Sennheiser MKH800 / MKH80 capsule (ks80)

[Tubetec]

Even though I cant see the scale of the FR plot ,its looks very very good ,
those smooth curves look perfect for compensation eq .

Its fairly difficult to find pot cores that small , I see in the Uwe Beis RF mic he uses pot cores also ,similar in style to the usual kind used in audio inductances .

Heres an RF inductance calculator I found ,maybe its of some use
https://coil32.net/online-calculators/multilayer-coil-calculator.html

 

[kingkorg]

Even though I cant see the scale of the FR plot ,its looks very very good ,
those smooth curves look perfect for compensation eq .

Its fairly difficult to find pot cores that small , I see in the Uwe Beis RF mic he uses pot cores also ,similar in style to the usual kind used in audio inductances .

Heres an RF inductance calculator I found ,maybe its of some use
https://coil32.net/online-calculators/multilayer-coil-calculator.html

Yes, unfortunate cropping. Here's an updated image. It's 5db per division. Thanks for the calculator.

 

[kingkorg]

@rogs @Tubetec i am super excited to say i wound a custom T1 inductor and it worked out flawlessly right from the first go!

(Edit, updated with better coils)
So, as the capsule has the capacitance of 50pF, i went for 30 26 turns (2x15 2x13 wound just like the Spectrum) on primary and 6 5 turns on secondary of T1. Stock is 20:4. I left T2 stock Spectrum, it doesn't seem to be that critical. This left me with the original ratio of 5:1, the primary gave me about 11uH 15uH at max setting. 105um (awg38?) Wire just like stock one.

The mic seems to have greatest sensitivity at maximum T1 setting, so it could be i should have gone a bit higher to get even better sensitivity and tuning range. Compared to Røde NT2a (7db noise) with applied correction curve the mic has comparable noise level. Without the correction it smokes the Røde noise wise, but sounds rather honky.

Important to remember C4 is replaced by front plate to diaphragm.

Sadly i don't have appropriate bandwidth scope to see what is going on. I have several different crystals on the way, maybe i get even better results at different speeds.

As the capsule is underdamped i don't know how high i can go in level before i overdrive the rest of the circuit. The signal at 1k is 15db hotter with RF mic vs Røde Nt2a. Røde has gain of about unity, Rog's mic has i believe the same as Schoeps of about 2 i guess? So i don't think i can go much higher, this could be already too hot.

This is RF mic uncorrected (red) vs NT2a (green).

 

[rogs]

That's excellent news! --

I've yet to try out my Aliexpress cores - although I have put the first one together (I used 0.1mm enamelled wire).... Not as difficult as I thought it would be. It doesn't take much to 'heat off' the enamel with 0.1mm wire ! )
This does seem to open the doors to those who want or need to try out differnt inductor values, and also gives a chance for builders to build the inductors cheaply. (Spectrum's shipping charge outside the UK is quite high )

It'll be interesting to see whether the Sepctrum tables on THIS PAGE give any useful pointers, regarding alternative inductive values and turns ratios?
These Aliexpress devices seem very similar to the Spectrum cores...

One thing you might like to try if you find the mic is too 'hot'.... You can affect the sensitivity by changing the oscillator amplitude. One added advantage in doing this is that the noise level should also drop as the oscillator amplitude is reduced.
Whether the drop in sensitivity from reducing the oscillator amplitude is directly proportional to the drop in noise level, I have yet to confirm?

Simplest way of reducing the oscillator level is to increase the value of R2

 

[kingkorg]

One thing you might like to try if you find the mic is too 'hot'.... You can affect the sensitivity by changing the oscillator amplitude. One added advantage in doing this is that the noise level should also drop as the oscillator amplitude is reduced.
Whether the drop in sensitivity from reducing the oscillator amplitude is directly proportional to the drop in noise level, I have yet to confirm?

Simplest way of reducing the oscillator level is to increase the value of R2

Wow, this is exactly what i was wondering about. Thank you

 

[rogs]

I have several different crystals on the way, maybe i get even better results at different speeds.

