Debenham Capsule Questions

[PeteSanders]

Hi,

Hope I'm not re-posting this.

In the Debenham, Robinson and Stebbings article, 

isn't it a  one double-sided backplate, sharing the same through-holes, so it's 1/2" thick altogether? 

Fig. 6 on page 239 shows a profile of the dual-diaphragm capsule.  The through-holes go through to the opposite membranes.  but on page 241, the side-view just shows single-sided drilling.

Other question: 
On p. 243, the left-side notes say all the .028 (or something) holes are drilled through, and the side-view looks to be a quarter-inch thick, with a vent-hole.  Is the idea to intersect with just one of the backplate through-holes?

I've hacksawed a piece of brass-bar, got capacitor-foil;

now I'm afraid to drill.

 

[gyraf]

Not the answer you're looking for, but:

Don't be afraid - experiment.

Get some more brass if that's what's needed.

Jakob E.

 

[shaggy]

Hi Pete. I've been wanting to try this project myself. The original blueprint is certainly not easy on the eyes.

As I read it, "all .040" (#60) holes drilled to a depth of .170". All .028" (#70) dia. holes drilled right through."

From the center out,  there are three (3) 'circles' of .028" (through) holes (of 3, 3, and 6 holes) for a total of (12) .028" dia. through-holes. (I think that's probably another through-hole in the center. although I don't see it marked.)

And (also from the center out) there are five (5) 'circles' of .040" (non-through/ .170" depth) holes (of 4, 6, 12, 12, and 12 holes) for a total of (46) .040" dia. non-through-holes.

(I hope I got all that right - every time I look at it I seem to see something I didn't notice before!)

Interesting that all the through-holes are specified 'equally spaced' around their pitch circles, whereas the non-through-holes are not so spec'd (some are offset in order to dodge through-holes positioned on the same pitch circle).

('PCD' on the original drawings means Pitch Circle Diameter, thus hole 'circles'.)


Dave (shaggy)

 

[PeteSanders]

Ha,

Thanks  Jakob,

I've got brass, and am no longer afraid!

Hi Dave,
Yeah,  he says the blind holes are 170 mils (about 4mm) deep. 
The profile picture just above the caption of fig. 7 is shaped like it's two-sided.  But the brass is only 250 mils thick;  don't we need something over 340 mils?

So what do we have -  is there something I don't get, or is the thickness mismarked.  If I were to make 2 single-sided ones and put them together, what's a reason to do that, instead of making it thicker? 
I wonder if there are discussions about it in a later issue of the magazine -  but what magazine is it?

In the meantime, I came across http://www.homefly.com/products.asp?id=31 which seems to have the elusive metallized mylar in micron-thicknesses.    It's wider than my capacitor-roll.

Pete

 

[shaggy]

There's only one backplate, and I don't see anywhere it's shown to be thicker than ~1/4". Add the two diaphragm rings (.055" each) and the whole capsule assembly will be under 3/8" thick.

Hey, what happened to making the backplate out of Dural, as specified 8>] I guess light alloy seems 'cheap' to people these days!

PS - Thanks for the Mylar link, Pete. There's a good source (in the UK) that also has 5 mil aluminised. So far I've not found any 6 mil.


Dave (shaggy)

 

[PeteSanders]

Ahaaaa....

I wasn't seeing that, on p. 243, "View A" and "View B"  must be the two different sides of the capsule.  I imagine that's how you can drill the 1/4" thick backplate on both sides without any hole-collisions.

D'oh.

I reduced the pdf page to 79%, to match the 1" backplate.  Then I printed two of them on tracing paper, overlayed View A and View B, viewed them  on a light-box.  It looks like a lot of the blind-holes would become through-holes, if I were to cut out the drawings and drill through them for a pattern.

On second try, it looks like you can rotate one a little, and avoid the problem.  I may do it that way.

Important edit: on third try, looks like you don't have to rotate.

Looks like  the thing to do is to put the sketches back-to-back when viewing, so the ink points outward.  Then it's like the capsule would be.

Looks like it's the way to go, but try it for yourself.


Yes, using aluminum sounds like it'd be nice to use - There's some corrosion after a long time; the paint from my aluminum bike is bubbling-up in places, from that.  I don't think of replacing it with a brass one, though. 

 

[shaggy]

Hi Pete - Smart idea with the lightbox. That could have saved you some serious frustration!

It makes you wonder why the backplate wasn't thicker, so there'd be no need to stagger the holes. Seems obvious to me - but then that usually just means I'm missing something(!)

Dave (shaggy)

 

[gyraf]

Backplate thickness is probably a function of 1) required internal air mass and 2) front-to-back distance

Jakob E.

