Debenham capsule directivity

[molke]

After building two (RF) microphones with hand tools, I got a small mill and an even smaller lathe, and as a rather overambitious first lathe project, I then attempted to build the Debenham, Robinson, and Stebbings capsule. The three capsules so far are put in circuits derived from the FET847, a C414 EB P48, and the usual Schoeps (the only multipattern one). The result after a few months and several failed attempts is surprisingly good and consistent on-axis. The quick measurement (at about 30 cm/1 ft with 1/3 octave smoothing) is certainly not absolute, but it may indicate a general trend. The article claims e.g. 23 dB attenuation at 180 degrees for cardioid, which is obviously not the case here, and the frequency response for the different patterns is obviously not as flat as ‘advertised’.

Does anyone (who built this type of capsule or others) have an idea what could be a reason for the rather low attenuation and ‘bad’ pattern response?

Diaphragm tension is probably still too high for a number of reasons. Apart from that, there seems to be a rather large discrepancy regarding the “clearance ring around the capsule” between the text (1.5 thou = 38 µm) and the drawing ((1 in - 0.94 in)/2 = 30 thou = 0.762 mm). I went with the latter as it seemed more realistic.

 

[soliloqueen]

Got any pictures of the Assembly and back plate? Even just pictures with the diaphragm on with lighting that shows the holes is fine. Just to knock out some basic problems and to give us a good idea of what you've built

 

[molke]

Thanks for the reply. Unfortunately, my digital camera is really bad, so I’m not sure that my pictures really help beyond the basic idea.

One picture is the measured Schoeps-type microphone. The other two are very similar construction-wise, and there are probably several areas for improvement. The membrane material (5 µm aluminized mylar from freeflightsupplies.co.uk) is not as translucent as gold-sputtered membranes (that is, not at all).

The other picture is a spare backplate that is unfinished. Despite not being able to remove all hack saw marks (bottom corner), it is already way too thin (5.7 mm) compared to the plan (6.25 mm). I used 36 µm Mylar for now as a spacer ring instead of permanently machining the distance in the electrode.

 

[kingkorg]

Your measurements show typical 180° response for an omni capsule. You are not getting enough cancelation from the rear. Delay network is not functioning.

A tip. You can use that wm61 response as a calibration/compensation in order to get better and more absolute measurements of the other capsules.

 

[soliloqueen]

after thinking on it a bit, can you replace the rear diaphragm with an unmetalized one to rule out the simplest explanation (a short)?

 

[molke]

Regarding the omni response, I thought so as well, but the capsule construction is clearly not meant as omnidirectional. So which factors in my execution make it an omni: membrane tension too high, too little/too much membrane spacing, …?

Thanks for the tip. I tried referencing the other responses against the WM61, but the general trend is not much clearer. This is a ten-year old microphone (with again a Schoeps-type circuit), which I used a few years ago for measuring my speakers, so I’m not convinced of the response there as well (not to mention my acoustically untreated room, different measuring distance, …). Therefore, I just included it as something that might be somehow known.

 

[molke]

Does the material (thickness) matter when replacing? The thinnest unmetalized Mylar I have here is 22 µm.

 

[soliloqueen]

Does the material (thickness) matter when replacing? The thinnest unmetalized Mylar I have here is 22 µm.

in general yes, but the difference will be so dramatic if it is a short that it may be fine to test the theory. not sure though, never used anything that thick. how are you terminating the diaphragms?

 

[molke]

What do you mean by terminating? And would a short not show with a simple multimeter as well?

 

[soliloqueen]

What do you mean by terminating? And would a short not show with a simple multimeter as well?

with the multimeter, check all of your sets of diaphragm screws for shorts between the front and back

 

[soliloqueen]

You can use the thinner backplate BTW. DRS is fundamentally just an edge-terminated k47, and those are generally about 5.7mm thick to begin with. the loss of thickness won't cause anything catastrophic, just might act a bit different.

 

[molke]

Then it’s probably good that I kept it as a spare.

I’ll try the multimeter first and will report back.

 

[soliloqueen]

how much space is there between the plastic outside ring and the metal backplate? it should fit with friction, so pretty close. any additional space will act as additional through holes.

 

[molke]

That’s what I initally meant with discrepancy. According to the drawing in the original article (as far as I understand it), the metal backplate is 0.94 in in diameter. The inner diameter of the clearance ring is exactly 1 inch (the hole to fit the electrode is supposed to be a rather tight 0.9395 in). That should be 0.03 in clearance: “This guard ring ensures that any small wrinkle in the diaphragm at its edge do not short to the central electrode.” As it goes on with “Since this forms a considerable addition to the cavity volume […]”, this seems to be accounted for. The accompanying text has “1 1/2 thou. clearance ring” though, which would be more along the lines of “pretty close”.

