec graphic eqs last year for a client.  Added bud boxes with XLRs and a bypass switch that kept the ~15dB loss consistent.  They can plug them into any preamp or line amp for makeup gain.  Works perfectly.
 

Attachments

  • Passive EQ Pad Switching small.jpg
    Passive EQ Pad Switching small.jpg
    76.2 KB · Views: 46
Thanks for all the comments!

Have never had to interface those Olde EQ's into a studio.

But, after decades of studio design work, I am now "gun shy" about the gear that shows up in a patchbay which may be randomly used by anybody at any time.  IOW, if only ONE person EVER uses the system, then they know whazzup.

Reminds me of an oddball situation when I was doing some work at the Hanson Brothers (yeah...the "MMM-Bop" guys) studio in Tulsa many years ago.    They were recording in/out via Digidesign 192 interfaces.

They added an API lunchbox with (IIRC) 550 EQ modules.

Long story short, the "Tascam Problem" Digidesign boxes did not play well with the unbalanced input API EQs in the patchbays.

6 dB signal loss and oddball crosstalk in the entire system.

I bandaided it with transformers into the lunchbox.

Hence my goal of bullet-proofing so that anything can be patched anytime by anyone without surprises.  I CANNOT know what a guest engineer may patch.

Bri


 

Latest posts

Back
Top