Debenham capsule directivity

[Scrappersa]

Hi Soliloqueen and Molke,

I've been playing around with this capsule for over a year now. I've machined a few myself and had a small run done with 10 of them. They tend to have very good rear rejection. They're definitely not omni.

I really wonder what's going on with your capsule. I did not machine the edge lip into my capsule. I opted to make the entire backplate flat and to use shims so I can adjust the height depending on the capacitance I'm targeting. I wonder if that surface isn't flat and there's a leak on either side.

I've tuned Debenham that I would consider gorgeous sounding. And a bunch that sound consistently good.

The issue that I'm having right now is with the tension of the membrane on the back side. I've been able to make consistent sounding capsules in cardioid with only the front membrane as long as there is no back membrane. The second I put a back membrane on, it completely changes the sound. I've experimented with different frequencies across the board on both the front and back.

The best sounding one I've made has a back membrane but I have not been able to recreate it yet. The other ones I've made sound almost as good but with slightly less top end- these ones all do not have a back membrane.

I'm wondering what the relationship between the front and back membranes are, specifically with this capsule.

I've also found that the Debenham capsule is not very forgiving in regards to its tuning frequency. If it's too tight, it loses a ton of low end. I've found it to sound best when it's tuned as loose as possible.

Any ideas for how the back membrane should be tuned?

 

[molke]

They look really good (I only managed four). Is that black acrylic or some other plastic?

Thanks for your confirmation. I did already suspect that something went wrong with my attempt. I used shims as well—I trust neither me nor my lathe to get the required flatness without grinding afterwards. From the picture, it seems you also left out the ‘clearance ring’ that accounts for almost half the volume of the blind holes (75/156 mm³) as far as I understand the drawing?

Not flat enough might very well be the case, though I suspect that my hand-carved shims may be more to blame than the actual backplate flatness. But how would I get omnidirectional pattern with an air leak? Regardless of the diaphragm tension, I almost always got a reasonably flat frequency response—just no cardioid pattern.

In the meantime, I improved my membrane tuning a little by measuring the resonance in the tensioning jig before mounting (targeting around 1 kHz with the final diameter), and I widened the through holes from 0.7 mm to 1mm on one, which didn’t change much, so the basic construction is probably not the reason. Next step for me was to finally try the specified diaphragm spacing of 2 mil.

I would assume that both diaphragms are ideally supposed to be tuned the same on both sides.

Just to be clear, do you get (good) rear rejection with both membranes or just with one (cause I haven’t tried this yet)?
Did you measure frequency response?

 

[Scrappersa]

Thanks! It's delrin! Soft yet durable.

In the original drawing, I believe it called for the delrin to be 0.002 inches above the brass section. I opted to make it the same height and to use plastic shims to set that height. I cut my plastic shims with a cricut. You can buy these on McMaster Carr in different thicknesses. I then measure capacitance and add or subtract shims based on the capacitance I'm aiming for.

I just bought a coaxial speaker so my measurements will be more accurate. But in my previous tests, the rear rejection is very nice. The low frequencies come through a little, but nothing unusual.

In an ideal world, the tuning would be the same on both sides but it's bizarre that isn't the case with my backplate design. I've now made a bunch of them and they are very consistent. It's so odd that it seems to sound better with a different tuning on the front and back. I'd love to understand why this is the case. Just happy I've found something that I've been able to recreate many times now.

Okay, in regards to the clearance ring (the little milled out channel in the delrin around the brass piece)- I missed this in the drawings when I originally machined my capsule. Funny enough, in my own Debenham thread, nobody else noticed it. Thank you for pointing this out. I'm so curious now how this would affect the sound.

I absolutely love this capsule in its current form but it's about 10dB lower output than other capsules. @soliloqueen any clue how this clearance ring would affect the output or the top end? I imagine it would have more of an impact on the high frequencies like the blind holes. I'm currently in the process of making another Debenham with the clearance ring implemented to test this out.

