Skip to main content ↑ ↓

Notice

Please note that most of the software linked on this forum is likely to be safe to use. If you are unsure, feel free to ask in the relevant topics, or send a private message to an administrator or moderator. To help curb the problems of false positives, or in the event that you do find actual malware, you can contribute through the article linked here.
Topic: Capsule for PC Mic input (Read 12169 times) previous topic - next topic
0 Members and 1 Guest are viewing this topic.

Capsule for PC Mic input

I needed a mic to attach to my Samson SR850 headphones. I used one of these Panasonic WM-61A Electret Condenser mic capsules http://www.ebay.co.uk/itm/Panasonic-WM-61A...=item1c14d3e197 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 http://www.asus.com/uk/Motherboards/Z97A/specifications/ 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. 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) but maybe it's not the ideal voltage to drive this capsule, or it needs other circuitry to drive it optimally?

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 but would I be able to improve things by running an external amp to boost the signal before it goes into the PC, so that I can reset the boost to 0db and thus prevent the onboard audio boosting the noise? Would an external amp be able to increase the signal without increasing the noise? I think I might be able to set a line input device as my default comms device for speech recognition and Teamspeak has it's own audio input selector, so I should be able to select it there, if it will be necessary or better to boost the signal to line level externally, to avoid using the mic input at all. I've actually got a semi-pro mic amp (I'll have to dig it out and check the make/model) I bought for my SM58, which I could connect the mic to and then connect the amp to the PC's line input. It only has 1/4" and XLR inputs, so I'd need an adapter to convert the 3.5mm jack to a 1/4" one but I don't know if this amp is likely to even be suitable for this capsule.

So will I be able to increase the level and signal quality using this capsule with my PC or should I just give up on the idea and get a different capsule and if so, what?

Capsule for PC Mic input

Reply #1 –
Quote from: doveman on
I needed a mic to attach to my Samson SR850 headphones. I used one of these Panasonic WM-61A Electret Condenser mic capsules http://www.ebay.co.uk/itm/Panasonic-WM-61A...=item1c14d3e197 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 http://www.asus.com/uk/Motherboards/Z97A/specifications/ 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. 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) but maybe it's not the ideal voltage to drive this capsule, or it needs other circuitry to drive it optimally?

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 but would I be able to improve things by running an external amp to boost the signal before it goes into the PC, so that I can reset the boost to 0db and thus prevent the onboard audio boosting the noise? Would an external amp be able to increase the signal without increasing the noise? I think I might be able to set a line input device as my default comms device for speech recognition and Teamspeak has it's own audio input selector, so I should be able to select it there, if it will be necessary or better to boost the signal to line level externally, to avoid using the mic input at all. I've actually got a semi-pro mic amp (I'll have to dig it out and check the make/model) I bought for my SM58, which I could connect the mic to and then connect the amp to the PC's line input. It only has 1/4" and XLR inputs, so I'd need an adapter to convert the 3.5mm jack to a 1/4" one but I don't know if this amp is likely to even be suitable for this capsule.

So will I be able to increase the level and signal quality using this capsule with my PC or should I just give up on the idea and get a different capsule and if so, what?


The biggest problems I see with the Panasonic capsule you seem to be using is that it is omnidirectional and has great frequency response. It's exactly what you want in a measurement mic, and probably not what you want in a mic for picking up speech. The one exception being how the mic is positioned. If you are using it on a headset wand just a few inches from your mouth, an omni is a good choice.

If you using the mic on a separate stand and thus a foot or more from your mouth then a cardioid capsule would probably be a better idea.

However, intelligibility is generally enhanced when a mic is high pass filtered in the range of 50 to 120 Hz with a 12 dB/octave or higher roll off. Many cardioid mics have this built in as it can easily be implemented acoustically when the mic is a cardioid.

Looking at the genesis of cell phones, the mics that are in common use appear to be very similar. They are typically wideband omnis such as the one you are using.  However, many manufacturers have discovered the advisability of adding a high pass filter which ruins the mic for picking up wide-range music and measurements, but optimizes it for speech. The later models of many cell phones and smart phones have the high pass filter built in.

Actually implementing a speech-optimizing  high pass filter for a PC mic may be possible with software, but I am unaware of the details of doing such a thing. Doing so by means of hardware would take some engineering,  soldering and general craft work. I am unaware of anybody who has stepped up to this requirement.

I notice that the link to the ebay ad includes schematics that appear to have been reproduced from Linkwitz's site for constructing a 10 dB preamp. That project might be a good starting point for the high pass filter project. A good starting point would be to resize the 2.2 uF electrolytic coupling capacitor marked "7 Hz" by a factor of from 10 to 20 which would probably include changing it to a film (mylar) cap.

