Its interesting how he now seems to be able to to use cheaper non-timing GPS's to discipline his hardware OCXOs.
BG7TBL has been able to push the envelope pretty far as what can be done with cheap GPS hardware.
As long as a GPS receiver puts out some kind of signal, usually 1 PPS, you can use it to discipline an OCXO. The question is what kind of performance will you see? The long term frequency error for any properly designed GPS receiver will be zero, regardless of what type it is. The differences, if any, should show up in short-term measurements.
For a navigation receiver, the 1 PPS output is something of a minor afterthought so it gets low priority. On some older navigation receivers, I've seen the 1 PPS accuracy specified in the microsecond range.
On a timing receiver, the 1 PPS output is the main output while the location is sort of a minor afterthought. The code is optimized to provide the lowest possible error on the timing output(s). I've posted numbers in the past that showed the period of the 1 PPS output with a Standard Deviation of a few nanoseconds and a maximum range of 50 nanoseconds or less.
TBH, I've never seen a direct comparison of the output of a GPSDO built with a navigation receiver vs. a timing receiver. It might be possible to adjust the PLL parameters for the nav-based GPSDO to minimize the short-term differences.
Ed