Tobias,
In the 4194A you can store 3 OSL cal data with frequency range, etc...
That's an nice feature since it's ready, right from power on, with up to the 3 cal/config you need for the current job.
Maybe HP felt this was less needed for typical VNA use.
Another reason may be that, AFAIK, the 4194A was designed byt HP/Yokogawa and the 4195A was in the US, but on the same HW plateform.
Two design teams hence two different machines. One thing in favor of this is that the 4195A and 4194A do not share much in matter of user interface.
I don't think there's any good reason to prevent it to start from low battery (except maybe marketing teams looking for opportunity to brick it).
But from experience don't underestimate design teams abitility to get something wrong from time to time, even at this level and also even from HP, Tek, and some other good ones.
I do have a very nice lab with lots of instruments accumulated over years, so had to reverse/fix quite a bunch of, and found my share of strangeness.
Hey they're humans, ain't they?
The last one is a fixed 4291A that did not start. The PSU was, I believe, a co-design from HP and Murata.
Apart from 2 toasted BJTs and a pair of dried Al caps (0.47uF for PSU startup timing) they also included fan rotation monitoring as a startup condition.
That's nice, but after I "painfully" closed the machine (lots of screws,...) it soon decided to brick again.
The fan grease (not the Grease fan...) had thickened a bit with time and sure enough, the startup window allowing the fan to speedup was dangerously too close to the limit (no need for half a second timeout. 2 or 5 or even 10 seconds would have been perfect).
No good reason for that, except it worked at design time and this was good enough...
Thanks,
Fred