Capsule for PC Mic input

Reply #2 –
Quote
I've actually got a semi-pro mic amp (I'll have to dig it out and check the make/model) I bought for my SM58, which I could connect the mic to and then connect the amp to the PC's line input. It only has 1/4" and XLR inputs, so I'd need an adapter to convert the 3.5mm jack to a 1/4" one but I don't know if this amp is likely to even be suitable for this capsule.
You'd need a battery to power the electret.  Computer soundcards provide 5V for "computer microphones", but your preamp won't provide that..

Your preamp may provide 48V phantom power for studio condenser mics, but that's not appropriate for a computer mic. 

Stage & studio electret microphones usually require a battery, although some use phantom power.  (Your SM58 is a dynamic mic and it doesn't need power.)

Capsule for PC Mic input

Reply #3 –
Quote from: DVDdoug on
Quote
I've actually got a semi-pro mic amp (I'll have to dig it out and check the make/model) I bought for my SM58, which I could connect the mic to and then connect the amp to the PC's line input. It only has 1/4" and XLR inputs, so I'd need an adapter to convert the 3.5mm jack to a 1/4" one but I don't know if this amp is likely to even be suitable for this capsule.
You'd need a battery to power the electret.  Computer soundcards provide 5V for "computer microphones", but your preamp won't provide that..


Alll good points.

Quote
Your preamp may provide 48V phantom power for studio condenser mics, but that's not appropriate for a computer mic.


Common inexpensive measurement mics such as the Berhinger ECM 8000 are based on a Chinese clone of the Panasonic Electret capsule. It contains the necessary electronics to run an electret capsule off of phantom power.

Probable Schematic of Behringer ECM 8000

Note there are several different mics that were sold as the ECM8000, at least one of which had a transformer. The above is probably representative of the currently sold product.

Quote
Stage & studio electret microphones usually require a battery, although some use phantom power.  (Your SM58 is a dynamic mic and it doesn't need power.)


IME not so much. Stage and studio mics  that have electret capsules of one kind or the other under their covers are actually pretty common, and as a rule they run off of regular phantom power.  Furthermore, stage or studio mics that have a battery power option, may not be electrets.  Some times you have to read the service manual and/or a lot of fine print to find out that a mic is actually has an electret  capsule because electret capsules don't have the best reputation, partially because so many of them are absolutely dirt cheap.

Note that standard phantom power voltages are 12 volts, 24 volts and 48 volts. As a rule these are approximate minimums for the lower 2, and the approximate maximum for 48 volts. Most rely on robust voltage regulators to cut the supplied phantom power down to what they actually need. For example, a lot of the 12 volt mics are based on 3 volt electret capsules (ECM 8000 being included as it is actually a 12 volt phantom power mic) but they all generally run fine to well above 48 volts.

Interestingly enough the DPA/B&K measurement mics (e.g. 4006) are almost all electrets.

Capsule for PC Mic input

Reply #4 –
Quote from: Arnold B. Krueger on
The biggest problems I see with the Panasonic capsule you seem to be using is that it is omnidirectional and has great frequency response. It's exactly what you want in a measurement mic, and probably not what you want in a mic for picking up speech. The one exception being how the mic is positioned. If you are using it on a headset wand just a few inches from your mouth, an omni is a good choice.


Yeah, I guess a unidirectional would be more suitable but it is on a headset boom/wand so probably doesn't matter as the gain just needs to be set to pick up my voice, which will be much louder than any background noise. The boom I'm using is a bit shorter than is ideal, so the mic sits at the corner of my mouth rather than directly in front of it (or just below, where'd I'd normally position it to avoid excessive popping/sibilance) but I think it's close enough.

Quote
However, intelligibility is generally enhanced when a mic is high pass filtered in the range of 50 to 120 Hz with a 12 dB/octave or higher roll off. Many cardioid mics have this built in as it can easily be implemented acoustically when the mic is a cardioid.

Looking at the genesis of cell phones, the mics that are in common use appear to be very similar. They are typically wideband omnis such as the one you are using.  However, many manufacturers have discovered the advisability of adding a high pass filter which ruins the mic for picking up wide-range music and measurements, but optimizes it for speech. The later models of many cell phones and smart phones have the high pass filter built in.

Actually implementing a speech-optimizing  high pass filter for a PC mic may be possible with software, but I am unaware of the details of doing such a thing. Doing so by means of hardware would take some engineering,  soldering and general craft work. I am unaware of anybody who has stepped up to this requirement.


It's quite simple to alter/filter the input with software, as the plethora of gimmicky voice-changing programs show. Having said that, I haven't come across such a program with a suitable preset for this purpose and the only way I've seen to setup custom filtering involves installing Virtual Audio Cable, which allows me to patch the input to a filtering program (I forget the name just now) and that outputs to a virtual device which can be selected as the input in Teamspeak, etc, so it's a bit convoluted and requires starting several programs in the right sequence.

