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.
You're right I didn't see that in the schematics I mentioned it has been added also the 1nF cap.. 🙈 It was in front of me! Thanks
 
My initial concept sketch did not optimise the capsule DC... Adding the coupling capacitor and second 1G resistor will maximise the capsule DC to around 48v The actual figure will depend on how low an impedance the 48v phantom power source actually is. As only about 2mA is drawn by the op-amp, it shouldn't be much less than 48v!
The OPA1641 amplifier has excellent noise and distortion figures. In this schematic it is likely to be powered with around 24v, which should provide plenty of headroom.
Should allow for a very simple - but high quality - LDC mic project......
(Modified schematic attached)
 

Attachments

  • OPIC.phantom.schematic.jpg
    OPIC.phantom.schematic.jpg
    155.4 KB
Last edited:
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.
I have made a schematic and am trying to size the components to fit into the body of a takstar cm60. The hardest part is finding electrolytic capacitors that are not too bulky. Maybe some can be substituted with 35V ones, for example C4 but I am not sure..
 
I have made a schematic and am trying to size the components to fit into the body of a takstar cm60. The hardest part is finding electrolytic capacitors that are not too bulky. Maybe some can be substituted with 35V ones, for example C4 but I am not sure..
C1,C2 and C4 on your schematic should be OK as 35V. C3 and C10 need to be 50V.
I would take R3 to C1 -ve (gnd) and not to C1 +ve, to help optimise the line impedance balance.
 
Here I share the schematic for a SDC microphone, thank you Roger for your corrections!
C1,C2 and C4 on your schematic should be OK as 35V. C3 and C10 need to be 50V.
I would take R3 to C1 -ve (gnd) and not to C1 +ve, to help optimise the line impedance balance.

This is the first pcb design for me and I basically adopted some ideas which Mic Sharf put into his pcb from which I started. I still have to print it on paper to see if the sizes are correct and then I'll order a first batch.
 

Attachments

  • Schematic_OPA1641 circuit_2022-12-17.pdf
    37.4 KB
  • PCB.pdf
    43 KB
I did wonder just how much difference using 48v as a polarisation voltage - as opposed to the more conventional 60 or 70 volts -would make, and the answer is - not a lot!

The sensitivity is reduced by some 2dB from a similar capsule and circuit polarised with 60v, as you would imagine, but it's still quite usable.

The exact polarisation voltage will of course depend on the details of the specific phantom power supply being used, but as the circuit itself only draws some 2mA from one side of the phantom power supply, the unused leg should present pretty close to 48v for use as a polarisation voltage.
The actual impedance of the power supply it self may drop the supply by a volt or 2?...

I built a version using this concept as a variation of my OPIC mic project. Some notes here: http://www.OPIC.jp137.com/OPIC48.pdf

I've made a quick unprocessed 2 track spoken word audio clip , using a simultaneous recording of this mic (left track) and the Rode NT1 (right track).

Copy of that track here: http://www.jp137.com/las/OPIC48.v.NT1.wav

I think the mic performs pretty well. It's about 3dB less sensitive than the Rode (which has a published sensitivity of -29dB) but the noise floor is pretty comparable. (Well, it is when most of the noise is environmental..... Not sure how much noisier it would be than the Rode in an anechoic chamber? )

When a decent mic preamp - or DAW - is used to pick up the missing gain, it makes me wonder if actually using voltage multiplier for this type of mic is strictly necessary? .....
 
I did wonder just how much difference using 48v as a polarisation voltage - as opposed to the more conventional 60 or 70 volts -would make, and the answer is - not a lot!
It depends on capsule, I think, the 70V or even 60V can be too much for a capsule with thinner spacer (bigger capacitance) designed and built for 48V bias. I got a second hand SCT-800 which couldn't take that about 65V from the PSU (maybe the capsule had gone bad).
I built a version using this concept as a variation of my OPIC mic project. Some notes here: http://www.OPIC.jp137.com/OPIC48.pdf
That circuit obviously has some unbalance (thus less CMMR), wonder how much it makes difference in practice, could it be measured and compared to another circuit where the om amp power has been taken from both (+ and -) outputs (phantom source inputs)?
 
