I agree that in-tank is the best solution. But it isn't easy to achieve with an already-built tank. Also, a single pump in each tank has the significant downside of trapping fuel in the tank if the single pump or anything in its control path fails. So, multiple source tanks would mean multiple pump pairs, and to preserve redundancy, a pair of switches for each pair of pumps. See how nasty it gets?

The little turbine pumps are just tall enough that it's really tight in a -7/8 tank, so they might not fit at all in a -9 tank. And they are difficult to orient horizontally, because the filter sock (fuel pickup) attach on the input is a little 3/8" tapered stub, with a little post beside it that accepts a spring loaded lock ring, which retains the filter sock. I've never been able to find a filter sock that is set up to let the pump sit horizontally. I suppose I could have found a way to bond an adapter to the pump, but then field replacement would become an issue. Another complication for height is that the outlet is also a minimally barbed fitting, and requires an output hose that adds to the height. So, I moved back to the 393s.

One of my 'iterations' was a pair of 393s on a subframe that I could screw to the stiffeners in the bottom of the tank. (pics available, if you want them.) I have a wrecked RV-8 tank (same as my -7) that I cut open and used it as a pattern, so I had it easier than most in designing it. Plumbing is done, but I stopped short of mounting fuel proof electrical pass-throughs in the access panel.

The in-tank pumps, by themselves, won't cure the return line problem. Returnless systems with turbine pumps need PWM control to regulate the pressure. (According to Walbro, you shouldn't use PWM on the gerotor pumps because it fatigues the gears in the pump.) Cars have the PWM controller built into the car's computer. I found a guy several years ago who created a stand-alone PWM controller for his motorcycle injection system and was selling them, but when I started adding up all the changes, unproven control hardware, etc, I decided that at least for now, our conventional setups make more sense for me.

I do have a pair of aux tanks, but I elected to plumb engine supply from one of the main tanks, and use the stock 3 in/1 out fuel selector to pick which of the 3 other tanks gets transferred to the supply tank. This resulted in the lowest part count I could come up with. A pair of injection pumps, a pair of transfer pumps (for transfer redundancy), and one valve. The injection pumps serve as my main fuel shutoff. I wanted to completely eliminate selector valves, but that would have meant redundant pumps for all 3 source tanks, along with switches for control. So retaining the single selector valve was as simple as I could get.

Charlie

On 2/15/2019 7:06 PM, Todd Bartrim bartrim@gmail.com wrote: