As an addendum to Draconis' excellent answer, one source worth trying if you're looking for attestations of a sign in a certain period would be CDLI, whose advanced search form allows searching by both sign and by period and/or time range.
For example, searching for the sign "PAN" (and any alternative readings it may have) in the "Old Babylonian" period currently turns up 52 results. A lot of them either have no photos or line art, or are composite texts whose "line art" is useless here (since it's just the text rendered in a modern cuneiform font), but filtering those out still leaves 28 results.
The first result I got is TCL 07, 011 (P386004), described as a "Letter tablet excavated in Larsa (mod. Tell as-Senkereh), dated to the Old Babylonian (ca. 1900-1600 BC) period and now kept in Louvre Museum, Paris, France", which has the sign PAN (read here as BAN) at the end of the third line ("ša-ad-da-aq-di-im RÁ.GABA.MEŠ LÚ.BAN"), as seen here in the "line art" tracing:
The photo is a lot harder to read (as usual; reading cuneiform from photos is rarely easy), but it's just possible to pick out the PAN/BAN sign at the edge of the tablet based on the surrounding signs:
So there, that's one example of the sign PAN in (cursive) Old Babylonian cuneiform.
Tangent: Based on a quick look through some of the other results from the CDLI search, the cursive sign form above seems perhaps a bit more simplified than usual (although I'd say it's almost certainly still a valid variant form, as opposed to a scribal error or a misreading).
Admittedly there don't seem to be too many clear examples of Old Babylonian PAN on CDLI in the first place; in most cases the PAN sign seems to be either damaged, right at the edge of a missing or damaged area, an uncertain (and possibly mistaken) reading, or just a mess of wedges that I have no idea how anyone could read. Still, if I squint at the most readable results just right, I can kind of see a semi-consistent sign form emerge. The general shape seems to consist of (from left to right):
- a (more or less) horizontal wedge near the midline of the sign,
- some (most often two, but sometimes just one or more than two) diagonal wedges (and/or Winkelhakens) in the middle, and
- another horizontal wedge at roughly the same height as the first.
There seems to be quite a bit of variation in the middle portion. In the example above, the middle part is simplified to just one diagonal wedge (and the two horizontal wedges are offset a bit, possibly because the sign was written right at the edge of the tablet). More often there seem to be two diagonal wedges, either both slanted down (as in the later forms of the sign shown by Labat, as quoted in Draconis' answer) or the first one slanted up and the second down. Sometimes the orientation of the first diagonal wedge can be ambiguous, making it look like a Winkelhaken. Occasionally there even seem to be some extra wedges crammed in there (or at least whoever traced the tablet into line art seems to have thought there were).
(The ambiguity between upward and downward diagonals and Winkelhakens is common in cuneiform writing: they're all part of a continuum of wedge shapes that can be formed with very similar stylus presses, the only difference being the tilt of the stylus around its long axis. Thus, especially when writing quickly in soft clay, it's easy for these wedges to blur together. This is sometimes even reflected in the evolution of sign forms, as seen e.g. in the variants and evolution of ḪI / ŠÁR and other signs containing the "diagonal square" motif, which all but universally turns into four downward diagonals in Neo-Assyrian cuneiform.)
One particularly nice example (courtesy of the comments below) is BIN 09, 124 (P236134), an Early Old Babylonian era administrative tablet written in Sumerian, dating from the reign of Išbi-Erra, which features both a fairly clean handwritten PAN and a cylinder seal impression containing a PAN written in a "monumental" (cleaner, more deliberate and usually archaic) style.
I've circled the handwritten sign in red and some of the seal impressions in green in the photo below (cropped and rearranged from the photo on CDLI, courtesy of the Yale Babylonian Collection):
There are a few more seal impressions of the PAN sign on the tablet besides the ones I circled, including one just below and to the left of the handwritten PAN (partly covered by the red circle). The full seal impression is a bit hard to reconstruct since (as usual) the seal has been rolled over the handwritten text, but fortunately both the seal and the handwriting have already been traced separately.
Alas, CDLI does not currently seem to have line art of the seal, but both can be found in the book Babylonian Inscriptions in the Collection of James B. Nies, Yale University, vol. IX: Sumerian Economic Texts from the First Dynasty of Isin (a.k.a. "BIN 9" in common shorthand), which is conveniently available as a PDF from the Yale University web site. Here's the tracing of both the handwriting (obverse only, from plate XX, tablet 124) and the seal (from plate XCIII, seal I), with the sign PAN circled in both:
The version of the PAN sign on the seal is basically the "W or M-shaped" Sumerian form, which isn't too surprising; the carved signs on seals, wall inscriptions, etc. tend to be "fancier" and retain archaic sign forms long after the signs became simplified in handwriting.
The handwritten form, meanwhile, is a nice example of the "typical Old Babylonian" form of PAN that I described above: a horizontal (or in this case somewhat upward-slanting) wedge, two stacked diagonal wedges (of which the upper one is somewhat ambiguous in direction) and a final horizontal wedge again.
(Note that the line art seems to show the upper diagonal wedge as somewhat more ambiguous than it looks to me in the photo; the upward corner of the wedge seems to extend further in the photo than it the drawing, making it look more clearly like an upward diagonal to me than the line art suggests.)
Anyway, the main point of this tangential digression (besides getting the details out of the comments and into the answer) is that when consulting primary sources like the CDLI for sign forms it's important to look at multiple examples to get a good idea of how much variation there was and what the "typical" form of the sign (in a particular period and region) looked like — after all, sloppy handwriting and scribal errors have always been a thing.
Also, while line art is typically cleaner and easier to read than photographs of actual tablet, when researching sign shapes it's good to at least check the line art against the photo (or multiple photos or a 3D scan or, if possible, the actual object). Drawing line art of a cuneiform text always involves some degree of subjective interpretation and educated guesswork, especially when the source artifact is poorly preserved, and it's easy to subtly "tidy up" the signs to better match what you think they should look like, rather than what's actually there in the clay.
