Monday, October 1, 2012

Concrete and dowel

This evening I filled my lathe bed with concrete. I said I was going to buy some 10mm dowels and grease them up before putting them into the bolt holes, but that wasn't possible. Because Magnet Mart are so good at stocking usefully sized lumber, I had a choice between 9.5mm and 12.5mm dowel...

Obviously, I chose 9.5mm. The process I used was only marginally more complex than the grease option. I wrapped each dowel (which was cut to just a little longer than the height of the bed) with a small piece of standard printing paper, fastening it with a piece of tape. Then I wrapped each piece with thin plastic packing tape, hopefully preventing the conrete from adhering to it.

The goal of this approach was two-fold: to make up the missing 0.5mm, and to facilitate the removal of the dowels with minimum fuss, ideally without leaving anything but air and concrete behind.

The concrete is curing now, so I'll let you know how it went in a while.

Tuesday, September 18, 2012

Lathe progress - Bed and bedways

The other day I headed back into the workshop to do some work on my lathe, which has been languishing for a while as a result of my working too much, and having uni work to do. Anyway, I bought some bolts from ebay (M8x1.25x80mm), and they turned up not long ago. So I set out laying the holes out on the ways and drilling them, ready for tapping.

I ran into a couple of...snags...during this process. Firstly, and worstly, I snapped not one, but two 3mm drill bits off in the bedways. After much swearing and a lot of dremel work, I managed to remove them. The second issue came as a result of misreading a drill-size chart. I read 7.2mm for M8 taps, when in fact the drill size is 6.8mm... No real drama, I just have shallower threads than I wanted. It seems to work fine though.

My camera was nowhere to be found during this process, but it was all quite straightforward, errors aside. I had 25 bolts, and a metre of bed to join to a metre of bedway. Casting one bolt aside as a spare, I set about spacing the 24 bolts evenly over the length of the bed, about 30mm apart, and centred along the length of the  ways. If you're doing this at  home, make absolutely sure that the bolts are spaced close enough together to go through the bed with no troubles. Had I spaced them 50mm apart, they would have been going through the walls of the bed...bad idea.

Anyway, pilot holes drilled, I clamped the ways to the bed and got them aligned, then proceeded to drill holes through the bed, using four of the pre-drilled holes as guides. I tapped the four holes I had used as guides, and bolted the bed and ways together with four bolts. Next, I drilled the remainder of the holes using a complex multi-step method of my own invention...

I drilled a small dimple into the bed, using the holes in the ways as a guide. Then I removed the ways from the bed and put the bed onto my drill press, which I used to drill 10mm holes through the top surface of the bed. If you have a nice big drill press, drill straight through in one go. Unfortunately for me, I had to drill my holes in two steps. I did all the top holes, then had to fiddle around a LOT to drill through the bottom, while maintaining alignment. I ended up doing this by chucking and unchucking the drill bit for every new hole, which worked fine, but was as tedious as all-get-out.

That done, I tapped the rest of the holes in the bedway and gave it a test fit. Brilliant!

The next step is to put a 10mm dowel through each pair of holes and fill the bed with concrete. This will add mass to the bed, which is always a good thing, but it will also help reduce flexing. Another good thing! The dowels will be greased so that I can just push them out (hopefully!). If not, I can drill them out easily.

Now, in case anyone was curious about why I drilled 10mm holes for M8 bolts, good question. Basically, it gives me about 2mm of wiggle room along the length and width of the bed for getting alignment right, and it absorbs any error I may have made in aligning the holes correctly to start with.

Anyway, onwards and upwards.

