Simple HT-312 (simple API312) with phantom power switch - cap value questions.

Rob H. (Handcrafted Tone) has a simple 312 layout with phantom power on board. Output transformer off board.
Project page is here.

The only thing I can't figure out is the values for C5, C6, and C7 which are not in the original API312 schematic.

Also, I can't figure out what R4 is for, and wonder if it should be left off. (Again, not in original schematic.)
Same with C7 - should I just leave it off? (It's a cap before the output - not in original - I assume not needed since I'll be using an  output transformer.)

Hi Kato,
C5+C6 in parallel to C1+C2 will probably be 100nF. R4 is needed for loading the input transformer. From pcb layout, the primaries are wired in parallel (a little mixed up in schematic). With the quoted cinemag CM75101APC this is a 1:10 ratio, setting the input impedance to 1k5 with R4 150K. The zobel network R8+C9 is probably not needed for this transformer. With output transformer, C7 could probably be shorted out. Expecting some offset from your DOA and without transformer I'd try C7 as 680uF with a 68R between DOA and cap, followed by a 10k to gnd for unbalanced out. YMMV.
regards
-Harpo

Harpo is just right!

Thanks Harpo (and 3nity.) That's what I thought.

I figured Rob was just including all the zobel options for people who choose to use other transformers instead of the CM75101APC.  It was just throwing me off because I expected R4 and R8 to be next to each other on the PCB. (It's an impressive, hard-for-me-to-follow layout.  I'm planning to use 150k for R8, jumper C9, and leave R4 out altogether - which I think accomplishes the correct loading for the CM75101.

Thanks!  Kato

I'd only use R4 and leave out R8 and C9. YMMV as your plan would have the same effect, but might lead to confusion if you try another transformer one day that might need a zobel network or other load resistor. If you follow the trace from transformer pin5, R4 is next to R8.

C1&C2 12.5uf ??? Where the hell he found that value? I read that like 10uf or 22uf 

Harpo said:

...If you follow the trace from transformer pin5, R4 is next to R8.

Click to expand...

Just you saying that made it all clear.
I'd been staring at it thinking it was in the wrong place - like my mind couldn't see that connection from pin 5 extend beyond R8. Now it's all quite obvious. Thanks.

Moby said:

C1&C2 12.5uf ??? Where the hell he found that value? I read that like 10uf or 22uf 

Click to expand...

The original API schematic. 
Almost everyone bumps that value today, now that caps are much much smaller (in size.)  Rob's just keeping it old school.  8)

I'll etch some boards this weekend and give it a go - if anyone's interested.

Almost everyone bumps that value today, now that caps are much much smaller (in size.)  Rob's just keeping it old school.

Click to expand...

i wouldn't search for NOS 12.5uf... After 40 years they are probably dried  ;D

Just thought I'd add my 1 cent (half-brained 2 cents, that is)..
I've been using/testing this layout for a bit now, it works well, but I've been playing a lot with the cap values. I'm almost 100% certain that I like it without any zobel network (with Cinemag and Altran trafos). I've been told that API now uses a bipolar lytic in parallel with a film cap and a 47R resistor both for the output (C7) and the cap in the gain loop (C3). (Check the Eisen Audio schem for a visual). This has been working better for me.  I also upped C1 and C2 values to 100uF.

Hey guys, it's been a long time since I've visited the forum, but it's good to see people are using my layout!

and yes, I need to go back and fix the transformer symbol in Eagle, the secondary winding should be reversed. I haven't had much time for DIY lately... new years resolution maybe???

mitsos said:

...I've been told that API now uses a bipolar lytic in parallel with a film cap and a 47R resistor both for the output (C7) and the cap in the gain loop (C3). (Check the Eisen Audio schem for a visual). This has been working better for me.  I also upped C1 and C2 values to 100uF.

Click to expand...

Those are some valuable cents!

You mean API and Eisen both use a C7/output cap - even with a transformer?
Perhaps I ought not skip that component, eh?

Handcrafted Tone said:

and yes, I need to go back and fix the transformer symbol in Eagle, the secondary winding should be reversed. I haven't had much time for DIY lately... new years resolution maybe???

Click to expand...

