Brickthus

I've done that. 5 56T turntables will fit in a Hailfire Droid wheel. I even had 8T cogs inside the middle turntable too You will find that the turntables in the wheel have a lot of friction. The numbers of teeth add up OK (3 x 56 = 168) but the wheel is a bit tight. When I used this system, driving from the sun wheel, I had to add extra motors in the gearbox to force the planet carrier and annulus to contra-rotate, because of the high friction. The friction was partly from the five turntables though, so not just the wheel's fault. Power Functions was really useful because I could drive the extra motors from the battery with infra-red control while it was all rotating! Some more planet gears, different ratios with the Power Miners wheels here A Power Miners wheel gearbox has less friction, little enough that a smaller version of the Hailfire Droid wheel application was able to overcome its friction and have the annulus turn the right way, without additional motors mounted in the gearbox. It still took quite a lot of input power though - four 9V 4000 rpm motors and 1.3A of current at 9V. The Power Functions motors also use planet gears. In this case 3 planets in each of 2 stages for the medium and XL motors. I took a stage out of a medium motor to increase the speed (no point gearing down and up again), but in the end there wasn't much more power available. Mark

My Apache is one of them The swash plate moves collectively and cyclically, with the 4 blade 7 stud propeller piece driving the pitch of the rotor blades. The individual rotor blade pitches change as the rotor rotates. The tail rotor has just a collective control. I used clip plate and rod parts in this case. Both joysticks link up with the main rotor controls and the rudder pedals are ready to be linked to the tail rotor. I built this before we had the powerful motors like the PF XL one. I might use 2 of them to power the main rotor. My original thought, long before PF, was to use two 12V train motors like this - 8 watts from each motor! I think for a future helicopter TLG might look for less-conventional styles, but ones that make a good solution to the problems that a helicopter solves and use Technic parts well. There is as much merit in the concept idea as in the modelling skill. As such, my Apache is a classic design and a scale mode (1:20) so there is no originality of concept apart from the attempt at a proper rotor system. Having made it work myself, I'll be watching closely to see how the entrants do pitch control Bear in mind also that some entrants in previous months have used lots of parts. A helicopter is meant to have minimum weight in order to get off the ground, so less is more! Make sure your rotor blades are big enough for the vehicle. When I put the Apache's blades on, they stretch right back to the tail - it needs a 6ft circle to operate in. PF IR would be the only safe way to control the motors! Of course the ultimate prize would be to make a model the flies. So far I achieved 39g of lift the the pictured models, up to 50g in the latest experiments. I need more aerodynamic parts, like the white fan blades from the educational kit. Mark

My layout is modular, with intention for exhibition. This means it would be useful for an extra pack of adhesive strips to be available. If the adhesive strips would have to be replaced at each module boundary at each build, this would be tiresome, probably taking 2 hours to replace 480 rail joints (assuming 2 tracks of 2 rails between adjacent modules and 120 modules in the layout). Since my hope was to still build the layout in 8 hours with more track than before by using a modular approach, this would not work. Maybe if modules were 4ftx4ft (10-20 modules), as with some model railways of other types, the adhesive strip time overhead would be small as well as saving build time. This would need more people (4 to carry a module) and a wider van (4ft between wheel arches), meaning that a club effort is required. Modular layouts would naturally use 48x48 baseplates, so triple straights would be useful. Alternating single and double straights might sound odd as trains pass over the unevenly-spaced joints. Real railways used 15ft and 60ft rails, with continuous welded rail now the norm, so it helps when modules are 64M long and can use quad rails. The module structure is often dictated by the location of the switch points, which need a 32x32 space to themselves, 48x32 if the siding-straightening curve is included for easier (straight) module connection. Is there enough spring in the standard track ends to avoid using adhesive strips where a standard track piece meets a pair of ME-models rails? If so, standard track pieces alternating with ME double rails could work, though it might look odd without full ballast. I applaud the concept of these rails, the aim of extending 9V longevity and the effort demonstrated but it seems there is some proving and ironing out to do. I look forward to seeing how your tests go, and whether the adhesive strips are required in all, some or no cases. Mark

