In article d.invalid,
Adrian Tuddenham wrote:
I've had similar experiences with static electricity on the drive belts
of some tape machines, but this is a hand-held digital recorder with no
moving parts.

It makes the noises even with a short double-ended 3.5mm stereo jack
lead in the input socket and nothing connected to the far end. If the
lead is plugged into a passive mixer, the result is just the same and if
it is connected to a grounded mains-powered mixer, the result is still
the same. I don't think static electricity or anything to do with
grounding can be the problem.

Is it possible to download the resulting digital recording, and
examine the waveform with Audacity or a similar tool?

A ticking sound might be the result of a one-sample error - either a
single bad sample (which would show up as a vertical "tick" on the
waveform), or a missing or duplicated sample (which would show up as a
vertical jump in the waveform, or a sudden flat spot).

In either case, "clocking errors" are what comes to mind for me... a
synchronization problem in the analog-to-digital/storage process.

Years ago, there used to be a fairly common problem of this sort with
long-distance phone calls, where the calls were being transmitted in
digital format between phone switches whose internal
clocks/oscillators were not properly synchronized. One would be
running faster than the other, and occasionally (every few seconds)
the difference between the clocks would cause one audio sample to be
dropped or duplicated.

This resulted in an almost-inaudible "tick" sound. Not a problem for
voice calls, really. However, for modem connections (e.g. V.32) which
were sensitive to signal phase, it was enough to cause a one-bit error
in the data, resulting in a spurious character appearing (most
commonly "}" if I recall correctly). If V.42 error control was turned
on, the error was detected and the data packet re-sent, so you
wouldn't see the bad data, but the retransmission slowed down the
connection significantly.