Detection was established on the very first attempt, using a new (not blue!) blade and solid hookup wire for contact. It was found that only printed areas of the blade were effective.
However, the local broadcast station sounded real weak with the blade than with the OA5. Results were even worse with a pencil as contact.
A spare carbon brush of a mixer/grinder was then tried out. It worked fine on most parts of the blade's surface and the signal strength went up multifold, though not as strong as with the OA5.
The spring made it possible to vary the contact pressure for best results.