Is philosophy of science possible without prior understanding of depth and complexity of modern scientific formalisms?
There must obviously be a distinction between the Philosophy of Science, and Philosophy resulting from the implication of scientific facts. The need for a good understanding of science in the latter is self-evident and so only the former is really in question.
The most influential figures in the Philosophy of Science certainly considered that they were more than just historians. Karl Popper was a founding member of The British Society for the Philosophy of Science with part of its mission being to “...approach through the various special sciences to the philosophy of science”.
Popper said in Conjectures and Refutations “All that is needed is a readiness to learn from one's partner in the discussion, which includes a genuine wish to understand what he intends to say. If this readiness is there, the discussion will be the more fruitful the more the partner's backgrounds differ.“
In contrast, Kuhn's work was overwhelmingly historical and although could be read as a critique of the scientific process, Kuhn himself made few pronouncements about how it “should” be done instead. His less involved approach, however, drew criticisms particularly with regards to his lack of consideration for the regular discoveries of normal science, the impact of which he was not familiar with.
For modern philosophers of science, there is certainly a consensus opinion within the British Society for the Philosophy of Science that there should be a two way dialogue between philosophers of science and scientists. In the words of Grüne-Yanoff, a professor in the philosophy of Science “too few of us and our graduate students are literate enough in the methodologies and norms of specific sciences to be able to offer an effective exploration of their limitations and path dependencies", which I think sums up the answer to the question.
To provide a simple example on which to examine this question see this recent article in the British Journal for the Philosophy of Science. http://bjps.oxfordjournals.org/content/early/2016/08/18/bjps.axw014.abstract. Being a paid journal, I can't paste the full article, but hopefully it's abstract will be sufficient to see that regardless of what has been the case, modern study of the philosophy of science would be impossible without a good understanding of the relevant science.
Rephrasing: is it mandatory to understand (all) the "depth and complexity of modern scientific formalisms" to make philosophy of science?
Well, no. In fact, probably all philosophers lack of some of this knowledge.
But the lesser the knowledge of the current state of modern formalisms, the greater the chances of proposing wrong, repeated or outdated ideas, therefore, losing time. Knowledge is a history of learners starting where their masters have ended. I personally state that creativity without knowledge on a discipline leads to poor results (easy to observe with science or knowledge, but also in jazz ;) Then, proposing new ideas without knowing the state of a discipline may lead to inconsistent knowledge, to knowledge that can be easily disputed.
 Socrates: "The only true wisdom is in knowing you know nothing".  See the concept of "Reinventing the wheel"  Simple deduction: formalisms, or knowledge in general, have been created for centuries, and it has grown. Therefore, learners have obviously contributed to what their masters already did.  See this post about a student that make a thesis about a subject which already was done by other person: those are the risks. https://academia.stackexchange.com/questions/27779/i-found-out-my-masters-thesis-topic-has-already-been-done-exactly-and-my-adv
Many of the field's most prominent practitioners have focussed on older science. Kuhn's work started from the caloric model of head and the Ptolemaic model of astronomy. Feyerabend's famous case study is Galileo, Popper started by comparing the historical evolution of science and political philosophies. Though all of them did understand the origins of quantum theory, they probably did not know it well enough to actually do any.
For them, the issue is not the depth and complexity of modern science, but the ways in which older scientific ideas did or did not match our own, and evolved into them. Of course, this involves truly understanding that science, in its original complexity, and not only the altered versions that survived into the present. The point is to understand why the process works, how it evolved, and what made it function better or worse in its application to different cases.
The kind of philosophy that most involves really understanding modern science is not really in the same field. Finding philosophical roots behind modern scientific ideas is often more the work of ontology or epistemology than 'Philosophy of Science'.
Yes, it is possible. Modern science biases itself on observation -- a powerful basis for an ontology of the world, but not infallible. You can start there, in epistemology and solipcism.
I'll also give you this little nugget: empiricism is highly dependent on measuring instruments. However, every instrument implicitly encodes an ontology -- a model of the world. The problems of this should be self-evident.