Quote
I notice that the link to the ebay ad includes schematics that appear to have been reproduced from Linkwitz's site for constructing a 10 dB preamp. That project might be a good starting point for the high pass filter project. A good starting point would be to resize the 2.2 uF electrolytic coupling capacitor marked "7 Hz" by a factor of from 10 to 20 which would probably include changing it to a film (mylar) cap.


Thanks for the suggestion, although 22uf-47uf film caps seem to cost £6+. Such values are available as cheap (8p-68p) electrolytic capacitors, so why do you suggest it would need to be a film cap?  Also, do you know what benefits are obtained by cutting the trace from terminal 2 to GND on the capsule and joining terminals 1 and 3, as shown in a couple of the diagrams on the e-bay page?

Capsule for PC Mic input

Reply #5 –
Quote from: DVDdoug on
Quote
I've actually got a semi-pro mic amp (I'll have to dig it out and check the make/model) I bought for my SM58, which I could connect the mic to and then connect the amp to the PC's line input. It only has 1/4" and XLR inputs, so I'd need an adapter to convert the 3.5mm jack to a 1/4" one but I don't know if this amp is likely to even be suitable for this capsule.
You'd need a battery to power the electret.  Computer soundcards provide 5V for "computer microphones", but your preamp won't provide that..

Your preamp may provide 48V phantom power for studio condenser mics, but that's not appropriate for a computer mic. 

Stage & studio electret microphones usually require a battery, although some use phantom power.  (Your SM58 is a dynamic mic and it doesn't need power.)


Yeah, I'd have to provide the power separately, although I could just take 5v from a USB hub or a portable, rechargable USB battery (which would probably be a cleaner supply). I don't particularly want to use my preamp though, as it's quite bulky, so it'll be better if I build the circuit Arnold suggested I try. I wonder if I could use 5v to power that, rather than the 9v shown in the diagram?

Capsule for PC Mic input

Reply #6 –
Quote from: doveman on
Quote from: Arnold B. Krueger on
The biggest problems I see with the Panasonic capsule you seem to be using is that it is omnidirectional and has great frequency response. It's exactly what you want in a measurement mic, and probably not what you want in a mic for picking up speech. The one exception being how the mic is positioned. If you are using it on a headset wand just a few inches from your mouth, an omni is a good choice.


Yeah, I guess a unidirectional would be more suitable but it is on a headset boom/wand so probably doesn't matter as the gain just needs to be set to pick up my voice, which will be much louder than any background noise. The boom I'm using is a bit shorter than is ideal, so the mic sits at the corner of my mouth rather than directly in front of it (or just below, where'd I'd normally position it to avoid excessive popping/sibilance) but I think it's close enough.


Miniature electret omnis on short booms are now very popular for  sound system and recording purposes.  The omni capsule is an advantage here because directional mics have inherent proximity and wind noise (including breathing) problems that the omnis avoid.  These mics need not be in front of the mouth to be effective, and work well even when on short booms that go only half way down the cheek.

Quote
Quote
However, intelligibility is generally enhanced when a mic is high pass filtered in the range of 50 to 120 Hz with a 12 dB/octave or higher roll off. Many cardioid mics have this built in as it can easily be implemented acoustically when the mic is a cardioid.

Looking at the genesis of cell phones, the mics that are in common use appear to be very similar. They are typically wideband omnis such as the one you are using.  However, many manufacturers have discovered the advisability of adding a high pass filter which ruins the mic for picking up wide-range music and measurements, but optimizes it for speech. The later models of many cell phones and smart phones have the high pass filter built in.

Actually implementing a speech-optimizing  high pass filter for a PC mic may be possible with software, but I am unaware of the details of doing such a thing. Doing so by means of hardware would take some engineering,  soldering and general craft work. I am unaware of anybody who has stepped up to this requirement.


It's quite simple to alter/filter the input with software, as the plethora of gimmicky voice-changing programs show. Having said that, I haven't come across such a program with a suitable preset for this purpose and the only way I've seen to setup custom filtering involves installing Virtual Audio Cable, which allows me to patch the input to a filtering program (I forget the name just now) and that outputs to a virtual device which can be selected as the input in Teamspeak, etc, so it's a bit convoluted and requires starting several programs in the right sequence.

Quote
I notice that the link to the ebay ad includes schematics that appear to have been reproduced from Linkwitz's site for constructing a 10 dB preamp. That project might be a good starting point for the high pass filter project. A good starting point would be to resize the 2.2 uF electrolytic coupling capacitor marked "7 Hz" by a factor of from 10 to 20 which would probably include changing it to a film (mylar) cap.


Thanks for the suggestion, although 22uf-47uf film caps seem to cost £6+. Such values are available as cheap (8p-68p) electrolytic capacitors, so why do you suggest it would need to be a film cap?  Also, do you know what benefits are obtained by cutting the trace from terminal 2 to GND on the capsule and joining terminals 1 and 3, as shown in a couple of the diagrams on the e-bay page?