 

[soliloqueen]

Thanks! It's delrin! Soft yet durable.

In the original drawing, I believe it called for the delrin to be 0.002 inches above the brass section. I opted to make it the same height and to use plastic shims to set that height. I cut my plastic shims with a cricut. You can buy these on McMaster Carr in different thicknesses. I then measure capacitance and add or subtract shims based on the capacitance I'm aiming for.

I just bought a coaxial speaker so my measurements will be more accurate. But in my previous tests, the rear rejection is very nice. The low frequencies come through a little, but nothing unusual.

In an ideal world, the tuning would be the same on both sides but it's bizarre that isn't the case with my backplate design. I've now made a bunch of them and they are very consistent. It's so odd that it seems to sound better with a different tuning on the front and back. I'd love to understand why this is the case. Just happy I've found something that I've been able to recreate many times now.

Okay, in regards to the clearance ring (the little milled out channel in the delrin around the brass piece)- I missed this in the drawings when I originally machined my capsule. Funny enough, in my own Debenham thread, nobody else noticed it. Thank you for pointing this out. I'm so curious now how this would affect the sound.

I absolutely love this capsule in its current form but it's about 10dB lower output than other capsules. @soliloqueen any clue how this clearance ring would affect the output or the top end? I imagine it would have more of an impact on the high frequencies like the blind holes. I'm currently in the process of making another Debenham with the clearance ring implemented to test this out.

it's the same as any other increase or decrease in air volume behind the diaphragm, i think? the capsule would just be a bit brighter and more sensitive.

the Debenham capsule is an edge-terminated k47 more or less. k47s are notoriously annoying to get sounding good on both sides. at neumann, there was a really low success rate in the beginning, so low that they invented the u47 fet just to have an economic place to dump all the failures. the k47 is a true pressure gradient from front to back, it doesn't have any chambers to decouple the two halves, so each diaphragm contributes much more directly to the formation of the pattern of the opposite diaphragm than in later capsules. you need to get the tension both exactly right and identically matched, otherwise you will have either a capsule that sounds really good on one side or a capsule that sucks equally in all patterns. are you tensioning the diaphragm with weights or a jig?

 

[Scrappersa]

That makes perfect sense. I'm looking forward to hearing the results with the increase in air volume.

I'm using a jig to set the tension and measuring the resonant frequency of each membrane.

 

[Murdock]

That makes perfect sense. I'm looking forward to hearing the results with the increase in air volume.

I'm using a jig to set the tension and measuring the resonant frequency of each membrane.

How do you measure the resonant frequency if I may ask?

 

[molke]

How do you measure the resonant frequency if I may ask?

I would be very interested in that as well. The best I’ve come up with so far is tuning the material in the tensioning jig’s clamping diameter excited with pink noise, measuring the resonance frequency with another mic, then calculating the resulting resonance frequency for the final diameter.

 

[Scrappersa]

Like you, I measure with another mic but I have it hooked into an Audio Precision. I have the AP run a sweep to see where the resonant peak is.

However, my tuning jig is larger than the diameter of the membrane once it's glued/screwed to the retainer ring. So I double check all of my membranes with the AP after they're glued to the retainer ring. I'll also double check by running a sweep and putting some salt on the membrane to see where it really starts to go crazy. This is a really effective way to get close to the actual resonance without any special measuring tools. I think this only works when the membrane is in the jig or glued to a retainer ring.

Without that channel in the capsule, I imagine my backplate is a bit different from the actual Debenham. I've found my favorite sound is when the front is tuned between 750hz and 850hz and the back is tuned at 1100hz exactly. Really bizarre, but it just seems to sound the best here. It seems that some capsules are less affected by the back membrane tuning- my backplates back membrane has a profound effect. I spent weeks chasing my tail giving myself a headache before I started playing around with just the back membrane.