 

[PeteSanders]

I edited my previous post,

found that the traced drawings make more sense if  viewed back-to-back, like the capsule-faces are.

Woops.

When I do that, it looks more likely to be the way it was originally thought of.

 

[Marik]

Backplate thickness is probably a function of 1) required internal air mass and 2) front-to-back distance

Jakob E.

The diaphragm/backplate spacing and the through hole diameter are by far dominating in this case. The first (along with diaphragm frequency tuning) would mostly affect frequency response, and the second--polar response. The blind holes do affect damping of the diaphragm--in the Braunmuhl-Weber design should be consistent on both sides for side-to-side uniform sensitivity.

I would not use originally specified Duralum for backplate--it is an Al alloy, but due to high content of Cu it is prone to severe oxidation problems. Brass (corrosion resistant grade, or passivated) would be a good choice here. Generally, use clamping ring of the same material as the backplate for the similar thermal expansion.

When drilling try to use very high RPM spindle (at least 10K), with the shortest stub drills you can find (carbide, or cobalt are good, but even twisted HSS installed short would still do the job) to avoid the drill bits "walking" on the backplate and avoid drill breakage. If possible, instead of drill chuck use collets, which are much more precise with runout, so the fragile drill bits will last quite a bit longer and the holes will be more precise. Also, instead of plunging through the holes, rather peck them (esp. smaller ones)--this will remove chips and help to avoid overheating--you don't want to set the hole depth every time over and over again when the bits break. Use plenty of cutting oil, esp. for tapping holes. If you use CNC then mark all the holes with the spot drill, first.

Sounds like a fun project!

Best, M

 

[ricardo]

PS - Thanks for the Mylar link, Pete. There's a good source (in the UK) that also has 5 mil aluminised. So far I've not found any 6 mil.

The dominant mass load is the air in this type of design.

The difference between 5um & 6um Mylar only makes a difference to how difficult it is to tension the diaphragm.

As Debenham et al don't specify the tension (best set by diaphragm resonance), the difference is moot.

For the fans of ultra thin diaphragms, you probably can't set a good resonance (about 1.4kHz) with 2um Mylar 

But the stuff is cheap enough to buy loadsa thicknesses & try.

Use a tensioning jig like http://www.10000cows.com/DiaphragmTensioning.htm.  A larger diameter weighting ring helps achieve even diaphragm tension.

Calrec's history is that my mike mentor, Clem Beaumont made one of these mikes and showed it to Bernhard Weingartner (who designed CK12 & C12 for AKG).  It was Bernhard's encouragement and friendship that led to Calrec .. and eventually the Soundfield and my involvement with microphones.

Bernhard says good things about the Debenham and admitted it did stuff his C12 did but much more simply.

 

[PeteSanders]

Thank you for your far-reaching posts,

The posts have expanded what I've been thinking about.  Jakob's comment goes with the article - makes me ponder how the thru-holes make the back-diaphragm behave, and affect the sound. 

I just drilled the thru-holes for the two capsules; they are a small diameter ( 0.028")  and their pattern is kind of sparse.  The  capsule is only 1/4" thick, so  I would think when a sound hits the front diaphragm, there is some air-resistance and a reaction from the back diaphragm- it must get a little reaction in the opposite shape that the front does.  So if you had more, bigger, or longer thru holes, what difference in sound would you get?

Marik, thanks for writing about how to drill - I needed that.  I had already gotten tin-tipped pointed bits from McMaster, before thinking about stubby bits or high rpm.  Since you mentioned the high rpms, i used my Dremel 4000 with its little drill-press, and a drill-vise.  The thru-hole bit (0.028") broke right away, but its remainder worked great through all the remaining through-holes, for two capsules.  The wobbly Dremel stand did the job, probably because the bit didn't walk.  I used about 13,000 rpm, and did the short plunges you mentioned. 

I used the little plastic  Xacto hole-cutter compass to cut the drill-pattern right out of the pdf-paper, then glued it to the capsule-brass.  To avoid wrecking the paper, I used a little candle-wax on the bit instead of oil.  It worked better than I thought, but maybe oil would still work. 

I had left the hacksawed brass a little rough, and the bit didn't try to walk.

I had also gotten a rod of Nylon for the capsule-body, before reading about its potential water-absorbency (means that water will enter or leave the nylon capsule-body as humidity varies, so the diaphragm's distance from the backplate won't be that stable) .  If I get sound, I'll be happy.

 

[ricardo]

I had also gotten a rod of Nylon for the capsule-body, before reading about its potential water-absorbency (means that water will enter or leave the nylon capsule-body as humidity varies, so the diaphragm's distance from the backplate won't be that stable) .

