OPA Alice in a split config for upright bass mic?

[joulupukki]

I'm wondering if anyone has done something like this in the DIY community...

There's a brand of mics out there using standard hardware parts to build mics. They've got one for upright bass and I figured it can't be that difficult to build. It looks like theirs is using a JFET for impedance transforming at the capsule. The OPA Alice seems like it could be a nice circuit to pair up with an inexpensive JLI-2555BXZ3-GP electret capsule.

Does something like this look like it'd work? I basically built a small board that'd attach directly to the capsule and then one that'd be for inside of a copper tube. The high impedance circuit would all be right next to the capsule and the signal would be transferred to the main body PCB via a microphone cable. Just trying my hand at EasyEDA with SMD components for space reasons (and the fact that the OPA1642 is SMD). Ignore the color bands on capsule PCB's through-hole resistor. I'd use a 1G resistor. I just couldn't find one in EasyEDA so I wouldn't order that part pre-populated.

 

[jp8]

Pin 5 and R1 connect to VGND, not to GND. Then it should work. Maybe you want to place R4, R5 and C3 on the capsule PCBA. But you don't want an elcap on that board because of its size. Change R4 and R5 to 1M and C3 to a 1uF X7R of sufficiently high voltage (> Vcc/2)

Jan

 

[joulupukki]

Pin 5 and R1 connect to VGND, not to GND. Then it should work. Maybe you want to place R4, R5 and C3 on the capsule PCBA. But you don't want an elcap on that board because of its size. Change R4 and R5 to 1M and C3 to a 1uF X7R of sufficiently high voltage (> Vcc/2)

Jan

Hi Jan, thanks for the glance at the project!

I need to re-label the capsule board to not be GND but VGND. The idea is to connect VGND from the body PCB to the GND of the capsule PCB.

This circuit is from here: https://www.instructables.com/OPA-Based-Alice-Microphones-a-Cardioid-and-a-Figur/

R4 and R5 are currently 47K. I figured that the high-impedance portions of the circuit are on my smaller board already. Is that not the case? R4 and R5 are part of the power portion, right? …creating the virtual ground?

All the el caps and MLCC caps are 50V right now.

Can you explain the reasons behind the recommended changes?

Yeah, the capsule PCBA has to be small to fit in the area right on the back of the electret capsule.

 

[joulupukki]

Another question I have is … on the ORS87 there are recommendations on changing values of caps and resistors to tune the frequency response based on potentially reducing hyped high end of any given capsule. I’m still trying to learn and understand all of that in circuits. Which components here would I need to tweak, for example, if the capsule I end up using seems way too hyped? Or, would I need to be adding additional components to pull that off?

Ideally this would be a mic I could mount onto my upright bass and be good to go directly into a PA system w/o much fuss.

 

[rogs]

If you want to minimise the component count, you might find it simpler to try out my OPIC circuit?

(Some notes for the Fetless version here: https://www.opic.jp137.com/index-fet.html )

Simular in concept to the OPA Alice, but without any 'frills'..... There is a single sided audio output (although the outut is still impedance balanced) and there is no regulation to the op-amp supply. The Zener is not really required if you ensure that the feed resistors for the opamp DC ensure the max op-amp supply value cannot be exceeded.
Just a thought, if space is at a premium ? ...

 

[Voyager10]

Looks like you have pins 2 and 3 on the OPA1642 swapped. The capsule input needs to go to 3 (the '+' input) and pins 1 and 2 should be connected.

 

[jp8]

Hi Jan, thanks for the glance at the project!

I need to re-label the capsule board to not be GND but VGND. The idea is to connect VGND from the body PCB to the GND of the capsule PCB.

No problem. As always, glad to help you!

The mic board needs two grounds: a power ground for the OPA and a virtual ground at Vcc/2 (= VGND). Capsule ground node will also connect to VGND, btw.

You don't want to run VGND through the cable as it could pick up noise and inject as common mode signal into the OPA. Being CM, most of it will be rejected by your preamp, but not all. And suppose you have a cell phone nearby...? The VGND wire inductance with stray capacitances and capsule capacitance could also become a resonant circuit. You would at least want the decoupling cap near the OPA. But then it's just a small step to also add the resistors and reduce the number of connections between the boards by one. But you would still need a 4 wire connection, so a standard miccable wont fly. How about shielded UTP? Not uncommon to use as signal wire in pro audio.

If the mic board and power board are to become detachable units, I would leave the zener in the supply. Suppose you power up the preamp with P48 on. Then C2 charges to 48V. If you subsequently connect the mic to the power unit, your OPA will instantly die. A 12V zener should give you ample headroom in this application. You could increase it to maybe 24V, but I wouldn't go higher if you have a detachable mic: upon connecting the mic, the wire inductance, C1 and C2 will constitute a resonant circuit and you'll see large voltage swings on the Vcc line, which could exceed the OPA ratings.

A single OPA would make the circuit simpler, but if the idea is to have all elcaps in the power unit because of their size, then the wiring between mic and power board will become unbalanced. Could be acceptable if the length remains short, dunno. It will also deprive you from the possibility to add a capacitor shunting R3 to add some high-end roll-off.

Lastly, depending on the length of the cable between mic and power board, the capacitive load on the OPA output may exceed 100pF, which could cause instability or overshoot issues. Consider splitting up the 47R resistors into 2 x 24R, with one 24R on the mic board.

