Portable mic preamp (ENG)

[abbey road d enfer]

You need to add the thermal noise of the source resistor.

My bad; I'm so used to calculating only the noise added by the other components.

 

[abbey road d enfer]

WRT Vactrol based limiter.  Here is a schematic for a rather old battery powered in line mic preamp (no phantom) using such a limiter.

Anyone know what that output transformer is - couldn't find the number anywhere?

IIRC, RTS used a custom xfmr by Reichenbach, I think now Cinemag. You could use any decent 1:1 xfmr there because the output level is somewhat limited by the +/- 9v rails.

 

[mjrippe]

Most RTS 1400 had a Frost Magnetics 600:600 output transformer.  I have about a dozen of these units sitting around and ONE of the later ones has Jensen in and outs.

P.S. - Thanks for the official schematic!  I drew it out a few months ago after I couldn't find one online ;-)

 

[bruce0]

Moved to http://www.groupdiy.com/index.php?topic=53072.msg677362#msg677362

 

[mjrippe]

The JE/JT designation is simply a matter of when they were manufactured.  I forget the year they switched from one to the other, but that is all it means.

Interesting, I have seen a couple with Jensen inputs only and Frost outputs.  The single one I have is the only Jensen In/Out that I have seen so far.  It is a later version with a darker grey (anodized?) case. 

We are digressing quite far from the original thread here.  Perhaps a new thread about these nifty preamps?

 

[bruce0]

OK, I moved my OT post... to here http://www.groupdiy.com/index.php?topic=53072.msg677362#msg677362

Also I posted the schematic there, and I put the full technical manual at the technical documents section.

All of mine had frost outputs, only 2 of 20 had jensen inputs.

They were all the light grey anodized aluminum

 

[ricothetroll]

Hi,

Thanx for this schematic (RTS) ! But IMHO playing with negative feedback is better, because it doesn't load the mic as the impedance shunt (like in the RTS) do. Furthermore, I intend to use it not only for ambience recording but also with my MKH416 on the boompole, and that particular mike has a low 25R output impedance...

I worked on the schematic of the new version :
http://img560.imageshack.us/img560/8819/8ztw.png

Instead of switching the source of the sidechain, I implemented a "dummy" VCA : if the main VCA volume is lower than 0dB, the sidechain VCA stays at unity gain and so the sidechain sees the same signal as right after the preamp. If the gain of the main VCA is higher than 0dB, and so the main VCA is prone to clip, the sidechain sees a copy of the post sidechain signal. I intend to thermally couple those VCA, well at least I'll try to route them as close as possibke.

Any comment is appreciated, as usual

I'll do the routing ASAP !

Best regards.

Eric

PS : the vactrol I'll use is VTL5C1, not VTL5C9 as noted on the schemo, I was too lazy to create a VTL5C1 model and the footprint is the same anyway...

 

[ricothetroll]

OK, so I breadboarded the preamp and the sidechain. I did my previous tests with visual analyser, a scope/FFT software I feed with my soundcard (RME Multiface). This time I used my old scope to have a look at the signal and I noticed some huge HF oscillations when I reduce the gain, that the RME couldn't see because of its audio limited bandwidth... I guess those are caused by the lack of compensation of my system. Those oscillations happen at about -21dB GR, which is already quite good for a peak limiter, but I'd like to make sure those won't start in more critical situations (higher temp, worse PSU...). The oscillations waveform is triangular, and freq is about 200kHz.

I tried to put a small cap across the feedback resistor (vactrol) but it seems to create some other oscillations, lower in frequency.

What else could I do to improve the preamp's stability ? I tried to read TI's paper "Stability Analysis of Voltage feedback Op-Amps" but I must admit that after a few pages it is far beyond my level of understanding  :-[

Thanx in advance for your advices !

Best regards.

Eric

 

[abbey road d enfer]

OK, so I breadboarded the preamp and the sidechain. I did my previous tests with visual analyser, a scope/FFT software I feed with my soundcard (RME Multiface). This time I used my old scope to have a look at the signal and I noticed some huge HF oscillations when I reduce the gain, that the RME couldn't see because of its audio limited bandwidth... I guess those are caused by the lack of compensation of my system. Those oscillations happen at about -21dB GR, which is already quite good for a peak limiter, but I'd like to make sure those won't start in more critical situations (higher temp, worse PSU...). The oscillations waveform is triangular, and freq is about 200kHz.

I tried to put a small cap across the feedback resistor (vactrol) but it seems to create some other oscillations, lower in frequency.

