Tuesday, 11 February 2025

MEGAphone MEMS Microphone Module

Next up, I'm looking at the MEMS microphone modules. We want to be able to have more than one, so I'll route out the left-right select line that they usually have. 

1. Requirements Analysis

But first, let's look at the requirements that relate to this module:

Requirement 3.1.1: One or more power-switchable MEMS microphones with power status indication.
 

This should hopefully be a simple module, as I believe it can just have a MEMS microphone, with very few passives.

2. Selection of Key Component(s)

Let's look at a candidate microphone: SPH0690LM4H-1
These are about $2. Digikey's link to the datasheet is broken, though. But I found it here.

Unlike the ones I have used previously, this one is happy with 3.3V VCC, which is convenient.

Like most of these things, they use a simple 2-pin data protocol, with data on either rising or falling edge, based on a SELECT line, which allows 2 microphones to share a pair of lines. 

 


I'll expose that SELECT line, to allow selection to be made by the mainboard, possibly with a 100K pull-up resistor on the module, so that it will default to one or the other, even if the mainboard doesn't select.

Apart from that, I'll just need the decoupling caps (which must not contain type 2 dialectrics) and output resistor on the DATA lines, and optionally also on the CLOCK line, which should be safe to include on each module.

The datasheet includes no guidance on the R_TERM values, nor for the bypass capacitors, which is a bit annoying. Maybe I'll look at a different part, like a IM69D130V01XTSA1, whose datasheet has all the info nicely laid out:


So it looks like Infineon wins out, because their datasheet is better. Note that the actual interface is identical, which is good for allowing future substitution of parts.

3. Schematic

Let's make a schematic for it in KiCad:


I've also made the 2x5 module footprint with the extra pins removed, so that we have only the 5 we need, while still keeping it impossible to mis-orient, and have all the pads line up.

4. Part Selection

The final step is to pick out parts for the passives. The three resistors are simple, but we have that note about the capacitor not being allowed to have type 2 dialectrics. That excludes most 100nF SMD capacitors, leaving just three options with Digikey: the AC1206JKNPO9BN104 which is cheap, but out of stock, or for 10x the price, I can get the CC1206JKNPO9BN104. Both are 1206, so that makes it easy -- I'll just add a 1206 footprint for it.

Using 0603s for the resistors, it looks like we might just be able to get away with the 2x5 small module size, which would be nice:

It's just a whisker out of spec for the resistor pads, but given that the resistors will sit within the pads, rather than right to the edges of them, it should probably be ok... Assuming I can route the nets... which I can.

5. PCB Layout

Oh, it turns out the footprints hadn't updated from 0805s to 0603s. With that done, I can get it to fit properly:


You really do get the idea of just how small this module has ended up. It's only 10x13mm, after all.

6. Bill of Materials

IM69D130V01XTSA1 - MEMS microphone

AC1206JKNPO9BN104 (if available, if not substitute with CC1206JKNPO9BN104) - capacitor

RK73H1JTTD1002F - 100 Ohm 0603 SMD resistor

ERJ-3EKF1003V - 100K Ohm 0603 SMD resistor

7. Review

So I think that's it for the module. Let's double-check with the requirement it has to satisfy:

Requirement 3.1.1: One or more power-switchable MEMS microphones with power status indication.

The only thing not on the module is power status indication, which we will put on the carrier board. 

This also gets us half-way through the schematic and PCB layout milestones for 2.5.1 Internal microphone and Speaker Module.  I'll work on the speaker module next, so that I can close those milestones out.  The actual driver for the speaker will be on a separate module that will use a standard cellular phone "audio codec" IC, and have the pads for the internal speaker. I'll tackle that in a separate post, though.

No comments:

Post a Comment