This is a good solution as a substitute for the 1x4 plates outside the rails, for the purpose of keeping the track straight. Having used jumpers and 2x3 inserts this way before, for ballast on a curve, I'm already using the technique with jumpers, 2x3 plates and tiles along the track on my turntable. The advantage for a turntable is not so much to keep the track straight (as it's on straight frames already) but that the length of the turntable can be other than a multiple of 16M long. I've settled on 72M for now. 1x4 tiles bridge between the jumpers, each keeping a 2x3 plate in place, then other 1x4 tiles bridge between 2x3 plates, giving the offset plank effect. The tiles along the track substitute the function of the clips on the curved ballasted section in keeping the 2x3s in place. I took some pictures a couple of weeks ago so they will follow soon, though I might decide to hide the handrails! Mark

I chose to build to 8mm:1ft scale, which is approximately 1 stud to the foot. This is because my original aim was to exhibit LEGO trains at model railway shows, where scale is important. This was before I joined the LEGO community. In 1996 I was building 6-wide trains in 12V when I thought "real trains are wider than this", and so it began. Initially my 8-wide Class 47 loco was too short, so I lengthened it to make the 2nd build. It then changed livery for the 3rd build and is due another refresh! Once I got some books from model railway shows, such as "British Main Line Diesel Locomotives", I was able to measure the 4mm scale drawings, double the sizes and reproduce the main dimensions to within about 6 inches (0.5 stud). Some locos I have built with a main body 8-wide but with accessories making up to the scale size. A few could be widened in the middle - for one, the width is 9'3" but tapered, so I might use some hinges to do that later. The beauty of LEGO modelling is that everything can be improved later! The track gauge is 37.8mm, compared to 1435.1mm for the real gauge of many countries (some countries have multiple main line gauges). Of course in reality the gauge on curves is up to 10mm wider on the real railway for tolerance reasons. Ever wondered why LEGO trains drag in the curves? With British trains, the train body size varies from 7 to 10-wide, but a scale build can be done. I put straights in between the curves on the layout, to widen the radius. This may benefit from ME models wider radii in due course, enabling faster running. If I built US trains, a typical boxcar is 10'5" wide, so the nearest dimension is 8-wide plus a non-studded brick (2 plates and a tile) on its side each side, depending how precise I chose to be for a particular model. This is a choice. It takes more time, more bricks, more consideration of power, motors, clearances and slopes to build to scale. It becomes more like model engineering. I'm still happy with the results, and model railway enthusiasts are surprised too, especially when they can't tell it's LEGO! Mark

I'd really like R104 too. This would replace a set of standard curves with two straights between each curve. 48M rails would go good for 48x48 modules. Eventually I would like a double-slip kit (tracks at 22.5 degrees), and an express point kit (right and left) (64-96M total length but no piece longer than 48M so that the layout can be modular, so maybe 48M for the rails to move across by 8M and another 32-48M for either turning straight or connecting to another point to make a crossover). Would be useful if any points used a similar slider interface to standard points, but anything versatile would be OK. Being able to use a motor or pneumatics underneath would be an advantage for scenic modules. This is a really good initiative and I hope it succeeds and grows in product range. Mark

You could add some 2x3 plates under each rail, to support two 1x1 plates each side of each rail. Alternate black 1x1s with grey or colour to get a sleeper effect. It would be similar to some 9V schemes, just with the 2 middle studs missed out. Everything outside the rails is just like 9V too. Consider the profile of how high you want the ballast to be in the middle compare to how it slops away from the track at the sides. That way the track looks embedded in its environment rather than stuck on top. Mark

The 7863 point motors can also make level crossing barriers, as in 7866. Use counterweights to keep the load balanced. They were also used for the decoupler in 7862. Or any MOC application that requires a 90-degree movement when the switch is pressed, and stays put by itself at other times. I would advise against keeping power on to these devices because, as solenoids, they could overheat. They are designed for momentary operation. They're not like moving motors, where the coils alternate in absorbing current. I haven't tried variable operation (driving with a variable H-bridge duty cycle) but it would probably overheat them. The solenoid units can operate with as little as 9 Volts, so they're not entirely 9V-incompatible. However they are not strong enough to change unmodified 9V points. I haven't tested whether they could move modified ones (that a micro motor could move) but it's touch-and-go because even modified 9V points have friction of the electrical contacts, where 12V points have almost no friction. Mark

The 8878 LiPo battery has the advantage of a re-starting circuit, which gets a heavy train to the top of the hill by retrying if it stalls. I think the auto-off time of a 6AAA battery box is shorter - not helpful if you have a tunnel on your layout! With really heavy trains, best to have a battery in each loco. I would like to have a larger capacity LiPo battery in one loco and run a cable to the "B" loco, using a single IR receiver to make sure both locos' motors got exactly the same command and power all the time. If not, there is the possibility of one loco pushing the other off the rails. I would have to make a slave H-bridge motor driver and run a few of them from a single IR receiver command in order to power all the motors - a single IR receiver can power only two train motors at once. The motor driver chip recommends 400mA per channel and 800mA total but a single motor can pull more than 200mA with a heavy train. Nevertheless, the 8878 LiPo is a really good product for light or average weight trains. Mark

