I can't think of any ways to modulate the output of a feed pump without horsepower losses besides a few. Varying the stroke of the pump like an axial piston hydraulic motor or using a linkage between the crank throw and the piston rod that you can vary like a stephenson linkage. Varying the shaft speed of the pump, perhaps with a manually controlled cvt setup or electric drive with a vfd speed control. Pretty much everything else requires a fancier pump or wastes energy.

Yep, and even the various methods of varying the pump stroke complicate things and add friction losses. In some cases possibly[?] adding more horsepower loss than they'd save. It is hard to beat the simplicity of the old steam cars' main-engine-driven feed water pump and bypass control Ron's reported "throttled-intake" approach, or maybe it could be called "induced cavitation", might have some promise; building and testing would be needed to see if it is OK with a car-sized feedwater pump.

Cavitation usually adds what Ken says the folks at GM call "noise/vibration/harshness" [NVH] to any hydrodynamic system. Look at marine propellers/"screws". Methods of designing/machining those to minimize cavitation/noise for high-performance naval stealth purposes, especially in submarines, were once -- maybe still are -- "extremely top secret". NVH is also often an indicator of power loss -- power producing effects other than useful propulsion, such as vibrating the ambient liquid or gaseous (water or air) environment, and/or the surrounding vehicle structure.

Vibration can be damped with flexible elastomeric engine/transmission mounts, but those fatigue and harden in time. I remember when I had the transmission rebuilt in my old air-cooled (1969) VW Beetle, over 20 years ago now. They had to replace the "motor mounts", which actually serve as both engine and transmission mounts in those peculiar vehicles. Immediately afterwards, and for a few years, the old car suddenly ran incredibly smooth and silent, fantastic gigantic improvement over the old/hard "motor mounts" which I had gotten used to. It was like a whole different car, like a new car. Fun to drive before, and an absolute joy to drive after. After a few years, though, the new mounts hardened up, and it was back to "Max-NVH" running. When I get the Bug running again, I am going to get the mounts replaced again, and maybe re-do it every 5 years or so. Though some shops might balk at doing it in a car that "runs fine". Same way that the guys who re-did the transmission & mounts tried to talk me out of replacing the old shot/crumbled-away rubber air seals around the engine. A big pain in the rump-roast repair job, btw; I don't blame the mechanics for not wanting to do it until I waved some extra cash in their faces. But essential to keep the air-cooled VW pulling cool air through the engine compartment, instead of re-circulating hot air from under the engine, which causes all sorts of problems, like power loss and overheating in extreme conditions.

If I ever build/test and discover some great improvements from advanced high-expansion steam engines -- which, despite my apparent rep as "Mr. Die-Hard Old-Steam Tech", I am not ruling out, BTW -- flexible/NVH-damping motor/drive mounts will probably be needed. If they can damp out the NVH in an air-cooled Old VW, then I figure they can damp out darn near anything. My idea is to somehow design the mounts so that they can be quickly, easily, and inexpensively replaced every few years. Preferably by a typical back-yard mechanic, or any local/independent repair shop. Carmakers today don't do that. Yet. I suspect -- no evidence yet -- that they want people to replace the whole car when the motor/drive mounts fatigue/harden. Consumers test-drive a new car, are impressed with the smoother quieter running of the new thing, and don't remember that their old thing ran just as nice when it was new. They just chalk it up to "improved new car technology". Kinda like the old joke about "I wear a magic charm to keep wild elephants from attacking me, and it must work because I've never been attacked by a wild elephant". The idea of replacing NVH-damping motor/drive mounts instead of the whole vehicle never occurs to them.

OK, that's kinda off-topic, but IMO something well worth considering by high-expansion "IC-style steam engines" folks. I may yet join their ranks. Use that old flex-mount gas car trick to smooth it out, but design it for quick/cheap/easy motor mount swaps.

