Member postings for Andrew Johnston

+1 - both in the lathe and mill.

Andrew

Stoning the cutting edges has never really worked for me, and ruins the drill for anything else. I prefer to use slow helix drills specifically designed for brass. I have a small collection of the sizes I use most often. Brass isn't too bad for snatching, bronze, and especially gunmetal, are much worse.

Andrew

For quick 'n' dirty work I use a propane torch. But for bulk normalising, hardening and tempering of cutting tools and heating parts prior to hot forming I use a heat treatment oven by Carbolite. It has a PID controller that goes to 1100°C and is intended for heat treatment, it is not a kiln in the sense of pottery firing. New they cost a couple of thousand, but I bought mine secondhand on Ebay. Working area is 6" x 7" x 15" and I have had it full on some occasions. The downside is that a harden/temper cycle consumes a fair number of kWh.

Andrew

No, in a word. A bigger tool will be more rigid. But controlling overhang is as, or more, important. It also depends upon the job. Big tools can get in the way, may need more overhang or make the cut impossible. On my 13"x40" centre lathe I use 12mm and 16mm shanks for carbide and HSS 5/16" to 1/2" square depending upon what is easiest to grind. My repetition lathe takes 5/16" square as standard. It will easily take a 1/4" depth of cut in steel. When I get my Pultra up and running I'll be using HSS 1/8" to 1/4" square tooling.

Andrew

Shim stock may be springy, but in my experience if you bend it too far it snaps. To bend shim properly it needs to be annealed. While a petal valve may not bend very far the fatigue life may well be limited.

Andrew

Agreed. Mine was an impulse buy, but it has turned out to be very useful; for spur gear cutting:

Flycutting:

Helical gears:

And shifting metal:

I have a vertical head, but it's now too heavy for me to lift, so I need a plan B.

Andrew

Fourteen years so far. However, there are some factors in my defence:

Since I live on my own I have to do all the shopping, cooking, gardening, housework and so on.......

I have another time consuming hobby

Building two engines takes longer than one; I reckon about 50%

The drawings have many errors and omissions, so a lot of time has been spent creating CAD models, and assemblies to check function and fits

I want to make everything myself, including cutting all the gears, making most of the nuts, bolts and studs and forming and welding all the wheels. The main items I haven't made are the boilers and rubber tyres.

I am adding significantly to the design, based on fullsize, such as the singling valve and prototypical regulator lever

In order to make some items, such as injectors, I need to understand the underlying theory in order to create a design I can then make

The list goes on..................!

In essence it's the engineering that really interests me. Making parts is secondary, unless it's a technique I haven't used before.

Andrew

Building an engine involves making large numbers of disparate parts, which suddenly come together in one assembly. The basic motion work for my engines is now assembled. Turning the flywheel results in the crosshead and piston rods sliding back and forth:

Close up of the motion work:

All the components fitted together, and worked smoothly, with no adjustment needed. The slidebars and cylinders slope backwards by about a degree, relative to the boiler, as per the full size engines.

The far engine has a sheet metal chimney, made according to the plans. It has been pointed out that it is too skinny, and doesn't look right. Burrell fitted cast iron chimneys to some of their engines. The engine at the front has a prototype "cast" chimney, which looks a lot better. The "cast" chimney was modelled in CAD based on pictures of the fullsize engines. It was then 3D printed in two halves, stuck together and sprayed black. Since making the "cast" prototype the top has been changed, based on a picture of a Burrell pattern for the cast chimney and a convincing fluid flow explanation of why the top looks as it does. i will be going with the "cast" chimney as it looks much better, and is nearer to prototype.

Andrew

No keyway is fairly common on arbors for smaller horizontal mills. Friction should be enough to hold the cutter for the width and depth of cut that can be taken. I suspect that B&S tapers are rare in the UK, so hang on to what you've got. All the arbors for my horizontal mill have keyways. But it's a heavier, more powerful, mill capable of taking large cuts.

Andrew

My Pultra isn't running yet, but I have bought a 1/2hp 3-phase induction motor and VFD, which I expect will be fine. This will be driving the headstock, and accessories, via a baseplate and multiple pulleys behind the lathe. I have an original motor with clutch, but there is no plan to use it.

Andrew

I've had another look at both of my collet closer headstocks, this time with the manual open at the adjustment page. It's not clear that there is an over-centre effect. The closure just seems to rely on friction from a spring loaded ball in the handle, on the extreme left, on serrations on the spindle.

The way I read the manual is that the handle, extreme left, is rotated to close the collet onto the work and is then locked in place. The vertical handle then simply opens the collet or closes it to the pre-determined place. Probably the best thing to do is look at a real unit while referring to the first and third photos posted by Colin.

