I've used one of these Panasonic WM-61A Electret Condenser mic capsules Panasonic WM-61A Electret Condenser MIC Capsule 10PCS | eBay as they're meant to have high sensitivity, attached it to a boom I took from a cheap headset and used the boom's wiring and 3.5mm jack but the results have been rather disappointing.
Whilst it works with my motherboard Motherboard | Z97-A | ASUS UK with Realtek onboard ALC892 it's too quiet and I have to turn up the boost to +30db so that people can hear me properly on Teamspeak, although Windows speech recognition seems to work better with it at +20db, or at least with the input volume at a lower setting than is suitable for voice comms, so I can't really use both. Naturally the boost also boosts the noise, which is mostly filtered out by a setting in Teamspeak but no doubt confuses the speech recognition, which doesn't work reliably at all. The PC mic input does provide the voltage needed to drive the mic (when wiring it up it wouldn't work at all with just the signal + and GND connected, only once I connected the third pole to provide the voltage).
As far as I understand, the realtek onboard audio is fairly good quality nowadays and it's unlikely that a separate soundcard will make much difference, so I'm hoping I can improve things by building an external pre-amp, either to boost the signal a bit before going into the mic input or up to line level to go into the line input. The latter might have the advantage of being less noisy than the mic input but maybe if the signal is boosted enough so that I can leave the PC's mic boost on 0db there won't be much difference between the two.
There's a "Microphone with 10db preamp" circuit on the e-bay page for the capsule but that's obviously not going to be sufficient if I need to use 30db boost on the PC. I guess I'd need to add some extra circuitry on the end of that to boost it further to line input level (775mv if I recall correctly).
I've been advised that a high pass filter in the range of 50 to 120 Hz with a 12 dB/octave or higher roll off would boost intelligibility and that I could try changing the 2.2uf capacitor for 0.1-0.22uf to achieve that.
The other problem is that ideally I want to power it from 5v, which I can take from a USB hub or USB rechargable portable battery and that circuit is designed for two 9v supplies in series and whilst the capsule itself will be fine with 5v, I'm not sure if there's any suitable op-amps that will work at that voltage, particularly if I want to boost to line level. If it's impossible, I'll just have to get another higher voltage PSU but I'd rather avoid the extra mess if possible.
Would a LM358 http://www.ti.com/lit/ds/symlink/lm158-n.pdf or LM321 http://www.ti.com/lit/ds/symlink/lm321.pdf be good alternatives to the OP2134, or is there something better anyone can suggest?
Also, what's the advantage of cutting the trace from terminal 2 on the capsule and then shorting terminals 1 and 3?
Whilst it works with my motherboard Motherboard | Z97-A | ASUS UK with Realtek onboard ALC892 it's too quiet and I have to turn up the boost to +30db so that people can hear me properly on Teamspeak, although Windows speech recognition seems to work better with it at +20db, or at least with the input volume at a lower setting than is suitable for voice comms, so I can't really use both. Naturally the boost also boosts the noise, which is mostly filtered out by a setting in Teamspeak but no doubt confuses the speech recognition, which doesn't work reliably at all. The PC mic input does provide the voltage needed to drive the mic (when wiring it up it wouldn't work at all with just the signal + and GND connected, only once I connected the third pole to provide the voltage).
As far as I understand, the realtek onboard audio is fairly good quality nowadays and it's unlikely that a separate soundcard will make much difference, so I'm hoping I can improve things by building an external pre-amp, either to boost the signal a bit before going into the mic input or up to line level to go into the line input. The latter might have the advantage of being less noisy than the mic input but maybe if the signal is boosted enough so that I can leave the PC's mic boost on 0db there won't be much difference between the two.
There's a "Microphone with 10db preamp" circuit on the e-bay page for the capsule but that's obviously not going to be sufficient if I need to use 30db boost on the PC. I guess I'd need to add some extra circuitry on the end of that to boost it further to line input level (775mv if I recall correctly).
I've been advised that a high pass filter in the range of 50 to 120 Hz with a 12 dB/octave or higher roll off would boost intelligibility and that I could try changing the 2.2uf capacitor for 0.1-0.22uf to achieve that.
The other problem is that ideally I want to power it from 5v, which I can take from a USB hub or USB rechargable portable battery and that circuit is designed for two 9v supplies in series and whilst the capsule itself will be fine with 5v, I'm not sure if there's any suitable op-amps that will work at that voltage, particularly if I want to boost to line level. If it's impossible, I'll just have to get another higher voltage PSU but I'd rather avoid the extra mess if possible.
Would a LM358 http://www.ti.com/lit/ds/symlink/lm158-n.pdf or LM321 http://www.ti.com/lit/ds/symlink/lm321.pdf be good alternatives to the OP2134, or is there something better anyone can suggest?
Also, what's the advantage of cutting the trace from terminal 2 on the capsule and then shorting terminals 1 and 3?
Thanks. Is it really that simple and will it boost it by at least 30db? I just wonder why anyone would use a far more complicated op-amp circuit, like the one shown on the e-bay page, if that's all that's needed.
What output level would I have with that, something suitable for my PC's line input or only for the mic input? If the latter, I can always give it a try and do something else if I still find it too noisy I guess.
The electret mic needs to be capacitively coupled, and the transistor to use a self biased stage which is a 1meg resistor between base and collector.
Shown here on the left of the image, clarifies how it is connected using a 3 v supply.