It sure does work. Delving into the range of "really old guy" now, I have to admit that I learned how to drive in a near-new 1973 Ford Grand Torino Station Wagon [think "Brady Bunch" car], with the [long story] Windsor 400cid V8 engine. The squooshy motor/suspension mounts, 50-gallon-drum-sized muffler, and several tons of "Detroit Muffle Magician" soundproofing made for a smoovey-woovey driving experience akin to piloting a giant wet sponge on a roadway paved with puffy-fluffy clouds. With a 3-speed Old-MoTown Hydra-Slush Auto-trans and something like 6-Liter engine [haven't done the math; my current pickup truck's MoPar 5.7L Hemi V8 is like 347cid if I recall correctly; right in the long-standing economy/acceleration sweet spot for a naturally-aspirated Otto-engined 4000-5000-lb vehicle -- VVT, variable-displacement, direct-injection, microprocessor-controlled 8/10/12-speed auto-trans and other effectively gubmint-mandated "0.1% improvement at any cost" IC engine tech be damned].

High-expansion/compression-ratio steam car engines have been advocated, and criticized for their "old-diesel-like" noise and shake. OK, but I remember an article in one car magazine [Road & Track or Car & Driver], late 1970s/early-1980s, entitled "The Diesel From Space". This was about an experimental diesel car from that era -- anybody remember how those ran? -- which was miraculously and massively muffled and sound/vibration-encapsulated/flex-mounted by a European carmaker to run with Otto-engine-beating smoothness and quietness. The reviewer was astonished by the bizarrely "impossible" combination of Diesel efficiency and Detroit-luxury-car-beating smoothness and quietness. Alas, it never saw production, to my knowledge, but it shows what is possible.

"Muffle magician"-packaged high-expansion/residual-exhaust-recompression steam car engines/drivetrains are not to my current taste, but who knows where any steam car tech will lead, if actually designed, built, road-tested, and -- dare I even suggest the unthinkable possibility? -- marketed.

Who knows? Not me.

Peter

Hydraulic pumps need to pump oil at super high pressures with extremely fast varying flow rates from 0 to full and back. Industry standards for the mechanisms of these pumps exist because the designs that work are highly reliable. There are several constraints to hydraulic pumps that are also applicable to our needs. First, you can't control the speed of the drive motor, since you can't turn on an off electric motors repeatedly. This is also true with a direct drive feed water pump. Second, you need to be able to deliver high pressure with the flow rate varying from 0 to full, without wasting horsepower, which is also what we want. Third, you have to maximize reliability and ease of maintenance, since if the hydraulics are down, lost time is lost money. There are several different designs that have become industry standard for common piston based pumps.

However, I would still recommend the standard bypass valve. When the bypass valve is open, the only power lost is from pump moving parts friction, and a little bit of pressure for moving the water back through the return line. You could always have a small cylinder and a large cylinder.

That's what I'm thinking, plain pump and bypass valve for the first build. For future variable-stroke pump experiments, maybe the Stephenson link mechanism, with the expansion-tube "Feed Water Automatic" moving the variable link via a step-up lever and perhaps a pull-rod or cable, instead of open/closing a bypass valve. Just a preliminary brainstorm there; that might turn out to be impractical or more R&D work that it's worth. Things that look or sound simple at first often end up much trickier, and very different from the initial concept, by the time they reach the "buildable and workable-looking" stage. At least that has been my experience, including with the feed-water & fuel pumps I recently designed/blueprinted, and with the cylinder oil pump that's now fully designed but currently driving me nuts with the component shop drawings. I'm about half way through. Blueprinting all the little parts, assigning part numbers and names, looking up materials, parts, tools & sources and ordering details, planning fabrication procedures, etc, and getting all this data into complete drawing form, is tedious. I enjoy inventing/designing and machining/assembling stuff, but the "million micro-details" shop-drawings process in between those two stages is a seemingly-endless boring brain-burn. I suspect that this kills most steam ideas before they ever reach the workshop. In many cases, this probably also answers the question "why did they stick with that design when they could have used X?". Paperwork, paperwork. OK, enough "venting". And procrastinating.

I think the way to do a feed pump for nowadays logic control to omit the need for a linear actuator is:

Use a three cylinder feedpump with each cylinder having a different length of stroke and use three solenoid valves for bypasses that manifold in to a return to the day tank. This would provide full control with simple logic signal over a wide range of constant output, redundancy as it would provide three individual pumps i.e low medium high and would provide 7 levels of output with combinations of binary commands to fine tune the output.