Friday, July 13, 2012

Tips for using NeoCat's Arduino Twitter library

This may be obvious to people who are better at programming than me, but I struggled to figure it out for a while. NeoCat's library can only post chars and char arrays to twitter, but readings are not in char format. To get around this I had to do some fiddly type conversions, which I lay out below.

itoa(var, varChar, 10);                               //converts 'var', which is an int, to 'varChar', which is a char
                                                                //If anyone can tell me what the '10' is for, please do

stringone = String(stringone + varChar  );    //make a string called stringone. Append the newly formed  
                                                                  //char 'varChar', which was the int 'var'.

   stringone.toCharArray(msg, 100);             //convert 'stringone' to char array (reads "value of the in
                                                                  //save the char array as 'msg'. 100 is the number of
                                                                  //characters in the char array.;                                       //posts data to twitter

'msg' must be declared as a char at the beginning of the sketch.

var was declared as an int, it could be, for instance, an analog reading.
varChar was declared as a char[].
stringone does not need to be declared until it is made.
More chars and char arrays can be added to stringone by using the + symbol and the char name.


I put down my first all-grain batch of beer the other day, and discovered that the arduino-based thermostat I had made for it had stopped working. I put the beer inside to keep warm, and it seems to be ok, but it prompted me to continue work on my brewing computer.

I now have a functioning network, so I've decided to add a reporting feature to my device. I've written the first revision of the code, but haven't yet got all the hardware going.

Here are the functions planned:

Internet enabled
Temperature sensing
Thermostatic control (hopefully with PID/PD)
Specific gravity sensor
Temperature calibration for gravity readings (reads gravity and spits out what the value would be if the liquid were at a set temperature)
Alcohol content calculation
Daily report of gravity, alc% and temperature via Twitter and email
Thermostatic control failure alert via Twitter and email
'Time to bottle' alert via Twitter and email

I'm yet to get the gravity sensor working, but that shouldn't be too much of an issue. I also haven't trialled the datalogging to SD card or PID yet. So far I know I can get it to send me an email with the readings, and I know it can control the temperature. More to come when I'm closer to completion.

Furnacey fun

I made the rest of my casting patterns the other day, and after painting them with glossy paint, set to finishing the furnace.

I lined it with the clay mix from before after grinding it down to a very fine powder. Because I knew I was short on volume I added all the sawdust I could find, most of a bag of perlite, some more sand, and the last of my bag of bentonite. I also filled the bottom of the body with dirt to fill up some of the extra space.

When lining furnaces it's important that your clay not be too wet. That was my problem the first time, and the loss of moisture caused a lot of shrinkage, and thus, cracks. This time I made the clay much drier. I spread the dry clay out as thinly as I could on a plastic ground sheet and sprayed a very fine mist of water over it, then mixed thoroughly with a rake. I repeated this until the clay was uniformly damp. It looked a bit like breadcrumbs and clumped together when squeezed, but did not feel wet. After that, I covered it for a while so it could percolate a bit, and to give me time to get the body ready to be lined.

The inner form was made by wrapping the metal from an old canola oil tin around a pair of wooden discs. Before doing this I rolled it up as tightly as possible and tied it like that for a few days. This meant that the metal was springy and would hold the discs, but also that it would pop in on itself when the discs were removed. I packed about an inch of clay down straight on top of the dirt, making sure there was a former for a drain hole. That done, I set the inner form on the clay bottom and began to ram clay in around it. When I got to the level of the burner I put the pipe in and continued to ram. You should build the lining up in very small layers, working around the furnace as you go.

To line the lid I just placed it on a piece of plastic sheet on the ground, inserted a piece of PVC in the middle as a former, and packed the clay in. I rand wires all through the lid to add stability. After using the concrete here it occurred to me that I should have used concrete in the bottom instead of dirt too...

Unfortunately, I was STILL short of clay, so the lid was only 2 thirds full, and the body lining stopped short about an inch from the top of the steel wall. To compensate for this I ground the steel on the body back to the level of the clay, I also filled the top of the lid in with quick set concrete. This is not a hotface, so concrete was fine. It was just to add stability.

I also began work on a lifting mechanism for the lid, but haven't finished it yet.

