How to publish analysis work of an approach worse that state-of-the-art?

Question

I am currently working on an engineering problem for which I have developed a new and quite effective solution. My research field is electronics (if that matters).

The problem is quite complex as it revolves about using a single circuit configuration for multiple voltage ranges (conversion of ac voltage to high dc voltage). State-of-the-art circuits offer unacceptable performance for input voltage range A and reasonable performance for input voltage range B.

My solution offers reasonable performance for voltage range A but the performance for voltage range B is slightly worse than state-of-the-art.

Since the problem can be broken to two individual parts (low voltage range and high voltage range), I plan to publish the first part on its own (e.g., novel high voltage step-up converter). However, I am debating how to publish the second part that on its own does not offer better performance than state-of-the-art. I have spent half a year developing, evaluating, and optimizing the circuit and I believe the engineering society would benefit from this work.

I am thinking along the lines of "Analysis and optimization of ...".

Answer 1

Perhaps it would result in a paper that is too long to publish, but you might consider, at least, putting the two results into the same paper and giving an analysis of each case.

Show how and why it works for one range, but, in a later section, show why it fails to work for the other.

The overall suggestion here is that the analysis of the result and the possible insight it might bring to the general case is more important than each result individually. That is the perspective of a mathematician, anyway. The insight of the analysis/proof is more important than the theorem itself.

Answer 2

However, I am debating how to publish the second part that on its own does not offer better performance than state-of-the-art. I have spent half a year developing, evaluating, and optimizing the circuit and I believe the engineering society would benefit from this work.

(emphasis added)

Why do you believe the engineering society would benefit from it? What does it offer over the existing approaches? If you can answer that question then you know how you should proceed with writing a paper for this second part.

You seem set on defeating the state-of-the-art performance by a particular metric, but this is not the only way to publish. You can pick your own metric to evaluate; you can even say that you do not perform as well, but your approach is novel because of X, Y, and Z. Perhaps you are the first to apply a certain technique in this domain, and you show that the technique is promising because it achieves almost as good performance as the state of the art. Or perhaps you can argue the the problem you are solving is more difficult or more general.

There are lots of ways to pitch a research contribution, not necessarily just based on evaluating the performance of your method. The success of publishing the work will depend on how well you can convince others that the work is interesting.

Answer 3

Why split the result into two papers since they are obviously quite closely related? It seems like salami-slicing and makes both papers weaker than they would otherwise be together. If it would be overlong to pack both voltage ranges into a single paper, you might consider putting the "worse than state of the art" range in the Supplemental Material, stating that the design gives acceptable performance in range B (since you say the results are only slightly below state-of-the-art) with the added benefit of a wider operating range. Getting a single paper out of 6 months work is already pretty good!