X = bypass open

1 X X
1 2 X
1 2 3
1 X 3
X X 3
X 2 3
X 2 X
X X X

Each combination would correlate to a value based on demand. If one cylinder/valve fouls, it would issue a higher command and basically ignore the dead cylinder. It would be very smooth running too with multiple shorter strokes.

-Ron

Way back when, one of the Flinn brothers built a feed pump with two rams and a single valve chest. The rams were connected by bevel differential gearing; this permitted both rams to be driven simultaneously while the phase angle between the two could be altered on the fly. If both rams ran in unison, they both drew water in and expelled it simultaneously, giving you full delivery. If the rams were set to run 180 degrees apart, then one ram discharged as the other drew suction and their actions cancelled out .. no pump delivery. Intermediate positions gave infinitely variable displacement.

The same patent specified an alternative method whereby the driving eccentrics were keyed to opposing helical splines on the drive shaft, sliding the shaft from end-to-end altered the phase relationship.

Hi Ken,

The issue with most of these mechanical designs that afford gradual control for use with modern day logic is that they require some sort actuator with positional loop controls, whether it be linear or some sort of rotary i.e. stepper or servo motor, the control scheme and design is much more complicated. Using three solenoid valves requires only three high/low logic signals from the control, there are no moving mechanisms to control the graduated valving and it is handled in the code. Finely tuning the feedwater input is by far more efficient than simple on or off feedwater control.

-Ron

P.S. I was accepted for (er the Locomobile was for Greenfield Village Old Car Festival at Dearborn, Mi Sept 8 and 9, hope to see you again this year. Anyone else interested in attending, come on out, there always a few steamers there. Oldest running antique car show in the US with about a thousand entries, second only to London to Brighton worldwide.

Hi Ron,

I'll certainly be at the show, am just as certainly looking forward to seeing you there!

Just thinking, if we want electronic feed water control, how about attacking it right at the pump inlet valve? We could replace the inlet check valve with a check valve having a solenoid function that holds it open....in other words, turn it into an electrically activated unloader valve. Now, you could control feed rate just like we control fuel injection, by varying the duty cycle of the solenoid. Solenoid off, pump delivers full rated output. Solenoid on, pump delivers nothing. Solenoid on 50% of the time, the pump delivers half its rated output. Unloaders are nice because the fluid or gas is drawn in and then pushed right back out without significant change in pressure ... there is some power consumed, but not too very much. Heck, you can do this even without electronics, a lot of monotubes, including the Doble, had burners that operated on and off with the intervals determining total output. Maybe Abner would have done well to have tossed out his elaborate thermocouple system and replaced it with a single probe and a variable intervalometer.

Regards,

Ken

Oh, yeah, that's a really good way to do it. I had a similar setup in the programming of my first boiler. It was programmed to only check the water level sensor every 15 seconds, and if it was off, it would turn on the pump until the sensor turned back on. This (usually) kept the water level quite steady. Your average industrial solenoid valve often spends its life actuating every 5 seconds anyways.

"Just thinking, if we want electronic feed water control, how about attacking it right at the pump inlet valve? We could replace the inlet check valve with a check valve having a solenoid function that holds it open....in other words, turn it into an electrically activated unloader valve. "

Ken

This is no bad idea. But there is one thing which made me to deny it. When the time of delaying in delivering of water is too long, the portion which is pounding this way may to boil.... How do You think?

With regards, Serge.

Hi Serge,

I don't see that happening. You are probably thinking along the lines of the heating that goes on in a hydraulic power system. In that case the fluid is being pressurized and then forced to move against massive resistance to perform work; this causes the fluid to heat up.
...
The pump unloader is on the inlet valve, meaning that there is no pressure being formed. Pump inlet valves must have very little restriction so that they can easily draw in fluid...there is very little restriction to the flow. Without restriction there is little energy being applied to the water … enough to overcome inertia and friction. Since little energy is applied, there is little energy that can be transformed into heat. If you thrashed the water back and forth long enough, and fast enough, you could probably produce enough heat to boil the water. Long before that occurs, the engine will have used enough steam to require more water being added to the boiler … drawing in cool water from the feed tank.

Honestly, I can't see this being an issue in real world operation.

