To begin the story at a time when anticipated results have eluded me seems inappropriate but modifications following a rethink have produced good results. The steam generator tubing had suffered severe temperature variations & was distorted, furthermore certain sections have been hotter than the material will tolerate & were in the process of burning out. That is now in the past.
The project was conceived because it presented new challenges, having constructed numerous conventional & having owned two Stanley steam cars. (See opposite) a new concept had to be explored.
This was not to be my first experience of once through boilers, many years ago as a power plant engineer I operated a number of these but they were in the gas circuit of a 1000MWt Magnox nuclear reactor. As a home project my ambitions were subtly more modest & the control system had to be less complex.
I considered a plant for a steamboat but this had already been successfully demonstrated by Wally Mounster in Tasmania & his achievement with the added complexity of using solid fuel. Another friend from the Antipodes has had a record-breaking steamboat. Rod Muller is currently re-vamping his Blackcrow project for another water speed record attempt. He has demonstrated to my surprise how much steam power it is possible to generate in very small bore tube.
I chose to pursue my interest with a road going vehicle, if I could cope with traffic lights, hills & other traffic problems I would consider the job done.
A Reliant was eventually selected as a donor vehicle, (wouldn’t Philip Hopes have been pleased) no not the 3 wheel Plastic pig of Only Fools & Horses fame but a less well known 4 wheel utility vehicle. The attributes of 4 wheel Reliant cars are lightweight, a galvanized chassis, a rot proof body & registration document.
An estate version was of initial interest but discounted in favour of a pick up utility vehicle known as the FOX. Intrinsically safer than a totally enclosed vehicle where products of combustion may be dangerous.
The performance of my Stanley Ex was limited by my instincts for self-preservation, most would consider the limit to a boiler or engine but its period brakes steering & suspension restricted its potential for safe high speeds. The Reliant would hopefully surpass these limitations.
The work content was much greater than envisaged although the Reliant is easy to work on being of a simple & straight forward construction. The end result was a pleasing reward. Photo 1 shows the chassis following extensive cleaning, renewal of steering joints, wheel bearings, brake pipes, cylinders, ect.ect.
Once removed from the body it was possible to roll the body around in the garden to remove 20 years dirt accumulation with a pressure washer. The grass was a funny color for a few weeks!! See photo 2.
Many monocoque vehicles were discounted; the envisaged problems of steel shell body rot were not for me, the fibre glass option would be a safe bet, NOT SO fibre glass can suffer osmosis damage. The Fox had been left in a field storing hay & other food for horses when I rescued it. Rubbing it down produced small specks where water was coming out of micro pores in the glass matrix. I had it in a warm dry garage for 18 months before the beads of water ceased to appear but this left a surface of micro lunar appearance. This is one of the reasons that the surface finish is not exhibition standard; other reasons are idleness & the low value of the finished vehicle.
In a previous project (a 5/8 full sized steam wagon) I experienced tremendous flexibility conferred by incorporating a 4 speed & reverse automotive gearbox in its drive train. Good performance on a rally field & road being the benefits, its swan song being a holiday in the Isle of Man where the roads of Snaefell were consumed with ease.
The success of earlier projects resolved that I re-use the gearbox and axle of the Fox bringing with it the added benefit of simplifying the engine construction, which in this application does not need reverse. This decision does not come without a sting in the tail; the rear axle conspires to make life difficult because it has a 4:1 reduction. It should be remembered that the transmission is designed for an engine that will go up to 5000 revs /min where as a simple steam engine will not be happy above 1000 revs/min. furthermore a gearbox designed for use with internal combustion engines will not tolerate the torque produced a steam engine if directly coupled.
To overcome this difficulty I grafted the final drive gear from a Vauxhall transmission onto the engine crankshaft & the mating pinion to a layshaft. See photo 3 where the wheel is mounted on its hub & a keyway is being cut. It is necessary to select components where the wheel and pinion are separate items & not part of a complex shaft. This arrangement provides a 3.9 increase in speed to the gearbox input shaft & identical reduction in torque. I chose the gearing so that the engine would make about 850 revs/min at a vehicle speed of 60mph. photo 4 shows the splines being cut in the pinion shaft, this being the shaft that provides an input to the gearbox.