I cast my first four lathe parts the other day, and they came out reasonably well. Some had little pockets caused by loose sand, but the pockets are in non-critical places. It took all day, but I got the carriage, cross slide, compound base and compound slide castings done, and turned all my scrap into ingots. I'll upload some photos soon, but at the moment I'm busy machining the castings on a friend's milling machine.

Tuesday, June 19, 2012

First lathe casting pattern - the carriage

So today I started making the first casting pattern for my lathe. I'm starting with carriage, because it makes up the base of the tool-holding system, which I will need to use when I get to the headstock.

There are two ways to go about this (assuming you want to make the pattern from wood): cut it out of a single block, or glue it up from separate parts. I chose the former, because it seemed easier to me after taking stock of my selection of stock. You can do it the other way, it would be the easier of the two options, providing you had stock of the right dimensions handy.

Anyway, here are some photos of it, and here is a link to a pdf with the cutting sequence I used. I hope it makes sense, if not, let me know.

The top of the carriage

And the bottom

Obviously the pattern isn't completely done, I need to finish plugging little holes in the surface, give it a good sand, and paint it with something glossy. But it has all its tapers and rounded edges.

Sunday, June 17, 2012

I cannot stress this enough

Support the steel you're using for your ways!

I realised this afternoon that I had made the same mistake again. Frustrating! This time it was on the narrower piece of steel that will become the ways on my cross slide and topslide. I had stored it unsupported, and it developed a 0.8mm bow. This would be acceptable in the slideways, as it would be a much smaller deviation across their length, but I want to use it as a scraping reference for the front bedways. A 0.8mm bow is certainly not acceptable in the front ways. I've clamped it to my workbench under the bedways to straighten out, but still, bloody annoying.

Tuesday, June 5, 2012

Getting into bed

So, I can't remember if I've said this or not yet, but I scrapped the steel I-beam as a bed, it was too bowed, and wasn't going to deal with the torsional strain. I've changed to a metre-long piece of 75mm box-steel. This will deal much better with the torsional forces, and also I can fill it with concrete to add to the lathe's mass, always a good thing.

The problem with it is that it's not very flat. The biggest milling machine I could get my hands on had a maximum pass of 700mm, not long enough. So I have to do it by hand. I filed it back to get the longitudinal bow so common in box-section out, but it's still way off flat. I started scraping as per Gingery's instructions, but soon realised that it would take years to do it that way because of the amount of material that needed removing.

Then a thought occurred to me: why not use an angle-grinder? Just for roughing-in of course. I had a lot of material to remove, and wanted to get it close to flat, as close as possible. So, here's my method of roughing in. It's not finished yet, but I've done enough to see that the method has legs.

Lay your surface plate out (mine is the cold rolled steel that will one day be my ways). spread a thick layer of oil paint onto it, keeping it as even as possible.

Clean your steel and remove any burrs. Place the steel upside-down, very carefully, on the painted surface plate.

Do not apply pressure! It's cheating. Instead, gently push the steel forward a few centimetres, then pull is back again. Remove the steel and place it on your workbench. (Your workbench should be a good distance away from the surface plate to prevent grit landing on it.

Get out your angle grinder and make sure it has a grinding wheel in it. Now gently grind the spots with paint on them.

Rinse and repeat, taking a break every three or four passes.

A few important warnings:

Make sure your surface plate is well supported - mine wasn't, so I spent a good amount of time diligently ruining my own hard work. I fixed the problem, but am still trying fix the results of the problem.

Make sure you take those breaks - they're not for you, they're for the steel. You don't want the steel heating up too much, otherwise it will deform. Then you will be wasting effort on a bent piece of steel that will cool back to a different shape.

Make sure you clean the steel thoroughly before you put it on the painted surface plate - if there is any grit or burrs, it will render your markings meaningless.

Repeat this process until you have a pretty even spread of dots on the steel, reasonably close together. Once you have that, you can start scraping and it should be a lot faster.