Regards,

Ken

Ken,

That is a good idea, essentially throttling the check valve to the pump. Instead of a throttle before the check valve, what you and Zimirken are suggesting, changes the volume of the pump based on demand. Yes, stagnant feedwater is not an issue, steam boilers use lots of water.

-Ron

Lots of interesting ideas here!

Theoretically, almost any kind of plunger pump could be set up to operate "hit and miss". Not to be confused with a rattlesnake crawling into a baseball catcher's glove, which is a matter of "mitt and hiss". This -- not the rattlesnake thing, but hit and miss pump control -- could be done electronically or mechanically, several different ways for each come to mind. Maybe just a set of cam wheels which turn at 1/10 the speed of the pump, with cam lobes holding inlet valves open. Depending on which cam is auto-selected, the inlet valves could be held open -- "miss" for, say, 9 strokes of the plungers, for every stroke -- "hit" -- where they are allowed to close. Or all the strokes "hit" for full flow, or 5 out of 10 strokes for half-flow, etc.. The cams could open/close switches to solenoid "valve holders", for electromechanical control, a la old-school washing-machine control boxes, or they could hold open the inlet valves directly for pure-mechanical operation. Or solid-state microprocessor control could be used, with an electronic revolution [or plunger stroke] counter and software actuating the "hold-open" solenoids, probably via relays.

Any of those approaches could be applied to darn near any plunger pump which is otherwise suitable for steam car use, anything from an off the shelf pump to a modified Stanley pump, or anything in between.

The horsepower waste with bypass-valve feedwater controls, like Stanley, Ofeldt, etc, is when the bypass is partly open [eg, expansion tube only partly filled with water, & part with steam], and part of the feedwater is being leaked from the supply line at boiler pressure. All that leaked water is being pumped up to delivery pressure and then dropped to zero psig; the horsepower put into pressurizing the leaked water is [mostly] wasted by converting the pump energy into water heat. No horsepower waste when the bypass is fully closed or fully open. Though with an economizer, having a bypass full open while burner is firing, which can happen with large-"reserve" boilers, could result in the economizer briefly heating up enough to reduce heat exchange -- more heat loss out the stack/exhaust-gas flue.

The pump hp loss could also be eliminated by having a bypass valve that "clicks" between full-open and fully-closed. The bypass would never be partly open. What I am wondering is whether that on/off bypass control, or the on/off "hit and miss" inlet-check control, could give rough running to the vehicle. Or would the vehicle inertia etc smooth out the variations of net torque delivery to the driven wheels? Not even an issue with the wastey classic variable-opening bypass valve, which runs "smoove" at all water delivery rates -- with crazy-simple control mechanisms to boot.
.
Also not an issue with an auxiliary engine or motor to run the feedwater pump. But that would probably need a governor or motor controller of some sort. Which then raises the issue: why not just control the pump's auxiliary engine or motor at continuously variable speed, and forget about bypasses and suction-check-valve deactivation devices?

On YouTube tonight I watched a half-hour episode of Jay Leno's Garage which was entirely about his Doble roadster, how it works, how to operate it, etc.. At one point, cruising smoothly and silently down the road with massive torque on tap and easily passing modern vehicles, he says something to the effect of "imagine what could be done with steam cars today, with modern computer controls".

By way of update, the previously-mentioned blueprint hassle for my oil pump has been finished, and there are only a few details of the final plunger drive & sealed splash-lube enclosure left to work out. The oil pump unit is not as compact or lightweight as I'd like, but inexpensive and very easy to build, with minimal machine tooling & fabrication skills. Water, fuel, and oil pumps in this design are entirely separate units, distributed to optimum locations around the vehicle. Cylinder oil pump is a dual-pump unit, with one pump to meter oil into the steam line for engine and the other to "skim" cylinder oil off the water surface in the water tank, sending it to a heated cylinder oil tank for dehydrating and recycling oil [& water] back into the system. No separate exhaust-steam/oil separator [which still sends a trace of oil to condenser/tank, and needs periodic draining] or flushing cylinder oil out of water tank, or periodic cylinder oil tank refills, as in old-time condensing steam cars; in this system cylinder oil is continuously removed and recycled on-board the vehicle. The trick is a simple system of internal tank baffles and a special ball-jointed surface float which keeps the oil suction inlet just below the floating oil layer, even with water/oil sloshing around in there on the road. Looks worth a try at least.

