KM84 revisited

[jp8]

If you have been following this and this thread of @joulupukki , you know that I have made two SMT variants of the KM84 circuit with several improvements compared to the classic KM84 circuit. For convenience, I have called these circuits KM84+ and KM84++. They are described in detail on my website, which I have now launched. You will find the links below. The KM84+/KM84+ article texts are not completely finished yet, but I plan to finish them during the Christmas holidays. I hope you'll enjoy reading the articles on my website and find them useful!

I will make KM84+/++ PCBs available for the Takstar CM-60 and CM-63 and Alctron T-02A on PCBWay, but at the moment of writing, I have only finished the building instructions for the KM84+, suitable for the CM-60 microphone.

Links to KM84 and KM84+/++ articles on my website:

KM84 circuit description: https://modimications.com/km84-style-charge-amplifier-circuit/

KM84+ and KM84++ description: https://modimications.com/km84-and-km84-head-amp-circuits/

Takstar CM-60 and CM-63 description and mods: https://modimications.com/takstar-cm-60-and-cm-63/

Available PCBs (at the moment only one, but you get an idea of what I'm working on: https://modimications.com/all-orderable-pcbs/

When more KM84+/++ PCB variants become available, I'll keep you informed via this thread.

Note: unfortunately, my website does not display correctly on mobile devices. I'll have to work on that.

Happy reading and Merry Xmas to all!

 

[kingkorg]

 

[rock soderstrom]

They are described in detail on my website, which I have now launched.

Congratulations on your new website jp8! It looks great and I'm looking forward to reading it. Bravo!

 

[gyraf]

Darn, that website is a really good resource on all things KM84 - Thank you!

/Jakob E.

 

[jp8]

Thank you, Jakob! And to all who liked my post. I'm happy to hear you find it usefull! If you think anything is missing or I have it all wrong somewhere, just let me know. Ain't nothin' perfect on this world...

 

[pmfalcman]

@jp8 - FYI.: No issues on my phone. The site displays and works fine.

Motorola, so it runs practically pure Android.

 

[RuudNL]

As far as I see, the schematic of the Takstar CM-60 is not correct.
In the DC converter C4 should be connected to the junction of L1 and C5, not to the emitter of Q1.
Also I have a feeling that C6 (usually 22 or 27 pF) is connected between collector and emitter of Q1.
In general C5 is 6.8 or 7 pF. The position of R9 looks 'strange'. (Collector Q1 not decoupled.)

 

[jp8]

Thank you Ruud for your valuable comments! I assume you were referring to the schematics here on my website. I checked the circuit diagram, and you were partially right.

• C4 should indeed connect to the C5-L1 junction.
• C6 is connected as depicted! You'd indeed expect it to be wired as depicted in the CM-63 schematic, i.e. between Collector and Emitter, but it isn't. I removed the capacitor from the PCBA and measured its value using my Smart Tweezers. It measured 2.7 pF.
• I also measured the value of C5. It's value is 22pF.
• While checking the circuit, I also noticed another error: trimmer P2, which is actually not placed, does not shunt R8 but R9 instead. I assume R9 was added to reduce the output voltage, and with P2, the capsule bias voltage can be adjusted. There is NO decoupling capacitor on the R9/Q1/R10/D2/C3 node!

I have revised the schematic accordingly and published it on my website.

 

[jp8]

I have made the KM84++ for the CM-63 available on PCBWay now. Link: https://www.pcbway.com/project/shareproject/KM84_for_Takstar_CM_63_Microphone_524d4f22.html

 

[jp8]

If anyone used the KM84++/CM-63 Centroid file, kindly note that the initial version I uploaded with the design on PCBWay used the wrong reference point. I have uploaded a new one using the right reference (bottom-left corner of the PCB, looking from the Top side).

 

[c1184701]

If anyone used the KM84++/CM-63 Centroid file, kindly note that the initial version I uploaded with the design on PCBWay used the wrong reference point. I have uploaded a new one using the right reference (bottom-left corner of the PCB, looking from the Top side).

Your website looks great Jan, thanks for all of your work on this. I missed the boat with some of the KM84 projects from a few years ago, so I'm going to have a go at these in the next couple of months.

On the KM84++ what components determine the amount of attenuation of the pad? I was wondering if I could increase the amount of attenuation for a dedicated close-snare version of this mic.

 

[jp8]

Thank you very much for the compliments!

R25 and R26 determine the pad, but you can set the sensitivity in various ways.

The gain of the JFET charge amplifier is controlled through C2. With C2 = 8.2pF and using a 3U cardioid capsule, the calculated max SPL equals 121 dBSPL with the pad off. With the pad on, it equals 141 dB, which I guess should be sufficient for most applications, including snare drums. But if you want more headroom, you could increase C2, which increases the max SPL, with and without the pad on.

The gain is also determined by the polarization voltage. The level of which can be set with trimmer P1, but you can leave this optional trimmer out and jumper it. If you want to reduce the gain, then I'd rather increase C2 than lower the polarization voltage, because a higher value for C2 will also reduce THD, while a lower polarization voltage would increase noise slightly.

As a Pad circuit, you can basically choose to either make C2 switchable as done in the classic KM84, or make the polarization voltage switchable to a lower voltage. I chose for the latter, as it was easier te implement and the increased noise is not an issue when the pad is required at high SPLs. The reduced polarization voltage is determined by voltage divider R25-R26 and equals 17.5 * R26 / (R25 + R26) Volts (Vdd of IC1 and IC2 equals 17.5 V). With P1 set to 0 Ohm, the polarization voltage equals ~69.5V. The sensitivity scales linearly with the polarization voltage, so for a 20 dB pad, you'll want a polarization voltage of 10^(-20/20) * 69.5 = 6.95 V. With R25 = 150k and R26 = 100k, you'll get ~7V. Not exactly 6.95V, but the error is less than 0.1 dB. Suppose you want to keep the max SPL without pad set to 121 dBSPL and want a 30 dB pad, then you'll want a polarization voltage of 10^(-30/20) * 69.5 V = 2.2 V. R26 can now be calculated as: R26 = (Upol / (17.5 - Upol)) * R25 = (2.2 / (17.5 -2.2) ) * R25 = 21.57k. The nearest standard E24 value is 22k.

I hope this helped!