A message to Gingery lovers/loathers

If you've arrived at my blog by googling 'Gingery' then you probably know that there's a small war going on between people who like Gingery and people who don't. Usually, the argument from those who don't like him is that his machines are inferior, and that it would be quicker and easier to buy the machine instead. I just thought I'd put my two bob in. I find myself agreeing with both sides of the Gingery war: true, his machines will never match commercial machines in terms of accuracy or longevity. At the same time, I completely sympathise with the idea of building a machine as a learning process. I find myself, however, coming out with more respect for the pro-Gingery party - not because they are correct about everything, not by a long shot - but because I feel the anti-Gingery mob are missing the point. People who build Gingery lathes do so for different reasons than people who buy lathes. Some of the reasons include: they can't afford to spend much money on a lathe, they relish the challenge, they need something to fill their time, they want to learn to cast. The list goes on. Some of the arguments against building these lathes include: it will never be as accurate as a commercial lathe, aluminium is a poor material to make a machine tool from, it will take much longer to end up with a lathe, etc.

To these arguments I would respond thus:

If you are careful, you can get decent accuracy, but at the same time, how accurate do you actually NEED your lathe to be? If you're building it yourself, you're probably a hobbyist, someone who can't justify buying a lathe, but wants one anyway. For this sort of person, accuracy to a tenth of a millimeter might be just fine. You have to take into consideration what you actually want your machine to be able to DO. If, like me, you tend to work by eye, then absolute accuracy is probably not important. If, also like me, you tend to make things on a lathe that aren't part of fiddly little engines, or of critical, life-saving devices, then likewise, how much accuracy do you actually need? Personally, I would be happy with 10th of a millimeter accuracy. Better is better, but I'd survive with that.

As for aluminium, yes, it's not a very good material for machine tools. So what? The builder can choose a different material, like ZA-12, or iron. These things are possible in the backyard. Another option open to the  maker is to cast only the parts he can't fabricate, and fabricate the rest out of steel. Personally, I think Aluminium is probably fine for small work, and the fact that with a furnace and the pattern handy I can whip up a new cross-slide pretty easily means that breaking parts is a hassle, but not much more. The other option open to the maker is to start with aluminium, and upgrade over time. I for one plan to complete it in aluminium, then slowly build up to iron casting and replace the pieces bit by bit.

And yes, it certainly will take longer. But if you're not a professional machinist, who cares? For hobbyists, it's all part of the journey. Some people are time-rich but money poor, this is perfect for them. Others aren't in any particular hurry, and would rather use the money for rent/food/grog.

My response to the pro-Gingery mob who are up in arms about these issues would be to think about what you need and modify Gingery's plans to suit. I feel certain he wasn't being prescriptive when he wrote his guides; feel free to change them! I am, my bed is made from steel box-section and will be filled with concrete, it's also a metre long! Do your own thing, using Gingery as inspiration and for basic methodology.

So there's my two-bob.

Monday, February 20, 2012

Disasters and discoveries in the alloy aisle, plus a new furnace

So, given my somewhat less-than-satisfactory results with the aluminium bronze, I started looking into other options. I looked at buying pre-alloyed aluminium bronze, but shipping will bite me. My next thought was Za-12, a zinc/aluminium alloy with excellent castability and hardness (google it).

While at a scrap yard looking for the piece of steel that makes up the front leg of my new furnace (read on for more), I discovered that the man who owns the scrapyard also sells ingots of zinc, 5kg each. AU50c/kg! At that rate, I'll be going with za-12, I hear it's easy to make at home given the low melting point of zinc.

Now, on to the furnace...

While 'making' the aluminium bronze, I overheated my furnace, driving the Portland cement reaction backwards, and making my refractory more like 'refractury'. A short while later, I found an abandoned water heater, one of the little ones. I also found a stock trolley in a skip. After putting my thinking cap on, I decided to build a completely new furnace, rather than relining the existing one.

I cut the water heater down, cut a ring out of it for a lid, and welded it onto the trolley, about 20cm above ground level. Then I welded a piece of 40mm box section steel -- the one that the nice guy at the scrapyard gave me -- onto the front, to provide support for the weight of the furnace.