Jan

 

[rogs]

....... Suppose you power up the preamp with P48 on. Then C2 charges to 48V.

I see your point, but as C2 can only ever charge via a series 6k8 resistor I'm not sure if the supply can ever exceed the limit? .... I've certanly never had a problem, and I have hot plugged the OPIC circuit to various P48 supplies over the last couple of years many hundreds of times.
I think a transzorb (TVS) rather than a zener might be more appropriate, if over voltage suppression is the goal....

 

[jp8]

I'm referring to @joulupukki 's case where he has separate boards for OPA impedance transformer and PSU. If they have a non-permanent, detachable connection, and you FIRST power up the PSU and THEN connect the impedance converter board to the PSU, the OPA will get the full 48V. Or actually 48V plus the LC voltage swing as described.

Jan

 

[mhelin]

Maybe a stupid question but do we really need those DC coupling capacitors C4 and C5, couldn't we just adjust the Vgnd (and the absolute resistor values + voltage divider, and also the R7 and R8 if neede) so that the voltage about equals the voltage before 47ohm resistors? Guess someone could design a servo using the other half of the opamp, too (Vgnd could be set to exactly follow the voltage at the 47 ohm resistors, could the 'servo' be just a non-inverting buffer from the other opamp output to Vg if the Vg setting resistors were big enough?). Also the zener (if needed at all) voltage should be set as high as possible to avoid it actually working and generating noise but just be there as a protection for the opamps (36V or so might be fine).

 

[jp8]

In principle impossible with this concept. Voltage on XLR outputs can only be higher than Vcc, because you need an impedance between outputs and Vcc. And Vgnd is half of that, so even much lower.

Zener noise is a total non-issue as it is effectively filtered by R6 and C2. And the high PSRR of the OPAs will also reject any ripple or noise on Vcc. Any residual noise on Vgnd appears as common mode voltage on the preamp input and is also effectively reduced.

Jan

 

[joulupukki]

I'm referring to @joulupukki 's case where he has separate boards for OPA impedance transformer and PSU. If they have a non-permanent, detachable connection, and you FIRST power up the PSU and THEN connect the impedance converter board to the PSU, the OPA will get the full 48V. Or actually 48V plus the LC voltage swing as described.

Jan

The plan is that the capsule and power boards will be permanently wired and won’t be hot swapped.

 

[joulupukki]

What if I were to use high def microphone cable? It has four wires plus a shield.

I could also move those two resistors and one capacitor onto the capsule board. I could reduce the size of the SMD components on that board. If needed, I could use a through hole radial, electrolytic capacitor and lay it down on its side. All of that might fit. I will experiment a little bit later today in the software.

 

[k brown]

What if I were to use high def microphone cable? It has four wires plus a shield.

I could also move those two resistors and one capacitor onto the capsule board. I could reduce the size of the SMD components on that board. If needed, I could use a through hole radial, electrolytic capacitor and lay it down on its side. All of that might fit. I will experiment a little bit later today in the software.

There's nothing 'high def' about quad cable; it just rejects interference better. For very long cable runs, it's higher capacitance could make it slightly 'lower derf'.

 

[joulupukki]

The cable between the board and the capsule would be something like 6-10”.

 

[joulupukki]

So this is a 15x18mm board and the 47uF cap is 7.7mm high. Combine that with the thickness of the board and it'd be under 1cm tall. That'd maybe/probably fit. What voltage rating would that cap need to be? I've currently got a 50V cap but maybe it could be a 25V?

 

[jp8]

What if I were to use high def microphone cable? It has four wires plus a shield.

Quad core has intended tight coupling between the signal wires. Not sure how that would work out with respect to oscillations due to audio being coupled back into the power supply lines. Or Vcc AC currents coupling into the audio signal. Twisted pairs of UTP cable have very low coupling, which is why crosstalk between different signal pairs in the cable is also low. I'd either use UTP, or a 2-channel mic cable, such as this:

https://www.markertek.com/product/m...wisted-pair-multi-channel-mic-cable-by-the-ft

Jan

 

[Voyager10]

R4/R5 can be significantly higher, e.g. 1M, because there are only high-impedance nodes connected to it. So C3 can be reduced correspondingly, 1uF would do fine. Its normal working voltage will be 1/2 Vcc, or 6V, but you might want to consider fault conditions. The op-amp will survive Vcc up to 36V, so maybe a 16 or 25V rated component would be best.

 

[jp8]

I've currently got a 50V cap but maybe it could be a 25V?

Depends on how much derating you want to apply, but 25V should do the job. Even at a Vcc of 36V, you'll have only 18V on the cap.

Jan

 

[rogs]

So this is a 15x18mm board and the 47uF cap is 7.7mm high. Combine that with the thickness of the board and it'd be under 1cm tall. That'd maybe/probably fit. What voltage rating would that cap need to be? I've currently got a 50V cap but maybe it could be a 25V?

View attachment 146271View attachment 146272

I'm assuming the electrolytic is C3? ..... in which case yes, it can be rated at 25v.

I see you've corrected the schematic, to connect R1 to the half rail (which you call 'VGND' ).
The op amp connections are still wrong.... As Voyager10 pointed out in post #6:
•The capsule /R1 junction should be connected to pin 3, and not pin 2.
• Pins 1 and 2 are connected together.... Not 1 and 3 as you have at the moment.