Bill speculates:
<<Given that two EFIS units with battery backup are more reliable than a single vacuum pump, your argument that people must have "TSO'd" equipment is logically ridiculous - especially if they also have as part of their panel an independent 2-axis autopilot.>>

The primary assumption here is false. It is not "given" that "two EFIS units with battery backup are more reliable than a single vacuum pump". Analysis and data show the opposite is true.

Having designed several Certified EFIS systems and sensors (AHRS, Air Data, Magnetometer, OAT etc.) over the last 15 years and shepherded those systems through DO160 certification testing I can say with the confidence of having empirical  data (Remember that one test is worth a thousand expert opinions) that I would NOT fly behind a panel that had ONLY electronic gauges, regardless of their certification level or lack thereof. Even the Starship, with a million dollar 17 tube Pro Line 21 integrated avionics suite, STILL has a mechanical Airspeed, Altimeter and AH.

Regan Designs was the first company to design equipment that passed the DO160 Lightning Induced Transient Susceptibility (section 22) and Lightning Direct Effects (section 23) requirements introduced in 2004.  Hamid engineered several test articles that he then subjected to simulated lightning strikes in a certified laboratory.  Based on those tests I can say with a high level of confidence that most GA certified EFIS and likely all experimental EFIS systems will not survive a proximal lightning strike, let alone a direct strike. Same goes for experimental autopilots. And that is considering just 2 of  26 sections. There is also Shock, Vibration, Temperature, Magnet Effect, Voltage Spike, Operational Voltage, RF Susceptibility, ESD,  etc. etc. etc.

One transient event can take out most of the digital electronics wired into the aircraft. A battery back up won't do any good if your EFIS is fried.

Here are some additional reasons for a spinning mass backup to an EFIS.

1) Compelling disaster. If your EFIS starts to roll (e.g. due to a long climbing departure turn), you feel compelled to follow it to your doom. Having a familiar AH in you scan will help you keep you wings level.
2) Different physics. Spinning mass and MEMS gyroscopes work on different principals and physics. The set of events that will kill both is small and most of those involve the pilot not surviving either.
3) Old faithful. Most of us learned to fly with an AH. The response to and AH display is nearly reflexive and may save you life during a helmet fire.

Some words on TSOs. TSOs are the Technical Standards that equipment must meet to to be considered as equipment on aircraft. Therefore, in order to have an "altimeter" in your aircraft you must have an instrument that meets the TSOs for an "Altimeter". You can either let the instrument manufacturer do the testing or, as an aircraft manufacturer, you can do the testing, and document same. Why? Imagine you took a rock and printed "8,250 feet" on it. You then "install" it in your aircraft and claim it is an altimeter as it will tell you your altitude during certain conditions of flight. One of those conditions must be that you are actually flying at 8,250 feet MSL. To prevent this type of thing the FAA has established standards that a device must meet in ordered to be qualified to function as a required device. See FAR 21.601.b.1.

So, Bill's statement that "...(the) argument that people must have "TSO'd" equipment is logically ridiculous..." is false. You MUST have at least one of each of the required instruments and they MUST meet the TSO. You can call it ridiculous, but it does not change the fact that it is the law.

Bill also postulates that " Experimental EFIS units work acceptably with either good pitot-static input or GPS input, removing the single point of failure in steam gages.". This statement is non sequitur. The most common Pitot Static problems are blocked ports (insects or ice), leaks or water in the lines, any of which will produce a similarly wrong reading in either the steam or electronic display. Redundant sensors are fine ONLY if you have a method for differentiating good data from bad data.

FWIW, Being a good pilot, as I am sure Bill is, does not make you a good engine mechanic or good at failure analysis. You can take or ignore the advice of those with experience. Fred has put a lot of thought into his system and has reduced the likelihood of a catastrophic electrical failure. Now if you could only do the same for rocker arms.......

Regards
Brent Regan