Having spent the last few years trying to research this question and the distinction between a counterpoise and a radial, I think this is as good an answer as any. Choose your own unit of length from nanometres to light years. Many writers conflate the two terms, confusing the rest of us, and making any sensible discussion of CPs or of radials (and how to optimize either) meaningless.
There are so many different opinions, answers and contradictory information on this subject in AR literature and online. What I’ve concluded is:
- A counterpoise is part of a ground plane to replace the ‘missing’ pole for antennas like EFHWs and bottom-fed monopole verticals
- It forms the top ‘plate’ of a capacitor and the ground on which it is lying forms the bottom plate
- (Unlike a radial) it is not tuned for the operating frequency(s)
- However, its minimum length (to be effective) is somewhat related to operating frequency and I’ve read mostly a recommended total length of the wire(s) between wavelength/4 and wavelength/5
- Given a total length of wire, multiple shorter counterpoises are more effective than one long one, e.g. I cut my original 16m-long CP wire (as supplied with my Cha MPAS Lite vertical), first into two 8m CP wires, and more recently into four 4m CP wires. This is more hassle and time-consuming to lay out. I’ve no objective way of judging any improvement but it comforts me because people say it should be better
- I never bother using a CP with my three EFHWs. I get good performance (i.e. reports given/received) with 10W of CW on mostly rocky summits. IMO I use EFHWs coz they’re a bit quicker to deploy. If I had to use a CP with them, I might as well erect one of my inverted-V linked dipoles instead
- Conclusion: don’t bother with a CP with the BandSpringer unless it comforts you or have time to spare