API 312 Thread!

[jsteiger]

Mitsos,

From something Bo has posted:
About the cap, so if you only have a small cap in series with the signal patch, you get a high pass/lo-cut filter, togheter with the 1 Mohm resistor, (the op-amp +input ground reference) but with a resistor in paralell with this cap, you get a "treble boost filter", (same topology as the bright switch on a Fender amp :grin:  ) and with a selected value it will compensate the treble loss you get in the 2,2 Mohm resistor (when it alone act as a pad-resistor) because the capacitance in the signal patch and op-amp input.

When I was working on my first VP312DI proto, I did not have a 18pf cap around. I remember a roll-off on the top end without this cap at all. I use a Cornell Dubilier mica cap that works pretty well. More or less back to "flat".

Sorry I can't be more help on your howling problem. Good luck with that.

Best, Jeff

 

[mitsos]

I figured it was some sort of filter. This is very curious.. Guess I'm going to swap the plug next... there isn't anything else left on that little board...

 

[jsteiger]

I am not familiar with those boards but I'm sure there is a 1M or something reference R hanging just before the DOA on the non-inverting input. Probably after this DI board though. Yes, you are right. There is not much else there. Can't be a DOA problem as you said the problem follows the DI board. Weird and curious.

 

[mitsos]

Hi jeff, yeah the 1M is on the mainboard. The DI board only has the 2M2/18pF, 100nF cap, and the 2 zeners. That's it.  I attached a part of the schematic.  The control (ring of the TRS) is the ground path for the transistor that controls the relay switching between mic/DI.  I even swapped the TRS jack and it still does it... so the only thing left is the PCB! haha! wt?  I cleaned the PCB but nothing.

It seems like the output trafo is ringing to me. While it's doing it, I measured 58V AC between pins 2-3 on the output XLR... which goes to zero when it's not at absolute max gain and goes up to about 1.x volts when strumming a guitar (again, this is when it's not at absolute max gain).

Worst case I'm removing the RC, because that seemed to make it go away, most guitarists I know always want more bass, maybe the rolloff will be a good thing!  ;D

anyway, that's what's going on... anyone know what else it could be?

thanks!

 

[jsteiger]

Bo's original schem has a 10k series R in it too. In between the RC pad and the .1uF cap.

 

[mitsos]

Just saw that... but at this point, I removed the RC.  between the two channels, I think I like it better. Funny, this has more high end to me.  Will test it out over the next few days and see.

cheers!

 

[jsteiger]

Interesting stuff. Would be good to know the cause of the howling. I'm pretty sure you are happy just to have it gone.

FWIW, I have ran across a schemo from 1987 of the 3124. Paul had a 47.5k series R into a 10uF coupling cap. No RC pad circuit there of course but the same 1M ref R. I have no idea what they are doing these days. Might be the same...might not be the same.

Also, I have noticed some DC offset, depending on the DOA, when in Bo's DI mode. No biggie though cause I think you have a cap after your DOA? The 3124 has one as well bypassed with a 1uF, according to the 1987 schemo.

 

[mitsos]

care to share  ;D

I've got 88mV DC on the channel with the intact DI board, and 99mV on the one I removed the cap from.  ha!  Fabio's boards don't have a spot for an output cap, so these don't have it.  But they've been working for ages.  I was going to use some unused section of the board (like unbal out) to rig an output cap but it hasn't killed the trafos yet, so I don't think it will.  These trafos are a bit different, I think they were made by fabio, they are stacked 2x2 (or 3x3 not enough light to tell right now) but they're not 1x1 like most others (including the 2503) which may or may not help with DC offset.

Maybe Bo could decipher this.

cheers!

 

[jsteiger]

care to share  ;D...

Oh, you know I will!  ;D

I just need to take it to the office to scan it. There is nothing here that would surprise you. I already mentioned the most of it. The only other thing that surprised me a little was, well 2 more things. First was the 1000pF decoupling caps before the JE 115K-E. I thought they were mostly to help tame the old 2622. Next was the 10k R before the 6k8's for phantom. I learned that this needed to be more like 200. Oh, and last thing, there is no IC buffer before the LM3915. I am pretty sure I remember seeing one in 512 pics.