The PCB artwork works correctly though, right?

mitsos said:

Just thought I'd add my 1 cent (half-brained 2 cents, that is)..
I've been using/testing this layout for a bit now, it works well, but I've been playing a lot with the cap values. I'm almost 100% certain that I like it without any zobel network (with Cinemag and Altran trafos). I've been told that API now uses a bipolar lytic in parallel with a film cap and a 47R resistor both for the output (C7) and the cap in the gain loop (C3). (Check the Eisen Audio schem for a visual). This has been working better for me.  I also upped C1 and C2 values to 100uF.

Click to expand...

More thoughts about C3 and C7.
Is there audio signal going through the gain loop? C3? (I guess I don't really get how the gain knob works.)

I checked the Eisen schematic at your suggestion. He puts a lot of stuff 'tween the opamp and output transformer.
Two resistors and two caps, and some kind of interference suppressing inductor (?) see part marked 'OLI.'

kato said:

More thoughts about C3 and C7.
Is there audio signal going through the gain loop? C3? (I guess I don't really get how the gain knob works.)

I checked the Eisen schematic at your suggestion. He puts a lot of stuff 'tween the opamp and output transformer.
Two resistors and two caps, and some kind of interference suppressing inductor (?) see part marked 'OLI.'

Click to expand...

The DOA is a non inverting gain stage. Voltage gain is set by 1+(R3/(R2+VR1)).
With R3=20K, VR1=~20K neg.log, R2=200R you get voltage gain varying from 2 (VR1 ccw=20K) to 101 (VR1 cw=0R). Make this a dB reading is log10(voltage gain)*20 = 6dB to 40dB. The non variable part of 20dB gain is comming from your 1:10 input transformer and maybe 6dB from whatever your maybe 1:2 output transformer is, giving a total range from 32dB up to 66dB. (a pad in front might be useful).
C3 keeps your pot VR1 from DC current to prevent it becoming scratchy. Unfortunatly this 250uF cap also forms a HPF with -3dB cutoff at 1/(2*PI()*R2*C3) = 3.2Hz. This introduces a phase error of -ARCTAN(3.2Hz/20Hz)*180/pi() = -9° at tested frequency 20Hz.
The Eisen Audio seems to have some optional parts, f.i.your C7 is his C9, bypassable by conn4, transformer out or bypassed by R19/R20, secondary termination, ..
Some good reading about 990, OLI, servo, .. might give this page by John Hardy.

Handcrafted Tone said:

Hey guys, it's been a long time since I've visited the forum, but it's good to see people are using my layout!

and yes, I need to go back and fix the transformer symbol in Eagle, the secondary winding should be reversed. I haven't had much time for DIY lately... new years resolution maybe???

Click to expand...

Primary winding in schematic is mixed up. Both middle taps connect to the opposite side, wired in parallel, not shorted doing nothing. The reversed dot on secondary winding is only a phase turn.

Wow! Harpo, thanks. Super helpful information. I can really learn something from posts like these.

I've read through it three times and I'm getting a better grip each time.
So I'll read it a few more times before I post any daft questions.  Just wanted to say thanks.

Also, for anyone else interested in how these parts work together. Eisen has helpful explanations of each component - from his kit - so you'll need to transpose part designations with whatever 312 variant you're building, here:

http://www.eisenaudio.com/products/diy500mkii/temp_components.html

Here's a summary of what Eisen says:

Output cap: C7 on Handcrafted Tone (HT), C9 on Eisen.
"...should always be 470uF or greater (1000uF if you can find a part that fits). Some op amps demonstrate no DC offset unless they fail, in which case you can leave out C9, but we don't recommend it."

still debating this cap, even after this warning. I know any cap in the signal path will add some phase shift. but do i want to risk wasting an output transformer in the event of op-amp failure.  or even consider the dc trickle, can that magnetize a transformer over time? Still, that cap was not in the original schematic. I assume hundred of channels were happily used in studios without incident. Fabio's version doesn't use an output cap at all.

Output cap's bypass cap: NA on HT, C10 on Eisen.
"Normally N/A. Would be 0.1uF same as above, but should be left out under normal circumstances."

Gain control pot: C3 on HT (250uF), C6 on Eisen(470uF), bypassed with C7 (.1uF).


And Rob, thanks for the layout.  [insert beer icon here]

kato said:

Output cap: C7 on Handcrafted Tone (HT), C9 on Eisen.
"...should always be 470uF or greater (1000uF if you can find a part that fits). Some op amps demonstrate no DC offset unless they fail, in which case you can leave out C9, but we don't recommend it."

still debating this cap, even after this warning. I know any cap in the signal path will add some phase shift. but do i want to risk wasting an output transformer in the event of op-amp failure.  or even consider the dc trickle, can that magnetize a transformer over time? Still, that cap was not in the original schematic. I assume hundred of channels were happily used in studios without incident. Fabio's version doesn't use an output cap at all.

