Some Transound electret microphone capsules from JLI Electronics - photos, and dismantling three of them

http://www.jlielectronics.com/transsound/electrets/index.htm?Submit2=Electrets

http://groups.yahoo.com/group/micbuilders/

Neumann's  Microphones for Studio and Home-Recording Applications (4th edition, 1999 http://www.neumann.com/download.php?download=docu0002.PDF) is a good reference for understanding microphones.

All these microphones are "back electrets".  There is supposedly a "front electret" design, which I don't understand.  The main alternative to the "back electret" approach is the "foil-type" or "diaphragm-type", which I attempt to describe in the next paragraph.  These pages at Google Books contain a diagram and explanation too: Standard Handbook of Audio and Radio Engineering.

Many electret microphones were made, and I guess some may still be made, with the diaphragm being an electret film.  The front was metalized and the back was not.  The metalized layer is connected to ground.  The electret polarization gives the back of the diaphragm a charge, much like that back being a conductive sheet with a voltage on it.  (I will assume a positive charge.)  Very close behind the back surface of the diaphragm is a metal backplate.  (Or a conductive plastic back-plate - I have an old Panasonic 9.7 mm omni microphone like this.)  When the sound moves the diaphragm closer to the backplate, a temporary positive voltage is induced. (Electrons are attracted to the front surface of the backplate, due to the closer presence of the positive charge on the back of the electret diaphragm.)  When the diaphragm moves forwards, there is less capacitance from its rear surface to the backplate, and a negative voltage is induced.  These voltage changes are like sending a signal through a capacitor of a few picofarads or a few tens of picofarads.  So there needs to be a very high impedance amplifier right there to buffer the signal so it can sent somewhere else.

The microphones shown here use the "back electret" principle.  The diaphragm is not an electret.  It is (I assume) Mylar with an aluminium layer on the front.  This layer is grounded and the back of the Mylar has no charge built into it.  So the back surface of the diaphragm is capacitively coupled to ground, with the mylar being the insulator of the capacitance. 

The backplate has on its front surface a thin layer of electret material, with the top surface having a negative charge.  This is somewhat like putting down an insulating layer and then a conductive layer, and connecting a steady negative voltage to that layer, with respect to the metal backplate.

I understand this electret layer is some kind of fluoro carbon polymer - Teflon and the like.  I understand that after it is deposited, some kind of process (such as corona discharge) is used to permanently alter the polymer so its front surface has a charge with respect to the backplate.  Teflon, or whatever is used, is a very stable insulator and I understand the charge can persist for hundreds of years.  I haven't yet found good information about charge stability of electrets.  Please point me to any such information.

Please see my recent message http://tech.groups.yahoo.com/group/micbuilders/message/11824 for how I measured the thickness of the diaphragms and spacer rings.  (I used a micrometer with mirror-polished tungsten carbide tips - very carefully.)
 

Model            Diameter   Diaphragm  Spacer   Sensitivity
                            thickness  ring

TSB-120A   Uni    8.0 mm    ~3 um      ~75 um   -46 dB +/- 4

TSB-140AO  Omni  10.5 mm    ~6 um      ~60 um   -45 dB +/- 4

TSB-2555A  Uni   21.0 mm    ~6 um      ~75 um   -40 db +/- 4

Googling reveals that 3 um and 6 um are popular thicknesses of diaphragms of (non-electret) condenser microphones.

I have some older, definitely not back electret, Panasonic 9.7 mm omni electret cartridges.  Their backplates have no insulating coating.  Their diaphragms are ~10 and ~15 um thick.

TSB-120A: http://www.jlielectronics.com/transsound/electrets/tsb-120a.htm

Transound has a datasheet for the TSB-120A1, which shows a slightly different diagram and a 2.2k impedance:  http://www.transsound.com/work_show.asp?work_id=397

Only the first datasheet  depicts the capsule as shown above, with a 2.2k load resistor to ground from terminal 2.