I'm an engineer. My first degree is in Electronics and I did an MSc in Safety. I learned most of my mechanical engineering with LEGO! Real mechanical engineering has a lot more to do with material properties, where LEGO fans usually build with a limited range of materials! The level of detail applies across topic boundaries. I also built a Rail Crane, as a scale model of a real one. With trains I try to make the model work for real, sometimes even more importantly than just how it looks. I have been able to emulate some real-world hydraulic system operations with LEGO pneumatics Three of my MOCs have become patent applications for real mechanical things too! Just shows how good LEGO is as a concept prototyping medium. Mark

OK, that's quite understandable, especially where you have a loco with one cab, or a steam engine with a tender. In that case you might consider a Triangular Junction as an alternative to a reversing loop. Both take up a lot of space in a layout, so the smallest way to turn a loco is to make a turntable. I have a turntable under construction. Indy Mine track supports the loco weight and a PF motor drives a Technic turntable in the middle. The Indy Mine track makes a good size circle to support a turntable near the ends, especially for 6-wide locos that are shorter than 8mm scale ones. For a visible turntable, the rotation speed should be limited but if it were behind the scenes, one without a motor could be spun quickly by hand. The bonus for a 2-cab loco is that automatically changing the lights on changing direction means there is no intervention required, and no points or turntable need operating! I have found that railway modellers like to reduce "Hand of God" intervention at shows. Reading magazines about real trains, the locos seem to operate either way round. One a preserved railway, steam engines often do the same because the maximum line speed is 25mph, below the maximum speed for driving tender-first. Tank engines were designed not to need turning. Lights are operated manually on real trains and sometimes the driver makes an error! Mark

Here are some milk tankers I made recently. They're based on a UK one from the 1930s, a 6-wheel design. The middle wheels move to allow the tankers to get round the curves. They have less drag than a lot of other wagons. I made 6 white ones and a purple one (with Zurg parts). Guess which one went to Cadbury's! I did make a parts list for the tanker, which I could post later for anyone who want to make one. Mark

This electronic circuit, attached in parallel with the drive motor and driving two sets of PF light bricks, will do the trick. It can be used with either PF or 9V trains, but with 9V the lights will go off when the train stops unless a battery is used for the 9V and 0V lines as well. Without a battery, a bridge rectifier could drive the supply rails from the motor input. Here it is in a loco. It works really well and the lights are bright even at the lowest motor power levels. Mark

Curved track, straights at angles and flexi-track can be fixed to a baseplate using the techniques here Apply ballast and sleepers to the track itself. Apply any cant (tilt) to the track itself. Use composite tiles to make any slope that is required in the trackbed. My example has a slope of 1 plate per 4 pieces of flexi-track, or 1 in 40. That's the same as 1 plate per piece of standard track. I support the track (or the composite tile) every 2 pieces of flexi-track. Begin by marking out the edge of the ballasted track on the baseplate, so you know where to put the supporting bricks. The supporting bricks will make the border between sloping track and the flat adjacent trackbed. Cheeses mounted on headlight bricks keep the track in place when a heavy train goes fast round the corner. The technique applies particularly to flexi-track as it has a lot of flexibility so that a train can move it, but also to standard track because that can open its joints under the force of a heavy train. It is best when the track is always at least 3 plates above the baseplate. Good scenery means this is not an overhead. For flat trackbeds, tiles will do, keeping the ballasted track at least 1 plate above the baseplate. The main attachment is at each end of the curve but adjacent ballast studs sticking up next to the track ballast can serve to limit track movement. Mark

Flexi-track is noisier than standard track but has less friction than standard curves, especially at a wide radius. It is also good for double bends, which allow trains to go round an island platform, or step across by 16M in the profile of a point and reverse curve. I have succeeded in ballasting, canting and sloping it too. Here's a of a train running on the result! (Yes, I know it's too dark, and there will be some still pictures to follow!)To be realistic, it would make no commercial sense for TLG to sell straights separately. They need to get flexi-track used by people and then they'll see whether it is a good product, once AFOLs and TFOLs have given it a proper try. At least all kids will be able to make a circuit this Christmas! Mark