The Spectrum inductors are specified with a centre frequency of 7MHz. I normally use them at 10MHz, simply becasue that suits most 65pF capsules best.
I have tried them with 12 and even 14MHz crystals, but the 'Q starts to drop off quite dramatically.
The highest 'Q' I've had is with 8MHz crystals -- but it needed a 90pF capsule to function with the tuning slug set roughly centrally (highest 'Q' position)

 

[molke]

The Spectrum inductors are specified with a centre frequency of 7MHz. I normally use them at 10MHz, simply becasue that suits most 65pF capsules best.
I have tried them with 12 and even 14MHz crystals, but the 'Q starts to drop off quite dramatically.
The highest 'Q' I've had is with 8MHz crystals -- but it needed a 90pF capsule to function with the tuning slug set roughly centrally (highest 'Q' position)

Have you by any chance ever tried one of the higher-value inductors, like 11u(H)L?
I just noticed the old ones I’ve got (5u3H) are apparently “not suitable for high stability oscillator circuits”.

 

[rogs]

Have you by any chance ever tried one of the higher-value inductors, like 11u(H)L?
I just noticed the old ones I’ve got (5u3H) are apparently “not suitable for high stability oscillator circuits”.

No -- the 11uH(L) inductors do not have a centre tapped winding, and are not really suitable for this particular task.
There are virtually no other suitable coils from the Spectrum range -- or even in the entire TOKO 10K range either !

The two types of 5u3H coil are both suitable..... In this circuit the oscillator stability is ensured by the crystal.
However, the two types are slightly different. I found that the earlier coils performed better with 8MHz crystals, for example, especially with larger value capsules.

I think I probably do prefer the later coils overall -- but there's really not much between them, in this application anyway....

 

[Tubetec]

The dimensions of the Sennheiser inductor I can measure are as follows ,
A 10.2mm
B 8.8mm
E 4.25mm
The adjustment slug measures approx 3mm wide and a little over dimension B in lenght approx 10mm .

 

[Tubetec]

I came across a mention of Deity microphones in conjuction with my searches for RF mics ,
Seems like they do a similar shotgun to the MKH416 for 350 dollars US , no idea how good or bad they are .
https://deitymic.com/products/s-mic-2/

 

[rogs]

I came across a mention of Deity microphones in conjuction with my searches for RF mics ,
Seems like they do a similar shotgun to the MKH416 for 350 dollars US , no idea how good or bad they are .
https://deitymic.com/products/s-mic-2/

They're not RF bias mics ... their 'weather proofing' comes from a spraying conformal coating onto important bits and the mention of 'RF' in the documentation is about rejection of external 'RF interference', rather than of using low impedance RF circuitry in the device itself.
The Deity may be a good mic - I've no idea? -
For a few dollar more I think I would go for a Rode NTG5 instead, if I needed a small(ish) outdoor shotgun mic...

 

[kingkorg]

@rogs if a push-pull capsule is used (the other half of one side of the cardioid replacing c4), should the <10pf imbalance be there? I guess when the diaphragm goes in one direction it creates the imbalance on the other side, so in this case it's not important?

I just ordered a cheapo oscilloscope with FFT and appropriate range, so i guess i'll have more insight into what i'm actually doing here

 

[Tubetec]

The Deity seems to be trying to claw at the segment of the market ocupied by both Sennheiser and Rode , your right though it doesnt specifically refer to RF bias on the website though . Still be nice to see the pcb from it ,but no pictures seem to be around .

I'll hang tough and keep an eye out for any spare case parts for the two 416's I have for now .

 

[rogs]

..... if a push-pull capsule is used (the other half of one side of the cardioid replacing c4), should the <10pf imbalance be there? I guess when the diaphragm goes in one direction it creates the imbalance on the other side, so in this case it's not important?