I have specified 6061 aluminum alloy billet for lots of the reciprocating parts in the pump drive system. Strong and cost-effective as cold-rolled mild steel, good machinability & corrosion resistance, and 1/3 the weight. On "Battlebots" [Science Channel & Discovery Channel; maker/tech-nerd ultraviolence on basic cable] the other night, one of the full-body spinner fighting robots [Capt. Shredderator?] got its spinning dome knocked off & thrown clear across the battlebox/arena, and the driver mentioned in a post-fight interview that the reason was an out-shopped cast-aluminum mount/drive shaft. He said that he will rebuild it with a shaft machined in-shop from 6061 aluminum billet. Sounds like tough stuff!

Peter



Edited 1 time(s). Last edit at 07/14/2018 01:30AM by Peter Brow.

Here's that Doble episode of "Jay Leno's Garage" which I mentioned:

[www.youtube.com]

Under half hour to watch; well worth it for steam car fans.

Peter

Hi Ken

You told me a good news! I returning to this design in my prototype water pump "blueprints".

Serge



Edited 2 time(s). Last edit at 07/14/2018 02:22AM by russt.

Well, if you want a pump that doesn't waste much power when you throttle the outlet, I can think of a couple of methods.

Edward C. Newcomb, the guy who invented the flow motor (the patent dates suggest Rolling White stole the idea) had such a pump. His valves were located on the cylinder walls and not in a cylinder head. The cylinder head had a second piston opposite that of the one connected to the engine. This piston was held down by a very strong spring under high compression. When the pump pressure rose above the spring setpoint, the spring gave way and started to allow its piston to move. As the piston moved, it absorbed the water that the driving piston should have displaced. Of course, this still absorbed energy. The kicker, however, is that as the driving piston went downwards into its suction stroke, the spring returned back to position and its piston pushed water against the driving piston with great force. This would be transmitted through the driving piston to the connecting rod and crankshaft. Thus, most of the "wasted" energy was restored to the engine.

My own theory is that this could be accomplished without a spring by keeping the dummy piston but by applying boiler water pressure above it. The boiler water would only allow the dummy piston to move when the pump pressure exceeded boiler (we'd use differential diameters to allow the pump to exceed boiler pressure a bit, first). This method works because there is steam above the boiler water and it's compressible …. like a spring. If you wanted, you could apply steam directly above the dummy piston. This makes for better action but increases odds of leakage a bit.

An alternative method of achieving all this without a dummy piston is a device I developed 35 years ago to accurately mill compound bow handles without CNC machinery. I call it the split rod. Basically, you cut a center section out of a connecting rod and redesign it so that the rod elements can be lengthened by sliding. Then you stick a compression spring in there. As long as the pressure on the rod remains below the spring preload, it works just like a normal rod. Otherwise, the spring compresses and returns throughout the stroke … absorbing and releasing energy. The same thing can happen with a normal connecting rod connected to a lever that in turn drives the piston through a second rod. The lever fulcrum can be attached to a preloaded spring and …. you get the point...

Ken, you're a genius! I like your "split rod" idea. Here's a handy place to shop for the springs:

[www.mcmaster.com]

They also carry "die spring stock", 3 foot long springs which can be cut to whatever length is needed. Use coolant when cutting, to avoid cutting-heat changing the temper of the spring. A professional tile-cutting tool would be ideal; giving cool cutting and nice flat spring ends.

