Sennheiser MKH mics in LTSpice

[Voyager10]

After a bit of playing I've got an LTSpice simulation which (approximately) emulates the RF circuitry of the MKH 405 / 415.

The schematic is as follows:


The essential elements are based on the MKH 405 schematic from this thread: Sennheiser

The heart of the original circuit is coil L2, whose various windings are represented here by L1-L4. I've simplified the oscillator section, so the original coil L1 is not shown. The capsule is represented by C3, which varies either side of a nominal 27pF.

Running the simulation shows output voltage Vo varying as C3 changes:



Now before anyone gets carried away this is not intended to give accurate values for any of the inductors in the circuit, or the actual voltage levels, it's currently just a demo of theory behind how the FM detector works.

 

[Voyager10]

How it works​

(Or at least, how I think it's supposed to work. There's an AES Journal article at AES E-Library » Self-Calibrating Condenser Microphones with Integrated r-f Circuitry for Acoustical Measurements which gives clues on the front page).

Here's an LTSpice trace of a few relevant signals:


The green trace (Vpri) is the oscillator output. It's applied to inductor L1, which is loosely coupled (K = 0.3) to L2-L4.

Because of the loose coupling, and the resonance of C3 and L4, the voltage induced in L4 lags the primary voltage by 90 degrees (the cyan trace, shown as Vcap/2 for clarity). This lag is extremely sensitive to the exact C3/L4 values, and indeed the other inductances and coupling coefficients. Just like the real mic, I had to tweak the L4 value to get this just right.

L2 and L3 are closely-coupled (K = 0.99) to L4, so their induced voltages are in phase with L4, and lag Vpri by 90 degrees.
The dark blue trace (Vs1-Vs2) shows the total voltage across L2+L3, which is arranged to peak at about 1.3V for a brief period each cycle (done by tweaking the oscillator voltage and the number of turns, i.e. inductance).

So, for a brief period each cycle, the L2+L3 voltage is high enough to allow D1 and, um, D3 to conduct. This point is shown by the red trace, which is the current in diode D1. When the diodes are conducting, the Vo voltage is about 0.65V less than Vs1, and 0.65V greater than Vs2, i.e. the mid-point of these two voltages.

Now, the mid-point of Vs1 and Vs2 is also the voltage at the L2/L3 junction, which is driven from the Vpri voltage. So, for a brief period each cycle, Vpri is connected (via the diode 'switch' and filtering resistor R1) to the output capacitor C1. The output voltage Vo is therefore a filtered version of the average Vpri value during the diode conduction period.


Remember that the phase shift between primary and secondary was highly dependent on L4 and C3 values? So, as C3 changes, the phase shift moves back and forth around 90 degrees, and the diodes conduct at a different part of the Vpri cycle. So the Vpri voltage "sampled" onto C1 is above or below 0V depending on the phase shift. (Larger C3 = greater lag = -ve Vo).

(I have a 405 as a willing experimental subject, so next time I have a free evening I'll see if I can measure some voltages, and tweak the simulation to match).

 

[Voyager10]

As a follow-up, here's an LTSpice version of the audio path for the MKH405 schematic, as posted in the Sennheiser thread (and attached again below, for reference):

And here's how the simulation came out...



Crazy huh? It looks a lot like C15 is wrong, it's taking away all the negative feedback at high frequencies so we get runaway gain.

Looks like that schematic needs to be taken with a pinch of salt. Perhaps I can track down the service manual for it.

 

[rogs]

As a follow-up, here's an LTSpice version of the audio path for the MKH405 schematic, as posted in the Sennheiser thread (and attached again below, for reference):
View attachment 117565
And here's how the simulation came out...

View attachment 117566

Crazy huh? It looks a lot like C15 is wrong, it's taking away all the negative feedback at high frequencies so we get runaway gain.

Looks like that schematic needs to be taken with a pinch of salt. Perhaps I can track down the service manual for it.

If you look at the comments and graphs in this post: Sennheiser MKH800 / MKH80 capsule (ks80) you'll see that kingkorg shows the frequency response of the underdamped Sennheiser RF capsule he was experimenting with.
As he comments, that would require some serious EQ correction in the following audio stages.
There may be a similar type of capsule fitted to the circuit you are investigating?... That might explain the response you have observed.

 

[rogs]

............ There's an AES Journal article at AES E-Library » Self-Calibrating Condenser Microphones with Integrated r-f Circuitry for Acoustical Measurements which gives clues on the front page).........

Thanks for posting that link. Useful information on that 'free' first page!

In his paper on the history of the MKH mics HERE , Manfred Hibbing explains how Sennheiser first used FM for their original RF mics - although they used a phase modulation variation, to allow them to use a crystal oscillator.
Interesting to note that by the 1980s Sennheiser had decided to change to Amplitude Modulation, with their new 'push pull' capsules.

When first tried my experimental DIY RF mic project (link in my sig) I read Peter Baxandall's Wireless World paper (copy HERE )
He had decided that AM was probably a better option, even as early as 1963?

