Questions for API 312-esque circuits - gurus??

GroupDIY Audio Forum

Help Support GroupDIY Audio Forum:

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

TomWaterman

Well-known member
Joined
Jun 4, 2004
Messages
1,151
Location
The Shire, UK
Hi!

I want to implement an output volume control trim to my API style pres and although was told back at TT that a switched output pad would work, I have been looking at the Seventh Circle A12 schematics as this unit will implement an output control differently.

Schematic is here:

http://www.seventhcirclestudios.com/SCA/A12/docs/IO.pdf

I have a few questions concerning this layout....

The two resistors R15 and R16 (series 10k resistors) in the feedback loop? replace the single 20k as seen in the other schems like Fabios etc. Why is the second series resistor (R16) a trim pot? What benefit does this add to the circuit - trimming gain stabilization??

The output trim I'm guessing is the 10k trim pot (R19)?? Unfortunately I'm not understanding that part of the circuit very well.....can anyone help with an explanation of that, the use of the parallel diode and why that part of the circuit is connect to B+ and B- though 22k resistors, which I am presuming is the +- power rails??

Thanks for any help with this - sorry for the newbie questions...

Cheers Tom
 
Well, I?m not a guru, but?

You are correct, R15 and R16 are in the feedback loop and do replace the single 20k as seen in the other schematics. Also, the second series resistor (R16) is a trim pot. This lets you hit all the in-between spots so you are not limited to the gain switch settings.

R19 is used to adjust the input DC offset voltage so the output DC can be set to ?0?. The LED drops a ?fixed? voltage and sets up a constant current through R17 and R18 so that the DC offset will not fluctuate with changes in the temperature. Jensen used a similar circuit (using a 1N914 instead of an LED) for the 990 that only had a positive voltage adjustment. Apparently 7th Circle saw a need to have the adjustment swing negative as well, which is probably why they used the LED ? to get twice the voltage drop.
 
Thanks Tim, just the sort of info i was looking for.. Which brings me to my next question. I'm confused on how he has his rotary switch hooked up. I thought when building a shunt type attenuator the resistor values would be from lowest value to highest value. with one side of each resistor going to ground?

-Jay
 
[quote author="fallout"]I'm confused on how he has his rotary switch hooked up.[/quote]After a quick look, I'm not sure if I understand it either. I'll give it a closer look tomorrow when I'm more awake if someone else doesn't chime in by then.
 
Superb thanks very much Tim!!! :grin: :cool:

So that 10k pot R16 is not actually an output control but a trim control so that you can get more ranges of gain from the limited 5 db/step rotary.

Well I'll be using 23 position switches for mine. ...maybe I won't need it?

How much range can you tweak with that trim pot?? about 6 dB I guess - by looking at the way he has implemented it in his other designs.

If I had small steps such as 2 dB I'm not sure it would be worth it to use one....but I'm really interested in being able to drive the preamp hard without hitting my AD converters with too much level.

Is the reason the output DC is kept at zero in the 7th circle design something to do with the output transformer performance?

Cheers Tom
 
[quote author="TomWaterman"]...Well I'll be using 23 position switches for mine. ...maybe I won't need it?...How much range can you tweak with that trim pot?? about 6 dB I guess...If I had small steps such as 2 dB I'm not sure it would be worth it to use one...[/quote]
Correct, the trim always gives you 0-6dB of gain range because it goes from 0 to 1 x the value of the fixed feedback resistor. You won?t need it with a 24-step/2dB gain switch.

...Is the reason the output DC is kept at zero in the 7th circle design something to do with the output transformer performance?...
All designs need to keep DC off the output whether a transformer is used or not. Since output transformers can tolerate a small amount of DC consisting of a few mVs and block it from reaching the output, the method used here works just fine. However, it?s probably not precise or reliable enough if there?s nothing between the op amp and the preamp?s output. (Would you want to be libel for blowing up the inputs on someone?s $$$$ A/D converter?)

For non-transformer outputs, either a capacitor or a DC servo circuit needs to be used. This is better insurance for keeping the output at ?0? volts DC.
 
Thanks for the info Tim.. Do you think it would be wise to add the DC offset adjustment to the API312 circuit? The reason I ask is because the original 312 schematics dont include this. Would the choice of opamp have anything to do with this?