Output cap's bypass cap: NA on HT, C10 on Eisen.
"Normally N/A. Would be 0.1uF same as above, but should be left out under normal circumstances."

Gain control pot: C3 on HT (250uF), C6 on Eisen(470uF), bypassed with C7 (.1uF).

Click to expand...

Depends on your output transformer (you never told us). A DC gapped one will prevent its core from polarization.
Fabio uses either a AP2503 transformer, a servo circuit to cancel out DC-offset or an output cap (CU1+CU2 220uF each, could also be 100uF bipolar) for unbalanced operation, so he took this into account. Bypassing this cap with a 100nF might help for better sleep, you probably won't notice, but who knows.
EisenJens uses C6 470uF with R13 100R. Handcrafted Tone's C3 250uF with R2 200R will give the same HPF -3dB cutoff point. Only the lowest resistor value in the shunt arm (R2+VR1) is important, as cutoff decreases to nearly DC 0.0315Hz with VR1 full cw.
In your case more important is the difference of Eisen C8 100pF across feedback resistor R12 10K compared with Handcrafted Tone C4 120pF across R3 20K.
This forms a LPF with -3dB cutoff @ 1/(2*PI()*R3*C4)=66.3kHz (Eisen @ 159kHz) in order to keep your DOA (you never told us) in stable operating areas.
You might pretend you're no bat, but the phase error at 20kHz is ARCTAN(20000Hz/cutoff frequency)*180/PI()=16.78° compared to Eisen 7.15°, so this might sound different. 

My parts:

I'm using input transformer: CM75101APC, opamp is a 2520 clone from fabio's group buy a couple years ago.
output transformer: CMOQ-2S - steel core from cinemag - not gapped as far as I know. I also have some Profile 4804 that I could use. But I think that's pretty much similar to the CMOQ-2S. I also have a pair of 990c from John Hardy but I was planning to save them for something more hardyesque / jensen twin servo-ish.

I am no bat.  I lose everything above 14k.  (Too many years of standing next to the drummer I guess.)
But I'm open to the possibility that shifts at 20k can manifest themselves as tangible effects down in the audible spectrum.

I'm also not that picky about slight changes in sound for practical recording. Here is a clear example of a newbie (me) reading that "caps in the audio path are bad." So I assume wrongly they should be avoided at all cost. I guess I just need some experience A/B listening with my own ears. So I'll go with C7. Perhaps I'll try the bypass caps too to see if they make a difference just so I can have an educated opinion about such things.  I'll see if I can weasel a couple small film caps into the layout before I etch.

Harpo, thank you.  I appreciate your detailed response. It's very kind of you to share your knowledge, and it makes this a much more enjoyable and worthwhile (learning) experience for me. It helps me to understand that there is a well thought-out purpose for each of these component values, and that I can't just sub out parts without a reason.

Found the "OLI" part on Eisen schematic at Jensen:
http://jensen-transformers.com/datashts/oli3.pdf

$4.65 each. Listed in datasheet above as equivalent to an inductor and a resistor in parallel.
I wonder if I can save a few bucks and just wire a resistor and inductor in parallel.  ???

Sure you could. In 80's everybody wound trafo-wire over 10 Ohm/2W resistors for amps here..

I don't think you would need a 2W resistor for the diy "OLI" though.

Great thanks - I'll just try resistor / inductor combo - or magnet wire around a resistor.

On another topic - I'm looking into adding a PAD to this circuit before etching.
I found the Jensen Standard Microphone Preamp Input Circuitry Schematic and it looks like it's as simple as adding three resistors. But the switch symbol as it's drawn doesn't make sense. Or maybe it does. (Looking at it right now.)

Is it, all three resistors in circuit = 20dB pad, bypass all resistors = no pad?
Or (as it appears) the two 619R resistors ALWAYS inline, but the pad switch just engages the 169R resistor?

Anyone know off the top of their head?  Kato

When the switch is in bypass position (toggle direction upwards or DPDT pushbutton disengaged), resistors R7 + R8 are shorted out and shunt resistor R6 is out of circuit, so no pad is engaged. When engaged, these three resistors (an unbalanced signal would only need two resistors) form a input voltage divider.