Unidirectional 12mm diameter, internal FET with separate connections for source and drain, and with a load resistor which could easily be removed.

TSB-140A: http://www.transsound.com/work_show.asp?work_id=395

Unidirectional (cardioid) 14mm with internal FET giving access to both source and drain.

TSB-140AO: http://www.transsound.com/work_show.asp?work_id=386

Omni 14mm with internal FET giving access to both source and drain.

TSB-1460A: http://www.transsound.com/work_show.asp?work_id=396

Unidirectional (cardioid) 14mm diameter, without internal FET.

TSB-160A: http://www.transsound.com/work_show.asp?work_id=392

Unidirectional (cardioid) 16.1mm, without internal FET.

These did not have a solder tag on the screw - so they would be hard to use.  There is no sign of a solder tag in the data sheet, but there is no sign of one in the TSB-2555A datasheet, and these capsules do have one.  One can't very well solder to the screw.  It would probably not wet to solder and I expect the heat transfer would melt parts of the capsule.

JLI Electronics wrote to me on 2009-05-25, including a PDF of the revised spec sheet for the TSB0160A.  Below is an edited version of a drawing from that PDF, showing the addition of two solder lugs to the new design.  It is my impression from these drawings that they have added a thin printed circuit board to the capsule.  It is not entirely clear from the drawing how Terminal 1 (the backplate) is to be distinguished from Terminal 2 (ground).  "Term.2" looks like it has a different shape from "Term.1".  It would probably be possible to see the PCB trace, and of course a multimeter would show which is which.

I assume they will do this for the TSB-160AO too.

There are 6 vent holes, two of which have been sealed.  I guess this is to tune the capsule in terms of resonances and/or its unidirectional response.

Through the holes a random-weave dampening material can be seen.


There is also an omni version of this - the largest omni microphone Transound makes -  which I did not get.  This has no solder tag either.

TSB-160AO: http://www.transsound.com/work_show.asp?work_id=385

TSB-165A: http://www.transsound.com/work_show.asp?work_id=391

Unidirectional (cardioid) 16.5mm, without internal FET.

This appears to contain the same kind of open-cell foam sheet which I found inside the TSB-2555A.

TSB-2555A: http://www.transsound.com/work_show.asp?work_id=388

Unidirectional (cardioid) 26mm diameter, without internal FET.

To use the microphone, wires would be connected to both terminals of the FET and to the gold-plated copper track on this rear printed circuit board (PCB).

The capsule is sealed at the back in terms of short-term air pressure variations.  Since the crimped aluminium retention of the PCB is not glued or soldered, air would seep through the gaps between the aluminium and the PCB.  So when there are long-term changes in atmospheric pressures, the internal air pressure would follow that of the outside.  Without this - if the back was completely sealed - the diaphragm would be pushed in or out if the internal air pressure was not the same as that outside.

The FET has its gate pressed against a gold-plated stud in the transparent piece of plastic.  The source and drain are available separately.

Its only identifying marks are a few dots and "J34".  Richard Crowley of the Micbuilders list told me that this was probably a 2SK1109 N-channel FET, specifically made to be an impedance converter for electret microphones.  Indeed, "J34" is one grade of 2SK1109s:

Here we see the transparent piece with its back towards us and the back of the backplate.  The backplate is brass and is plated with something silvery.  The yellow appearance in this photo is not realistic.

There is a small spring - stainless steel, I guess - which connects the front part of the gold-plated stud to the backplate.

The backplate is 0.53 mm thick and 12.0 mm in diameter.

It is somewhat recessed in the transparent piece, which has an outside diameter of 13.0 mm, to match the 13.1 mm inside diameter of the aluminium shell.

The internal diameter of the cavity behind the backplate, which is the main area of the transparent piece visible here, is 10.5 mm.  The depth of the cavity, once the backplate is in position, is 0.7 mm.  Some of the volume is taken up by the central part with the gold-plated stud.  Also, there is some extra volume of air in the 17 holes in the backplate.