Not 10 or 20 times larger, but 10 or 20 times smaller!  I'm thinking 0.1 - 0.2 uF which is a cheap fairly small film cap.

Linkwitz's circuit would need to be re-engineered to operate off of a single polarity supply were you to operate it off of +5 volts. The TI OPA x134 family of devices are near their low voltage limit when operated off of a single polarity +5 volt supply, and choosing a different op amp might be in order.

Capsule for PC Mic input

Reply #7 –
Quote from: Arnold B. Krueger on
The omni capsule is an advantage here because directional mics have inherent proximity and wind noise (including breathing) problems that the omnis avoid.


Although the level may typically be lower, I'm not sure, some omnis definitely still claim to suffer from both proximity effect and wind noise, according to Behringer. Example:

Capsule for PC Mic input

Reply #8 –
Quote from: mzil on
Quote from: Arnold B. Krueger on
The omni capsule is an advantage here because directional mics have inherent proximity and wind noise (including breathing) problems that the omnis avoid.


Although the level may typically be lower, I'm not sure, some omnis definitely still claim to suffer from both proximity effect and wind noise, according to Behringer. Example:



Please check the Wikipedia article on proximity effect. It pretty well covers the relevant science.  There are cardioid mics (example "Variable D" mics from EV)  that have vastly reduced proximity effect due to a sophisticated method for terminating the rear port.  I forgot that dual capsule cardioids often lack proximity effect because they lack the rear porting. They are almost a separate class of mic than regular rear-ported single-diaphragm cardioids.

A true omni mic like the ECM 8000's that I've used don't have any notifiable proximity effect. Maybe they changed it lately. I dunno.

Capsule for PC Mic input

Reply #9 –
Quote from: Arnold B. Krueger on
Miniature electret omnis on short booms are now very popular for  sound system and recording purposes.  The omni capsule is an advantage here because directional mics have inherent proximity and wind noise (including breathing) problems that the omnis avoid.  These mics need not be in front of the mouth to be effective, and work well even when on short booms that go only half way down the cheek.


Ah yeah, they're quite common to see being used by musicians, comedians, etc on stage but I didn't know what type of mic they were. I also often see them used in movies by operators in control rooms but they're probably less suitable for that purpose in reality, where there's other people sitting next to you talking, so they're probably just used because they look cool.

Quote
Not 10 or 20 times larger, but 10 or 20 times smaller!  I'm thinking 0.1 - 0.2 uF which is a cheap fairly small film cap.


Ah, woops! Those values are also available as electrolytic for as low as 3.6p each (in packs of 5) http://cpc.farnell.com/webapp/wcs/stores/s...14718,110053017

The cheapest film caps are a bit more expensive, around 12p each (with some very expensive ones that I don't need), although there's at least one option to buy a single one of those. It's always handy to have a few spares though. http://cpc.farnell.com/webapp/wcs/stores/s...53017,110066864

Whilst the difference in cost is insignificant, is there actually a reason to prefer the film cap over the electrolytic for this application?

Quote
Linkwitz's circuit would need to be re-engineered to operate off of a single polarity supply were you to operate it off of +5 volts. The TI OPA x134 family of devices are near their low voltage limit when operated off of a single polarity +5 volt supply, and choosing a different op amp might be in order.


Hmm, yeah a single 5v is quite a drop from two 9v in series. I already ordered a OPA2134 as it was only £2, so I guess I could try it but what alternative op amp might you suggest? How about the LM321http://www.ti.com/lit/ds/symlink/lm321.pdf or the LM358 http://www.ti.com/lit/ds/symlink/lm158-n.pdf

It doesn't look like that circuit will output a high enough signal to feed into the PC's line input though, as I think line level is meant to be around 775mv and from the diagram it looks like it only outputs 20-40mv. Maybe it would be OK to feed the mic input though, if the signal is high enough that the noise inherent to this input won't be an issue? Otherwise, is there something else I could add to the end of the circuit that would run off +5v and boost the signal to line level, with less noise than the PC input?

Capsule for PC Mic input

Reply #10 –
Quote from: doveman on
Whilst the difference in cost is insignificant, is there actually a reason to prefer the film cap over the electrolytic for this application?


A general rule is that film capacitors are preferred for filters wherever practical because they are generally higher tolerance and more stable with changes in time and temperature.

Quote
Quote
Linkwitz's circuit would need to be re-engineered to operate off of a single polarity supply were you to operate it off of +5 volts. The TI OPA x134 family of devices are near their low voltage limit when operated off of a single polarity +5 volt supply, and choosing a different op amp might be in order.


