A May 2006 update to the Build a Miniature Lempor Ejector Page.

The update notes on this page follow work done in the winter of 2005-06 to modify Romulus "Lesley's" draughting and run some tests to attempt to evaluate the results. Throughout this period, correspondence and collaboration took place between myself, Richard Stuart and Jos Koopmans.

Changing a blast nozzle "hot" during a stationary nozzle-testing run with the seven-degree diffuser. The hydraulic load pump is visible at bottom right. There are other photos of the test rig here.

The assortment of test nozzles. From left to right, a normal 3/8" one-hole; a 4-hole "blast cap"; 4x1/4" at 7 degrees; 4x1/4" at 5 degrees; the winning "Koopmans/Goodfellow" 5 degree 4-nozzle; and a 3/8", 7 degree 4-nozzle.

 Notes on Lempor Diffuser Design

L. D. Porta and David Wardale have published data that the diffuser angle should be less than or equal to 12 degrees and 10-10.5 degrees respectively. Reference to current diffuser design charts suggest that each diffuser will have a different efficiency depending upon the proportions of the cone used. It is my opinion that once the throat diameter and length have been determined, the exit diameter and therefore the angle used should be determined from a Cp (static pressure recovery coefficient) chart such as that shown below. The designer should aim for a non-dimensional length value and angle that intersect as closely as possible to the ideal Cp line. In the case of Lesley, that angle was 7 degrees for the second trial diffuser with a Cp of  slightly above .75.

The angle between multiple blast nozzles should be determined graphically after the diffuser dimensions are determined. The intent being that at both the diffuser choke and exit, the area is evenly filled by the steam jets. Once these have been drawn the correct distance apart, a section view will show the correct inter-nozzle angle to use. Click the diagram for a larger view.

This diagram also shows where the individual nozzles must be placed in relation to the mixing chamber end or "petticoat" flare. The nozzle centres must coincide with the drawn centrelines. The importance of getting the inter-nozzle angle correct cannot be overstressed. If the angle is too great the effectiveness of the ejector will be severely limited as the exhaust streams bounce off the mixing chamber sides and choke the system leaving no room for smokegas. The performance of my first nozzles fell below expectations for this reason.

Blast Nozzle Notes

Porta's Lempor Theory yields a value for a single-nozzle area. This value may be split up into as many smaller nozzles as may appeal to the designer, four is commonly used. It is a fact that multiple nozzles are better than one and this is so because of the mechanics of the interaction between the blast jet and the surrounding smokegas.

Smokebox vacuum is created when stationary or slow moving smokegas is accelerated and ejected from the chimney by the blast jets. Much research has been done on this subject and it is a fact that the blast jets entrain smokegas by the action of a turbulent boundary layer around each jet. It naturally follows that the more boundaries there are and the more turbulent they are, the more the vacuum will improve over that created by fewer jets with less turbulence, for a given amount of exhaust gas flow. This is fully explained in a steam locomotive context by Jos Koopmans in his new (2006) book "The Fire Burns Much Better".

Since the efficiency with which the exhaust steam jets entrain the smokegas is dependent upon the turbulent boundaries between them, Dr. Koopmans suggested that the addition of "Goodfellow Tips" to each nozzle should improve matters. Each corner of an object in a high-speed gas stream creates a vortex, or rotating turbulence, in the gas flow. The addition of four square tips to a nozzle creates 4 x 4 = 16 vortices. If the nozzle itself is square that adds another four, for a total in a four-nozzle system of eighty vortex generators. This should and does improve the entrainment of smokebox gasses by the blast jets. Below is a sketch of one of Lesley's nozzles and a photo of a trial nozzle fitted for comparison with a normal, round, non-Goodfellow nozzle array. Note that the tips are square in section with no streamlined "Vee" to the bottom side. Such a "V" shape is counter-productive as it removes one vortex-generating angle.

Initial comparative tests of these two nozzles suggested that the benefits of the square, Goodfellow nozzles might only be apparent at higher gas flow rates. A track test of the locomotive however, confirmed that the Goodfellow nozzle did create a higher vacuum, perhaps a quarter of an inch of water additional at 10 km/hr.

Vacuum levels created by all the test nozzles was dependent on gas flow, i.e. the faster the engine ran, the more exhaust gas flow was available and the higher was the draught. Correspondingly, the higher the draught the higher the blast backpressure. At a "normal" track maximum speed of about 8-10 km/hr the smokebox vacuum was under two inches of water and the backpressure below 1/2 psig. With the two nozzles shown above, six inches of water vacuum attained still only created about three psig backpressure, far below the 12-18 psig created by the single 3/8" diameter nozzle previously fitted.

An amusing side effect of draught created by a smokebox vacuum of five inches of water or more was that Lesley's brick arch would be "sucked" off it's supports and flap up against the end of the fire tubes effectively blanking them off and reducing the grate draught considerably. This effect was the cause of much head scratching before the reason was noticed.

Final update May 17th 2006.

If the success or failure of a locomotive draughting device is judged by it's ability to provide enough draught to keep the safeties lifted nearly 100% of the time when running, while at the same time showing near zero exhaust backpressure to the pistons, then this one is a success. Today's test run (at Golden Horseshoe Live Steamers track in Hamilton, Ontario) of Romulus "Lesley" was like that. If the engine was running, the boiler pressure rose to maximum and stayed there. The slightest crack of the blower steam valve was sufficient to brighten the fire and pop the safety valves within a moment or two even after a stop of several minutes. The exhaust backpressure gauge never rose off zero psi unless I deliberately slipped the engine when it would go to 3 psi. For the record, the most successful configuration was the Koopmans/Goodfellow 4-nozzle, a mixing chamber of 1-1/2" diameter and a seven degree included angle diffuser.

That is the end of these notes, I hope they were useful, or at least, amusing. Jos Koopmans book "The Fire Burns Much Better" is due to arrive here this week, and who knows, it may change everything. I will let Jos have the last word on L. D. Porta's Lempor Exhaust Theory:

"The Lempor calculation is garbage in, garbage out. Using unrealistic numbers does provide an answer which I would not prefer to use on a locomotive. However, starting with proper numbers, the answers do provide a basis upon which to work." - Dr. J. J. G. Koopmans, May 2006.

******