That circuit obviously has some unbalance (thus less CMMR), wonder how much it makes difference in practice, could it be measured and compared to another circuit where the om amp power has been taken from both (+ and -) outputs (phantom source inputs)?
I can't see why there should be any impedance imbalance that will degrade the CMRR to any real extent?
Although DC power is only drawn from one side of the supply I can't see that should affect the impedance very much -- At least not with AC coupled preamps?
I can see there could be some difference with a transformer coupled input - as you observed in post #14 - but I'm not sure this type of hobby mic is likely to be often used with such an exotic preamp! :) . Unfortunately I don't have access to such a preamp so I can't check experimentally to see if there are any obvious problems
I have tried the mic with an 11m cable (well, a 5 metre with 2 x 3m cables in series! ) and I cannot detect any difference in performance or noise levels....

The DC on pin 2 measures c.33v, while the DC on pin 3 measures 47v when used with my SD interface, which corresponds to the expected current flowing in each 6.81k phantom power feed resistor. I can't see that the difference in DC current drain will affect the impedance of either side?

The op-amp output impedance when used as a closed loop buffer is only likely to be in the milliohm range I would suggest, so the output AC components connected to pin 2 R2 and C2 are impedance balanced by R3 and C3 on pin 3.
In addition, the DC feed impedance is determined by R4 and C1 on one side, which is matched by R5 and C9 on the other. (I don't think the additional 1M resistor R8 in series with the capsule DC is going to make much difference to the impedance.)
It is certainly intended to be impedance balanced, and the idea is 'borrowed' from similar techniques adopted by manufacturers like dbx, Rode and Neumann (and I'm sure many others!).
They all use this passive impedance balancing concept to help maintain a useful CMRR, when used with single sided audio outputs...
 
Last edited:
Impedance balancing is OK but the DC levels at preamp makes my worry at least with all transformer input mic preamps where the current caused by the DC difference goes through the primaries of input transformers, though as the transformer primaries have so low DC resistance the difference will not be that high causing also lower bias voltage. Don't know what this current will cause to transformer cores either, might be harmful. This scheme only works properly with capacitor coupled mic preamps which have those 6k8 resistors connected to +48 in P48 supply.
 
...... This scheme only works properly with capacitor coupled mic preamps which have those 6k8 resistors connected to +48 in P48 supply.
That is the intended 'marketplace' :) ..... I's only really offered as a very simple introduction to the world of LDC hobby mics.
I think my LDC circuit on this page: OPIC LDC might be a better choice for those using a transformer coupled preamp?
A balanced DC feed, and a wider range of polarisation voltages available from the voltage multiplier.
It is useful to remember to add extra volts to compensate for the DC bias on the op-amp input.....
It was an error on my part in not doing that on one of my mics that prompted me to try out this version, using only 48v as a polarisation voltage.
As I say, this is only bit of fun, and might provide a cheap way of introducing those more used to using electret capsules to the world of LDCs? ....
 
Yes it is fine as it is. I wish there was a simple way to generate the capsule bias voltage. Even simple voltage doubler circuit would be enough if you have some 25-30V to get some 50-60V out depending on topology, if you use diodes these will drop output voltage at least a 0.3V per a (Schottky) diode. That's a single transistor oscillator + two diodes and capacitors. Are there any simple higher voltage switched capacitor charge-pump IC's available, maybe you could construct one from discrete components (MOSFETs)? Most like MAX1682 are low voltage parts for mobile devices.

Edit: Somehow related study on voltage doubler topologies: https://backend.orbit.dtu.dk/ws/por...52/28827_Article_Text_98102_1_10_20210426.pdf
 
Last edited:
If you're not opposed to China-sourced SMD parts, things like MT3608, a small inductor and a couple passives could make for a compact boost-converter. Actually, looks like that's good for up to 28v out, but i've seen such low(ish)-voltage parts used with a voltage multiplier tacked onto the output in some MOTU audio interfaces, for providing the phantom power rail.

Stand by for some examples soon, as soon as i get back to a computer...

As it turns out, there's not a whole lot to choose from, with over-60v output capabilities. "Even" LCSC only shows a handful of possibilities, such as MIC2172 (which sucks about 7mA quiescent!!!) , and then much pricier ex-LinearTech offerings (but which also suck 6-9mA quiescent).