Hmm, yeah a single 5v is quite a drop from two 9v in series. I already ordered a OPA2134 as it was only £2, so I guess I could try it but what alternative op amp might you suggest? How about the LM321http://www.ti.com/lit/ds/symlink/lm321.pdf or the LM358 http://www.ti.com/lit/ds/symlink/lm158-n.pdf

It doesn't look like that circuit will output a high enough signal to feed into the PC's line input though, as I think line level is meant to be around 775mv and from the diagram it looks like it only outputs 20-40mv. Maybe it would be OK to feed the mic input though, if the signal is high enough that the noise inherent to this input won't be an issue? Otherwise, is there something else I could add to the end of the circuit that would run off +5v and boost the signal to line level, with less noise than the PC input?


I don't have any experience with circuits running at really low supply voltages. I'd end up using one of the parametric device selectors in the TI or other competitive chip manufacturer (National, Maxim, THAT etc. ) web site. Have fun!

Re: Capsule for PC Mic input

Reply #11 –
Embarassingly, due to poor health and being busy with court cases, I haven't had a chance to do anything with this before now.

I've just knocked up this circuit (second picture) http://tymkrs.tumblr.com/post/9168179996/diy-microphone-preamp on protoboard but then I realised that it doesn't indicate which way round the capacitors (I think most capacitors are polarized) need to be connected, so can anyone advise me please?

Another issue I've run into is that my PC's mic input has 2.7v on both the tip and ring, although I've measured across them to confirm that they're isolated from each other. I'm going to try powering the circuit with this (from the ring) instead of 5v, as it will simplify things if I don't need to tap another power source but won't I need to do something to stop the voltage feeding back into the output of the circuit from the tip connected to it, or will C3 suffice for that?

I don't really understand how it can "hear" the output from the mic/pre-amp when there's 2.7v sitting on the tip already but maybe it's because that's DC and the audio signal is AC?

Re: Capsule for PC Mic input

Reply #12 –
I don't consider poor health to be embarassing, but rather as unfortunate. ;)

Your PC's mic input is probably stereo, so two microphones could be connected to it. This isn't always so, but it seems to be the usual configuration nowadays.

I only became aware of this thread just now, so I couldn't comment earlier, but I should say that I'm not convinced yet that you actually need a separate preamp. In fact I don't exacly see how you connected your capsule to the PC in your first tries. Maybe there is the reason for your poor sensitivity, especially when I look at the various confusing ways of connecting it on the ebay auction page.

A bit of background is perhaps in order:

An electret microphone contains a FET as an impedance converter, which needs some operating current to actually amplify. This is provided through a pullup resistor connected to a voltage source, which in a PC would be typically 3.3V or 5V, but could be anything in the range of typically 2-12V. More would endanger the FET, less would cripple the performance. The resistor value of 2.2k seems to be used almost universally. The signal gets coupled out via a capacitor to get rid of the DC bias voltage. See the conventional hookup picture on the auction page.

In a PC, both the resistor and the capacitor or its equivalent are built into the microphone input. These days the circuitry is part of the audio codec chip. So you can connect the capsule directly, between ground and the signal pin. No further components needed. Use the tip of the jack as the left signal pin, and the ring of the jack as the right signal pin, with the sleeve being ground for both. If you have only a single capsule to connect, just use the tip and forget about the ring.

There is a possibility to connect the capsule to both tip and ring in order to feed both channels from the same capsule, but that feeds twice the current into the FET. That doesn't necessarily make it fail, but is likely to reduce its sensitivity.

So before you build your own electronics, try first to connect the capsule without using further components between sleeve and tip, leaving the ring unconnected. You should then get a signal on the left channel only. If this can be made loud enough, with sufficient quality, I don't see a need to build anything.

If you are unhappy with the result, we will have to look further, but only then.

Re: Capsule for PC Mic input

Reply #13 –
Hi pelmazo, thanks for the advice.

I just tried wiring only the tip and sleeve and as you suspected, the input is stereo as that gives me a signal on the left but only buzzing on the right. There's a slider in the Realtek control panel for the mic and if I move that all the way to the left it cuts out the buzzing from the right.

So that's why I would have joined both the tip and ring to the capsule +ve.

Whether just the tip or the tip and ring are connected however, the performance/volume is the same and I have to set the mic input's boost to +30db to really make myself heard, which of course increases the noise floor as well, which is why I wanted to build a pre-amp so that I could feed a higher signal in and avoid using the soundcard boost.

Re: Capsule for PC Mic input

Reply #14 –
Given the technical data of the capsule, I don't quite understand why it is still too low in level at 30 dB mic gain, but if it is so you indeed need an amplifier.

However, the web page you have hit upon offers a very crude amplifier design. I believe you would be better served by a design using an OpAmp, which should be overall better behaved.

See for example the very detailed application note tidu765 by Texas Instruments, which explains just about everything in the design of such an amplifier, including schematics, parts lists, calculation of component values and measurements.

Re: Capsule for PC Mic input

Reply #15 –
Thanks for the link, that document has a lot of good information. The diagrams still omit the capacitor orientation information though and whilst it describes this at s5.1 for C3 and C5 (both with +ve towards the op-amp) , it doesn't for C2 and C6.

