I wouldn't use this technique for much beyond a 2:1 ratio so using a 10K:16 is good for a 5K:8 equivalent. Using a 2.5K:4 would also (technically speaking) work, but the LF response would be poor In all cases the impedance ratio is 635:1, and the turns ratio, 25:1. The latter case would be OK for SE guitar amp usage, as the low E string is around 82 Hz, so the LF response limitation is of less relevance.
Similarly, with the 6CW5 / EL86, a good PP OPT primary is 3.3K A-A, so I'd just use the more commonly available 6.6K:16 ohm as a 3.3K:8 ohm replacement. The Hammond 1620 comes to mind here. This trannie has secondary values of 4, 8 and 16 ohms available, and an UL tapped primary.
For non-UL PP applications, the hard to find Sansui 1000 (NOT 1000A) OPTs would also be near perfect, with 3200 ohms A-A and 8, 16, 32 ohm secondary. THese OPTs were likely produced by Tango or Tamura, for Sansui.
I haven't yet run the "10K" tranny in an actual working circuit, that may be a while yet to come... The primary DCR as compared to a GXSE15-8-5K is nearly twice as high, 301 ohms vs 156 ohms. If Edcor uses a fixed number of turns per K of primary inductance, then the "10K" would have 1.4 times the number of turns than the "5K" primary. But as more turns are added, each layer of primary winding has a slightly larger diameter, and this adds up pretty quickly when there ar multiple layers of windings.
(think of a roll of paper towels, or of toilet paper).
So, there would be more "copper" losses, but not unreasonably so, I see several other "5K" SE OPTs with around 300 ohms DCR on the primary, and all are decent to excellent, as well as really cheap ones with under 100 ohms DCR (meaning les turns, significantly less inductance, and a resulting poor LF response) .
But then, the GXSE15-16-10K's primary inductance is also nearly twice as high 29H vs 17H), and that definitely will help the LF response. The DCR is a product of the wire gauge used in the winding, as well as the number of turns. Just think of it as a big wirewound resistor, for DC considerations.
The first time I played the "ratio game" was when I cloned Jef Larson's "Abraxas" 12B4A amp, which uses the small German "EL95" SE OPTs with 11 to 12 K primaries and 4 ohm secondaries, as a 22 to 24K, primary, to 8 ohm secondary. This combo may seem extreme for a 12B4A (with a 1.09 K Rp, but then the sound quality is excellent, and about 900 mW max, so it is a strong contender in the "Darling" sub-watt SE amp class.
FWIW, these trannies are even better with 6S4A. MUCH better in fact, probably one of the sweetest sounding amps I've ever heard, but also in the sub-watt class.
/ed B