Hi Viv,

> So is it possible that the error is actually 'bubbling up' from an underlying rowsource/controlsource problem?

Yep.

> Maybe it's safer to look for a Windows API alternative to INKEY() ?

I'm not sure about this one. It really depends on how the idle loop is implemented. But if it's part of the regular message loop, you will likely see the same effect no matter if VFP waits for a key or you do. The fact that you are waiting might trigger the code that causes the error.