[b] API 312 Overall [/b]

[mikets]

API 312 -Some considerations


Ok, in search for the "perfect circuit" - Forselltech JFET990-2, inTx/Lundahl1636 (amorphous core), or 1538xl, outTx/ the ones we order with Fallout. Questions arise:

1. The ratio of 1538xl is 1:2.5 , or 1:5 for (200Ohm/10k)
details: http://www.lundahl.se/pdfs/datash/1538_8xl.pdf
The ratio of 1636 is 1:5 for (400Ohm/10k), 1:10 for (100Ohm/10k), 1:20 for (25Ohm/10k)
details: http://www.lundahl.se/pdfs/datash/1636.pdf
How they will change the schematics? Or they will just fit in straight away?

2. I still wonder how accurate the out trafos will be compared to orig 2503? Are there 3 sec windings!?

3. What would be the value of R4 and C5 in accordance to in/Tx Lundahl?
Can they be removed!?

4. With Lundahl's and good PS do we really need C6 and C7?

5.What would be the best value for Gain Pot ? (what type?)

6. What would be the most suitable PS Voltage in regard to best SN ratio and proper Gain
(in the Forssell schematics is set to +/- 24V) http://www.forsselltech.com/JFET990-2a.PDF

7. The PS and Phantom PS may be http://www.forsselltech.com/MikePrePS1.PDF
Any other suggestions?

8. What wold be the proper value of R3/C4 ?
In the Forssell schem of 990-2 as Non-Inverrting Amp they are 2k/10pF
http://www.forsselltech.com/JFET990-2a.PDF
in the orig API 312scheme - 20k/120pF

9. What are the BEST QUALITY resistors and Caps to use in the circuit? I know some will say the autentic sound..etc, but I think they didn't designed it to be Vintage at the time, and if it had to be designed today it would've been using the best parts and topology (after 30 years would still have this Vintage 2003 sound

10. How/where to connect 0dB overload LED or simple Vol.meter (or 3LED - power/signal/overload ) circuit? Any schematics?

11. Any other considerations in this regard!?!?! [img]graemlins/up.gif[/img]

[sginn]

Originally posted by Mike T:
API 312 -Some considerations
6. What would be the most suitable PS Voltage in regard to best SN ratio and proper Gain
(in the Forssell schematics is set to +/- 24V) http://www.forsselltech.com/JFET990-2a.PDF

7. The PS and Phantom PS may be http://www.forsselltech.com/MikePrePS1.PDF
Any other suggestions?

11. Any other considerations in this regard!?!?! [img]graemlins/up.gif[/img]

Isn't the PS requirement for the 1073/1272 also +/- 24V? I thought someone in an earlier thread said the PS requirements for the two units were different.

Steve

P.S. never mind, I just checked Thomas 1290 site and it is +24V not +/-. How many API312's could the above power? Could it be used to power a combination of API's and 1290s?

[rascalseven]

According to the 2520 specs I've seen (there's a copy over at Dan Alexander's site too, I think) the 2520 is meant for voltages of +/- 12VDC to +/- 20VDC max. Forssell's +/- 24VDC PSU will be too much.

api's current PSU's provide +/- 16VDC for the op amps in addition to the +48V for phantom power. I believe Brent Averill's PSU's are the same. Probably a good place to start, though I, too, wonder how different the performance would be with slightly higher voltages? Higher headroom, lower noise???

Joel

[mikets]

"...Forssell JFET 990-2 OpAmp is a high performance, wide bandwidth, low noise amplifier with a 2 megohm input impedance. It is designed to operate with a bipolar power supply voltage of 12 to 24 VDC per polarity. Power consumption is approximately 30 milliamps per polarity regardless of supply voltage. A low noise regulated power supply is highly recommended."

[BjornZetterlund]

I haven't really looked at this circuit yet, but I thought since I've ordered the output trannies, I might as well jump in. I hope someone will set me straight if I'm wrong on any of this.

RE. the input tranny...if you're using the JFET 990 op-amp you're looking at an input impedance of 2M, so a step-up transformer of say 1:7 should handle that just fine without loading the mic too much. For an op-amp with a lower impedance, you need to have a think about it; I think the Jensen twin-servo topology uses a 1:2 for least noise. A JFET is noisier anyway, so you may as well take advantage of its high impedance. That's just the electronics, though, what the sound will be I don't know. If you're really going for a no-compromise re-creation of the original, presumably you wouldn't be using the JFET?