What else could I do to improve the preamp's stability ? I tried to read TI's paper "Stability Analysis of Voltage feedback Op-Amps" but I must admit that after a few pages it is far beyond my level of understanding  :-[

Thanx in advance for your advices !

Best regards.

Eric

Have you SPICE'd the schemo?
I guess the problem comes from having an active element in the NFB loop, so the loop has excess phase-shift. I've done a quick sim that shows a big hump at 2MHz. It doesn't show oscillation, but it's pretty close. Reducing the 33pF caps to 3pF almost gets rid of the hump.
Anyway, I don't think your arrangement has more advantages than flaws. It reduces the gain by loading the source, just the same as putting the vactrol on the xfmr's secondary, and the operation is not very predictable, giving different results for different gain settings.

 

[abbey road d enfer]

Hi,

Thanx for this schematic (RTS) ! But IMHO playing with negative feedback is better, because it doesn't load the mic as the impedance shunt (like in the RTS) do. Furthermore, I intend to use it not only for ambience recording but also with my MKH416 on the boompole, and that particular mike has a low 25R output impedance...

As I wrote earlier, your arrangement loads the mic just like (and maybe even more than) the simple vactrol-in-parallels-with-the-secondary connection.

Instead of switching the source of the sidechain, I implemented a "dummy" VCA : if the main VCA volume is lower than 0dB, the sidechain VCA stays at unity gain and so the sidechain sees the same signal as right after the preamp. If the gain of the main VCA is higher than 0dB, and so the main VCA is prone to clip, the sidechain sees a copy of the post sidechain signal.

I don't fully understand what you mean. Basically, the side-chain is operating on the input signal, not on the VCA output, which makes it a feed-forward structure. But the side-chain is in a feedback arrangement, like a GSSL.

Any comment is appreciated, as usual

I see that when the limiter is engaged, the VCA volume control is disconnected; is it really what you want?
Then do you really need the offset trims? Is it just to save a coupling cap?
Also I see there's no input EMI/RFI protection. Just a pair of 1nF caps on Pins 2 & 3 to ground (chassis) would be nice.

 

[ricardo]

I haven't looked but ...  surely the THAT limiter/compressor + preamp chips must have been done in anticipation of just this requirement. 

"ENG mike preamp with heavy AGC"

Have you checked all their Datasheets/Application Notes for the relevant devices ... or even asked THAT?

 

[PRR]

> some huge HF oscillations when I reduce the gain

That plan is obscene.

The LDR runs back around *two* stages.

It butts right against the mike+transformer source resistance.

With such high (also variable) gain, the LDR resistance is *divided by* forward gain. So for gain of 100, LDR at 10K, the transformer secondary sees 10K/100= 100 ohms. If tranny is 1:4, that's 6 ohms seen by the mike!

Also your gain and frequency response are set by ratio of R(ldr)/R(mic) instead of R1/R(mic) (plus iron factors). Mike impedance is rarely flat, so frequency response changes when limiting happens.

Meanwhile it oscillates. Cut to the final act. All amplifiers have a treble roll-off (and associated phase-twist). If a single 6dB/oct rolloff, you can apply any amount of NFB. Common opamps are rigged to have a 6dB fall up to the frequency where gain vanishes. They actually have 3 or more roll-offs but 2 are pushed-up and one is pushed-down so only one matters.

Here you have two such opamps inside the NFB loop. Each is 6dB/oct, so both are 12dB/oct. Both are high-gain so the 12dB/oct slope extends over a wide supersonic range. Phase shift will ultimately approach 180 degrees. Positive becomes Negative. Gain increases to a peak above the desired band.

But wait. We have two dominant rolloffs, but rolloffs are everywhere. And we got so close to 180 degrees that even another 5 degrees from some far-out rolloff will push it over the edge to 180 degrees and outright oscillation.

I also think this is way too-too complicated. 16-bit recorders can capture nearly any sound desired. Two gain settings. Bit zero equals 20dB SPL (lowest likely ambient noise; take 12dB SPL if you like) which works to ~~110dB SPL (LOUD); and bit 15 equals 125dB SPL which is really the loudest sound field the recordist should work in. Sure most interviews etc won't fill-the-meter, but that can be adjusted in post. That's also a far better place to smooth out levels for good reproduction.

It lacks elegance.

 

[ricothetroll]

Hi,

Thanx for your criticism, made me realize my mistakes ! I actually didn't think that the NFB could load the input, now I understand why.

I did some tests with Aphex's method, and that sounds quite good actually. I kept the whole sidechain though, because their method with the comparator introduces a lot of distortion a low-low mid frequencies (sidechain feedthrough to the signal).