This is an omnidirectional microphone and there is no damping material in this cavity.  The two other microphones which I dismantled are directional and have damping material in their cavities.

Here we see the components in order from the front.  The grid and then the ring with its diaphragm.  The aluminized side is topmost and is visible here.  The other side is the plastic side.

The backplate with its electret layer is pressed up against the blue spacer ring, which presses against the non-conductive side of the diaphragm.

The spacer ring is 60 um thick.

The diaphragm is 6 um thick.


The two most important dimensions, apart from the diameter of the active part of the diaphragm: 10.5 mm, and the charge of the electret film on the front of the backplate are the thickness of the spacer and the depth of the cavity.

A 0.001 instantaneous increase in air pressure will compress the air in the cavity and so (in a simplistic view of the motion of the diaphragm) cause it to move 0.001 x 0.7 mm towards the backplate.  This is 0.7 um.  This displacement is a significant fraction of the distance from the back of the diaphragm to the backplate: 0.7 / 60 = 0.012 of the distance.  So I can imagine sound pressure at 0.001 of atmospheric pressure causing an instantaneous change of voltage on the backplate of 0.012 of whatever the effective voltage is on the top of the electret layer. 

Increasing the depth, or the total volume, of the cavity would increase sensitivity - but potentially lead to resonances.  Decreasing the spacer width would increase sensitivity, but limit the diaphragm's ability to move towards the backplate without hitting it.

The diaphragm's metalization (I assume it is aluminium) is thin enough that some light passes through it - I measured about 10%.  So it is possible to see clearly through it when looking at light fittings, light-globes etc.  A similar level of transparency can be found in aluminized food packaging films.  The diaphragms of the two older Panasonic 9.7 mm omni capsules which I dismantled had a similar amount of light transmittance.  Those diaphragms were thicker and were electrets.

How does this thin layer of aluminium survive?  It is on the front surface - this can be tested with a multimeter.  Aluminium oxidizes instantly (or very quickly at least) on contact with air, producing a thin layer of aluminium oxide, which is a ceramic, and protects against further oxidation.

Hundreds of millions of electret microphones in cell-phones have people talking into them with moisture, dust, vibration and whatever else.  Yet these electret microphones are extraordinarily reliable.

This capsule appears to be identical to the TSB-140AO except that:

1. The clear piece of plastic has some holes in it - five holes, of which one is blocked.
    
   
2. There are two pieces of damping material in the cavity.
   
   
3. The PCB has five large holes.

The cavity is the same - 10.5 mm internal diameter and 0.7 mm deep. 

The first layer of damping material is some kind of plastic foam, as described below for the TSB-2555A.  This annulus is 10mm wide.  It is about 0.8 mm thick and compresses to about 0.2 mm, or a little thinner if forced hard.

The second layer is some kind of open-weave synthetic material.  It is 100 um thick.

For a larger image similar to this: TSB-140A-dismantling-c.jpg .

Behind the clear plastic piece a second cavity is formed between it and the PCB.  This is not much of a cavity, due to the large holes in the PCB, but here are the measurements.

I estimate the PCB is 2.6 mm behind the rear rim of the clear plastic piece.  The internal diameter of the aluminium case is 13.6 mm.  Most of the back of the clear piece is recessed about 1.2 mm from the rim, so most of the second cavity has a depth of 3.8 mm.  The FET takes up some room and around the edge the depth is less due to the rim of the clear piece.  The inside diameter of this part of the cavity, due to the rim, is 11.0 mm.

I guess the volume of this second cavity, taking account of the FET and the rim, when adjusted to the diameter of the active diameter of the diaphragm (10.5 mm), would be about  3.5 to 3.8 mm. 

