Alice dual OPA circuit may not work with Micparts SDC Omni capsule?

GroupDIY Audio Forum

Help Support GroupDIY Audio Forum:

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

antocu

Active member
Joined
Feb 19, 2022
Messages
33
Location
Italia
Hi there! I built a pencil mic with the dual OPA circuit with an omni capsule by Micparts and what I get is bad sensitivity and high noise floor.
If I use the same capsule with the original circuit of the Takstar CM-60 (of which I used the body) the microphone works very well with good sensitivity and low noise. I was wondering if the omni capsule needs a voltage polarisation that the OPA circuit can't provide.. but if I use the cardioid stock capsule is the same story. What am I missing?

Here's a link to the Alice OPA pencil pcb I used: OPA "Alice" Pencil Microphone - Share Project - PCBWay
 
Listing says nothing about it being an electret, so yes, requires polarization voltage.
Looking at the km84 circuit I was thinking of trying to use the same bias circuit. Taking the 46V by two 2k2 resistors between pin 3 and 1, following with the two RC filters and finishing with the 1G resistor to the capsule. Do you think it might work?
 

Attachments

  • KM84 bias.png
    KM84 bias.png
    111 KB · Views: 3
Technically yes, but you'll be left with considerably less than 46V there, since the opamp will very likely draw more current at idle, than the sub-1mA the KM84 circuit needs.
 
Oh you're right the opa1642 will draw 3.6mA.. I didn't consider it. Do you know any alternative? Always if it is possible.. I'm kinda trying to refunctionalize a circuit designed for electret capsule so maybe there's nothing I can do about it. (My other idea was to implement the Hex inverter voltage multiplier but It'd mean to redesign the pcb)
 
3.6mA through 3.4kohm (the two 6.8k in parallel, inside the preamp) plus 1.1k (2x 2.2k in parallel) would drop about 16v, leaving you with about 30V to play with. It should still work fine with a condenser capsule, just lower output and slightly higher noise floor than at 40-60v.
 
Use just the other half of the phantom supplies (like the pin 3) to supply the capsule bias and the other (pin 2) to supply the opamp (split the two 2k2 resistors). That way you will get full 48V to capsule (you need two 1G resistors and single 1nf capacitor to input, if you connect the capsule to virtual ground you save the cap and the other resistor but get lower bias voltage).
 
Use just the other half of the phantom supplies (like the pin 3) to supply the capsule bias and the other (pin 2) to supply the opamp (split the two 2k2 resistors). That way you will get full 48V to capsule (you need two 1G resistors and single 1nf capacitor to input, if you connect the capsule to virtual ground you save the cap and the other resistor but get lower bias voltage).

... And totally mess up the CMRR of the balanced connection, in the process... :confused:
 
... And totally mess up the CMRR of the balanced connection, in the process... :confused:
One possible way of avoiding that might be to try and adapt this version of using the 164* series op amp as an impedance converter:

https://www.opic.jp137.com/index-ldc.html
It only draws about half the current of the differential 'Alice' version and has a single sided audio output, but is still impedance balanced to help maintain a reasonable CMRR.
It also doesn't use any zeners to regulate the opamp DC supply. (The DC 'half rail' is automatically adjusted by the supply resistor divider), and so maximises the available phantom power voltage.

By disconnecting R5 from the DC 'mix' point and decoupling it with a further 47uF capacitor - to maintain the impedance balance - you could use the DC from pin 3 of the phantom power to supply the polarisation voltage to the capsule directly.
You might need to drop the value of R4 to optimise the DC supply to the op amp to improve the headroom. A matching value of R5 will (in conjunction with the additional 47uF de-coupling cap) help maintain the impedance balance. A value of 4k7 seems reasonable.

There will be almost no current drawn by the pin 3 side of the phantom power, so it will maximise the voltage available for capsule polarisation, but still retain the impedance balance.....
 
It would be nice to know how much the DC imbalance affects to the CMRR, I guess it depends on the mic pre. AC coupled mic pre front end does not see the DC anyway. However if there are no blocking capacitors like with transformer coupled preamps there would be some current though the primary (and a drop in the higher voltage). If both phases have the same terminating output impedances it should be quite OK for at least domestic use though maybe not OK enough to be implemented in a commercial product anyway.

Capsule biasing doesn't actually draw any current, so you could build a circuit which first charges some big capacitor and when it is done (no comparator etc. is needed, just a simple time delay & relay) switches off the charging and on the preamp circuit. Only problem would the capacitor self-discharging, don't know for an example the self-discharge rate of a 1000uF capacitor, I guess it can hold the charge quite a long (must depend an the dielectric).
 
Thanks to the members of this forum and the documentation they generously share, I am catching up on my (very basic and rusty) electronics knowledge. Before I became a musician I went to technical school and returning to the study of schematics and theory as a function of musical devices is really stimulating. So, I made a small draft on Rogs' circuit (I'm not sure about the 1G resistor), just to understand it better.. Is this what you meant here?

By disconnecting R5 from the DC 'mix' point and decoupling it with a further 47uF capacitor - to maintain the impedance balance - you could use the DC from pin 3 of the phantom power to supply the polarisation voltage to the capsule directly.
You might need to drop the value of R4 to optimise the DC supply to the op amp to improve the headroom. A matching value of R5 will (in conjunction with the additional 47uF de-coupling cap) help maintain the impedance balance. A value of 4k7 seems reasonable.
 

Attachments

  • opic sdc.png
    opic sdc.png
    113 KB · Views: 4
Thanks to the members of this forum and the documentation they generously share, I am catching up on my (very basic and rusty) electronics knowledge. Before I became a musician I went to technical school and returning to the study of schematics and theory as a function of musical devices is really stimulating. So, I made a small draft on Rogs' circuit (I'm not sure about the 1G resistor), just to understand it better.. Is this what you meant here?

C9 will cause a large impedance imbalance between pins 3 and 2. I don't think that is the connection that rogs meant.
 
The attached is what I had in mind. Should keep the line balanced, while allowing one side of the phantom power to act as capsule bias.
R8 and C9 just add a further low pass filter to get rid of any residue HF noise that C8 might miss.
Pretty Hi-Z, so shouldn't affect the line balance.
( Concept schematic only - not actually tried so far )
 

Attachments

  • OPIC.phantom.schematic.jpg
    OPIC.phantom.schematic.jpg
    143.1 KB · Views: 2
Sorry for my silence but I've been really busy. So I think I'll try Rog's circuit but I was thinking that there's no "y" termination as the capsule is already grounded to the body, as shown here for the Schoctava redrawn of Henry Spragen circuit (here). Couldn't be this be solved by adding a 1G resisitor between pin3 and C9-R8 junction?
 
Connect a (new) 1nF (1000 pF COG/NP0 ceramic or polystyrene) capacitor to point X and to a (new) point Z. Next connect a (new) 1G resistor from Y to Z, and connect the capsule between Z and GND. So you end up having the bias voltage going from Y to capsule and the capacitor AC coupling the capsule from the op amp input X.
 

Latest posts

Back
Top