Lots of good discussion...So Is there a pump that performs similar to how you guys are describing that doesn't have to be placed inside the well and can just replace the jet pump in my garage?  Is the fact that the jet pumps are 1/2hp and 3/4hp enough to size this magical higher efficiency pump appropriately w/o knowing the well specifics?  Or is this basically beyond the DIY approach of just replacing the pump in the garage and instead I need to call a Well guy out to evaluate the system and see if there is something more efficient which is down in the well?

1/4 of max flow at the same pressure (which is usually what people want). Very practical but very inefficient.

There is a link to non submersible variable speed pumps above.

For well specifics, you'll need to know the draw down level of the water table at maximum draw. If your water table drops below about 20 feet (measured below the level of the pumps, not the surface) you will either have to go submersible or stick with jet. A centrifugal pump at the surface can only "suck" so far. I also don't know anything about the efficiencies of a centrifugal designed for intake vacuum - might or might not be more efficient than your current setup.

If you decide to drop in a submersible, you'll still need to know your water table variation under pumping. There are 1/2 hp pumps designed to pump 1 gpm from 400 ft down, and there are 1/2 hp pumps designed to pump 50 gpm from 5 feet. "Artesian" implies water at the surface, but then you also mention the previous owner could over pump the well under certain circumstances.

If you know how far down your water is and the flow rate you need, you have enough information to size the pump. There is more flexibility with variable speed, but its not one-size-fits-all by any means.

JonR - where is the link you are refering to?

Blake- You are running at less then 1 amp at how many volts?

I am looking into "fountain" or "pond" pumps for a well with a water table of 5ft and less then 10gpm. Well depth is 40 ft and it is 50ft to brook. I am also sizing "booster" pumps. Most of these are very inexpensive and efficient. Slightly less so than the sqe perhaps but still at less than 140watts. One SQE pump costs about the same as 10 of these units and the surface install makes replacement/service simpler.

Any thoughts?

How about this option: a dewatering pump/ transfer pump, with high flow but only at lower head pressure and 690 watts of power consumption, 60hz, 1ph, 230v, 3 watts; rated for continuous duty, $400 w/shipping. Requires yearly maintenance too. The pump is placed down a 8" hole. On my project, the water tabel is very high; I plan on using a 6" Multi-quip trash pump to suck a 8" casing down; also with two discharge wells. $135 rental for 8 hours on the trash pump, plus parts.

1 amp at 240v - keep in mind my water table is usually 100 feet down and I'm pumping 7 GPM. (And, O.K. full disclosure -- today it's pulling 1.12 amps. Power factor is .95, so as we speak it's pulling 255 watts The water table dropped a little)

Fountain/pond/dewatering/transfer are basically the same pump architecture with varying degrees of quality/dependability/efficiency depending on price and brand - not all can produce suction head, so do your homework. Of course if you buy one for $150 and keep a spare that's easy to change out durability might not be critical. Most of the higher-end pumps I've seen in this category don't quite meet specs (too much flow/head) without throttling -- though that doesn't mean a good fit is not out there.

Booster pumps might be a very interesting option. Check to see if you can adjust the pressure down low enough to make them work efficiently, otherwise, you'll have to throttle which pretty much negates variable speed in terms of efficiency. Not all will work with suction head and I don't know if there are any efficiency penalties with those that can.

Trash pumps are notoriously inefficient - the impellers are designed to pass and grind solids, not necessarily to pump efficiently. I knew one guy off grid who quickly found his sewage pump was one of the biggest draws on his system. (His unpleasant solution was to clean a filter bag weekly and use a normal submersible)

If I'd had a near-surface water table (wouldn't that have been nice!) I would have gone with a Grundfos 22SQE-40. That means it's rated at 22 GPM at 40 feet of head. Slow it down and the math happens. It will pump 10 GPM at 8 feet of head using roughly 1/10 of it's rated power - that's 10% of 1/3 hp. (FYI with pumps, the math is a bit fuzzy, 1/3 hp with a 2X duty rating is really 2/3 hp...blah...blah...blah, lots of marketing to weed through to get to real numbers)

Unfortunately, the standard Grundfos CU301 controller doesn't make it obvious for you to set it up this way. Fortunately, a trip to Radio Shack and a $1.99 resister is an easy way to get a Grundfos to pump at whatever pressure you want.