I will be interested to test these directional cartridges modified to be an omni by blocking the five holes in the PCB.  There would then be a smaller, damped, first cavity, four small holes and a second larger cavity.  I estimate the effective depth of the first cavity to be about 0.45 due to the space taken up by the damping materials.  If the rear holes were sealed, and the resistance of the four internal holes ignored (which is not realistic, except at lower frequencies) then in theory the two treated as a single cavity would have an effective depth of about 4.0 to 4.3 mm, which is about 5 times the effective depth of the single undamped cavity in the omni version: TSB-140-AO.

I would expect a larger cavity to lead to greater sensitivity, since for a given air pressure variation, the diaphragm must move more in order that the pressure in the cavity matches that on the top face of the diaphragm.  However, as the cavity becomes physically larger, this will be frequency dependent and there may be resonances.

This microphone has no FET.

Omidirectional microphones such as the TSB-140AO are pressure sensors. The cavity behind the diaphragm has no openings to the outside world.

The TSB-2555A is a "unidirectional" microphone, meaning it tends to sense sound waves coming from the front, but not from behind.   It senses pressure gradient between its front and back, so it has holes at the back.  It also has two damping materials in its cavity.

Its main body is based on turned metal - brass I guess - rather than pressed aluminium.

I did not have any special tools for unscrewing the ring via its holes - I simply used a 1mm drill, some pointy-nosed pliers and a small vise.  First, I used a vise to squeeze around the back of the case in multiple positions, to break the bonds between the metal ring and the case.  Then I gripped the front of the case in the vise and levered the drill in order to turn the ring. 

Now we see the white plastic piece, with the backplate firmly pressed into it. In front of the backplate is the spacer ring.  On the left we see the non-conductive back side of the diaphragm.

On the right, we are looking at the metal of the back-plate, which is coated in a thin, transparent, not entirely smooth, layer of electret material.  I assume this is Teflon or something similar. It seems to be slippery like Teflon, but the surface texture is not precisely flat.

I scraped a little of this from the surface and applied a soldering iron to it.  It softened, I think, but did not melt or burn.  I reckon it is Teflon or some other fluoropolymer film.

I have taken off the spacer ring.

The internal diameter of the spacer ring, which sets the effective diameter of the diaphragm, is 21 mm.

I have dropped out the diaphragm and its ring and placed the diaphragm front (and ring) downwards.

The diameter of the front hole is 16mm.

Here we see the front of the three parts - the case, the diaphragm's metalized side and ring and the front of the backplate with its associated white plastic piece.

The backplate is plated brass, 0.56 mm thick.  The 52 holes are about 1.2 mm diameter.

The diaphragm is ~6 um thick.

The spacer ring is 75 um thick.

As for the capacitance calculation for the TSB-140AO above, here us an estimate for the  the capacitance of the diaphragm to the backplate.


The cavity is 1 mm deep, but is filled with damping materials.

There is a little spring which goes from the head of what I will call a "screw", since it has a metal thread protruding out the back of the capsule.  There's no need to loosen the nut, but I did - and then it is hard to tighten it since the "screw" has no slot.

If this was an omni, there would be no holes in the white plastic piece.  There are 12 holes which cause the entire microphone, to some extent at least, to measure pressure gradient and therefore be directional.

The cavity is 1mm deep from the back of the back-plate.  However, this is not free air - it is filled with two discs of dampening material, both 20 mm wide:

1. At the front, a disc of open-cell foam (I guess this is the best description).  This is a foam plastic of some kind, which can be stretched somewhat, and which has cellular foam where the walls of each cell are largely open to the next cell.  So air can be blown through the disc quite easily.  I can't tell what sort of material this is - though perhaps I could be burning it with the soldering iron and comparing the smell with those of other samples burnt the same way. 
   
   If it is a polyurethane foam, then I imagine it will turn to sticky crap at sometime in the future - maybe decades from now.  However, there are many other types of foam which I think are more stable than this.
   
   This foam disc is 0.8 mm thick and can be squashed to 0.2 mm.  When trying to look through this disc at a bright light, it a few pinholes which appear to let through direct light, but the holes are too small and/or too few to be able to see an image through.
      