The recipe for the refractory that I picked cannot be found here, because something has gone wrong with my link. Instead, I will just list the ingredients.

40% bentonite
25% 200-mesh silica sand
20% fire clay
10% alumina
5% borax

It has no cement in it, so is essentially a ceramic reaction, which I won't be able to drive backwards, no matter how hard I try. Hopefully.

I haven't done the lid yet, because I ran of of clay... But I won't be able to cast again for a while anyway, so that's fine.

[EDIT] I made the clay too wet, so it cracked and buckled all over the place. As a result, I had to let it dry and powder it right back down. Given how little extra volume I actually need for the lid, I've decided to add stuff to the mix willy-nilly to make up the volume. I have some perlite, which will go in, and I'll put in some sawdust as well, which will burn out, leaving a stronger, honeycomb-like structure. I might add some more clay and sand too, depending on how I'm feeling. More news later.

Sunday, January 29, 2012

Experiments with aluminium bronze

Today I made some aluminium bronze, and it was a partial success I think.
I learnt a few things about the material, but I'm sure there's more to learn.

The first thing that I learnt is that it takes a surprisingly long time to melt copper; I was still trying after 20 minutes at full blast. Its melting point is 1084.62 degrees celsius, but either I couldn't get that hot, or there's some trick to melting copper. I got it to a very bright red-heat, but it wouldn't go any further.

So here's the lesson in that: according to The Complete Handbook of Sand Casting by C.W. Ammen, you should melt the copper first (p189), then add the other metals. I got fed up with w

aiting for the copper to melt, so I added my aluminium, which began to melt almost instantly, forming a puddle in the bottom of the crucible. Once this puddle had formed, the copper began to melt really quickly. If anyone can explain this, please do. Anyway, in future, I'll put copper in the bottom, aluminium on top, then more copper, then fire it up.
The second thing I learnt is that either aluminiun bronze is not nearly as golden as the internet would have me believe, or that it's a lot harder to get the ratios right than I thought. I aimed for 10% aluminium by weight, so I put 9 parts copper and 1 part aluminium in (2kg copper, 222g aluminium). One thing that leads me to believe that the ratio is hard to get right is that there was an awful lot of dross (or something that looked like dross) stuck in the bottom of the crucible after the pour. Thus, much of either or both metal oxidized. At what ratio, I can't say, but surely it is likely to skew my intended ratio.

In all honesty, I don't care about the colour, and the colour came out being pretty nice anyway. It looks almost like steel, only shinier. What I care about is the mechanical properties.

Here's an idea of the colour: the watch is stainless steel, the light has made everything a little yellower, though there are some patches about the place on other ingots that actually are that yellow. Also, notice how smooth the side is, and the two big flat bits on the top. The bumpy bits were loose bits of sand. This implies that I can get a pretty good finish with this alloy even with reasonably coarse sand.

The third thing I learnt is that aluminium bronze needs to cool slowly, otherwise it becomes very brittle. I quenched most of the ingots that I cast, with the result that I could break them by hand, or by dropping them from a height of about a metre. This had me worried, until I found an ingot I had forgotten to quench. I couldn't break it at all, I didn't try with a hammer though. In future, NO QUENCHING.

Here is a n ingot that I quenched, it was so brittle that I could snap it in my hand. It leaves a very sharp edge, I discovered the hard way. Notice how shiny the inside is? Also, you can see how it's quite crystalline. This has to be the result of the quenching, as the alloy ratio is almost certainly similar to the range of 'real' aluminium bronze, and I have one ingot, which wasn't quenched, which isn't brittle.

Wednesday, January 25, 2012

My theoretical pet metal

People who work with metal tend to have a favourite; under normal circumstances, I would say my favourites are steel and copper. When talking about casting, however, my (theoretical) pet is aluminium bronze. I say 'theoretical' because I haven't yet cast with it. It is, therefore, my favourite only by virtue of what I have seen and read about it.