"For DIY, Perspex (Lucite, Plexiglas) for insulators is stable, reasonably non-hygroscopic & easy to machine (Google manufacturer’s recommendations to avoid chipping & stress crazing).  Sticks well with Epoxy & Cyanoacrylate.
http://en.wikipedia.org/wiki/Poly(methyl_methacrylate)

Les Watts (ex Shure) prefers Acetal / Delrin. http://en.wikipedia.org/wiki/Polyoxymethylene
Even easier to machine but harder to stick.

Brass is best for the metal bits but heavy.  You can Gold plate brass easily too."

This is from my Fig8.doc in the Yahoo MicBuilders Group which has loadsa other useful stuff.  You have to join.

I would use the method described in the Debenham et al article ... epoxy backplate to Perspex and then machine as one unit before drilling.

I second all of Marik's drilling recommendations.  Carbide tipped PCB drills are good at as high a speed as possible.  Definitely spot drill or punch positions.

The Blind holes adjust the acoustic resistance.  More & smaller holes = more resistance.

 

[Marik]

Les Watts (ex Shure) prefers Acetal / Delrin. http://en.wikipedia.org/wiki/Polyoxymethylene
Even easier to machine but harder to stick.

Delrin has higher melting point and is more "slippery", so Indeed, much easier to machine, but I could not find a satisfactory adhesive for the job. With machining Perspex it is the best to use a flood coolant, but even a mister would work. Otherwise it will just melt on you, or you will be forced to take very easy passes.

Brass is best for the metal bits but heavy.  You can Gold plate brass easily too."

You could actually Gold plate even plastic. For DIY experiments it might be easier--you can mask unwanted area and use it as an insulator. Also, you can somewhat reduce capacitance by plugging the holes and leaving them un-plated.

I always prefer to machine all the spacers into the body of the capsule. You will need to use an indicator and not rely on the dials (esp. Chinese machines). Use very good quality tooling--Iscar, Sandwik, Kennametal, etc., with sharp new inserts. More expensive, but in a long run give you by far superior results. Look for tool grades specifically for the material you are machining--they all have different chip breaker shape and will give you the best finish possible. For the last pass use the fastest RPM and very slow feed. The flood coolant (or again, mister) will greatly help to take the chips away and get you a mirror like finish.

The Blind holes adjust the acoustic resistance.  More & smaller holes = more resistance.

More resistance--flatter response, but on the expense of sensitivity. After that make a few different diaphragms with different tunings (start low, say, 800-900Hz and then go to some 1300-1400Hz)--that's where the whole fun starts...

Best, M

 

[tskguy]

Hey,

I thought I would chime in here regarding a few points, In terms of drilling holes I prefer to use bits from these guys, http://drillcity.stores.yahoo.net/newtuncardri.html. They work very very well and I do not need to spot drill at all or use coolant for that matter. My cnc runs at around  8000 rpm and at around 400mm per min feed rate.  I do this with bits all the way down to .4 mm with out issue. Peck drilling is essential though!! I peck at .5mm at a time. And yes collets are a must, get a good one with as close to zero run out as possible.  Using this method works very well for me, of coarse I am using my very purpose built CNC machine built to do this...

O spacers, Look on the web for color coded plastic shim material. It works great and is very consistent. Cutting it is another story. This is what I use...
http://knkusa.com/shop/knk-zing-air/

And here is a hint regarding tuning with weight, The beginning tension is critical!

Eric

 

[PeteSanders]

Eric,
Thanks!

I only own a few basic power-tools,  am finding out about lathes/milling/cnc through youtube and web-pages.  The microphone metal-work got me interested - I'm window-shopping. 

When you mentioned the diaphragm's beginning tension - do you mean when you're first putting it on a jig?

 

[ricardo]

And here is a hint regarding tuning with weight, The beginning tension is critical!

A large diameter weighting ring helps a lot with consistency.  It makes the initial stretching of the Mylar onto the weighting ring less critical.

At Calrec, the ID of our weighting ring was 4" while our capsules were 25mm max diameter.

Note the tension in the diaphragm is set ... not only by the weight .. but also the angle the Mylar makes with the edge of the capsule.

 

[PeteSanders]

Thanks Ricardo,

I had wondered about the size of the ring before, and if a weighted ring was the method to use.  Did you also use something to vibrate the diaphragm and test the pitch?

 

[ricardo]

I had wondered about the size of the ring before, and if a weighted ring was the method to use.  Did you also use something to vibrate the diaphragm and test the pitch?

We used a weighted ring and I think so do Neumann today.

The jig could measure capacitance and also resonance (by impedance) but, IIRC, the production method was to measure the change in Capacitance when 60V was applied.  Sorry to be vague but this was more than 30+ yrs ago.

The tension was adjusted by adding washers for extra weight.