I'm assuming that Senneheiser's most famous MKH 416 mic uses phase modulation?.....
The schematic I've seen for that model (copy attached to THIS POST ) shows a different configuration from the AES Joirnal article -- but with some similarities. Certainly different from the later 'push pull' AM MKH mics.
Whether the latest verson of the 416 still uses the same circuit?......

 

[Voyager10]

That 416 schematic seems to have the same basic arrangement around L1. It looks like D3/D4 are doing the same thing as D1/D2, but conducting on the opposite half of the cycle. So the total output swing (between 'A' and 'B') is twice that of the 405 circuit, for the cost of two diodes. The L2 coil arrangement seems to be different, but I'd guess you can get the same 90-degree phase shift by tweaking the coupling coefficients appropriately.

Regarding 405 circuit's frequency response, it seems the C15 cap is right - I've found a 405 service manual (now posted to the Sennheiser thread), and it shows the same thing. It would make sense if the basic capsule response drops off in the same way as kingkong measured for the MKH800/80 one.

I'll try to make some measurements to see how a real 405 compares with the simulation...

 

[Voyager10]

Just to follow this up - yes, my 405 does indeed have that feedback network as shown (in fact C15 was 0.22uF, not 0.1uF), and - within the limitations of measurement - basically matches the simulation. There's a big bass lift starting at ~200Hz, and rising at the top end starting at 4KHz or so.

As it happened, I thought the mic was pretty low on sensitivity, and the HF sounded rather harsh, so I removed C15 altogether, and put R6 = 330R and R5 = 510R. Seems much better (at least to my tastes) now.

 

[Icantthinkofaname]

Too technical for me so I'll ask like a noob. Any chance this can (eventually) help with old 405s and 415s that need servicing if the capsules still work? My 415 T is a bit noisy even using Sennheiser's BP2 power supply.

 

[Voyager10]

@Icantthinkofaname - so by 'noisy' do you suspect there's a fault, or are you wanting to mod it? I guess you've already done the 'alignment' procedure on the 415T (adjusting the L2 coil for 0V difference between the A and B test points) from the service manual? That will affect sensitivity, and therefore the effective noise level.

The 415T manual gives 70dB SNR (relative to 1 Pa, i.e. 94dB) so equivalent noise would be 24dB. The A-weighted equivalent is usually quite a bit lower, so it should be entirely comparable to other decent SDC's, and noticeably better than, say, a miniature electret.

 

[Icantthinkofaname]

@Icantthinkofaname - so by 'noisy' do you suspect there's a fault, or are you wanting to mod it? I guess you've already done the 'alignment' procedure on the 415T (adjusting the L2 coil for 0V difference between the A and B test points) from the service manual? That will affect sensitivity, and therefore the effective noise level.

The 415T manual gives 70dB SNR (relative to 1 Pa, i.e. 94dB) so equivalent noise would be 24dB. The A-weighted equivalent is usually quite a bit lower, so it should be entirely comparable to other decent SDC's, and noticeably better than, say, a miniature electret.

Hard to say, could be a fault, or maybe I just have to adjust the coil. With the PSC adapter it performs similarly to my NTG3, but with extra noise, audible amounts that shouldn't be there. With the Sennheiser BP2 it's similar noise with much lower output, though that might be the BP2, unless the PSC also acts as an inline mic boosting preamp like a Fethead Phantom.

 

[Voyager10]

If you have some audio analysis software (Room EQ Wizard works well) it should be possible to compare the noise levels of two mics, without much extra equipment.

Basically - set the mics up side by side, the same distance in front of a speaker. Run the mics into two channels of an audio interface, then play a sine wave through the speaker and adjust the channel gains to set the same level (e.g. -6dBFS) on each channel. You can then turn off the sound & put each mic in as quiet a place as possible. Measure the residual noise signal (e.g. using REW's real time analyser screen) for each channel. The results can be compared directly, as they will take the different sensitivities of the two mics into account.

I have an NTG1, which has a spec of 18 dB(A), and seems roughly the same as the 415T noise-wise. I'd expect the 415 to be a bit noisier than the NTG3, as the latter has a 13 dB(A) noise spec. If I get a chance I'll do a comparison.

 

[Voyager10]

Here's some comparative noise measurements I've made using the above procedure, for a selection of shotgun mics.

It's a bit rough and ready, but here we go:


From worst to best, we have
1. (Red trace) Beyer MCE 86, spec gives 25dB(A) noise level,
2. (Green trace) MKH 415T, spec gives 24dB unweighted
3. (Blue trace) Rode NTG1, spec gives 18db(A).
4. (Orange trace) MKH 416T (more recent P48 models given as 13dB(A) - this is an older one)

For reference, the sound source was 1V rms @ 1Khz, fed into a small test speaker at 30cm distance from the end of each mic, and the gain in each case was adjusted to give -10dBFS level as measured by REW.

 

[Tubetec]

Heres a BBC report on the MKH that might be of interest .