I'd still like to try the circuit I linked to first, as I've already got the components and have built it and would like to have a usable mic for this Thursday if possible and I can try and improve on it later. I also want to see if I can get away with using the 2.6v from the mic input whereas the op-amp circuit calls for 9v (or 5v using the modified design in s.7, which I could take from USB but that's still more complicated than using the supply from the mic input).

Taking my lead from that note, I'd guess that all the capacitors in the crude design should have +ve towards the transistors. I'd also guess that having 2.6v DC on the tip and thus on the output of the circuit won't be a problem, as this sits on the output of the capsule when no circuit is used and that's OK but I'd appreciate confirmation on both points before I try it.

Re: Capsule for PC Mic input

Reply #16 –
Looking at the datasheets for the OPA322 http://www.ti.com/lit/ds/symlink/opa322.pdf and the LMV796 http://www.ti.com.cn/cn/lit/ds/symlink/lmv796-q1.pdf shows they can operate as low as 1.8v, so in theory the modified circuit in the note might work using the 2.6v from the mic input, although what difference there might be in terms of maximum gain, noise, THD, etc. compared to running at 5v I don't know.

Another option might be the MC33502 http://www.onsemi.com/pub_link/Collateral/MC33502-D.PDF which can operate as low as 1v. It has higher noise and THD than the OPA322 but whether that would matter would depend on the soundcard circuitry I guess, as that might already have higher noise.

It occurred to me however that I'll still need to short the tip and ring as otherwise the right mic channel will just have noise on it and I doubt it's possible to make my comms software only use the left channel. This means that if I'm using the ring to provide the +ve supply then the output (going to the tip) and the +ve rail will be joined but as I mentioned before, I'm not sure this matters as the supply is DC and the audio signal is AC.

EDIT: I can't find any of those op-amps at the consumer suppliers like CPC Farnell, although the OPA2322 and LMV797 are on e-bay for £2-3 each
http://www.ebay.co.uk/itm/351591053501
http://www.ebay.co.uk/itm/172196250244

Unfortunately they're both SMD components, which I have no experience using but I guess if I have to I can just solder the other components to the legs and not worry about soldering the op-amp to the board. I thought I might as well get the dual versions, as then I'll have a spare op-amp I could use to build a switchable filter circuit if I fancy.

Re: Capsule for PC Mic input

Reply #17 –
If you think about using the bias voltage from the PC's mic input, remember that there is a series resistor in the PC (most likely 2k2), which will limit the amount of current you can draw. Furthermore, it is the same wire that will have to carry the audio signal. This seriously limits your options, and rules out many of the available OpAmps due to their current demand.

In practice, if you want to use an external amplifier, you will have to provide a power supply for it. Since you need at least 2 V for the capsule bias, you may be able to use a coin cell, if small size is a factor.

Re: Capsule for PC Mic input

Reply #18 –
Yeah, I guess I'll have to use an external power supply. I've actually got some rechargable USB 5v battery packs, so I can use one of them and I think they'll work in passthrough mode whilst charging so I won't have to worry about them going flat and my mic going dead.

I've actually been using one to power the transistor circuit but even with the battery the signal is quieter with the circuit than without it! I'm not sure what could be wrong, other than the orientation of the capacitors as I'm still not entirely sure which way round they should be but I don't know if that would make the signal decrease in volume.

Re: Capsule for PC Mic input

Reply #19 –
There's an easy way to find out if a capacitor needs to be reversed (for aluminum caps): Measure the DC voltage in-circuit, and in operation, across the capacitor. If it is negative, and beond a few hundred millivolts, it is the wrong way up.

In reverse direction, an electrolytic cap will start to conduct significant amounts of leakage current somewhere around 1-2 V of reverse bias. This will also reduce the part's lifetime. If you force enough current in this state, the part will disintegrate.

Re: Capsule for PC Mic input

Reply #20 –
OK, I measured across the caps with them orientated as shown in this picture and it seems OK, with voltage only on the +ve sides, except for C3 which has +2v on the -ve/output side (which is coming from the mic input) and -2v on the +ve/emitter side  (I'm not actually sure that the tantalum cap C3 is polarized, it's hard to tell).
https://s31.postimg.org/r5lk6m94b/Mic_Pre_amp_Voltages_across_caps.png

I had to use a voltage divider of two 100k resistors in series to ground to get 2.5v from the USB battery's 5v in order to power the electret capsule. To simplify things, I removed R3 and put C3 on the output of the first transistor to test the first half of the circuit. With the battery off, the circuit seems to just generate a lot of white noise but once I turn it on, there's definitely a boost to the signal and it seems to give my voice a bit more presence but unfortunately there's also a significant increase in the noise level, so there's not really any benefit compared to just using the soundcard's boost.