It was necessary to modify the gearbox, I no longer needed the bulky bell, housing & its input shaft required a revised bearing to provide additional support having been designed to locate at the rear of the crankshaft. It would now be driven via a short propeller shaft. See photos 5 & 6
The engine size of 2½” bore x 3½” stroke double acting twin was chosen because of its size successfully used in vehicles pf similar weight e.g. Locomobile cars & the many Likamobile replicas. I chose to make it with piston valves to enable a much higher pressure to be used if necessary. Subsequently experience has shown that side valves would have been adequate because higher pressure is not required & performance, at times, is now more than I can manage. An early set back occurred when warming through to remove condensate accumulation in the cylinders, the actions of a poor operator resulted in a broken piston. I have steam operated drain valves that also act as hydraulic relief valves but theses have a relatively small bore. Slide valves would possibly have been capable if relieving the hydraulic lock. The operator has now learned his lesson & 1 now warm through with first gear engaged so that the engine is restrained.
The cylinders were fabricated from mild steel using silver solder, see photo 7, The assembly was finished bored see photo 8 & the ends faced on a mandrel see photo 9 using a Harrison M300 lathe, finally iron liners were inserted in the bores. The steam ports were designed to be as short & direct as possible to prevent excessive clearance volume & simplify fabrication. The additional benefit being that this resulted in a useful steam reservoir between the heads of the piston valve. See photo 10.
Cast iron rings were made from spheriodal iron & ground to thickness using a tool & cutter grinder & a magnetic chuck on the surface plate of my lathe. See photo 11.
The cylinders were blanked & hydraulically tested to 40 bar. See photo 12.
The cranks are overhung in Stanley fashion & were turned from solid En8 bar; the intermittent cut shook the house!! This was a lot of work & produced a sigh of relief when the final operation of keyway cutting using a shaping machine was completed. See photo 13.
The main bearings mount on a boss on the inside of the crankwebs with the final drive gear mounted between, in Stanley fashion. The whole assembly is then pressed onto a keyed shaft made from En24T, an operation that could result in a disaster if the fit was too tight & galling resulted before it was pressed home. My disaster was a small one, a trial with scrap material proved the tolerances for an adequate press fit but I did not sufficiently remove the sharp corner from one end of the shaft & it sheared a sliver from within the web. I knew it would not press apart with my 10 ton capability so I decided to TIG weld the end to give belt & braces securely.
The choice of deign for a steam generator was not easy. So much information is available about good ideas & “what I would do if I had time”. Eventually I adopted the principles published by Peter Barrett as guidance incorporating my own ideas. Photo 16 shows the final coils assembly.
I contrived the following specification:
1. Use the same sized tube throughout to simplify & minimize the number of formers required.
2. Use a relatively thick wall tube to produce a thermal reservoir & make welding easier.
3. To avoid coil welds (I could reasonably access the centre space & coil periphery for TIG welding).
4. To minimize wasted material a pancake or individual coil would be 6 metres in length. (I.e. as supplied).
5. The hot end would be 2 coils & made from grade 316 stainless steel. The lower 10 off coils would be from carbon steel.
6. The free gas flow would be 50% of the tube diameter (I.e. ½” o.d. tube with ¼” gas space) & secured using laser cut combs.
7. The physically disposed top coil facing the burner would be the penultimate coil in the steam flow.
8. A normaliser (More commonly know in the UK as a desuperheater) would be introduced between the carbon steel & stainless steel coils.
9. The normaliser would be made in carbon steel in recognition of the propensity of stainless steel for stress corrosion due to chlorides.
10. The material transition would be effected using Ermeto compression fittings. My power plant experience material transition welds deemed they were best avoided if possible.
11. Top firing with water inlet at the bottom was chosen to avoid the fabrication complexity required by incorporating the anti drain loops that are a feature of the White steam generator if top fed and bottom fired.
The outlet steam conditions were initially chosen to be 500lbsf/ in² & 365ºC the former in the sure knowledge that this pressure would be sufficient with the chosen cylinder size & the flexibility conferred by a 4 speed gearbox. The latter being the value tolerated by compound cylinder oils.