I just ordered a cheapo oscilloscope with FFT and appropriate range, so i guess i'll have more insight into what i'm actually doing here

I think what you're describing is essentially a figure of 8 mic? ...... I have used one of Ari's flat K47 (double sided) capsules in that mode with an RF mic.
The outputs from both sides of those capsules are well balanced, so in essence the bridge is actually balanced (well, probably almost balanced! )

As the capacitance of one side gets larger with any modulation, the other side gets smaller. So the capsules act a bit like the wiper of a 'capacitive' potentiomenter. No need to introduce any imbalance....... It's the optimum configuration for this particular RF circuit.

With a 'standard' cardioid configuration the capsule has C4 in series, but needs to remain unbalanced enough so that any change in capsule capacitance doesn't ' force' the bridge to be unbalanced in the other direction......( i.e. if the capsule capacitance is larger than C4 it doesn't want to become smaller than C4 in any event....... That produces very strange results (the signal keeps trying to change polarity! )
Also, if the difference between the capsule value and C4 is too large, you get too much RF carrier - and an increase in noise level.
If you make C4 much larger than the capsule value you get lower sensitivity .. and more noise.
It it's much smaller you probably can't tune the core slug effectively an the 'Q' drops dramatically .. so again more noise. (that's with Spectrum coils of course)

The optimum cardioid version is much more difficult to specify accurately. Which is why - as other folk have mentioned - you can get quite a difference in performance between apparently identical circuits (many cheap Chinese capsules can vary quite a lot in capacitive value - and even in membrane tension !)

An oscilloscope can help with testing -- although most probes will still have a load capacitance of around 15pF, even on the Hi-Z x 10 setting --- and that will load the measurements of some of the RF test points. Getting a reliable RF carrier reading on the gate of the FET is particularly difficult!

 

[rogs]

The Deity seems to be trying to claw at the segment of the market ocupied by both Sennheiser and Rode , your right though it doesnt specifically refer to RF bias on the website though . Still be nice to see the pcb from it ,but no pictures seem to be around .

I'll hang tough and keep an eye out for any spare case parts for the two 416's I have for now .

The Deity might even have an electret capsule?..... If they're recommending outdoor use for a condenser mic, electrets are less susceptible to moisture than conventional SDC capsules. ( RF SDCs are of course low-Z mics)

The 416 is in a different league - especially those examples fitted with an external casing!

 

[kingkorg]

I think what you're describing is essentially a figure of 8 mic? ...... I have used one of Ari's flat K47 (double sided) capsules in that mode with an RF mic.
The outputs from both sides of those capsules are well balanced, so in essence the bridge is actually balanced (well, probably almost balanced! )

As the capacitance of one side gets larger with any modulation, the other side gets smaller. So the capsules act a bit like the wiper of a 'capacitive' potentiomenter. No need to introduce any imbalance....... It's the optimum configuration for this particular RF circuit.

With a 'standard' cardioid configuration the capsule has C4 in series, but needs to remain unbalanced enough so that any change in capsule capacitance doesn't ' force' the bridge to be unbalanced in the other direction......( i.e. if the capsule capacitance is larger than C4 it doesn't want to become smaller than C4 in any event....... That produces very strange results (the signal keeps trying to change polarity! )
Also, if the difference between the capsule value and C4 is too large, you get too much RF carrier - and an increase in noise level.
If you make C4 much larger than the capsule value you get lower sensitivity .. and more noise.
It it's much smaller you probably can't tune the core slug effectively an the 'Q' drops dramatically .. so again more noise. (that's with Spectrum coils of course)

The optimum cardioid version is much more difficult to specify accurately. Which is why - as other folk have mentioned - you can get quite a difference in performance between apparently identical circuits (many cheap Chinese capsules can vary quite a lot in capacitive value - and even in membrane tension !)

An oscilloscope can help with testing -- although most probes will still have a load capacitance of around 15pF, even on the Hi-Z x 10 setting --- and that will load the measurements of some of the RF test points. Getting a reliable RF carrier reading on the gate of the FET is particularly difficult!

Nope not f8. That's the tricky part to describe.

One diaphragm, but two plates. Diaphragm is sandwiched between two plates.