Thanks!
-Jay
 
[quote author="Flatpicker"][quote author="fallout"]I'm confused on how he has his rotary switch hooked up.[/quote]After a quick look, I'm not sure if I understand it either. I'll give it a closer look tomorrow when I'm more awake...[/quote]Well, I?m awake now (I think?) and looks like for part of the gain control SW2a is attenuating the signal by changing the resistance in parallel with the input transformer (understand that it?s not using both transformers simultaneously, it?s showing them like this because the PCB will let you use one or the other). For the other part of the gain control, SW2b is changing the resistance in the op amp gain circuit.

As far as the resistor values not being in incremental order on the drawing, I can only attribute that to the switch not having it?s positions associated with it?s pin numbers. If you could get the Grayhill switch part #, you could look it up and see.
 
[quote author="fallout"]...Do you think it would be wise to add the DC offset adjustment to the API312 circuit?...[/quote]
I would. Doesn't hurt a thing, doesn't cost much, and it's there if you need it.
The reason I ask is because the original 312 schematics dont include this.
Save a nickel, save a dime... :wink:

Would the choice of opamp have anything to do with this?
Definitely! Some op amps exhibit more DC offset than others. Another good reason for the DC nulling circuit - you might want to try different op amps later.
 
Great stuff Flatpicker - thanks again for sharing your knowledge.

What opamps are people using for the pres??

I will be using the Melcor 1731.....possibly the JFET992s at a later date in another version.

:guinness: :sam: :guinness:

Tom
 
[quote author="Flatpicker"]Well, I?m awake now (I think?) and looks like for part of the gain control SW2a is attenuating the signal by changing the resistance in parallel with the input transformer (understand that it?s not using both transformers simultaneously, it?s showing them like this because the PCB will let you use one or the other). For the other part of the gain control, SW2b is changing the resistance in the op amp gain circuit.[/quote]

Thanks again Tim! I'm definately going to go w/ the basic shunt type attenuator for my 312, because it's a lot easier for me to understand. However, do you see any advantage to doing it the way SCA does theirs? Also, you are correct, I checked the datasheet for the Grayhill switch and now the resistor values make sense! You rock! :thumb:

-Jay
 
Hooo boy! I got my mojo working! I need to find my wife before it wears off! :green:

Go to http://www.johnhardyco.com/index.html , go to the ?Products? page and read the pdfs on the 990 and the M1. These are full of schematics, application notes, and lots of good stuff. Definitely a must read!

...do you see any advantage to doing it the way SCA does theirs?
I was just discussing this with Gus last night on the phone. Some op amps are not very stable until you are running them at 15-30dB of gain. Couple that with a 1:5 to 1:10 ratio input transformer and the gain ends up being way up there before even turning up the gain switch. In this case you can use shunt resistors across the input transformer to attenuate the signal back down to a good starting point so you don?t over load the op amp?s input with those hot condensers. The SCA circuit has 4 attenuation steps before it starts increasing the gain. It's really not that complicated at all. I'm not sure how stable the Melcor is at low gains so it really depends on that.
 
I haven't had a look at this 7th Circle board, but this Grayhill switch is an adjustable-stop switch which is adjusted by inserting one or two pins into the front of it to determine where it stops. They're using all of the positions, so the one pin is inserted between the "high" setting and the next clockwise position which is the "low" setting. The pin prevents you from turning "down" the gain one step too far and landing at maximum gain. That's all pretty obvious, but the thing that's not immediately obvious is that Tim mounts his circuitboards vertically, and if you want "min" and "max" positions to be where you expect them to be on the faceplate, then you need to put the pin between pins 3 and 4 rather than between pins 12 and 1. It's just rotated 90 degrees. We can see that on the input attenuator portion of the switch (SW2a), positions 1,2,3, and 8, 9, 10, 11, 12 all have zero attentuation, but the other positions all tap into the series string of resistors that adds up to about 78k ohms. If you think about it, this is exactly like a potentiometer that has only 8 discrete positions to which it can be turned. Let's call this attenuator a "series connected" resistor string.
The gain control (SW2b) on the other hand applies a different resistor from the amplifier's inverting input to ground for each gain setting (although positions 17-21 all use the same 7k5 resistor because they all use the same amount of amplifier gain). Let's call this gain control a parallel-connected resistor set.