I'm afraid it's a case of "pick your poison" after all 🤦‍♂️
 
Last edited:
Yes it is fine as it is. I wish there was a simple way to generate the capsule bias voltage. Even simple voltage doubler circuit would be enough if you have some 25-30V to get some 50-60V out depending on topology, if you use diodes these will drop output voltage at least a 0.3V per a (Schottky) diode. That's a single transistor oscillator + two diodes and capacitors. Are there any simple higher voltage switched capacitor charge-pump IC's available, maybe you could construct one from discrete components (MOSFETs)? Most like MAX1682 are low voltage parts for mobile devices.
I think once there needs to be any additional circuitry for generating a higher polarisation voltage, then a simple voltage multiplier like the one I use on the LDC design on this page: OPIC LDC is probably all that is needed?.... Simple, cheap and easily adjustable .
Meanwhile, I have added notes to the schematic and description for this '48v' version recommending it only be used with transformerless preamps..
Many thanks for pointing out the shortcomings of asymmetric DC loading with transformer inputs -- I should have noticed myself of course! :oops:
 
It would be nice to have a lower BOM count voltage doubler as the current multiplier has quite a many components and required additional zener regulator, it can't be powered from the same voltage as the op amp can be (~30V or so). I think an astable multivibrator with two phase clock outputs + Pelliconi cascade charge pump (needs 2x PMOS + 2x NMOS FETs + 1 capacitor + RC) would be a nice solution, at least if you happen to find a MOSFET array with proper FETs (not low-voltage parts like on CD4007 etc.).
 
There are some 60V PMOS and NMOS pairs available from Panjit (https://www.mouser.fi/ProductDetail...user.fi/ProductDetail/Panjit/PJS6839_S1_00001) and others (https://www.mouser.fi/ProductDetail/Diodes-Incorporated/DMN65D8LDW-7) so you could build a discrete voltage doubler quite easily using two P-channel (for Pelliconi cell and two complementary inverter stages) and three N-channel ones (for Pelliconi, inverter and multivibrator which cannot drive the Pelliconi directly). In addition a couple of resistors and the capacitors are needed. Doesn't take much space on PCB. Maybe the Dickson type charge pump using the diode ladder would be simpler though not as good at least on paper.

Edit: Indeed if you build a Dickson type voltage doubler circuit using a complementary MOSFET pair as inverter you will come with a very low parts count circuit.
Complementary MOSFET pair:
https://fi.farnell.com/diodes-inc/bss8402dw-7-f/mosfet-np-ch-60v-sot-363/dp/1713834https://www.tme.eu/en/details/bss8402dwq-13/multi-channel-transistors/diodes-incorporated/Circuit: Fig 2A, just replace the inverter with the complementary pair:
https://gyraf.dk/schematics/Voltage_multipliers_with_CMOS_gates.pdf
 
Last edited:
There are some 60V PMOS and NMOS pairs available from Panjit (https://www.mouser.fi/ProductDetail...user.fi/ProductDetail/Panjit/PJS6839_S1_00001) and others (https://www.mouser.fi/ProductDetail/Diodes-Incorporated/DMN65D8LDW-7) so you could build a discrete voltage doubler quite easily using two P-channel (for Pelliconi cell and two complementary inverter stages) and three N-channel ones (for Pelliconi, inverter and multivibrator which cannot drive the Pelliconi directly). In addition a couple of resistors and the capacitors are needed. Doesn't take much space on PCB. Maybe the Dickson type charge pump using the diode ladder would be simpler though not as good at least on paper.

Edit: Indeed if you build a Dickson type voltage doubler circuit using a complementary MOSFET pair as inverter you will come with a very low parts count circuit.
Complementary MOSFET pair:
https://fi.farnell.com/diodes-inc/bss8402dw-7-f/mosfet-np-ch-60v-sot-363/dp/1713834https://www.tme.eu/en/details/bss8402dwq-13/multi-channel-transistors/diodes-incorporated/Circuit: Fig 2A, just replace the inverter with the complementary pair:
https://gyraf.dk/schematics/Voltage_multipliers_with_CMOS_gates.pdf
I think we may be entering the world of 'swings and roundabouts' here? (UK slang term for equal gains and losses)...
For use with a transformerless mic preamp, I think using 48 volts from one side of the phantom power supply in conjunction with a second 1G resistor - as I have in my schematic - probably offers the simplest option.
Selecting the feed resistors for the op-amp as 2k2 should give you around 30v DC to work with. Doubling that to 60v - but taking off 15v to allow for the input bias to the op-amp - will only give you around 45v as capsule bias. So no real advantage (except for transformer based preamps)
Making the feed resistors any lower than 2k2 is going to start attenuating the op-amp output, I would suggest?