My backplate is still about 10-12db down from a CK12 and K67 capsule. I'm waiting for some 3U K47 Capsules to arrive to see if I'm at least in the ballpark of sensitivity compared to other single backplate designs.

My circuit is just a basic KM84 circuit and the polarization voltage is set at 48v. So I'm also waiting on a board to arrive that will allow me to play with polarization voltage between 60v and 80v. At its current voltage, the mic sounds great and works on most instruments just fine. With very soft vocals, it's a little bit noisier than I want it to be.

It's a very different sounding capsule. It doesn't have the tizz of the CK12. It has a unique midrange and top end compared to the K67 capsule. It has a great low end. I love it on bass and on overheads. It's beautiful on vocals.

I really wonder what is up with your backplate. But since I left the channel out of mine, I don't know much about how that would change it- theoretically, it would make it less dampened giving it more top end and more output. But I don't see how it would change the polar pattern. The thru holes are what should be dictating that I believe. How did you get the brass piece into the center of the acrylic piece? Are you sure it's completely sealed between the two and there isn't any leak around the diameter of the brass?

 

[soliloqueen]

I really feel like this problem could only be electrical. The response of the capsule looks exactly like the diaphragms are connected somehow. I'm kind of at a loss about it

 

[Scrappersa]

Yeah that makes sense. What kind of circuit are you using?

 

[molke]

Like you, I measure with another mic but I have it hooked into an Audio Precision.

Thanks for the detailed description (although I'm a bit jealous of the AP).

Yeah that makes sense. What kind of circuit are you using?

The three circuits I used for testing are given here. Maybe the different parts are not clean enough so that there is some electrical connection, but I cannot measure any (up to 20 MΩ). I already wondered whether it is wise to keep one diaphragm just unconnected, but maybe that’s unrelated (and wouldn’t apply for omni and figure 8 in the Schoeps variant):

In some, quite common topologies, one side of the capsule can stay charged even though the pattern is changed, which could mean you are using some random pattern between F8 and omni without realizing, yet thinking you are in cardioid.

Meanwhile, I made some slight progress towards something resembling a cardioid. The main differences are spacers cut from 50 µm Mylar and different diaphragm material (not sure if this is even Mylar, but it’s much easier to flatten). One capsule has the through holes extended to 1 mm and both are too thin (5.7 and 6 mm). To see if this isn’t just a measurement problem, I tried different distances.

 

[molke]

How did you get the brass piece into the center of the acrylic piece? Are you sure it's completely sealed between the two and there isn't any leak around the diameter of the brass?

The brass piece is a press fit, glued to the acrylic with epoxy, as proposed in the original article. How would I be able to verify that there are no leaks?
I didn’t glue the membrane to the retainer rings (yet), so maybe there’s some chance for leaks.

 

[soliloqueen]

if there was a leak, the capsule would get more figure 8, not more omni...

 

[soliloqueen]

can you test 90 degrees?

 

[molke]

can you test 90 degrees?

This should be included in the last two plots (the red line).

 

[soliloqueen]

This should be included in the last two plots (the red line).

thanks, sorry. i just woke up haha

 

[Scrappersa]

Thanks for the detailed description (although I'm a bit jealous of the AP).

The three circuits I used for testing are given here. Maybe the different parts are not clean enough so that there is some electrical connection, but I cannot measure any (up to 20 MΩ). I already wondered whether it is wise to keep one diaphragm just unconnected, but maybe that’s unrelated (and wouldn’t apply for omni and figure 8 in the Schoeps variant):


Meanwhile, I made some slight progress towards something resembling a cardioid. The main differences are spacers cut from 50 µm Mylar and different diaphragm material (not sure if this is even Mylar, but it’s much easier to flatten). One capsule has the through holes extended to 1 mm and both are too thin (5.7 and 6 mm). To see if this isn’t just a measurement problem, I tried different distances.

You could probably use your DAW with a plug-in like Fab filter Q3 to achieve the same thing.