Avoiding big caps on solid state preamp inputs

GroupDIY Audio Forum

Help Support GroupDIY Audio Forum:

This site may earn a commission from merchant affiliate links, including eBay, Amazon, and others.

bjosephs

Well-known member
Joined
May 10, 2021
Messages
104
Location
Massachusetts
Hi All,

I got the bug to build a few more preamps and was going the direction of a monolithic IC like That 1512. I started a board layout based the implementation in figure 2 in this demo board data sheet but when I got to the input caps I found that high quality film caps are just enormous. Even at 2.2u Wimas take up a bunch of space and I've been warned can be a source of noise on such a sensitive input. I'd rather not use electrolytic or tantalum and I'm not doing another transformer input mic pre this time. I googled bit and found a mega thread here with all things mic pre but so many of the links were dead. I'm looking for thoughts on the following:
  1. Going with much smaller caps and living with a permanent high pass filter around 40Hz. I understand that this can effect low frequency CMMR as well.
  2. Skipping the phantom power altogether since dynamic mic are what I need more pres for. I would simply DC couple the XLR straight to the input pins of the 1512. The DC servo would be moved or omitted as well.
  3. I hear legends of a way to DC couple the input but float the whole circuit on 48v to achieve phantom power without input caps. You still need an output cap but obviously a much smaller one.

Side note: The intended use here is for tracking guitars and I want to be able to pre-blend the mic before getting to the recorder. I don't have a mixer on the front end so I was thinking of adding a 2:1 mix function on pairs of preamps. Any thoughts on how to do that well with a low parts count are welcome too I attached how I was planning to do it.
 

Attachments

  • Screen Shot 2023-02-24 at 3.01.48 PM.png
    Screen Shot 2023-02-24 at 3.01.48 PM.png
    29.7 KB · Views: 3
You're making a rod for your own back by having an irrational fear of electrolytics; as long as they're big enough they're as good as film. Using smaller caps will indeed trash your CMRR right when you need it most -around 50Hz- unless you select them to be precisely matched.
 
And don't go with the misguided fashion of using low-ESR caps (made for switching PSUs and the like) in the signal path.

Can you (or anyone) please teach me more about this?

I’ve heard it alluded to before, but have never really understood the disadvantages of low-ESR electrolytic caps in signal coupling applications (where low ESR film caps are usually just fine)?
 
Yes, just use quality bipolars (much better low end distortion than polar), I like the green Nichicon Muse caps with a WIMA bypass cap in parallel. And don't go with the misguided fashion of using low-ESR caps (made for switching PSUs and the like) in the signal path.

What's wrong with low ESR 'lytics in the signal path?
 
Going with much smaller caps and living with a permanent high pass filter around 40Hz. I understand that this can effect low frequency CMMR as well.
CMRR degradation is the main issue there. This degradation is directly related to the impedance difference of the input caps. Since their nominal impedance is inverse to the nominal capacitance, the larger the better. In that respect, I would not sustain the AN, which suggests 2.2uF with teh servo on. Now you have to make sure that the matching of input caps is good enough.

  1. Skipping the phantom power altogether since dynamic mic are what I need more pres for. I would simply DC couple the XLR straight to the input pins of the 1512. The DC servo would be moved or omitted as well.
I wouldn't skip the servo, because with a direct connection , input offset voltage is amplified with the same gain than signal.

  1. I hear legends of a way to DC couple the input but float the whole circuit on 48v to achieve phantom power without input caps.
Not a legend. Member 57sputnik has a huge thread about it
https://www.proaudiodesignforum.com/forum/php/viewtopic.php?t=598
 
Cap preferences aside, this is just an inverting follower. No voltage gain. Am I wrong?

The link at Pro Audio Forum has no pictures. Too bad. . .
For TMI, I have been thinking about this for decades. For a historical note, I literally joined this forum to follow Wayne's original musings on this subject here in this forum. I made lots of my own late night circuit design scribbles but never melted solder to make a working proof of concept. Wayne carried this approach much further than I ever did. At one point I bought some high voltage capacitors to make a high voltage rail to support my floating preamp, but they are still sitting in bags back in my lab.

For the record there are any number of ways to skin this cat but the basic approach involves floating the mic preamp up on the phantom supply voltage. Then level shifting back down to nominal supply circuits involves some thought.
===
One idea I chewed on some time ago but dismissed was to fly a high performance A/D convertor up to the phantom referenced mic preamp and optically shift the digital output down to real world voltages.

The problem about these exotic approaches to "solve" the hypothetical audio problem ( electrolytic capacitors in the signal path?) trades old issues for new ones, and a thoughtful examination of the real limitations to modern audio path performance do not include this.

JR
 
Jeez! That's a lot of work to get rid of a few capacitors. I work with tubes and fets. Harder to blow them up. I have an idea for a phantom power cascode cathode input mic pre that I've never built. I think I'll start a new thread.
 
Last edited:
Silmic i use in the power supply only. Decoupling with Os-Con and MLCC directly on the IC pins.
 

Attachments

  • FE41DB3F-CB5A-47D3-8236-ABD99BE179FE.jpeg
    FE41DB3F-CB5A-47D3-8236-ABD99BE179FE.jpeg
    1.2 MB · Views: 0
  • 82B181F0-9C4E-412F-A588-C9488BA1E39C.jpeg
    82B181F0-9C4E-412F-A588-C9488BA1E39C.jpeg
    1.4 MB · Views: 0

Latest posts

Back
Top