Spy Circuits
Attachments
Hang on a microsecond. Stock (unmodified) WM-61As should only ever need two connections. Or did you perform the Linkwitz mod?
That is quite possible, since guess where I've found lesser Realtek chips to fare worse? Actual amplifier noise level wasnt that different (still miles away from a decent low-noise opamp, but generally good enough for an electret), but mic bias voltage on an ALC898 proved much cleaner than on an ALC269. And even that wasn't perfect.
LM358 is a noisy, distorted piece of garbage. They're about the cheapest kind of dual opamp money can buy. With some Class A bias they're no longer distorted but still awfully noisy. At the very least, get an NE5532. Or if power consumption is more critical, NJM2068.
Read the Linkwitz Lab article.
Besides, while a source follower provides lower output, it also provides some power supply rejection.
I haven't done the mod. I think what I had to do was connect all three wires coming from the PC mic input to two terminals on the capsule.
Ah, the NE5532, I remember them fondly as I used them for a project a long time ago. I might even have a couple buried in my cupboard somewhere, I'll have a rummage.
Thanks, that explains it nicely.
Yes that is all that is needed as the PC/laptop has a gain control. Just ignore the second transistor and use the USB supply for your power.
.
You might be interested in this, especially because it runs on the 5 volts you're interested in: DIY Microphone Preamp
The 2N3904 is on eBay for, dunno, a dime or so.
Human speech ranges from around 300Hz to 3.5kHz. Filtering out higher and lower frequencies is said to greatly improve intelligibility. Here's a calculator: Guitar Pedals: R-C Filter Calculator
The (in)famous Linkowitz mod is said to improve SPL, and flatten the frequency response of an electret capsule. However, Mr. Linkowitz discusses reference microphones in his article, whether his mod helps anything in more general useage is in the ear of the beholder.
Computer power supplies are not always as clean as we might hope. Maybe some additional power supply filtering in the form of, say, a 470uF electrolytic paralleled by an 0.1uF ceramic would be a good idea.
.
You might be interested in this, especially because it runs on the 5 volts you're interested in: DIY Microphone Preamp
The 2N3904 is on eBay for, dunno, a dime or so.
Human speech ranges from around 300Hz to 3.5kHz. Filtering out higher and lower frequencies is said to greatly improve intelligibility. Here's a calculator: Guitar Pedals: R-C Filter Calculator
The (in)famous Linkowitz mod is said to improve SPL, and flatten the frequency response of an electret capsule. However, Mr. Linkowitz discusses reference microphones in his article, whether his mod helps anything in more general useage is in the ear of the beholder.
Computer power supplies are not always as clean as we might hope. Maybe some additional power supply filtering in the form of, say, a 470uF electrolytic paralleled by an 0.1uF ceramic would be a good idea.
.
If I'm using the PC's gain to reduce the signal that's fine, I just want to avoid using it to boost the signal and increasing the noise.
Thanks, those links are useful. I'll grab some 2N3904 and capacitors (I've already got enough resistors) and give it a go. I guess I should probably add a PTC fuse in as well, just to protect the motherboard from any shorts.
"I've used one of these Panasonic WM-61A Electret Condenser mic capsules Panasonic WM-61A Electret Condenser MIC Capsule 10PCS | eBay as they're meant to have high sensitivity, attached it to a boom I took from a cheap headset and used the boom's wiring and 3.5mm jack but the results have been rather disappointing."
Here's what you need:
Recording and Measurement Microphones
Here's what you need:
Recording and Measurement Microphones
I've been so frustratingly busy, I've only just found the time to start experimenting with this. I built the circuit on protoboard to test it but then I realised that the 10uf capacitors I've got for C1 and C2 and polarized like these https://www.amazon.com/uxcell-Polarized-Aluminum-Electrolytic-Capacitors/dp/B005X251AE and the circuit doesn't indicate which side is positive, so I'm worried about blowing them up!
Also, I checked the voltage supplied from the PC's mic input and it's 2.6v. Do you think I might be able to run the circuit using that, as it will simplify things if I can take the power from the mic cable rather than having to find another source?
Another issue I've just found is that testing the 3-pole mic input socket on my PC shows that both the ring and tip have 2.7v on them.
I expected it to just be on one or the other, even though the mic only actually has two connectors and so they'd end up being joined at that end anyway, as otherwise it seems pointless using a 3-pole socket rather than a mono 2-pole one.
It means that when I connect the output of the circuit to the mic input, there's going to be 2.7v going into the output capacitor C3 and if I take the voltage from either the ring or tip to the voltage rail to power the circuit (assuming 2.7v is sufficient), the voltage rail and the output will be connected.
So I'm not sure what I should do now and would appreciate some advice.
I expected it to just be on one or the other, even though the mic only actually has two connectors and so they'd end up being joined at that end anyway, as otherwise it seems pointless using a 3-pole socket rather than a mono 2-pole one.
It means that when I connect the output of the circuit to the mic input, there's going to be 2.7v going into the output capacitor C3 and if I take the voltage from either the ring or tip to the voltage rail to power the circuit (assuming 2.7v is sufficient), the voltage rail and the output will be connected.
So I'm not sure what I should do now and would appreciate some advice.
EDIT: I just double-checked and although there is 2.7v on both the tip and ring (with the -ve probe on the sleeve/ground), they seem to be isolated from each other as confirmed by measuring the resistance between them.
So whilst I think I still need advice about blocking the voltage from the tip feeding back into the output, I don't think there'll be any connection between the voltage rail (fed by the ring) and the output.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.