Tell me more about the $1.99 resistor.

10K-Ohm 15 Turn Cermet Potentiometer/Trimmer $3.19 at your local Radio Shack.

O.K. - here's why and how it works. Grundfos uses a 4-20 mA control loop. The stock transducer sends a 4 mA signal to the controller at 0 PSI and a 20 mA signal at 120 PSI. Let's say you set the controller for 60 PSI - the controller is "looking" for a 12 mA signal from the transducer. If it's over 12 mA, it slows the pump down. If it's under 12 mA it speeds the pump up. Point being, the controller is only reading the control loop signal, not the actual pressure.

The resistor is a simple way of modifying the control loop signal. (The control loop is 24V) A 5K-Ohm resistor will have 4.8 mA of current through it at 24V. When wired in parallel with the transducer, you are effectively increasing the current in the loop. Since in our example the controller is still "looking" for a 12 mA signal, and the resistor current is 4.8 mA, 12mA - 4.8mA = 7.2 mA. In other words, the controller will now "see" 12 mA total when the transducer is sending a 7.2 mA signal, which corresponds to 24 psi. At 3.5K-ohm, the resistor would draw 6.8 mA and the transducer would send 5.2 mA, which corresponds to 9 PSI. The practical limit to this approach is probably 9 PSI, given that Grundfos needs to see a 7PSI drop before it will start the pump. You can do even crazier stuff (like pump into a vacuum) if you swap out the transducer, but it gets a bit more complicated.

The only downside is that if (when) the transducer eventually fails after many years of cycling, the controller won't know what to make of the constant signal from the resistor. Most likely, it would start the pump and run it at full speed until someone noticed. A fail-safe overpressure switch might be advised. Also, you're on your own with adjusting the small bladder tank required in the system to prevent oscillations - Grundfos only gives guidence down to 40 psi.

It will pump 10 GPM at 8 feet of head using roughly 1/10 of it's rated power

Where do the Grundfos manuals show this?

Grundfos literature doesn't tell you much about pump physics. They're trying to sell you a very good pump using very sketchy and somewhat misleading graphs. When I was setting up my system I had to be pretty dang sure that I was getting the math right, so I dug through some pretty obscure sources - including tables that Grundfos publishes for it's SQE-Flex line for solar applications. In that application, available power is the limiting factor, so the tables detail how much each of its pumps will pump at what head given a solar panel of given watts. Also, some (also obscure) technical notes on the efficiency of Grundfos PM motors. (Which is really the only reason Grundfos might be able to claim the efficiency prize - its permanent magnet motors maintain their efficiency over a wider RPM and load range than an induction motor).

Grundfos impellers follow the same affinity law as everyone else's. There is a rating number - in this case 22 GPM at 40 feet head - that is the design spec for this pump. It will pump most efficiently at this spec. Being variable speed, however, it will pump equally efficiently on a curve that follows the affinity law. At 10 GPM along this curve, the RPMs will be 45% of its rated speed. The head at this speed is 20% the rated head, or 8 feet. The power is 10% of the rated power. This is physics, all centrifugal pumps do this. This is not to be confused with the sizable efficiency drop-off you would see if you tried to pump 10 GPM at the full 40 feet of head - that's the part that Grundfos is trying to obscure with its graphs.

Why I said, "roughly" (and included the disclaimer about fuzzy math) is there's another player, namely the motor. "Pump power" is the wet head and it follows the affinity law above. "Total power" includes motor inefficiencies and I know that Grundfos plays pretty loose with their motor power "ratings" so I'm not sure exactly what the final result would be.

But here's the thing - with zero inefficiencies, it only takes 15 watts to move 10 gallons 8 feet straight up in one minute. Even if the pump/motor combination were only 20% efficient (I'm pretty sure Grundfos can beat that mark) the pump would draw 75 watts. This pump is rated at 500 watts, so at 75 watts it's drawing 15% of rated power. If it could eek out 30% efficiency (upper end of what is probably possible) it would be at 10% of rated power.

And for those whose eyes have glazed over -- Here's theory in practice: My Grundfos 10SQE-240 is rated at 7.8 amps. It's pulling 1.12 amps, 15% of rated power. See lesson above for more info :-)