Question about console out/monitoring amp input. Pads. Bridging impedance etc

[JW]

In another post I had going on in the Lab, someone recommended I build a pad (for a different reason, which is beside the point right now) But anyway, it got me thinking about pads. U/H/L etc.

Specifically, pads that I have in a very critical area in my studio which I don't totally understand.

Let me explain. I have a console where I consistently monitor from one of two sets of output busses (depending how I'm feeling) One set of busses is 600 ohm transformer balanced outputs (PM2000 main outs) and the other is the CAPI ACA buss, which is like an API summing buss. Anyhow, that is also a transformer balanced output and I believe Jeff Steiger said at one point they were around 50 ohm output impedance (? Don't quote me.) Regardless, these are both balanced transformer outputs at 600 ohm or less output impedance.

My speaker amp is an Adcom 535. The manual says this has a 22k input impedance. This is an unbalanced input.

What I've found is that when I get the console at a good level for feeding a 2 track mixdown machine, the monitoring level is way too loud and overloading the Adcom amp.

My solution for awhile has been to use some inline pads between the console and the amp. These are "U pads" with a pair of 680K resistors, each in series with the balanced output leads (pins 2 and 3 of the console's transformer outputs), strapped with a 150ohm in the middle. This is supposed to be a 20dB pad.

This is where my question comes in. I built these pads a long time ago and now I go and look at the page that I gleaned the info from ( http://www.uneeda-audio.com/pads/ ) and I realize they were in reference to balanced microphone levels. So,  these would  go between say, a 150ohm mic output before being plugged into a mike preamp of something like 1.5K input impedance.

Also to be noted, I'm going from a balanced transformer output into an unbalanced high impedance input.

So, I think, first, I'm probably wrong to use a U pad anyway, because this is technically an unbalanced run. After the pads I have the cold pin connected to ground for the unbalanced input of the amp.

Can someone enlighten me as to how to calculate a pad for this sort of setup? (Between a balanced 600 ohm out and an unbalanced 22K ohm input)

The existing pads have definitely been padding the signal down, but I want to verify whether they're right or wrong or neither, and also, I think I want maybe about 25 to 35 percent less padding than what I have right now.

???

 

[radardoug]

Your pad values must be different not 680k.

 

[JW]

Okay, thanks,

Do you mean different because I want less attenuation, different because we're dealing with 600ohm into high impedance, or different because of the unbalancing that's taking place?

 

[radardoug]

That pad would be more than 60dB, so have you got the wrong values?

 

[ruffrecords]

I think you meant the pad was a pair of 680 ohm resistors and a single 150 ohm, that will make a 20dB pad. The attached schematic shows it. This is exactly what I use for a 20dB pad in my mic pre designs. It works just as well at an output. The amp driving it sees an impedance of about 1K5 which any amp should have no difficulty driving.  Its output 'looks like' it comes from a 150 ohm source. It is also  balanced.

Since both your outputs are transformer balanced and should therefore be earth free there should be no problem connecting the pad to an unbalanced input.

Active monitors always tend to be too sensitive. I am not sure why this is but it may stem from the old days when separate power amplifiers were used which were often set to 0dBu for full power output. Back in the 70s at Neve, monitor signals were usually derived via a 10K:600 step down transformer which also reduced the level by 12dB. A 600 ohm balanced pot could be added to the secondary side as a monitor level control.

Cheers

Ian

 

[ruffrecords]

I thought you might appreciate some info about source and load impedances. In the old days everything was 600 ohms. Outputs actually had a source impedance of 600 ohms and inputs actually had a 600 ohm input impedance. Everything was hunky dory and you could connect any two bits of gear together and be confident they would work.

There were two problems with this approach. First, every time you connect two bits of gear together you lose 6dB of level which means there needs to be an extra 6dB of gain in the system. This raises the noise level by 6dB. Secondly you can only connect one output to one input. If you want to drive several inputs from one output you need to go via another piece of kit called a distribution amp which has one 600 ohm input and several separate 600 ohm outputs.

The reason things were this way was because it was all derived from telephony which needed to work over long distance and it was essential that the driving and driven impedance were the same because transmission line tells you that if you don't do this you get nasty reflections that destroy your signal. For outside broadcasts there was/is still a need for properly sourced/terminated 600 ohm lines but over the short distance inside a studio it is not necessary.

So sometime in the 60s/70s studio gear began to change. Instead of adding resistance to make the output impedance 600 ohms,  the output  impedance was whatever the design produced - usually in the region of 50 to 100 ohms. Input impedance was also increased to 10K. This meant there was no longer the 6dB loss between connected gear. It also meant you could drive several inputs from one output without the need for a distribution amplifier. This simplified studio interconnect and improved noise levels.

Cheers

Ian

 

[JW]

Oops,
Yeah I meant 680R, not 680K,

Thanks Ian! All that helps a lot. I forgot to mention that additionally I do occasionally monitor after my digital recorder's outputs which are not transformer isolated. They're electronically balanced (less that 50 ohm output) Would the lack of transformers change the situation with the pad then? Should I then try an L pad?

