Ok, I will try to share what I have found out, but I will share a bit of background first. I will try to be clear when it is just my opinion, versus when it is information that can be verified or a known fact.
I have found that there is actually more than one opinion camp when it comes to isolation manifold, independent doubles, and related redundancy/ risk management approach.
There is the DIR philosophy, and there are many other philosophies. TDI as a training agency is much more open minded and I have talk about different configurations using my double hose with some of their instructors, when their central office was here in Maine.
I actually did some technical diving classes about 10 years ago with a double hose and the Sherwood manifold ( with independent regulators). The instructor had no issues with my setup and he actually liked it. An interesting side note, the instructor actually cut me about a 50% discount since he felt that he also learned a lot during the class. I was doing some extended range diving long before those classes.
Note: the Sherwood manifold is classified as an isolation manifold by some since it does isolate both regulators, but some groups do not consider it an isolation manifold since it does not split the gas from the two cylinders.
There are also other manifolds (some modern and some vintage) that can also be configured with two independent closing/ isolating regulator outlets, but that do not contain the central isolating valve. Therefore, for this discussion I will refer to the “modern isolation manifold” as the one that contains the central valve intended to isolate the two cylinders.
In contrast I will call my preferred manifold as the Sherwood two outlet manifold.
Some pros and Cons.
The Sherwood vintage manifold will isolate and be able to close either regulator if one is malfunctioning. Closing the valve to the malfunctioning regulator (free-flowing or leaking) will preserve your gas supply to be accessed with the other regulator.
Regulator malfunction (free-flow or leak) is the only common form of malfunction where gas will be lost.
The manifold is simple with minimal number of O-rings and connections.
The Pros of the modern isolation manifold:
The Modern isolation manifold does the same function as the vintage Sherwood manifold, but it adds the potential of isolating the cylinders (which in theory will preserve half of you gas).
There are only four types of additional gas leak malfunctions that the modern isolation manifold is capable of mitigating.
1. A burst disc release underwater
2. A tank neck O-ring failure underwater
3. A manifold joint release underwater
4. A high velocity projectile puncture of a cylinder (like from a bullet)while underwater
Let’s look at the probability and how many cases have been documented of any of the four modes of failure. All four are theoretically possible, but some have never happen and (IMO) all can be avoided with minimal care.
Of the four, number 4 is the only one that I have seen, but only in the movies. I have not heard or seen any documented cases of this type of failure actually happening.
Number 3, I have heard of one documented case when someone dropped a set of double and it only started leaking after they got in the water. The explanation for the incident is that most of the damage was done during the drop, but it was not visible at first. The thermal shock after entering the water was the last straw the caused the manifold joint to leak.
Number 2, I have not seen or heard any documented cases of this type of failure with a substantial leak underwater. I have seen many rental AL 80 in the Caribbean with minor leaks around the neck, but nothing catastrophic and they are always visible from the beginning of a dive.
Number 1, I have read of only two incidents, but I have only kind of confirmed one case. There may be more , but both cases that I read about seem similar enough that they might have been the same or not. In any case it was a burst disc letting go shortly after entering the water. To me, the basic explanation for this is thermal shock on a burst disc that has not been service or was at the very edge of its stress level (probably an over fill).
Note: These are the only confirmed incidents that I have read about (or somewhat confirmed). If anyone knows of other confirmed incidents, please share them. A description of the incident would be helpful or a link would be even better.
IMHO, the first three modes of failure can be totally avoided with proper equipment service and care against damage. Again, this is my opinion and it is only based on observations during several decades of servicing dive gear and my technical background, but no actual study.
In a cave or wreck environment I can see the potential for an impact to the manifold, but it would require a fairly severe impact to cause a catastrophic damage. The Sherwood manifold is actually much less susceptible to impact due to the type of metal to metal joints that it uses (no O-rings).
Thermal shock is the only other extra load that can affect a burst disc or a manifold, but this type of stress will always happen at the moment of immersion.
The issue is are we adding extra isolation and complexity for an unrealistic risk? That is the question.
The Cons of the modern isolation manifold.
1. It adds mechanical complexity with a few more points of failure and bit more delicate manifold (IMO, this is a relatively minor issue)
2. It adds procedural complexity were the diver has to decide what to close first in the case of a failure.
I have not trained on this type of manifold so I am only going to write about incidents that I have read.
Item 1, seems like a minor issue, but I have inspected this type of manifold and some do seem to be somewhat more delicate, but nothing that I would be concerned.
Item 2 is where I have read more of the controversy about this manifolds. I have read of several incidents (but I don’t have direct confirmation) where the diver wasted a lot of air/gas fumbling with the isolation center valve first, before closing the outlet to a leaking regulator.
As I recall the worst case issues was when the leaking regulator was behind the divers head and he could not easily identify the leak source.
I recall reading and hearing of several similar incidents incident were an unnecessary amount of gas was lost due to fumbling with the wrong valve.
It seems that good procedures and training will mitigate this fumbling issue, but I find it troubling that I have heard or read of it from several sources. Again, no firsthand experience.
The manifold is not easy to reach and the question that needs to be answer is: does that extra valve that needs to be reached really add any reasonable protection.
On my readings I have also found others that will only accept independent doubles, no manifold, but that seems to be a smaller group of divers.
I have also recently read of some that believe the safest kit is a very large single cylinder with two outlets (Y or H valve).
Now the modern side mount has brought back the concept of independent doubles with the major advantage of easy reach of the valves.
As I mentioned, very little here is my opinion, a lot of it comes from different schools of thought that I have been reading about. I am not going to defend or argue about someone else’s opinion or point of view. The incidents that I have read about are also from many different sources and covers years so I am going by memory.
Risk management is all about calculating the amount of risk versus the consequences. A risk management table is easy to read, the hard thing is figuring out the probability of the risk and figuring out what are the real interim consequences (not just the potential final).
BTW, the mix-mount configuration I have been playing with does have a few extra (minor) points of failure, but it does add one more independent redundancy easily isolated gas supply.