I have more problems with coupling detachment because I run heavy trains. In this picture the two similar wagons have a 1x5 liftarm and pegs instead of the couplings. Coupling magnets will weaken over time. It is possible to strengthen them by stroking them with a more powerful magnet, but the plastic surround might prevent this being very effective. I might go for liftarms for the whole train because a train of 8 wagons detached in 2 different places on different occasions. I would have used plates or tiles on the new couplings but I don't find the buffer beams compatible with the scale I build to, so I'll keep them on the train sets rather than using them in MOCs. Emerald night couplings have a tube that protrudes too low too. The later ones have been trimmed so they don't foul the crossing rails on the points. I haven't found attachment problems with either old or new types. Sometimes repeat shunting is necessary to get the magnets in the right range or orientation but that was always the case way back into the 1970s with red and blue magnets! Using sections of the chain with a stud on each end would look the best for old wagons but would not be nearly strong enough! Mark

Flexi-track with a minimum radius of 56M would enable the bumpiness to be reduced. It would continue to satisfy the diverse requirements of AFOLs, who need curves at 56M, 72M, 88M and 104M radii. However, one good use for the existing flexi-track pieces is in making a double bend equivalent to the curved road of a point and its reverse curve, for use in stations. It would not perform this role without the tighter turn capability. "Just go back to the prototype, but with the production hinge standard" would be my solution. The prototype was much less bumpy, had no check rails and was easier to ballast, though the rod-sized holes in the middle of the production standard hinges make it easier still. The best solution would be a single piece: a 1x8 plate with studs 2 and 7 replaced with tile areas. Then we would stick on O-gauge track clips, get O-gauge rails and slide the sleepers onto them, cutting the rails to length. Those sleepers could be cut up to make bespoke switching tracks and crossings, with the clips attached at an angle to the part sleepers where required. I wonder if, replacing the 2 and 7 tile areas with clips, the clips could be used to hold rails? 1x1 tiles with clip could be used on their own, but it might be better to hold the rails together with a 1x8 piece in order to maintain the track gauge. A new standard track feed connector, smaller than the 9V one, would be useful. The feed part should be capable of blending into a trackbed so that it's barely noticeable. Mark

I've posted a of the MOCs in my loft:- PF trains and testing, including the first showing of my Class 14 loco working, and a new brake van. - Flexi-track ballasted, canted and sloped successfully at wide radius. - Scenic modules. Channel: http://youtube.com/mbellisbrickmocs Hope it inspires you! Mark

I have succeeded in ballasting, canting and sloping flexi-track, but it takes a lot of work. It also needs more support than standard track if you have a hill. I boxed in the ballasted track by using 1x1 headlight bricks and cheeses to follow the curve - it took quite a while to build for a 72M radius curve! The need to fix it down actually makes it more realistic like a real railway unlike the check rails, which are only realistic for tram tracks in a street. I'm over the aesthetics but I have more problem with the function of the tracks. They're noisy and bumpy, especially with larger-wheeled trains such as Emerald Night. They do have less drag than standard curves, when used in a wide radius instead of a curve-straight-curve corner. Flexi-track was always going to be a compromise, trying to satisfy kids' need to make a circuit with adults' need for wider radius curves. Unfortunately a redesign of the parts is most unlikely, despite the face that the prototype parts were better in all but one respect (no check rails and less bumpy but they didn't have the useful rod-size hole in the hinge). The demise of standard curves in favour of 32 flexi-track pieces in 7499 is a risk for TLG. Do we want flexi-track any more or less than standard curves? Maybe if we got Indy mine track straights... but I've heard nothing. Mark

