I'm away from the shop now, but I will post pictures of DC coupled waveform tomorrow and scope screen at higher frequencies.
Jon, I had thought that the problem might be with the single supply, but firstly the circuit is from a pretty good source so I had confidence in it. Secondly, the output is really clean, and conforms to calculated values (I measured frequency of the square wave output with microcontroller), at higher frequencies. Finally, I was avoiding a split supply as the output is used for a microcontroller (arduino) which cannot take negative voltages, or at least not below -0.3v I believe. I will try a split supply tomorrw for sure, thanks.
Bud, the part is from Farnell, TI I think, or at least a reputable manufacturer, and I've switched it out several times for new ICs, plus it only suffers at low frequencies so fairly sure this is not the problem.