I'll take this guy to work with me tomorrow for some scan action.  8)

 

[mitsos]

Cool, it's always fun to see those schems!  I left 1000pF spots on my boards for Ed's trafos.  I've meant to try them with some Cinemags and Altrans as well but never did (yet). Maybe next weekend....

I'm curious to see the phantom circuit. Lately I use a 200R with a 68uF cap before the 6K8s. 

cheers!

 

[jsteiger]

Here's that 3124 skiz as promised. I am also gonna send it to the group email for archiving. Like I said, this is dated 1987 so I'm sure it has changed a little. The image I got was pretty light so I did what I could without fixing anything. I think it's pretty much all legible.

Enjoy!



Best, Jeff

 

[mitsos]

hey man, been almost a month.. forgot to say thanks for sharing!

 

[jsteiger]

No problem mitsos.

Did you sort out your DI problem?

 

[mitsos]

umm sort of.. just left it without the RC filter ;D

I had another DI issue though.. etched a "Bo 312 DI mod" with a relay and PNP switch(like fabio and dan kennedy have) onboard, wired it up and was working for 10 seconds, then all I got was loudass hum and the 47R on the output started to smoke. This was on my own 312 PCBs, which work fine with the guitar going straight to the opamp.  But I'm wondering if it was because of the 47R being in parallel with the output caps, and sending DC straight to the trafo.. I'll test some more this week.

btw, there is a thread somewhere that mentions the 47R in parallel... seems it's to cure a bump in the low frequencies from the LC filter created by the output cap and transformer winding.  Was going to post about this issue there as well, now just have to re-find the thread.

 

[mitsos]

Jeff and everyone else, here is the thread I referred to, in case you're interested to know why API uses a 47R in parallel with the output cap. See Abbey's and PRR's answers. Abbey even has some graphs.

http://www.groupdiy.com/index.php?topic=40582.msg502696#msg502696

I did get an answer to my new DI "problem" and it is in fact due to the high DC offset caused by the Bo DI mod.  I will try my home-etched board again, without the 47R in parallel to see what happens.

cheers!

 

[jsteiger]

Hey Mitsos,

Thanks for that link. I did find that earlier. Abbey is quick with Spice!    Very interesting indeed.

I found similar results when in Bo DI mode. It is also very DOA dependent from my testing.

I wonder how the DC offset would be if using the values from the 3124 for DI? They have no pad either.

People with a stock 312 card or a true 312 circuit should be careful when using a DI mod-circuit into the DOA. That is why I added the output cap on the VP312DI. No need to waste a good 2503.

Hey, maybe that bump below 10Hz is part of that nice API bottom end we all seem to love?!?!  ;D

Best, Jeff

 

[mrcase]

hi!

I´m building 4x Api312 with mitsos boards. how much VA would you suggest for the toroid?
would this be sufficient? http://at.farnell.com/multicomp/mcta060-18/transformer-60va-2-x-18v/dp/9530444
or better off with 80VA?

thank you!

 

[mitsos]

that one is way overkill. 30-35VA should be OK according to Harpo's math, which I trust more than mine. That said, I've got 2 channels running on 10VA and 2 others on 18VA, both running for a couple of months now, with no noise or heat from the power toroid

Of course, you also need to calculate the current consumption of the LEDs. V=IR -> R=V/I where V=Vrail - Vdrop of LED. Both green and red LEDs have a voltage drop of 1.7-2.6(at 20mA), I just use 2V for an average.  To give you an idea, I built my last ones with 3K9 resistors and they are the right brightness for me. It's about 2.5mA per LED with a 12V supply.  So, that's another 10mA per channel when all LEDs are lit. Pretty minimal but it adds up when you add channels or run LEDs brighter.

hope that helps!

 

[mrcase]

thanks a lot mitsos. that cleared things up!

 

[idylldon]

BOM for 312 with trafo output
BOM for 312 with 1646 output

These links seem to be switched.  When I click on the first one, I get the BOM for the 1646;when I click on the second one, I get the BOM for the trafo.  Maybe it's something with my browser, but I thought I'd mention it.

Cheers,
--
Don