Given that I run the LiPo battery to its maximum output of 800mA, one option for more powerful trains is to make a larger battery, perhaps 2S2P, with a 2 Amp output current. This would need cell balancing, more so than the 2S1P battery we have because the cells are in parallel. However, the current capacity of PF leads is unlikely to be much more than the 800mA output of the existing battery, since all elements of the system have been designed to work together. This one would have the same capacity as a set of alkaline AAs at one third of the weight, but the current capability is very high, requiring a circuit to keep it safe, as well as a standard non-LEGO charger to sort out charging and balancing. Therefore a more suitable way to get more power would be to use slave H-bridges. An IR receiver output should go to a number of H-bridge inputs, one per pair of PF train motors, each pair being supplied from a different battery source. That way, all locos in a train get the same control signals, with no chance of a speed signal going to one loco and not the others (causing derailment). Unfortunately it means more batteries, so hopefully the price will drop. I see the 2100mAh 2S3P one is just $14.99, though I expect the charger circuit and speed control circuit in the LEGO battery cost more to make. I would have thought a price of $25 would be possible though. $30 with the recent 20% price hike The way the IR receiver H-bridge chips are rated, using one per motor could allow more power if the battery has the capability. Just need to make sure that things are not run outside their ratings, including using a heatsink where necessary. The chips can be used with both sets of H-bridges paralleled up, though this is not possible whist they are inside the receiver, unless the correct IR codes are sent to keep red and blue outputs at the same speed as each other all the time i.e. NXT and IR Link sensor control. It would be useful to sort out a PWM reduction circuit to equalise speeds between train motors and the Emerald Night loco. This would need some sort of table in a chip, to set the right PWM at each speed setting. The diodes I'm using at the moment work up to a point, but they limit the power to the train motors when they need all the power they can get in order to pull the heavy train up hills and round corners. Maybe a system based on the bar code truck set 8479 would work, whereby the speed sensor wheel would be attached to an unpowered wheel of the train and the programming of the bar code unit would cause more power to be sent to the loco motor in proportion to the input frequency. Maybe replacing the current limit device would be an option, for those prepared to void the warranty and take the risk. The LiPo cells are easily capable of faster discharge. I would consider a 1 Amp device rather than the 800mA one, but it's just the speed and regulator circuit ratings that might be exceeded. The risk of cooking anything is up to you! I recommend you get a CO2 fire extinguisher if you want to try this sort of thing, so you don't burn anything but the battery! Make the train roof easily removable, so you can get the CO2 in if required. A slave H-bridge of higher rating might be a better way to go (1 per motor), so all the power components are non-LEGO and receive the command from the IR receiver. Mark

Absolutely and, while we're at it, bring back micro motors (8082) and the Flex system (8074)! 8082 used one motor to vary the switch setting of another in one model. I did that here with Power Functions too; it allows a few machines to activate each other remotely. We miss these sets and the ideas books that encourage exploration of new ideas and inventions. LEGO education needs to encourage that kind of thinking in kids, rather than just pandering to the instant gratification of this generation. The NXT is one good thing here, answering the need whilst also including modern technology, but something for a lower age group should be considered too. Mark

What is summed up in the two pictures is "complexity contained". It's obvious from the first picture that there's a lot going on in this model, so that builds the excitement. Pneumatic models are good for this because the switches are often a long way from the cylinders. The second picture shows the containment of the pipes to the switches, as well as a novel (then) switch mounting scheme. Each great Technic set has to use a new technique to overcome a new engineering problem, continuing our education! The second picture also shows the routing of the pipes through the turntable. This is the essence of how to get the functions working together without them interfering with each other. These are the additional requirements beyond the number of functions for your money and the new parts, both of which are prerequisites on the road to Technic greatness. 8455 backhoe has the same build-up of expectation as 8868 with the pneumatic switches. 8480 Shuttle has it with its gearbox. 8479 has it in the truck because the grab arm and tipper bucket are some of the last things to be built. The key is to show the builder how many functions there are. 8485 Control Centre had it to a certain extent but was always limited to three functions, so the NXT overcomes that limit, especially with an IR Link sensor and Power Functions (now up to 11 motors from one NXT!) 8868, 8455, 8479 and 8485 have obvious functions, where 8480's are neatly hidden away. Many MOCs are great but find it difficult to show off their internal functions. This might be one reason why the biggest Technic cars haven't had complete bodies (apart from cost of parts and weight vs. suspension spring strength). For looks, I think 8480 and 8856 are the best aircraft. 8836 scores with its ailerons as a simpler set. Another aspect is versatility of application. 8455, 8868 and 8479 are multi-tools that can interface with the non-LEGO world. 8480 is self-contained with its satellite and 8485 is self-contained with its helicopter and dinosaur. By contrast 8094's drawing machine is a real-world application, and its crane can pick up anything so I think that makes it better than 8485, despite the lack of power socket. This is where 8443 and 8049 score well, with their cranes. Another aspect is imaginative stimulation, what it leads the imaginations of builders to produce. 8868 led to a whole host of pneumatic Finite State Machines and beyond. 8856 was an open invitation to do a proper helicopter rotor, as well as pre-8880 4-wheel drive (the swash plate doesn't hold wheels on very well but I bet TLG were doing the same with the parts before 8880 as made). This occurs mostly in larger sets but occasionally a smaller set will open things up. Substance must also be considered. Few small sets struggle with this, so they don't have to demonstrate it, but a large crane with spindly stabilisers is bad. 8479's grab arm was close to the limit but it would lift a tyre well enough. Studless models may need more parts in places. How about the latest sets? 8043 (with the gearing issues fixed) has the makings of a classic. It has the expectation as its PF body is built, looks like the real thing and applies widely in the real world. We have already seen gearboxes and multi-PF controls but it will push those techniques forward for anyone who wasn't around for 8480. The concentric shafts technique is where it scores heavily, so I expect to see lots of turret vehicle MOCs soon! 8052 scores best on looks because it is like the real thing. Not so many functions, but better value than 8264. 8051 is good as motorbikes go. Obvious limitation on number of functions but we have to have a good motorbike in the range and this is it. 8053 provides imagination for those who missed 8421 but looks spindly. Not as solid-looking as 8460. Mark