Also, after a short while the signal drops significantly. I wonder if this could be caused by one or more of the capacitors discharging for some reason?

I then connected the second half of the circuit up again but that actually resulted in the signal being lower than with just the first half. I did find whilst checking the caps that I'd put the second transistor round the wrong way though, so maybe that was enough to damage it? I've got a couple of spares so I could use one of those to test but it seems that the transistors just boost the noise too much, or introduce noise of their own, making the circuit pretty useless. I'm seems strange though, considering that the website I got it from describes it as being as good as their £500 pre-amps (albeit they weren't using it with an electret capsule).

Re: Capsule for PC Mic input

Reply #21 –
I found a datasheet for my Realtek ALC892 soundcard http://www.hardwaresecrets.com/datasheets/ALC892-CG_DataSheet_1.3.pdf

It says on page 17 that the mic input bias voltage is software selectable between 2.5v/3.2v/4.0v but I can't find any setting for that in the Realtek control panel and it seems to be fixed at 2.5v on my board. Maybe they did this because that's what electret capsules need, which seems to be confirmed by the table on p.,3 of this PDF http://educypedia.karadimov.info/library/aes2003.pdf and the e-bay page for my capsule seems to specify 2v as well.

I'm not really sure I needed to use a voltage divider to provide 2.5v to the capsule though, or if I could have just used a single 10k resistor like in this diagram http://www.instructables.com/id/Pre-amp-to-electret-mic/

I tested with a single resistor (I tried 100k and 480k) and they only dropped the 5v to about 4.8v, so I was worried that'd be too much for the capsule. However this page, which specifically discusses using electret mics with the Soundblaster soundcards, shows the bias voltage as 5v so maybe it would be OK. I don't know what difference, if any, the voltage would have on the signal quality or level http://www.epanorama.net/circuits/microphone_powering.html

The only other potentially useful information I found in the ALC892 PDF is on p.82, where it says the analog input impedance is 64K, although that might be referring to the line input rather than the mic input. If it's the latter though, perhaps if I match the circuit output impedance to that it will give the best results. I need to get the circuit working properly first though!

Re: Capsule for PC Mic input

Reply #22 –
Quote from: doveman on
https://s31.postimg.org/r5lk6m94b/Mic_Pre_amp_Voltages_across_caps.png
This loads as an empty file here.

Quote
I had to use a voltage divider of two 100k resistors in series to ground to get 2.5v from the USB battery's 5v in order to power the electret capsule.
That resistance is almost certainly too high for powering an electret microphone. How do you arrive at such values?

I can't comment further without seeing the pictures.

Quote from: doveman on
It says on page 17 that the mic input bias voltage is software selectable between 2.5v/3.2v/4.0v but I can't find any setting for that in the Realtek control panel and it seems to be fixed at 2.5v on my board.
It is unusual for driver programmers to expose such settings to a user. Most users will not know what to set, and why.

Quote
Maybe they did this because that's what electret capsules need, which seems to be confirmed by the table on p.,3 of this PDF http://educypedia.karadimov.info/library/aes2003.pdf and the e-bay page for my capsule seems to specify 2v as well.
2V is commonly the minimum for analog electret microphones. You have to take into account that the bias is applied through a resistor, which is dropping a voltage as soon as the capsule draws current (see Ohm's law). Hence the bias voltage as supplied by the PC isn't going to be seen by the capsule in its full extent in practice. For example, assuming the typical 2kOhm resistor value, a current of 300 µA will drop 0.6V, so that the capsule sees 1.4V out of a 2V bias supply.

I have yet to see a general purpose capsule that doesn't support a 5V bias supply through a 2kOhm resistor, so in actual practice you would find that whatever the selection of your sound chip bias voltage, it would work. Furthermore, you have seen the specs of your capsule, haven't you?

Quote
I'm not really sure I needed to use a voltage divider to provide 2.5v to the capsule though, or if I could have just used a single 10k resistor like in this diagram http://www.instructables.com/id/Pre-amp-to-electret-mic/
What's wrong with using the recommended 2.2kOhm value? Not that the value is very critical, but it is usually a good idea to use the values recommended by the data sheet.

Quote
I tested with a single resistor (I tried 100k and 480k) and they only dropped the 5v to about 4.8v, so I was worried that'd be too much for the capsule.
The maximum voltage the capsule is specified to work with is 10V, so I don't understand why you are worried.

Quote
I don't know what difference, if any, the voltage would have on the signal quality or level http://www.epanorama.net/circuits/microphone_powering.html
It depends on the capsule internals. If it just uses a JFET, as most capsules do (the data sheet doesn't say explicitly, but it is likely), the resistance may have an impact on the level, but the overriding concern is to operate the capsule within its specified operating range. This concerns the bias voltage and bias current.