Thinking about this, there are a number of variables which determine how much "water throttling" an automatic feed water bypass control would do. In some cases, something like the Stanley or Ofeldt system, with gradual-opening expansion-tube bypass and more-stored-energy water-level boiler, could operate pretty much full on/off, with "throttling" of bypassed water only occurring for a few pump strokes, depending on vehicle speed and engine torque. But with a very light monotube steam generator, controls would have to be set up for continuously variable water flow. Heavier monotubes, with more hysteresis, like Dobles, could operate pumps on/off, with the heat stored in the tube stack giving a more constant steam temperature. If fast-on/off bypass valve operation resulted in car slowing down slightly when the water started feeding to the steam generator, then the driver would unconsciously open the throttle a hair more. Kinda like driving an automatic transmission IC car; the driver modulates the throttle to keep the desired speed when it shifts. But in the real world, you're always modulating the throttle anyway to compensate for changes in the slope/grade or roughness of the road -- or just leaving the throttle alone and ignoring slight vehicle speed changes. Depending on the vehicle and propulsion system, the difference between a click on/off and gradual-open/shut bypass control might be like the difference between old-school "firm shift" Lincoln and "soft shift" Cadillac automatic transmissions. The "click" bypass might be more perceptible to the driver, and maybe to passengers, like the Lincoln.

With all the variables involved, I think it boils down to test results. I plan to instrument my steam car to see first-hand how much on/off and part-open feedwater bypass valve opening actually occurs in this specific vehicle while driving on the road. If the bypass valve runs part-open most or all of the time, or even a significant percentage of the time, then something like your split-rod power-saver might be a good idea. But if it goes from full open to full closed relatively quickly, with only a small percentage of time spent at part-open, then the small horsepower savings might not be worth the extra complication. Then again, it is a very simple modification...

I'm thinking of fitting something like a pressure gauge on the water delivery line, with a pinhole aperture to the gauge line to avoid the gauge needle bouncing all over the place & wearing out the gauge like Ron described, and some kind of flow meter in the bypass line, to see how much water is being bypassed back to the water tank. High pressure in the feedwater line and no bypass flow equals full feed; no horsepower waste other than usual fluid friction through valves, passages, and lines. High pressure plus bypass flow indicates "water throttling" and horsepower waste. Low pressure in feed line equals bypass full open, minimal HP waste.

With thermal-expansion-tube feedwater bypass controls, one factor is the angle of the expansion tube. The more horizontal the tube, the faster it drains and refills, giving faster open/shut times. The more angled the tube is, relative to horizontal, the more time it spends part-full and with bypass valve partly open. I have seen these expansion tubes installed at various different angles. The slope/grade of the road, tube orientation, plumbing details, pump size, etc, also affect how the feedwater control behaves while driving.

Peter



Edited 1 time(s). Last edit at 07/14/2018 06:43PM by Peter Brow.

I just checked out the "Corrosion Resistant Die Springs" at the McMaster link I posted above. Neat stuff! Those would be a good way to damp out any harshness in all sorts of machinery, including pump drives. Might also be good for Ken's "split rod" mechanism. Sort of a high-tech space-age alternative to old-school metal springs. Bit pricey though, and I'm not sure about durability. I will be keeping these in mind for experiments...

Belville washers make a more compact spring in applications like this. You can stack them in series or in parallel to get the force and travel you need. I used them to lift a heavy turret tool post for a large NC lathe decades ago. It needed to disengage from a Curvic coupling to rotate.

Lohring Miller

Belleville washers are cool! I have a faint memory of designing them into some component, I think in the control system, many years ago. I can not remember if it was in a design that is still on my "plan to build" list; have to dig up the shop drawings. It was in something that I won't be building for a while anyway, unless my car build suddenly kicks into "ludicrous speed" mode -- which is not impossible. I spend huge amounts of time designing, drawing, fabrication planning, etc, then for some reason expect that the actual build will take a long time too, then I'm surprised that the finished component takes almost no time to build. Forgetting that I did like 95%-plus of the figure-out work [the hard part] ahead of time, at the drawing board instead of in the workshop. Including figuring out how to build it with minimum time/labor/tooling/cost. Anyway, yep, Bellevilles are tough as nails and great for places where you need a lot of spring in a small space. They are even available in Inconel, for high temperature applications.

Lots of really neat stuff and amazing possibilities out there. Like that laser plasma coating that Jay Leno used in his Doble fire box liner; see video link above. He said that you could put a piece of sheet metal coated with that stuff on the palm of your hand, heat the other side with a welding torch, and the side facing your hand was barely warm to the touch.

Peter

All files from this thread

File Name	File Size		Posted by	Date
test pump.jpg	129.4 KB	open | download	IronChief	06/18/2018	Read message