Here is the resulting schematic :

http://img59.imageshack.us/img59/3407/lcod.png

To be continued...

Best regards.

Eric

 

[abbey road d enfer]

Hi,

Thanx for your criticism, made me realize my mistakes ! I actually didn't think that the NFB could load the input, now I understand why.

I did some tests with Aphex's method, and that sounds quite good actually. I kept the whole sidechain though, because their method with the comparator introduces a lot of distortion a low-low mid frequencies (sidechain feedthrough to the signal).

Here is the resulting schematic :

http://img59.imageshack.us/img59/3407/lcod.png

To be continued...

Best regards.

Eric

R37 & 37 are useless. They would be necessary if you had coupling caps.

I still don't understand the switch at the bottom right.
In one position it increases the output gain with a range of about +24 to +60dB.
In the other it goes to a mysterious Link1...

Regarding the output stage, you may want to read this thread
http://www.groupdiy.com/index.php?topic=38473.0
In particular
http://www.groupdiy.com/index.php?topic=38473.msg677311#msg677311
I would rather have impedance-balanced outputs.

Are D1 & D2 really necessary? I don't think so.

 

[ricothetroll]

R37 & 37 are useless. They would be necessary if you had coupling caps.

Well that's a remain of one of the previous versions where there was some coupling caps... Thanx for pointing it !

I still don't understand the switch at the bottom right.
In one position it increases the output gain with a range of about +24 to +60dB.
In the other it goes to a mysterious Link1...

I must admit that I haven't been too explicit about it ! On the left channel, the 3 pins of the switch are shorted and the "link1" pin is connected to the "link1" of the right channel. On the right channel, the switch is wired as on the schematic. When the switch is on the normal position, both channel are independant (dual mono), and when the switch is on the "link1" position, the volume pot of the left channel controls the volume of both channels (stereo).

Are D1 & D2 really necessary? I don't think so.

Those diodes perform some protection/soft clipping as on Samuel's schematic :
http://www.sg-acoustics.ch/analogue_audio/microphone_preamplifiers/pdf/D_r1.pdf
I also doubt those would really be useful in my case, but who knows, maybe when recording a plane taking off with a really sensible microphone

Regarding the output stage, you may want to read this thread
http://www.groupdiy.com/index.php?topic=38473.0
In particular
http://www.groupdiy.com/index.php?topic=38473.msg677311#msg677311
I would rather have impedance-balanced outputs.

Maybe then I'll finally go with a THAT1646, as I intended in the first place

Thanx for your help !

Best regards.

Eric

 

[abbey road d enfer]

I still don't understand the switch at the bottom right.
In one position it increases the output gain with a range of about +24 to +60dB.
In the other it goes to a mysterious Link1...

I must admit that I haven't been too explicit about it ! On the left channel, the 3 pins of the switch are shorted and the "link1" pin is connected to the "link1" of the right channel. On the right channel, the switch is wired as on the schematic. When the switch is on the normal position, both channel are independant (dual mono), and when the switch is on the "link1" position, the volume pot of the left channel controls the volume of both channels (stereo).

OK, got it.

Are D1 & D2 really necessary? I don't think so.

Those diodes perform some protection/soft clipping as on Samuel's schematic :

Why not, they can do no harm anyway.

 

[carlmart]

I still don't understand the switch at the bottom right.
In one position it increases the output gain with a range of about +24 to +60dB.
In the other it goes to a mysterious Link1...

I must admit that I haven't been too explicit about it ! On the left channel, the 3 pins of the switch are shorted and the "link1" pin is connected to the "link1" of the right channel. On the right channel, the switch is wired as on the schematic. When the switch is on the normal position, both channel are independant (dual mono), and when the switch is on the "link1" position, the volume pot of the left channel controls the volume of both channels (stereo).

OK, got it.

Are D1 & D2 really necessary? I don't think so.

Those diodes perform some protection/soft clipping as on Samuel's schematic :

Why not, they can do no harm anyway.

Why, after all that work with the THAT did you end up using the transformer path? It's a lot more expensive, isn't it?

 

[ricothetroll]

Yes it is ! But the goal is to have a good machine, right ? Good limiters are essential for ENG production. BTW I still didn't find enough time to build it, maybe this summer...
To be continued...

 

[Luax]

Hi Rico,
I admire your ambition and will follow the progression of your project.
Just came across this:
http://www.filmtontechnik.de/611.html
It kills the diy but it might just be what you are looking for ( or at least can get inspiration from )
All the best