Zoom H4n Pro Repair

A Zoom H4n Pro audio recorder found its way to me, non-working. Recordings seemed to be silent, but with loud enough input you could just make out some distorted sound in the recording. After (unneccessarily) disassembling it completely, I found out that the power supply of the preamplifier board was sitting at around 0.8 V, which thankfully/boringly turned out to be the only fault in the unit.

The supply voltage was supposed to be provided via IC4, which sits on the first PCB you come across when removing the back of the case. This SOT-89-5 with package marking ‘3HN’ is an XC6227C501 5 V LDO regulator, which was fed by 5.5 V from a step-up, but wasn’t really doing its job. The preamplifier seemed fine as well, drawing about 13 mA, so no short circuit there or anything. Feeding in 5 V externally seemed to fix all the issues, so replacing the LDO was all I had to do.

The XC6227C501’s job is to provide some post-regulation to the step-up to make sure any ripple generated in the first stage doesn’t make its way into the power supply of the pre-amplifier op amps and into the audio signal. I didn‘t have this particular part on hand and didn’t feel like waiting with this repair for the next opportunity to order one. Could I substitute something else?

IC4, the LDO in question with no obvious signs of anything being wrong. Do excuse the crappy quality, this was made on a phone through a watchmaker’s glass and I turned up the contrast to make the IC marking clearly legible.
My first replacement attempt with an LM2940, which sort of fits on the original IC’s pads (the tab is just not soldered to anything). Though, it did turn out to have a too high dropout voltage after all.

The XC6227C501’s has an enable pin, which could be used to save power, but that is just tied to the input voltage—switching the power supply seems to be handled elsewhere (I didn’t check). It doesn’t have to supply a lot of current (as previously mentioned, 13 mA), so all that remains as relevant for the particular choice of this IC is a low drop-out voltage, since there only is a difference of half a volt between input and output, and a good ripple rejection to sufficiently isolate the 5 V output from any noise present on the step-up’s output.

At first, I wanted to use a LM2940IMP-5.0, which sports a higher quiescent current, but should have good ripple rejection. Its dropout voltage should also be fine for currents this low. On top of that, it still fits on the PCB neatly, for some definition of neatly. However, mine refused to output 5 V at 5.5 V input and sat at 4.5 V even without load. The datasheet’s a bit ambiguous regarding the input voltage range (there also is no chart for the low-voltage performance of the 5 V variant), so either TI and I have different opinions what ‘dropout voltage’ means or I got sent fakes. Either way, I went on to try the next LDO.

I ended up using an HT7350, which doesn’t spec its ripple rejection and almost certainly has worse regulation, since it also has a way lower quiescent current. However, the step-up operates at just under 1 MHz and it seemed unlikely to me that the switching frequency itself negatively influences the pre-amplifier. On the preamplifier PCB, the power supply of the microphone capsules—they seem to have some internal amplification, as they have a separate power pin—is filtered with an 100 Ω/220 µF network. Only the op amps are supplied directly from the LDO, with the first amplification stage handled by an NJM2100.

Photo of a new LDO in a very different package than the original. It has to sit its back and at a weird angle, so that all three pins could be soldered down.
Topological fact: the pinout of a replacement IC doesn’t really matter as long as it has only three pins.

I did quickly check with an oscilloscope FFT that anything in the audible range at the LDO’s input is significantly reduced at its output, which seemed to be fine, especially together with the NJM2100’s PSRR. I also recorded a few seconds of silence by shorting the external microphone input, which still leaves the preamplifier in the signal path, and verified with an FFT of the resulting WAV that nothing heineous was occurring. Perhaps I’ll order the original IC at some point and do a more in-depth comparison, but in the meantime, I’ll declare this recorder fixed.

Published 2025-01-25 in #repair and #consumer electronics.