I had already decided to use a commercial pressure atomizing burner, what could be more reliable than one used in millions of homes for central heating boilers. This meant that a 240v a.c supply was required which has a safety downside but the additional benefits of this choice were that many other components were more readily available including the all important digital temperature controller. My ideas for electrical safety were to use RCD protection but inverters do not provide an earth in the normal way. Essential for RCD protection. Eventually good separation as used in d.c. distribution was considered adequate.
My experience with steam cars of this size indicated that fuel consumption was about 3 gallons per hour (approximately 500,00Btu/hr or 145 kW); this formed the basis for sizing the burner to be purchased. The burner chosen fell just short of this output because of physical size limitations & its electrical power consumption.
As previously mentioned my ides for control of the final steam conditions came from information published in Moldtec & the journal of The Steam Automobile Club of America (SACA) by Peter Barrett, a very competent exponent of the hobby in the USA. Following considerable deliberation I decided on the following:
1. The burner would be operated by a simple on/off way using a pressure switch.
2. Steam temperature would be measured using a K type thermocouple feeding a commercial digital temperature controller.
3. The temperature controller would arrange for on/off feedwater supply to the generator. i.e. to feed when the outlet temperature is high & to cease when it drops below the desired value.
4. The temperature controller would provide an upper temperature shut down to prevent material damage (One will read later that it did not prevent such damage because at the time I was measuring steam temperature and not metal temperature).
5. The control of the feedwater supply was originally to be driving a feedwater pump intermittently via an electromagnetic clutch obtained from vehicle air compressor but the drive line complexity mitigated in favour of a solenoid valve dumping the output from a continuously running feedpump.
6. A safety valve would be equipped on the inlet manifold of the steam generator because if fitted to the outlet it may suffer seat damage from scale.
Initial trials gave unsatisfactory results, steam temperature control was poor & on occasions the steam outlet pipe was glowing dull red (700ºC) but the steam leaving this pipe was at saturation temperature (185ºC). reference to text books described this as a thin film boiling where a steam boundary layer on the inside of the tube presents a thermal barrier & poor heat transfer. The solution was to increase steam velocity & turbulence, I chose to follow the practice used in vaporizing burners by introducing stainless wires into the bore of the tube. This resulted a considerable improvement but excessive cylic times were giving an unacceptably wide temperature control range, or to put it another way the temperature controller was not seeing the anticipatory effect of the normaliser.
It was necessary to remove the steam generator from the car & dismantle it; this revealed distortion of the two top coils & could not remove the stainless wire. The laser cut combs used to position & support the tube pancake coils were not permitting sufficient freedom to allow for expansion & contraction. The upper two carbon steel coils were showing signs of serious overheating. See photo 17.
The remedial measures were to scrap the top four coils & replace them with smaller tube using 316 stainless steel. The reduction size to 10mm o.d x 1.5mm wall thickness would provide the increased steam velocity to ensure good heat transfer. Re-use of the combs would give good clearance to permit expansive movement.
The normaliser water injection point was moved to precede the temperature measuring the thermocouple by about 26”
The compression fittings used to avoid a transition weld did not remain steam tight due to the large temperature variations experienced in early trials. With more stable conditions I am sure that they would be up to the job but a weld would simplify matters so a material transition weld was made using iconel filler wire.
At the time of writing (Oct 2006) the Fox has 325 miles on the clock & it will consume a 25 mile trip with ease before a water stop is necessary. Steam temperature control of +/- 20ºC prevails most of the time although occasional traffic congestions conspire to exceed this band. The choice of relatively thick wall material is questionable, it made welding easy but may be the reason that a sudden stop from high speeds results in a pressure excursion. This is not a problem for the steam generator but the partially balanced throttle valve requires a greater operating force, also the pressure switch design limit is 425lbsf/in²
It is possible with experience to change gear both up and down with relative ease, I was advised to use a clutch but this has proved not to be necessary.
The Fox has potential for much further development, I have tried a larger burner nozzle but found the increased steam generator performance required new control settings. With this increased firing rate it was in steam from cold in about 30 seconds! Performance on the road demanded serious concentration so the earlier nozzle was refitted to restore a more leisurely driving background.
The steam generator pressure & temperature could be modestly increased to give grater performance & to much higher values permitting the use of a high speed poppet valve engine.