mhelin

OEP Z21805 app note
« on: October 07, 2010, 02:59:10 AM »
Another newish OEP transformer for mic pres. The mic pre in data sheet first looks like a conventional transformer followed by two opamp gain stages except that the first stage is quite an unusual. Could someone explain how it works (both inputs of the first stage's opamp are connected to Vee, and it seems like the Vcc and Vee pins are used as inputs), and what is the function of S1?

http://docs-europe.origin.electrocomponents.com/webdocs/0d72/0900766b80d72895.pdf

« Last Edit: October 07, 2010, 03:03:10 AM by mhelin »


Kingston

Re: OEP Z21805 app note
« Reply #1 on: October 07, 2010, 03:34:09 AM »
S1 is just a switch to turn off phantom for mics that might be destroyed from it. Most ribbon transformers can't take phantom without dying.

But the use of NE5534 here is one of those rare cases where they take advantage of the compensation pins. Usually this is done to skip the "weak" NE5534 output design, or boost it for headphone use for example, but oddly enough this time they use the pins to skip the differential input instead! Looks to me they take advantage of the transformers CMRR capability, skip differential input with only Q3 (single ended to input to pin8).

There are some interesting headphone amps with the output skipped but this is a real oddity you found. See NE5534 datasheet for the schematic overview. You can see where the pins are connected to.

I can't find the thread now but here on the forum PRR was talking about a headphone amp with the compensation pins used to skip the outputs. Also on the diyaudio forum if I remember right.
« Last Edit: October 07, 2010, 03:52:42 AM by Kingston »

gyraf

Re: OEP Z21805 app note
« Reply #2 on: October 07, 2010, 05:10:58 AM »
Most ribbon transformers can't take phantom without dying.

This is a common myth in our business, but not true.

As long as the mic is attached through a balanced connection, there will be no problems. We use plenty of ribbons in our studio, and our old SSL4044E had no phantom-off switch. Never had a problem (and shouldn't have a problem).

As for S1(a,b,c,d), it is the Mic/Line switch - when "line" the transformer is sorta in current- or "zero-field"-mode.

S2 is the Phantom power switch.

Jakob E.
..note to self: don't let Harman run your company..

Kingston

Re: OEP Z21805 app note
« Reply #3 on: October 07, 2010, 05:24:28 AM »
S2 is the Phantom power switch.

whoops. I was reading the simplified version on page 3. I should have mentioned that.

S1 is kind of pointless on page 2, isn't it? We already select input type with the actual physical connector. If worried about R15-R16 affecting line input load (negligible anyway), might as well remove S2 and have a DPST switching out the R15-R16 (and RZ if being completely pedantic).
« Last Edit: October 07, 2010, 05:31:45 AM by Kingston »

Samuel Groner

    Zürich, Switzerland
  • Posts: 2935
Re: OEP Z21805 app note
« Reply #4 on: October 07, 2010, 07:01:16 AM »
Indeed an interesting design. I didn't got to the bottom of how they use the NE5534 after a quick look; but S1 switches the transformer to "zero field" mode which greatly increases the level handling capability to accept line levels.

Samuel

abbey road d enfer

Re: OEP Z21805 app note
« Reply #5 on: October 07, 2010, 07:16:23 AM »
Very refreshing indeed, compared to the usual "cookbook" types often found on xfmr manufacturers sites.
It seems that OEP, now under the tutelage of a large group, are really willing to be a major player in their segment.
Who's right or wrong is irrelevant. What matters is what's right or wrong.
Star ground is for electricians.

Samuel Groner

    Zürich, Switzerland
  • Posts: 2935
Re: OEP Z21805 app note
« Reply #6 on: October 07, 2010, 08:12:05 AM »
Quote
It seems like the Vcc and Vee pins are used as inputs.

This comment first led me to a wrong path. Pin 1 and 8 are not the supply pins but rather the offset trim pins. They replace the input stage of the NE5534 with a discrete, low noise front end (Q1--Q3 etc.). This has been done e.g. in some Studer gear, although using a different opamp.

I wonder how they go about stability; the discrete differential pair has much higher gm, and there is no compensation capacitor fitted. If I'd implement this I'd make provision for both a compensation capacitor and a LR network for Q2/Q3 (as e.g. used by the JE-990).

Samuel

Kingston

Re: OEP Z21805 app note
« Reply #7 on: October 07, 2010, 08:18:01 AM »
Here's the headphone amp with good discussion on similar "hacks" for the NE5534 output stage.

http://gilmore2.chem.northwestern.edu/ubb/showpage.php?fnum=3&tid=7773

mhelin

Re: OEP Z21805 app note
« Reply #8 on: October 07, 2010, 10:36:13 AM »
Quote
It seems like the Vcc and Vee pins are used as inputs.

This comment first led me to a wrong path. Pin 1 and 8 are not the supply pins but rather the offset trim pins.
Oops, my bad .. So they bypass the NE5534 input stage via the compensation pins.  So Q1 makes a current sink and Q2/Q3 the diff stage. Would be interesting to try separating the S1C/S1D from the S1A/S1B so you could switch the current mode on and off alone. Quess the R7 value should be increased for mic input position on S1A/S1B though.
 

mhelin

Re: OEP Z21805 app note
« Reply #9 on: October 07, 2010, 10:40:28 AM »
Here's the headphone amp with good discussion on similar "hacks" for the NE5534 output stage.

http://gilmore2.chem.northwestern.edu/ubb/showpage.php?fnum=3&tid=7773

Now if you take the OEP's input and this output you are supposed to have a superiour opamp compared to NE5534? Guess not..


abbey road d enfer

Re: OEP Z21805 app note
« Reply #10 on: October 08, 2010, 07:32:02 AM »
Here's the headphone amp with good discussion on similar "hacks" for the NE5534 output stage.

http://gilmore2.chem.northwestern.edu/ubb/showpage.php?fnum=3&tid=7773

Now if you take the OEP's input and this output you are supposed to have a superiour opamp compared to NE5534? Guess not..

This scheme is optimised for a different set of parameters than for a mic pre.
It works around the fact that the 5534 is not capable of driving low-value loads.
The 5534's output stage has a current protection system that limits the current when it senses  a lower load.
Although it is capable of delivering about 50mApeak into a 300R load (15V peak), it is limited to about 40mA at lower loads, yielding about 20mW in a 32R load, typical of headphones.
Bypassing the output stage with an unprotected class-A push-pull pair gives it the capability to deliver about 150mApeak or a whopping 300mW into 32R. (that's with +/-15v rails; as per example, which uses only unipoalr 18V, power would be about 75mW @ 32R).
And in addition, it avoids the dreaded class-B operation  :-\
Who's right or wrong is irrelevant. What matters is what's right or wrong.
Star ground is for electricians.

mhelin

« Last Edit: October 08, 2010, 11:47:03 AM by mhelin »


 

Related Topics

  Subject / Started by Replies Last post
0 Replies
1349 Views
Last post February 17, 2005, 02:14:53 PM
by okgb
2 Replies
2511 Views
Last post August 06, 2005, 06:10:53 AM
by jensenmann
0 Replies
1846 Views
Last post July 16, 2006, 04:24:10 PM
by mikep
25 Replies
5290 Views
Last post August 19, 2011, 01:08:15 AM
by PRR