   
2. At the back of the cavity, a lightweight synthetic random weave material 80 um thick.  This is partially open, since it is possible to look through it and see the filament of a light-globe.
   

The holes in the white plastic piece are between 0.8 and 1.0 mm - I guess 0.9 mm.

The combination of the small holes and the damping material (depending on how the capsule is housed in an external body for the whole microphone assembly) causes delay and filtering of the sound which reaches the rear of the diaphragm.  This is explained with a diagram on page 17 of the above mentioned Neumann microphone handbook.

I didn't take photos of some earlier parts of the process.  The rear part of the capsule unscrews easily. This contains the FET, soldered to a small PCB.  The FET's identification markings are:



Eric Benjamin corrected my mistake in reading these numbers.  He informs me that this is a 2SK596 of the 'C' grade: IDSS = 210 to 350 uA. Here's the datasheet: 2SK596-datasheet-Sanyo.pdf .

The two parts of the body are made of turned brass.  This is mechanically quite a complex capsule.  I don't understand how it would screw into anything while making electrical contact at the same time, without soldered wires.

After the back part has been removed, a small ring with two slots can be unscrewed and the inner cartridge removed.  This is 10mm in diameter:

The body is not aluminium - it is a thin metal such as steel, however it is not magnetic.  It may be brass, but if so, it must be a very light yellow colour.

There are 8 holes. One of them is blocked with white material.  I think the holes are about 0.5 mm in diameter and that they are about 0.5 mm deep.

Here are all the parts after I had almost completely dismantled the inner capsule:

Just inside the seven holes in the front there is a piece of black synthetic random-weave material.

Above is the electret-layer front of the backplate, the spacer ring (75 um thick) and the diaphragm and its ring.  I think I can measure the thickness of the diaphragm with an accuracy of +/- 1 um, and my measurements suggest it is 3 um thick.

The backplate pops off easily.  There is a small spring connecting it to the conical front end of the plated metal stud in the middle of the white piece of plastic.

There is an annulus of foam material which looks similar to that in the TSB-2555A.  It is about 0.7 mm thick when not compressed and about 0.2 mm thick when squashed.

There was also an annulus of white random-weave material at the back of the cavity.  As I prised it out with a needle, it sprung away and is still AWOL somewhere in my workshop.

The cavity is about 7.5 mm wide and about 0.7 mm deep.

There is a lot of careful turning, painting and assembly in this microphone.  I think this is quite remarkable for a product which sells for USD$6.20 in quantities of 100 or more.

This is the only one of these microphones I bought, so I won't be able to test it.

The TSB-120A has been discussed on the Micbuilders list.  Searching for it at:

turns up various discussions.

While I was sure that the three Transound microphones I dismantled were back electrets, and the two older Panasonic 9.7 mm omni capsules I dismantled were "foil type" or "charged diaphragm type", I wanted to see the elecret's effects directly.

Here are two techniques which worked fine for me.

Firstly, the front surfaces of the Transound backplates and the back surfaces of the Panasonic diaphragms all picked up cornstarch and ordinary wheat flour. 

The second test was more satisfying.  I cut a strip of kitchen aluminium foil (10 um thick) 70 mm long and 2 mm wide.  I suspended this so one end hung free.  I grounding the top end of the foil by touching it with a finger of my left hand.  Then I held the Transound  backplate or the ring of the Panasonic diaphragm with pointy-nose pliers in my right hand.  With a finger I touched the metal of the pliers.  This ensures the alfoil strip has the same voltage (presumably zero, unless I am charged) as the metal of the backplate or the aluminized side of the Panasonic diaphragm.  (I found it not good enough to use a strand of hair or even a strip of paper, since these could develop static charges of their own.)

I could attract the dangling tip of the foil strip with the electret layer.  The attraction was only visible at a range of less than 1mm, but once the foil came close, the force was stronger and it tended to stick to the electret layer.

~~~ooo000ooo~~~