Andrew

Having read the manual after posting I realised that I was wrong about setting the collet position. Should have read the manual before posting.

Colin has saved me from needing to scan and post the relevant pages.

Andrew

Not for me - "a fart a day keeps the greens away".

Andrew

Had a quick play this morning. As expected there's an over-centre lock. Here's the headstock and collet closer in the open position:

As the vertical bronze lever, on the left, is moved to the left it comes up against resistance. With further movement it 'clicks' over and locks. The locked position is shown here:

I suspect that the over-centre lock is provided by the linkage under the lever and yoke. I don't know if my example is worn but everything is very loose in the open position. But it still seems to close and lock properly. Under the brass cap over the spindle nose there is a finely knurled bronze ring. I expect this is used to set the position of the collet so that it closes properly within the small amount of lever movement. The closer will definitely operate with the spindle running.

Andrew

I've got a couple of 1750 headstocks fitted with collet closers. I'll have a look tomorrow evening and figure out how it works. Current assumption is that something goes slightly over-centre as the collet grips. My collet closers look like they are intended to be used without stopping the spindle, as one would expect.

Andrew

Wow, looks good!

I thought I knew the Pultra accessory range, but I'm not familiar with the item shown in the second picture. Some sort of tailstock?

Andrew

Consumer products often use conductive rubber strips (aka Zebra strip) to make connections rather than a physical connector. They're very common on displays. If the strips get knocked they may distort enough to lose contact. But being rubber over time they can return slowly to their original shape, thus restoring contact.

Andrew

Post Office? Telephone exchanges used to have racks of lead acid batteries to provide the DC voltage down the phone lines that drove the telephone handset in case of mains failure and hence loss of the normal supply.

Old Post Office joke: Each telephone exchange was manned overnight by a man and a dog. The man was there to fix the equipment if it went wrong. The dog was there to bite the man if he attempted to fiddle with the equipment.

Andrew

Looks great and nicely made. I dream of actually finishing a model engine one day.

Andrew

Lead acid battery chargers are voltage sources. The ideal charge voltage is 2.45V per cell at 20C, ie, 14.7V for a 12V battery. The charge voltage should vary with temperature, but often isn't. The above figures are for fast charging when the battery is disconnected after charging. For float charging the voltage should be lower, around 13.8V. Float charging is less common these days. Many of the new charger ICs work on a cyclic basis - fast charge, disconnect, wait for the battery voltage to drop and repeat.

Current acceptance is dependent on state of charge (SOC) and temperature. The last pack I designed for a hybrid EV would accept a couple of hundred amps of regen, provided SOC was less than 60%. The 12V batteries used were only 25Ah nominal capacity.

The quickest way to kill a lead acid battery is to over-voltage it. The problem of accepting regen energy at high SOC is one of over-voltage. The battery will accept the current, but the battery voltage will rise to damaging levels. Over-voltage leads to disassociation in the battery and the production of hydrogen. When I first started working on EVs in the early 1990s our battery test room had to have hydrogen sensors fitted, as did our vehicles.

A deep discharged battery can be recovered but putting it on charge immediately. It is leaving the battery discharged that kills it. Short term low voltage excursions are not a problem. Many years ago I was involved in the design of an engine management system for US heavy duty trucks. The power supply input specification was 6V to 60V. Trucks use 24V batteries. The 60V was needed because jump starting was normally done series, not parallel. The 6V arose from cold cranking at -40C. Cranking current was 1500A and that will drag a 24V battery down towards 6V if you're in Wisconsin in the winter months.

Two additional problems with battery packs are temperature control and cell balancing. Batteries like to be warm to perform to maximum ability so ideally there needs to be a means of heating the battery. Conversely one also needs to be able to get rid of heat when ambient temperatures are high. Measuring battery temperature is tricky. Ideally the sensor would be deep in the battery. The best we managed was to put sensors directly on the terminal posts and hope that gave a reasonable approximation to battery internal temperature. However well battery manufacture is controlled individual batteries will have slightly different capacities. Over many cycles the battery SOCs will diverge resulting in a battery being under, or over, charged. The ideal is to add cell balancing. The passive method simply puts a power resistor across the most highly charged battery. The last pack I designed used active cell balancing, ie, a resistor, or a low current (1A) power supply, could be connected across any cell under software control.

In a lead acid battery terminal voltage is a good measure of SOC, provided the battery is left to "relax" after charging. On our battery packs we let the batteries sit for 5 hours before making SOC estimates.

Andrew