Regardless of what kind of pad, I'm gonna try 12dB attenuation and see if that gets me more in the ballpark for where I want my amp.

 

[ruffrecords]

A lot depends on the design of the electronically balanced output. Some are designed to sustain a short caused by connection to an unbalanced input but others are not. If you use a pad then neither of the outputs will be shorted; instead they will be grounded via one arm of the pad. So as long as the output is rated to drive one arm of the pad it should be OK and the attenuation will sril be the same.

I wonder why you have used a fixed pad. Why not use a variable one and use it as a monitor level? If you replace the 150 ohm resistor with a 1K log pot, then at the centre position, the pot will give between 10 and 20dB of attenuation depending on actual pot law.

Cheers

Ian

 

[JW]

Thanks Ian,

The answer to that is because I already have a 10K log passive pot on the input, inside of the Adcom amp. That amp had no attenuator so I wired one at the inputs. Sorry, another failure to divulge pertinent information.

So, that 10K pot usually sits around 2 o'clock or it doesn't sound as good (It's a dual gang cheapie) which is where the pads come in handy.

I have a conductive plastic alps blue velvet pot on the api buss that I really like the sound and feel of, so I'm intending to replace that cheap pot inside the adcom amp with one of those as well.

So the situation is the inline pads + the volume pot at the input of the amp are serving the same function, to attenuate volume. So, I should probably just combine these?

As in, essentially the schematic you posted above, but with a 1K pot (a nice one hopefully) in place of the 150 ohm, and I scrap the the inline pads as well as the 10k pot?

See, those blue velvet pots are expensive, so I only wanna buy it once. I can replace the 680R resistors if the listening volume is still not right.

Thanks for all your help Ian.

 

[gridcurrent]

The problem (s)  would be resolved if you installed appropriate step-down transformers on the input of your amplifier.

 

[JW]

Something like 10K:150 ohm? I don't want to color the signal much at this stage, so they would probably need to be very flat and expensive.

 

[ruffrecords]

So the situation is the inline pads + the volume pot at the input of the amp are serving the same function, to attenuate volume. So, I should probably just combine these?

As in, essentially the schematic you posted above, but with a 1K pot (a nice one hopefully) in place of the 150 ohm, and I scrap the the inline pads as well as the 10k pot?

See, those blue velvet pots are expensive, so I only wanna buy it once. I can replace the 680R resistors if the listening volume is still not right.

Thanks for all your help Ian.

Getting the right pot for a monitor control is not easy. You really want a log law but these are difficult/expensive to make such that the channels of a stereo one will track. For this reason, at Neve back in the 70s we never used log rotary pots at all (we did use P&G slider faders). OTOH linear pots tracks quite well so at Neve we used slugged linear pots for monitor level controls (when we didn't use stepped ones) If you slug a linear pot  with a resistor  equal to half its value connected between the slider and the anti-clockwise end you will get an approximate log law which is 10dB down at the centre point. At Neve we used a 2K lin pot slugged with a 1K resistor so that with the level fully up, the load was 2K in parallel with 1K  which is about 666 ohms. You could combine this with the two 680R resistors so it would be safe with an electronically balanced outputs. And a stereo one just needs a dula LIN pot. Sketch schematic attached.

Cheers

Ian

 

[JW]

Very cool. Thanks Ian. I'm gonna give that a shot. So, do I tie the shield (coming from pin 1 of the console) to the negative input at the 1K resistor, or do I just float the shield?

Edit: Well hell, it's actually pretty hard finding a 2K dual gang linear pot. Uh jeez. You're spoon feeding me  here Ian. Can I do the same thing with a 1K or 10K pot?

Maybe I should be thinking about a 12-24 position switch. . . .

 

[Monte McGuire]

What you're building is known as a 'monitor controller'. Such a device is designed to switch among several inputs and provide the ability to maintain high bus levels while being able to drive one or more modern, high gain power amps with the right amount of attenuation.

I don't want to discourage you from solving your monitoring problem using only pads, which are nice and clean, but at the end of the day, it is nice to have a control room pot, a mono switch, and some means to calibrate the levels from various sources and to various destinations. You can cobble this together with various pads and the like, but this is not a new problem, and your solution is not generally what people do.

While you don't have to buy a possibly over-complicated monitor controller, it'd probably help to look into monitor controller designs, and not just pads. Many of the issues you'll run into have been examined in great detail, and this will help you to solve your problem better and more quickly.

Best of luck!

 

[ruffrecords]

A 1K pot slugged with 510R might be too low. With the pot fully up its resistance will be 338 ohms. If we change the 680 resistors to 620 then the minimum attenuation will be 338/(338+620+620) = 13dB. If that is acceptable then 1K will be OK.

Alternatively you could try 5K slugged with 2K4 which would only attenuate the signal by 5dB with the pot fully up.

You should treat the whole thing as balanced right up to the point where it has to become unbalanced i.e at the amp input. So the screen goes up to the pot and continues tot eh amp where at that point is gets connected to the unbalanced ground.

Cheers

Ian