As discovered by members of LMR, there is a bug in one version of one kind of Picaxe.
Phil and Limpa described their problems in two different nodes on the site. This thread aims to bring all info together, so we can write a walk through about it.
Recognising Buggy Version:
I don't use the Rev-Ed program editor so I can't give specific instructions but when your program is downloaded there is a confirmation dialogue. I'm using the command line compiler on Linux which reported "PASS - programmed PICAXE-28X1 (40X1) vA.4 successfully.", I assume that something similar is seen in the Windows Program Editor.
I suggest that if that confirmation string contains vA.4 then you have buggy firmware. I assume that any future purchase from Rev-Ed will contain a later version of the firmware, and the A.4 firmware seems fine for other things - I have successfully connected up an SRF005, switches and LEDs.
The only chip I am aware has a servo/servopos problem is the 28X1 with v A.4 of the firmware. I assume that different chips have different firmware but it is possible that, say, 28X1 and 40X1 share similar code and therefore the 40X1 might have the same problem.
The servo and servopos commands move the servo clockwise and the server buzzes; however if you use the appropriate pulsout commands the servo works correctly.
Additionally, as both Limpa and I discovered, if you perform a servo command on another output then servo commands on the proper servo port work. To get port 0 to work you need to fidde with port 7, port 1 fiddle with port 0. This is not a solution in most cases and the servo port is in a funny state so if you stop issuing commands it will turn on after about 12 seconds, and stay on for over two minutes.
I emailed Rev-Ed technical support (and posted to the picaxe forum) describing the problem and, after they investigated and fixed the problem, they shipped me replacement chips.
Limpa is still waiting for a replacement chip.
(R)ik will see which version of the 28x1 ik bought back in October.