Aluminium bronze is an alloy of copper (~90%) and aluminium (~10%). It takes on a vaguely brassy look, and I believe that Australian gold coins are made from it. Here are the reasons I like the idea of aluminium bronze so much:

great material properties - it conducts heat well, is highly resistant to corrosion, and is very strong and hard. It's used to make bearings in aeroplane landing gear, boat propellers, engine parts, and many other things.

Great aesthetic properties - it looks a bit like brass, but not as 'bright'. It's also antimicrobial , apparently, as a result of the copper content.

And finally, great castability - aluminium bronze is a 'short freeze' alloy, meaning that it turns from liquid into solid almost instantly. The solidification happens from the outside face and progresses inwards. This means, according to this paper, that, assuming you have enough risers and feeders etc., you can obtain almost maximum density when casting aluminium bronze. Virtually pore free!

These reasons have all combined together to make me want to cast my lathe parts out of aluminium bronze instead of plain old aluminium. I may cast them in aluminium first, to make sure everything goes all right, and then re-cast them in aluminium bronze. But we'll see...

Tuesday, January 24, 2012

Lathe bench

I'm thinking ahead now. I'm going to need somewhere to put my lathe when it's done, and given that it's going to be bigger and heavier than Gingery's lathe, moving it around will not be an option. I've drawn this bench up, it's made of 25mm box section painted RHS, welded together. The small box shaped cavity in the top left corner will house the motor at the back, and have a little cupboard or drawer or something at the front. It's a metre long, so I mah have to shorten my lathe a bit, I'm planning on a 1m bed at the moment... we'll see.

A new project: ambitious, but possible

OK, I have a new major project. I'm going to build myself a Gingery-inspired lathe. If you don't know who Gingery is, google him. In short, he was a man who needed a machine shop, but lacked the funds for one. He had plenty of time though, so he decided to build one. He made many machines out of cast aluminium, which he melted in a furnace in his back yard. I admire his ingenuity, but feel that his lathe design is somewhat lacking in some areas. Those areas are:

size - his lathe is very small
fittings - his lathe has no standard tooling fittings
engineering - there are numerous engineering points that I disagree with, I'll cover these as they appear though

So, the first part that needs doing is the bed and ways. (Engineering flaw #1) Gingery cast his bed out of aluminium, making a 60cm hollow ribbed block for rigidity. Being aluminium, though, it is not particularly rigid. I imagine this lightweight casting introducing many inaccuracies, mostly through flexing and sagging. The other major downside to the bed is that it's a very big casting, especially for novices, at whom the series is mostly targeted.

To counter both of these problems, I have decided to skip casting the bed, and opt instead for a steel I-beam. This should give me the extra weight and rigidity that I yearn for, while also allowing me to scale the whole lathe up a bit (this addresses flaw #1, size). I plan on scaling it up to a 75mm swing, with around 60cm between centres.

Here is my basic idea for the bed: The black I-shaped section is, strangely enough, the I-beam. The thin grey piece is an aluminium match plate that I will cast up, and the red is a piece of bright flat bar.

The purpose of the match plate is simple. I will first scrape one side of it to fit the I-beam perfectly, then I will scrape the other side to match the bright bar perfectly. This will mean that the inaccuracies on the surface of the I-beam won't matter, as they will be averaged out by the aluminium plate. Then, given the very close tolerance of the bright flat bar, the matched aluminium surface will be almost perfectly flat. This will mean that I can bolt the bright bar onto it and not have to worry about it deforming.

Now, the only problem I can see with using I-beam is that it is not designed to withstand torsional force, which is exactly what I will be putting on it. I have two thoughts on this:
1) it may be so grossly over sized that it doesn't matter
2) I could weld straps into it, between the two parallel surfaces, and perpendicular to them. This would change the side view from being a long, wide channel, to a series of boxes. I feel that this should strengthen it sufficiently against the torsional forces.

So, those are my current plans. I'll update you as I build. But for now, google Gingery and see what he's all about.