Diaphragm > front plate instead of c4. Same diaphragm > back plate connected as capsule in your circuit. There is a chamber behind back plate that forms cardioid response.

Anyways i got the answer i was looking for. The diaphragm moving towards front plate will always go away from the backplate.

 

[rogs]

.......One diaphragm, but two plates. Diaphragm is sandwiched between two plates.

Diaphragm > front plate instead of c4. Same diaphragm > back plate connected as capsule in your circuit. There is a chamber behind back plate that forms cardioid response.

Anyways i got the answer i was looking for. The diaphragm moving towards front plate will always go away from the backplate.

Look like the ideal arrangement...... No concerns about an indeterminate unbalanced bridge set up! Conversion from figure of 8 to cardioid pattern determined by capsule chambering..... got to be better than relying on C4 !
With the value of the inductor optimised to allow a high 'Q' resonant tuning, I think you've probably created the best use of this circuit we've seen so far!

 

[Tubetec]

The fundamental Im not getting ,
Isnt the original Baxandall design and the later Sennheisers based on a four section coil , with a bifilar and two other feedback windings ,

The actual construction of the Spectrum coils ,Im finding hard to visualise ,
is it screened U ferrite core , thats not the same as pot core ,
if you look inside the earliest examples of the MKH to the most modern ,
the pot core is the common element all along ,

I misplaced my eye hock but upon closer inspection , at least one section of the main multiwound inductor in the 416 is litz , if Im not mistaken ,


The sad faced FR responce of the RF mic might be equally well compensated for digitally in a modern context + =


Cheers for the inspiration ,

 

[rogs]

The fundamental Im not getting ,
Isnt the original Baxandall design and the later Sennheisers based on a four section coil , with a bifilar and two other feedback windings ,

The actual construction of the Spectrum coils ,Im finding hard to visualise ,
is it screened U ferrite core , thats not the same as pot core ,
if you look inside the earliest examples of the MKH to the most modern ,
the pot core is the common element all along ,

The actual structure of the Spectrum coils is almost identical to these Aliexpress coil formers: https://www.aliexpress.com/item/1005003604363316.html

The secondary of the Spectrum 5u3HH coil is a centre tapped winding, with 10 turns wound around the 2nd (next to top) section of the bobbin, and terminated between pins 1 and 2.

The second half of that winding is a further 10 turns wound round the top section of the bobbin and terminated between pins 2 and 3.

The primary winding is 4 turns wound on top of that second section (i.e. around the top section of the bobbin) and terminated between pins 4 and 6.
So not true bifilar windings....

My circuit is based on Baxandall's concept sketch 2(b) on page 539 of his paper HERE.
I have coupled the primary of my second transformer across the bridge, between the centre tap of the first transformer secondary and the junction of the capsule and its balancing capacitor.
That central section of the inductor assemby would normally be galvanically isolated, but I have connected the secondary centre tap to ground - along with the associated end of T2 primary - so the centre section is actually referenced (symmetrically) around ground.
It doesn't seem to like 'floating' with no reference -- With no DC reference, the RF seems to do behave slightly 'strangely' after a while --although it's difficult to be more specific on that point at the moment.

I then use the step up ratio and the tuned 'Q' of that second transfomer to 'amplify' the bridge output signal.
The tuned secondary winding of T2 is connected to the gate of the JFET infinite impedance detector, which I use in place of Baxandall's illustrated 'switch'.
The gate of the JFET is effectively held at ground from a DC point of view, and this enables the JFET to self bias, so that it 'sits' naturally at Vp.

The 'multi winding' Sennheiser inductor assemblies seem a little more complex -- and I have no idea how they are actually physically constructed?..

 

[kingkorg]

@rogs
Update. Today i received the right gauge wire, just like in Spectrum ones. I got best results with 2X13 turns on primary, and 5 turns on secondary. This got me inductor range of 5uH to 15uH. It was extremely easy to dial in the inductors and i got even better snr with this one compared to previous DIY coil.