#19 Takstar / Gear4Music WPM-200 Wireless Monitor System part 3: transmitter mods

And at long last(?), we've reached the point where we take a look at the transmitter end of this wireless monitoring system. Not a whole lot to see on the outside, really: power switch, power LED, channel select rotary switch (1 to 6) on the front, and the DC input barrel socket and the 1/4" stereo input jack socket on the back.

On the inside, the circuitry's considerably less involved than the receivers. Not in the least because there's no digital single-button channel selection, no LCD to drive, no volume control - almost nothing, really. That being said, the input stage is somewhat interesting, in some ways, at least.

But first thing first, an overall, top-down view. DC input goes into a 78M09 9V voltage regulator (the DPAK-packaged one), which supplies the lone dual opamp, and also the SOT89-packaged 78L05 5V regulator, which provides power to everything else on the board.

The signal path is perhaps a smidge less... conventional - the input signal gets padded down considerably (a 22k / 680ohm divider on each channel), and then, around the dual opamp (a YouWang brand - no, i'm not kidding, see the link - UTC4558E) there is an arrangement that forms a crude BJT-based limiter. An LTspice recreation of the circuitry would seem to indicate indicate it has a very fast (sub-millisecond range) attack time, and a comparatively long (10-20ms, depending where you consider it to be "back to full amplitude") release time.

Next in line are two TA31101 knockoffs (LT31101S), of which only the compressors are used.

The signal from those goes into an anonymous (unmarked) SOIC14-packaged chip - the modulator? - which has an associated 7.6MHz quartz crystal connected to it. A few signal lines go off to an unmarked TSSOP16-packaged chip (remarkably similar arrangement to the ones in the receivers), which is connected to the RF-black-magic-voodoo going on in the shielded can. After that there's some filtering / clamping going on (caps, resistors, inductors, diodes), and then there's the antenna.

Right then, on to the mods. Right off the bat, i'm not 100% sure just how happy those two voltage regulators are, with next to no bypass capacitance around them (save for an 0805 and an 0603 cap on the 78M09 output / 78L05 input). Just for peace of mind, that's the job for a 100u/25V Rubycon YXF on the input of the whole thing, and a 1206-sized 8.2uF/16V on the output of the 5V regulator.

Perhaps at some point, i might get around to perhaps bother with swapping out the opamp with a lower-noise one (4558's are pretty much glorified dual 741's - literally ancient design), but i have my doubts that's the main contributor to the noise floor. The more pressing matter (no pun intended, you'll see why) would be perhaps lowering the attack & release times of the compressors. From the factory, they're quite slow, and don't sound particularly great. At least in the datasheet, the "default" seems to be 2.2uF (connected to pin 16); i wouldn't be surprised if they went with the same value.

Hunch confirmed - upon removal of one of the relevant caps from the board, my meter says it measures around 2.38uF (2.2uF +/-10%, pretty standard). Replaced it with a 100nF cap, but a brief listening test showed some noticeable "extra" distortion, possibly overmodulation(?) due to louder peaks in the output signal. At least for the time being, i think i'll just put the stock cap back in, and consider tweaking it later, if need be (maybe dropping it down to 1uF or 470nF).

In the mean time though, especially since the plug-packs these came with are... let's just say, "of questionable quality", i think i'll stick some extra filtering on the supply line. A 100nF film cap, right next to the input jack, as well as some series inductors. Two of'em, in fact - one before the power switch, and one after (since the tracks do run parallel to the signal inputs, for a couple cm).

A quick test on the bench supply shows the transmitter draws about 75mA, so i think some 100uH / 150mA inductors should do just fine. The resulting 200uH and 100uF LC low-pass filter has a cut-off frequency of 1.1kHz or so, so that should attenuate whatever high-frequency garbage might be coming through / from the power supply. Since linear regulators don't have TOO fabulous rejection from the kHz range upwards, they can use all the help they can get.

A brief listening test was encouraging. Where before, there was a clear whistle / whine audible in the earphones (in addition to the somewhat high-ish noise floor), with no signal applied to the transmitter's input, now all one can hear is the noise floor itself. Thus, i'd dare call this a success...