I looked at the data sheet for a Texas Instruments bq24103 LiPo charger chip (hence the TI and BQ markings, the CK identifies the 24103 from the series). The circuit is fairly standard, though the status LEDs are driven by the Alpha chip rather than by the LiPo chip outputs. The output inductance is substantial, being 220mH rather than the smaller inductances in most of the data sheet circuits. The current sense resistor is larger than standard (0.22 ohms rather than 0.1 ohm). The components in front of the chip supply input are the D11 diode, L2 inductor and a few capacitors, so there's no SMPS in front of the chip, just a single diode drop. A buck converter would have required a switching component for oscillation and there is no evidence of one. I think L2 and the capacitors are just a Pi-network to minimise noise and ripple. The chip's absolute maximum voltage input is 20V, so no wonder it played dead! The chip's recommended maximum voltage input, including all inherent switching noise spikes, is 16V. Therefore, if you insist on exceeding the standard 10V, I recommend 12V as the highest standard voltage level (multiple of 1.5V up to 9V, or 3V thereafter) to be used from an alternative power source, so that spikes are less likely to exceed 16V. An alternative power source should be capable of supplying 700mA but not more than 1 Amp per battery. Check also the level of regulation required on the power source because both regulated and unregulated wall warts are available. I guess the maximum voltage could be 1.414 (root 2) times the nominal voltage for an unregulated source. 14.14V is less than 16V but 16.97V is more than 16V. Therefore only go up to 12V with a regulated source. I could try it from my bench power supply, which I use for powering 9V main lines anyway (more current than a standard controller). Given that the chip includes its own FETs for switching, you might also like to try a 9V source, which would not hurt anything. See if the charging rate is any slower than with the standard 10V. The minimum voltage is 4.35V so you could try 6V too. You might find that a 13.8V power supply is tweakable via an internal potentiometer. Always have the lid on when it's powered (live mains), but turn off, allow to discharge capacitors and then tweak it. I used my 3A one to charge my car battery through 4m of 3A mains cable (resistance 0.3 ohms) once I had tweaked it up to 14.75V no load voltage. It picks up well where a standard charger gives up, for longer slow charging. Sorted without taking the lid off! Mark

I agree about the safety margins. I would like to determine how close the design intent is to them, to optimise the solution. I have a candidate for the charger chip, so I'll look into is a bit more and see what to recommend. I don't think it would charge any faster by upping the voltage, so maybe rectified +/-12V, giving 10.6V, is a good start. I'm reluctant to open up a battery, since these days I like to be able to put things back together properly afterwards! I'll see how far I get with Philo's photos. Logically in the PF product line design, the power supply would not be the weak link, so its 700mA rating is just above the 550-600mA used. That makes sense since it's no bigger (and more expensive in components) than it needs to be for manufacture. Mark

I tend to load the battery box to almost its maximum 800mA output. This gives about 1/2 hour running time for an average 1 lap per minute of a 16ft x 12ft looped eight with 1 in 30 slopes and a heavy train. I had thought about the 9V motor conversion to double socket, for track charging or hybrid trains, but I have 60 motors to convert I knew the official transformer was well overpriced but I hadn't had time to investigate other options. 10V at 700mA is the power input spec, so a rectified 12V, as suggested, should work. The current limit means I would recommend a power supply of no more than 1 Amp. I think Maplins used to do an 800mA one. By all means have a multi-tap power supply for charging multiple batteries at once. I'd recommend individual current limits on them though. I'm not so sure that higher voltages are such a good idea with the LiPo battery. Perhaps cell life is shorter the further above 10V we go, due to stress? I would need to look into it. How do laptop power supplies compare? Mark