R16 will in fact provide a 6dB trim of the total gain. This is a lot better than a full attenuator or a gain pot because the entire range of rotation is used for only 6dB of range, which makes the pot much more accurate. I think it's a very good idea because on the one hand you can eyeball two channels and be sure they're matched to within a small fraction of a decibel. On the other hand, if you're using a "matched" pair of ribbon mikes that sound the same but have output levels that differ by a couple of dB, it's also easy enough to match them up with these pots if necessary. For another thing, if you find you need to adjust the gain slightly in the middle of a performance, you don't ruin the take with a popping gain switch. It's really the best of both worlds. The Great River MP-2NV has both a stepped switch and a partial attenuator, as do a lot of other very useful preamps. It costs a bit more to implement but it's worth it. If you're going to have 2dB steps on your gain switch, you could easily use perhaps a 5k trim pot in series with a 15k feedback resistor, for a total of somewhere in the neighborhood of 3dB trim range.

The DC nulling circuitry would be fairly simple to add to your own circuitboard layout, but would be rather difficult to add to an existing API card. If you're doing your own layout, it makes a lot more sense to use a DC servo amplifier to null the DC rather than this crude trim. A servo samples the actual DC error and corrects it, rather than applying a pre-set correction that may or may not be correct at any given moment. There are plenty of simple servo circuits around that you can find. The one in John Hardy's literature is good, and the one in the Symetrix SX301 compressor is also a good place to start. Compare them, and try to figure out how they differ and why, then you should be able to toss $2 worth of parts into your own project to eliminate that pesky output capacitor. Or you can just buy a decent output capacitor and not worry about it. Either method works fine.
 
Hi Ulysses!

Thanks a lot for the explanation - I think I am actually warming to the idea of adding a trim control - even if I will use a fine resolution stepped attenuator as well. My thinking behind it was because of products such as the GR-2NV etc.

Would adding a DC servo to an API 312 (new board layout) be kind of 'against the grain'?

I mean its supposed to be a colourful preamp - would it remove a touch of the magic of the API?

It seems to me it would make more sense if it was a build with Hardy 990s or JFet992s - you know, a cleaner preamp?

Could anyone explain what the R23/C9 - R22 network is on the secondary of the 2503 TX? and the little (inductor/resistors) ferrite bead at the primary - isolator??

Thanks again Tom

Flatpicker - I'll read through the Hardy PDF - cheers
:guinness:
 
[quote author="TomWaterman"]Would adding a DC servo to an API 312 (new board layout) be kind of 'against the grain'?...I mean its supposed to be a colourful preamp - would it remove a touch of the magic of the API?...
...Could anyone explain what the R23/C9 - R22 network is on the secondary of the 2503 TX? and the little (inductor/resistors) ferrite bead at the primary - isolator??...[/quote]The servo doesn't have anything to do with "color" - it's just another way to null the output DCV. Personally I wouldn't use the servo with a transformer coupled output unless I wanted the option of switching the transformer out of the circuit, but that's just me. Also, I've heard some folks say that a servo messes up the low-end but I don't know that for a fact. You can bet there's always going to be tradeoffs, though. The "manual" DC nulling circuit and transformer coupled output should be just fine for an API type circuit, but feel free to experiment as that's the spirit of DIY!

R22, 23, and C9 looks like compensation for the output transformer, but I'm as puzzled as you because I didn't think it needed compensating. Anyone know about this?

Read about the JT-OLI-3 here:
http://www.jensen-transformers.com/datashts/oli3.pdf
 
I've finished my new layout for API312. Well, it ended up being a little more that a API312 board. Among other things it has 3 different output options:

- original transfomer out
- transformeless with caps
- transformeless with servo circuit.

Motivation: lots of folks are getting Forssell 992 opamp (myself included) so I thought it was good to have this option. The space was there, empty anyway....

I can post the untested version in a couple of days if you like.

cheers!
Fabio
 
Hi Tim.

Yeah I thought that a servo should be inaudible but I also thought it may also clean up the sound a little? and hence make it less API-like i.e the colour.

I don't think I will bother with a servo as I'm using 2503's on the outputs.
I also didn't think the 2503 needed compensating....but I have seen similar things elsewhere, I wonder if Tim from 7th circel knows about this board...maybe he could shed some light???

Does anyone have any information on servo design and implementation...there was lots at TT but....??? I would like to understand more about their design so that I can figure some out for my JFET992s for another design.

Fabio...it sounds interesting. I'm sure everyone would love to see it!

Cheers Tom
 
What with the JLM cards, I imagine a bunch of people will be doing something twin-servo-ish. Since I'll have a bunch of 990's and 992's on hand, I'll probably do at least one. The interesting thing is that Joe designed the board with his JLM99V opamp in mind, and I get the notion it's pretty colorful. His input here would probably be very illuminating. Joe?

Bear
 

Latest posts

Back
Top