It might be superfluous now, but obviously I have built the EB-63 and already commented on the results on this website at eham.net/reviews/detail/1223, as you have.
It is very easy to dramatically improve the RF-sensed T-R switching circuit. I didn't use the Motorola circuit at all, instead used a 5082-2800 hot carrier diode detector (decoupled from the RF input line with a .001uF cap and 1K resistor or so) driving an MPS-A13 (I think it was) Darlington amplifier which pulls in a sensitive Radio Shack 12V relay "instantly" and reliably, with the decay time set by an adjustable R-C circuit in the charge line to the Darlington. This whole circuit is about half the size of a postage stamp, built on a tiny piece of perfboard supported by the relay itself, and probably cost $2 to build. Extremely reliable, no drop outs, no glitches, works fine on all bands, as it has for about 11 years since I installed it.
As stated in my review from early December, I never bothered with the bandpass filters at all, and frankly, I'm glad I didn't. After writing the review on this site it raised my curiosity about spectral purity of this amp (which I can't recall ever checking initially), so I ran some tests. With 5W drive and 100W (approx) output, the amp is still in its linear range and purity is excellent without filtering -- worst case was 2nd harmonic of 14 MHz, down about 42 dB at +50dBm Po (100W). 3rd harmonics are well below this, across the spectra. My particular amp enters the non-linear region at about 120W out (7W drive) and is compressed 3dB at 150W out (10W drive). The obvious degradation in purity is easily noted on the spectrum analyzer....at +52dBm Po, 2nd harmonic at 14 MHz is down only 29 dB, a 13 dB degradation. But again, at this point, the amplifier is in compression, and something's got to give.
Signal source for this test was my TS850S/AT (adjustable Po down to <1W) run through DuneStar bandpass filters to the input of the EB-63; output of the EB-63 to 1.2kW Bird Termaline passive load, through -20dB HP 778D Dual Directional Coupler. -20dB decoupled port on the coupler run through 20dB 50 Ohm inline attenuator (Narda) to HP 141T storage display w/110 MHz spectrum analyzer plug-in. Purity test on source without amplifier indicated 2x down >65 dB, 3x down >75 dB at 5-10W drive levels.
Based on this, I'd recommend NO EXTERNAL FILTERING at all is necessary with the EB-63 provided you stay within its linear range, which appears to be ~100W Po @ 13.8Vdc supply voltage.