How many samples of each one did you do?

I'd have thought that variations in background noise would swamp any minor changes caused by the fuse? Or am I missing something? Obviously if the variation is consistent over many samples you might be onto something. Interesting experiment.