High pressure pump for water storage

There are products on the market by Dankoff and Grundfos for pumping water at high elevations and targeted for solar. I have a friend with a need for such a pump and thought I'd take on the challenge. In his case there will be a mile of line and about 550 of elevation to deal with. To keep cost down it is desirable to use 1/2" line. As long as we stay below a gallon a minute there is no real issue here. From this calculator

Liquid Friction Pressure Loss

Pressure Loss (psi):  27.17   Head Loss (ft):  62.8
Line Number:   
Date:  7/7/2016
Nominal Pipe Size:  0.5
Pipe Schedule:  SCH 40 
Flow Rate (gpm):  1
Viscosity (cP):  1
Specific Gravity (water=1):  1
Temperature (F):  70
Pipe Roughness (ft):  0.000016
Actual Pipe ID (in.):  0.622
Fluid Velocity (ft/sec):  1.06
Reynolds Number:  5085
Flow Region:  Turbulent
Friction Factor:  0.038
Overall K:  3623.12
Piping Length (ft):   5000

About a 10% cost in energy over the cost to pump the water. And the pump that would just squeeze by from Dankoff is a model #1408. The specs are at this link. What is interesting is that at 560 feet and 1.67 gpm this pump demands 435 watts. A foot-pound of energy per minute is equivalent to 0.0226 watts.

560 ft. * 8.35 lbs. * .0226 = 106 watts.

That makes this pump ~40% efficient. Not great. I don't know if they just way under spec the pump or there is really that much friction. I don't know what style the pump is, they don't say. I'd guess rotary vane from the looks of it. If we want to keep the flow down to 1 gpm, then we would have to pulse width modulate the Dankoff pump. The loss goes up at about the square of the flow and one gpm meets our tolerable loss.

But what really hurts with this kind of efficiency is the added cost of solar panels. If our pump really turns out to be more than twice as efficient it will mean as much as a $1000 savings in required panels. It works out that way because there may only be some 25% utilization of the panels in the winter. A battery will be charged and would handle a majority of the demand.

Our Pump

Radius RF Tool

Yes, this is a PM treadmill motor and a pressure washer pump, belt driven. The dark fitting is a pressure relief valve that is set at 500 lbs/in2. That is equivalent to 1160 feet of head. I cranked it up until it was delivering a gallon per minute, pressure meter steady at 500 lbs/in2. The motor required 40 volts and loaded the source at 6.75 amps, 270 watts. These measurements were taken with a pure DC supply.

1160 ft. * 8.35 lbs. * .0226 = 218 watts.

80% efficient, that is very satisfying. And it ran like there was no load at all, very smooth and quiet. The transmission losses and other friction won't decline linearly with a lower head, so the efficiency won't be that good at 550 feet of head.

OK, so we can't drive this from solar panels and/or batteries. At these power levels 12 volts would be nice. So we need power conversion...

The Power Supply

Radius Tool in use

On the right is a salvaged 12 volt DC to 120 volt AC 1000 watt converter. I got it because it would shut down with just a few watts of load. The primary, (above the red outline), was fine putting out some 140 volts DC. It was the secondary that was faulting. I rewired the secondary so that the MOSFETs could pulse width modulate the motor. The circuit on the lower left of the inverter, inside the red outline, was the output driver and no longer part of the circuit. The two boards to the left are the new power and driver circuit for the output. The breadboard to the far left is supporting an MPS430F2012, the new brain for the output. With a microprocessor, the likes of acceleration, rate of flow, possible faults, etc., can all be easily dealt with. I've tested this with a load on the pump and it works great. I didn't take numbers yet to see what kind of efficiency I was getting, but soon to be done.

Expected MTBF and/or regular maintenance, unknown. But when you can replace the pump for less than $100, there may be no concern. Another acquaintance recently bought a Dankoff #1408. It was over $1200 in the door. And it didn't have a brain.

That is it for now. I'll keep posting as we progress. Everything I'm doing here will be open source.

Thanks, Dan.

Copyright 1996 - 2016, Highlands Electronics All Rights Reserved.
Page created: 7 jul 16   rev: 8
jul 16