 

[soliloqueen]

That’s what I initally meant with discrepancy. According to the drawing in the original article (as far as I understand it), the metal backplate is 0.94 in in diameter. The inner diameter of the clearance ring is exactly 1 inch (the hole to fit the electrode is supposed to be a rather tight 0.9395 in). That should be 0.03 in clearance: “This guard ring ensures that any small wrinkle in the diaphragm at its edge do not short to the central electrode.” As it goes on with “Since this forms a considerable addition to the cavity volume […]”, this seems to be accounted for. The accompanying text has “1 1/2 thou. clearance ring” though, which would be more along the lines of “pretty close”.

i think you did it right, but just in case, to elaborate,
0.94 backplate
0.935" inner ring
0.005" negative interference fit. the backplate should be slightly bigger than the plastic ring. the backplate fits in with pressure.
Then, shorting is avoided by way of a 0.03" wide "trench" to 1" around the backplate

but i think the text made some errors as to the depth of this trench..? one moment.

 

[molke]

The depth of the clearance ring is clearly marked in the text as 0.05 in, and drawing: (0.25 in - 0.15 in)/2, which leaves the question of the space. As this is supposed to be done with “a special tool ground for this work” (probably something which is called “Planeinstichmeißel” in German), I doubt that this tool would have a width of 1.5 thou and a depth of more than 0.05 in as this would probably be a rather flimsy razor blade?

The press fit is what I tried to achieve.

I can’t find a(n electrical) short between both diaphragms. I did try to measure all screws, but a short would probably show everywhere since the clamping ring is in direct contact with the metalized side of the diaphragm. This probably does not rule out an acoustical short, like covering the through holes?

Actually, I did get a cheap Chinese LC meter a few months ago, intended to test the finished capsule, but unfortunately did not check it until recently. The actual values below 100 nF are probably garbage. I tried some ceramic caps which are supposed to be 100 pF (5% tolerance) that showed values all over the place but in the right ball park. The capsule used here measures around 80 pF each side to the backplate, and half of it between both sides.

 

[soliloqueen]

The depth of the clearance ring is clearly marked in the text as 0.05 in, and drawing: (0.25 in - 0.15 in)/2, which leaves the question of the space. As this is supposed to be done with “a special tool ground for this work” (probably something which is called “Planeinstichmeißel” in German), I doubt that this tool would have a width of 1.5 thou and a depth of more than 0.05 in as this would probably be a rather flimsy razor blade?

The press fit is what I tried to achieve.

I can’t find a(n electrical) short between both diaphragms. I did try to measure all screws, but a short would probably show everywhere since the clamping ring is in direct contact with the metalized side of the diaphragm. This probably does not rule out an acoustical short, like covering the through holes?

Actually, I did get a cheap Chinese LC meter a few months ago, intended to test the finished capsule, but unfortunately did not check it until recently. The actual values below 100 nF are probably garbage. I tried some ceramic caps which are supposed to be 100 pF (5% tolerance) that showed values all over the place but in the right ball park. The capsule used here measures around 80 pF each side to the backplate, and half of it between both sides.

give me a second, i'm parsing what all this stuff in the drawings is trying to say

 

[soliloqueen]

I have the same meter here and it gives the same values for a (working) k47. the issue here is that very few things are going to cause a capsule like this to literally just be omni. differences in air volume from something like the clearance ring would change the frequency response, not the entire pattern. The polar pattern is based on delay and attenuation from the through holes mostly, and you've got the amount of through holes you're supposed to have, a diaphragm set up the same way on the other side etc...

It's difficult to believe that this issue could be in construction and not electrical. You've got an airtight seal between the plastic ring and the plate, right? So no additional venting between sides other than the through holes. You have no other through holes. and...of course, you didn't drill the larger blind holes all the way through.

does the problem persist on cardioid-only circuits? what's the 180 degree response like with only one diaphragm and the backplate connected?

 

[molke]

There is one “vent hole” with 0.0135 in (#80) = 0.35 mm. The rest seems to be tight. Going by the drawing, I wasn’t really sure if the center is actually supposed to be a through hole because it’s nowhere indicated as such.

 

[soliloqueen]

There is one “vent hole” with 0.0135 in (#80) = 0.35 mm. The rest seems to be tight. Going by the drawing, I wasn’t really sure if the center is actually supposed to be a through hole because it’s nowhere indicated as such.

it is, but even if you didn't, that wouldn't cause something like this. i am perplexed.