R4 and C5 are there for the transformer, not the op-amp. With these components, you would need to follow whatever the tranny manufacturer recommends for loading. It may be used to provide some HF roll-off as well, maybe for RF, but I haven't worked out the 3dB point yet. Anyway, the emphasis here is that you need to load the transformer with whatever it likes to see. I assume C6 and C7 are for RF rejection. You could probably use ferrite beads here if you wanted to.

R3, C4 and the gain pot are just a normal gain-setting op-amp type of thing. The gain pot will probably work best if its a reverse log type. As far as the overall impedance of the potential divider in the feedback loop and the value of C4, that's where you need to consult with the manufacturer of the op-amp. First of all you need to establish how much gain you need, and how much of this is already provided by the passive gain of the transformers, and how much active gain is required by the op-amp. The overall impedance seen by the op-amp can perhaps be lowered if you use a 990. If you look at the Jensen designs, I think it's more like 2k rather than 20k+. Bear in mind that C3 has to be higher if the overall impedance is lowered in the feedback arm. I'm considering a DC servo to eliminate it, but we'll see.

As far as PSU goes; whatever the op-amp dictates. I think the JFET 990 will want 30ma per rail, and can handle +- 24V. Easy to do with modern regulators. Higher supply rails will get you more headroom, but probably not less noise. Perhaps even a little more noise as you're likely to have slightly hotter components, but probably won't make a difference. As far as noise goes, it's difficult to see how you would have a noise problem in this design. I you go with a JFET input op-amp you're already choosing sound over noise anyway. Perhaps it would be as quiet as a bipolar input if you have a higher ratio input transformer, but probably not.

The thing is, there really isn't much to this design! There are basically two transformers and an op-amp, so whatever you pick will be your sound. All the other components are dictated by the transformers and the op-amp you choose.

[img]graemlins/beerchug.gif[/img]

Bjorn

[elco]

FYI, my Brent Avrill 312's have 10K Lin pots instead of 1K, but considering the way the gain tracks, I think a 10K RLog pot would be best.

Cheers,

elco

[sginn]

So now that we have ordered the output transformers, what is the next step? Do we have a formalized parts list yet that we can follow to build our inventory for this project? Has a decision been reached as to the design of the project? I know the consensus has been that most people want a modular approach. However, there have been differing opinions such as whether or not the main card should have card edge connectors. I know everyone won't agree on all aspects of the project (for one reason or another), but it would sure help those of us who are less experienced at this stage of the project to have at least a definitive starting place.

Thanks,
Steve Ginn

[ArtCriminal]

Steve,

We're looking at nearly 175 channels of this project being built based on the number of transformers that are being ordered. With that large of an order it would surely make sense to maybe come up with some common board styles and have them mass produced. Perhaps an ExpressPCB order or something similar. Newbies would love that. The experienced aren't so keen to it because they want to make things the exact way that they want them. I'm not sure of where everyone else sits on the issue.

I started a parts list but have stopped as there are a few things that I need to learn. The current list can be found on my page but it is far from complete. There could actually be a number of variants on this circuit but I'm thinking that there will be three common ones.

Current API with 1122 opamp and Reichenbach/Jensen input transfo.

Modded API with 1122 opamp and Lundahl 1538XL input transfo.

Clean variant with JFET-992 opamp and Lundahl 1636 input transfo.

There are different needs for the 1122 and the JFET-992 opamps. This is mostly impedence oriented. I need to do some diggin' and figure out just what needs to happen. If anyone else has already done that then please step forward with some details. I'll post them up on my 312 project page for the world to see.

Then there's the pot/switch gain control debate. [img]smile.gif[/img] For me I think I've changed my goals up a bit. I'm going to make 6 channels of "Modded API" with two channels of "Clean." The Modded channels will probably be pot based and the clean channels will be switched.

Fabio has been working on a layout. I'll see if he can't share it with the rest of us. I don't know exactly what he's done with it though. I've gotta go look at it!

Peace,

James

[mikets]

You know, maybe I shouldn't started this topic, since there is one already, but I did so with the idea of summarizing the things out. As I thought after the long chat out there, here should've been the place to put the cleared out decisions and schematics, maybe I was wrong but...anyway.

Good as we can see the ideas are taking shape in the
"current, modded and clean" versions.

- As for the Lundahls, I found that both 1538xl, and 1636 do not need RC load at the secondaries.

- most probably each OpAmp would need different Gain correcton (Forssell JFet reccomends R3/C4 2k/10pF)

- The Gain Control variants so far are : 10kLog/Rlog, 10kLin+10k resistor, 25k or Switch

- The question about the Volume control is not yet cleared

- PS between 15-24V, we need definately 48v+ (and eventually some 5-9V for LED indication or some)

- Parts - Metal film Resistors, Polystyrene cap at the gain control ( how important it is ?!?)