Quote
The only other potentially useful information I found in the ALC892 PDF is on p.82, where it says the analog input impedance is 64K, although that might be referring to the line input rather than the mic input. If it's the latter though, perhaps if I match the circuit output impedance to that it will give the best results. I need to get the circuit working properly first though!
This is going to be the AC input impedance, and it will most likely be with the bias supply off. In other words you are looking at a line level input. With the bias supply connected, such a high input impedance would only be possible with a constant current bias circuit, which is unlikely. The datasheet of your soundchip is quite sketchy and far from revealing all relevant detail. There's no mentioning of the resistor value, for example. It is going to be something between 2k and 10k, but you'd need to measure to be sure.

For a line level input, 64 kOhm is within the usual range, which goes from about 22kOhm up to 100k in consumer gear, with the occasional device falling outside this range. There's no point in impedance matching, you should rather keep your driving device's source impedance way below the input impedance of the receiving device. A factor of 5 or more is recommended and common.

By the way, I just pulled a somewhat larger electret capsule out of my junk drawer and hooked it up to my laptop's microphone jack. It needed only 10dB of microphone gain to be well loud enough when spoken into from close range. My laptop uses a different sound chip, but allows for 30 dB microphone gain, too. Too bad I don't find any information on the capsule anymore, as it doesn't show any markings.

Re: Capsule for PC Mic input

Reply #23 –
Quote from: pelmazo on
This loads as an empty file here.

Sorry, not sure what happened there. Try this one https://s8.postimg.org/8xwh50l6d/Mic_Pre_amp_Voltages_across_caps.png

I've being trying this circuit today http://www.instructables.com/id/Pre-amp-to-electret-mic/ and it seems to work better. Using 100k for R2 the gain was too high and likewise with 47k but with 10k it's about right but it doesn't have the hiss/white noise I was getting with the other circuit. It does have some annoying hum though, so I need to find a way to eliminate that.

Quote
I had to use a voltage divider of two 100k resistors in series to ground to get 2.5v from the USB battery's 5v in order to power the electret capsule.
Quote
That resistance is almost certainly too high for powering an electret microphone. How do you arrive at such values?
I was just looking for a way to drop the battery's 5v down to the same 2.5v that my PC's mic input supplies and using a voltage divider of two identical resistors halves the input voltage. I don't see why the resistance should matter or be "too high", surely what matters is the voltage going to the mic, which would be the same if I'd used 10k or even lower?

Quote
What's wrong with using the recommended 2.2kOhm value? Not that the value is very critical, but it is usually a good idea to use the values recommended by the data sheet.
I hadn't realised the datasheet recommended that, I see it now.

Quote
The maximum voltage the capsule is specified to work with is 10V, so I don't understand why you are worried.
Yeah, I only just noticed that too!

Quote
It depends on the capsule internals. If it just uses a JFET, as most capsules do (the data sheet doesn't say explicitly, but it is likely), the resistance may have an impact on the level, but the overriding concern is to operate the capsule within its specified operating range. This concerns the bias voltage and bias current.

It does seem that the circuit I'm testing today which provides 5v from the battery through a 10k resistor instead of 2.5v directly to the capsule (as the PC's mic input does and as I did using the battery and voltage divider) results in higher signal, as with both today's circuit and the previous one I was trying I've used 10k for R2, which I understand is what controls the gain and the other components that differ or are omitted only affect the HF filtering.

Quote
There's no point in impedance matching, you should rather keep your driving device's source impedance way below the input impedance of the receiving device. A factor of 5 or more is recommended and common.

Okay, thanks.

Re: Capsule for PC Mic input

Reply #24 –
Quote from: doveman on
Sorry, not sure what happened there. Try this one https://s8.postimg.org/8xwh50l6d/Mic_Pre_amp_Voltages_across_caps.png
This one works. The circuit doesn't provide capsule bias, so I assume your divider was connected in front of C1. The combination of C1, C2 and R1 is weird. It almost seems as if C1 was meant to be the capsule itself.

Quote
I've being trying this circuit today http://www.instructables.com/id/Pre-amp-to-electret-mic/ and it seems to work better. Using 100k for R2 the gain was too high and likewise with 47k but with 10k it's about right but it doesn't have the hiss/white noise I was getting with the other circuit. It does have some annoying hum though, so I need to find a way to eliminate that.
Both circuits are crude, to say the very least. They are likely to distort, even if you can make them work. An amplifier based on an OpAmp would be easier to get working properly.

Quote
I was just looking for a way to drop the battery's 5v down to the same 2.5v that my PC's mic input supplies and using a voltage divider of two identical resistors halves the input voltage. I don't see why the resistance should matter or be "too high", surely what matters is the voltage going to the mic, which would be the same if I'd used 10k or even lower?
Your reasoning assumes an unloaded voltage divider. However, the capsule loads it, which makes your calculation wrong. You should be able to measure this easily with a multimeter when you compare the voltage of the divider without and with the capsule connected.