Using a single chip multiplier - using a 40106 based on the circuits in your 'gyraf' link above - along with a few extra diodes and small caps gives you the option of providing a much wider range of polarisation voltages, with the option of retaining the simpler DC biased op-amp input, and retaining the use of a single 1G resistor.

The difference between 48v and 60 v polarisation results in something like a 2dB change in sensitivity, so probably doesn't warrant any extra components fitted, just to get those extra few volts.
Multiplying to higher voltages than 60v - if necessary - could call on a whole range of multiplier options.
That might include single bipolar transistor inductor based oscillators (as used by Schoeps), to CMOS multipliers like those in the 'gyraf' article you link to, through to multipliers based on some of the circuit configurations you have suggested. .
There are lots of options. And probably not a lot of cost difference, for individual mic builds in the hobby world?....
The creation of one off PCBs to allow for the mounting of SMD components is going to be more expensive than individual 'though hole' builds using stripboard....
Selecting the best option for multiple mic builds - or for commercial use - will present a different set of cost options of course...
 
Selecting the feed resistors for the op-amp as 2k2 should give you around 30v DC to work with. Doubling that to 60v - but taking off 15v to allow for the input bias to the op-amp - will only give you around 45v as capsule bias. So no real advantage (except for transformer based preamps)
If you put a 1nF cap between op amp input and the capsule and use another 1G resistor to bias the capsule you will have the full 60V on the capsule which is obviously connected to the 'real' ground (0V) instead of to the virtual ground (at halfway of op amp supply).

Btw. noticed on the FB micbuilders group that homero.leal (here) has already designed a PCB with integrated 60V bias on the same PCB as the op amp (no separate PCB). Wonder how did he do it (circuitwise)...

Anyway, it's nice to have options, for many the through hole design fits better. For mass production SMT might be better choice, though. You can also combine both (OPA164x's are not available in DIP package either so you have to use SMD).
 
Hi there!
I was finally able to assemble the pcb and test the microphone. Everything works, it is necessary to turn up the gain a bit but for close applications it can be used without problems. i don't think it makes so much sense to replace the schoeps circuit that is inside the takstar cm60, since it provides the 60V and eventually does the job, however i had a lot of fun and learned even more. I thank you all.

P.S. if anyone is interested in trying them out I have some pcb's left over ;) I don't want any money, just the few pennies for the mail

msg1851901527-5276.jpg
 
Impedance balancing is OK but the DC levels at preamp makes my worry at least with all transformer input mic preamps where the current caused by the DC difference goes through the primaries of input transformers, though as the transformer primaries have so low DC resistance the difference will not be that high causing also lower bias voltage. Don't know what this current will cause to transformer cores either, might be harmful. This scheme only works properly with capacitor coupled mic preamps which have those 6k8 resistors connected to +48 in P48 supply.
Although we were discussing this almost a year ago!! 😮 I've only just got round to addressing the DC imbalance problem, with my simple 'no voltage multiplier' OPIC mic.
I didn't like the limitation of the project not being suitable for use with transformer coupled mic preamps, so I decided that for the sake of sacrificing an extra 1dB of sensistivity it was worth making the system balanced - for DC, as well as the existing passive impedance balancing for AC.
So I re-did the schematic and stripboard (not a lot of change, so no big deal ) and have reposted the project as the OPIC42 version. (Copy of the schematic attached as well).
Project is renamed 'OPIC 42' because 42v DC is the maximum voltage that can be presented across the capsule with this schematic.
It can often be a volt or 2 less than that of course - depending on the impedance of the phantom power supply being used.

Quite a low value for polarisation voltage of course, but I still get around -34dB sensitivity when I'm using THIS CAPSULE .
The noise figure from the OPA1641 is around 7dB(A), so quite low. The 30v DC supply to the opamp gives plenty of headroom!

It's a simple 'starter' LDC project, and the OPA1641 op-amp really does perform well, as a simple impedance converter...
 

Attachments

  • OPIC42.schematic.jpg
    OPIC42.schematic.jpg
    198.8 KB
Back
Top