Member postings for Andrew Johnston

Hi Pete,

Thanks for the description of the zeroing method. I think it's beginning to fall into place. I'll have to re-read the Haimer leaflet in the light of your notes, which I think make things clearer.

The next task is to test the Haimer again in the CNC mill, but this time taking care to clean everything beforehand. If I get a reasonable result (say less than 0.02mm) I'm going to assume that in due course I can zero that out with the adjustment screws on the unit. Once I'm happy with the CNC measurements I can concentrate on the Bridgeport. I'm pretty sure the appalling results on the Bridgeport are due to the spindle nose, but there are some tests I can do that may prove that. After that it's a question of cleaning up the spindle nose on the Bridgeport so that it runs true.

Regards,

Andrew

A recent thread discussed driving a 230V DC motor, **LINK**

I can't see any reason for changing to a stepper motor, a lot of mechanical work, as well as changing the drive electronics, and fitting a hefty power supply.

Regards,

Andrew

Hmmm, I suspect that the A and F on the nameplate stand for armature and field, which implies it's a DC motor. The fact that the armature and field list different voltages and currents may mean that it's a shunt wound DC motor. Therefore driving it with a VFD will not work. It's not uncommon to use DC motors for table drives; certainly a lot of Bridgeport mills do so.

Regards,

Andrew

Here are a couple more simulations, using a transformer rather than a coil. The output graph in each case represents the output of the secondary. The inductance of the primary is 100µH as before, for the secondary 10mH, on the naive basis that the turns ratio is 100:1, and the mutual inductance is 600µH, implying a coupling coefficient of 0.6. Here's the picture for the basic transformer:

and for the capacitor with transformer:

No great surprises, bigger voltages all round! I'm not sure why the oscillation in the case with the capacitor/transformer appears to be truncated for a short period after the first oscillation. I suspect a limitation of the simulation. In the real world both the diode and transistor would have let out the magic smoke long before. It seems pointless trying to simulate a spark gap, so I'm not going to try. Instead onto the real world and I'll try some tests with my experimental ignition coil. I might get time this weekend as it looks like the weather is going to degenerate into showers, so no flying, , but that does mean more time for playing at engineering .

Regards,

Andrew

White LED?

Andrew

Hi Pete,

Further good information, thanks! Yes, I've read the manual, mostly makes sense but I'm still none the wiser on how to adjust the runout. The translation certainly leaves something to be desired; like understanding.

Both my CNC and vertical mill are R8, not ideal but that's what I've got. My Tormach CNC mill utilises what they call TTS 'quick change' tooling which uses a modified 3/4" collet. All end mills, drill chucks, ER collet chucks etc are mounted in holders that have a 3/4" parallel shank. They also have a shoulder that is pulled up hard to the spindle nose by closing the collet, giving repeatability of the tool height, unlike a normal R8 collet. I use the same collet system on my Bridgeport. The Haimer is fitted in a special holder that converts its spindle (10mm) to 3/4". Again not ideal, but that's life.

I've done some measurements on my Bridgeport mill with the Haimer. Truely awful! With the same procedure of measuring at the 4 quadrants I get a total runout of 0.12mm. If I turn the Haimer through 180° then the runout follows it. Runout of the spindle measured in the R8 taper is a bit under 0.01mm. However, measuring on the flat underside of the spindle nose I get a runout of ±0.01mm. By contrast I can't see the needle on the DTI move when measuring the underside of spindle nose on the CNC mill, so probably less than 0.001mm?

Simple geometry shows that an error of 0.01mm at the spindle nose multiplied by the length of the Haimer and holder (120mm) gives an error at the measuring ball of ....... 0.066mm. That seems pretty close to what I'm seeing, not sure it's a coincidence.

The next task is to check the tram of the mill, and if that is ok, contemplate skimming the bottom of the spindle nose to remove the runout.

The info on the Renishaw probes is interesting, but I'd like to make clear the Haimer is not a Renishaw style probe. If nothing else Renishaw probes are waaaaay outside my price range. The Renishaw website doesn't mention prices, so I can't afford them!

Regards,

Andrew

Quite so, but presumably not including the spindle? Perhaps Wolfie can elucidate on this point.

Regards,

Andrew

Hi Wolfie,

Good grief, that's a mess and a half! The 'notches' on the end of the cutter are a consequence of the clearance angles ground on the flutes. The fact that 3 of the notches are missing is down to non-existent quality control. Personally I'd say that the poor finish is due to poor quality cutters.

My advice would be to bin them and buy a cutter of a recognised professional make and try it before mucking about with mill itself. Assuming the workpiece is steel, what feeds and speeds were you using?

One other point; has the mill ever produced a good finish, and if so what changed when you started getting a poor finish?

Regards,

Andrew

Hi Pete,

Thanks for the info. I hadn't really thought about the issues of runout when using the indicator on two different mills.

I did a few tests using the indicator on the CNC mill this evening. I zero'd the indicator in one position and zero'd the relevant CNC axis. I then turned the indicator through sucessive right-angles, zero'd it and noted what the CNC readout said. Total deviation across all four points was about 0.04mm. Not particularly good. The runout of the spindle and collet is less than 0.01mm, probably about 0.006mm by estimate. At this stage I'm assuming that most of the extra runout is caused by the holder that the indicator sits in. It uses a single side screw to hold the indicator, which is bound to make it run slightly eccentric. I need to play around and see what results I get. I also need to try the indicator on the Bridgeport, once I've swung the ram end for end to replace the slotting head with the milling head.

At some point down the track I might fiddle with the setting screws on the Haimer, but definitely not at the moment!

Regards,

Andrew

PS: Dooooh, of course what I should have done is noted how the deviation varied with the position of the screw.

On edit: Just tried it again; total deviation now 0.025mm as shown on the computer screen. I'm not doing precision work, so I'd be happy with 0.02mm or better. I need to give everything a good clean and then try again. I'm not convinced about the screen readings versus the slide move either, as the CNC mill is driven by stepper motors and is thus open loop. It'll be interesting to try the indicator on the Bridgeport against a quality (Newell) DRO. The position of the screw didn't seem to matter that much, although I will replace the screw with a home made one. While bought from the US the holder is Chinese, and one thing the Chinese can't do is use decent quality set screws. This one is brass and it's still rubbish. Oh, and the hex socket didn't match any Allen keys I've got.

Edited By Andrew Johnston on 31/07/2012 21:54:51

Hansrudolf: That sounds a very interesting lathe; any chance of posting a picture of the whole machine?

Neil: I assume you mean meaningless, rather than just meaning less? Possibly a more interesting question is what accuracy could be achieved if we assume the gear train is exact. In other words how good is the leadscrew? From a chart I have for Harrison lathes (kindly supplied by KWIL) for between centres of less than 2 metres it quotes less than 0.015mm over any 50mm, ie, 0.03%, and less than 0.04mm over any 300mm, ie, 0.013%. I've also found the following **LINK** on the accuracy of ground taps. For instance the pitch error for a 1mm pitch thread should be less than ±8µm measured over 9 pitches, ie, 9 mm. I make that 0.089%. Most of the thread tolerance stuff in my Machinery's Handbooks concentrates on diametric errors, not specifically pitch errors. May be that is because a pitch error translates to a diametric error?

Now we have some numbers let battle commence!

Regards,

Andrew

Edited By Andrew Johnston on 29/07/2012 22:05:45

Hansrudolf: Thanks for posting the chart. It seems odd that a Swiss lathe should have an imperial leadscrew. I would also surmise that somewhere in the drawing office some-one had a fetish for the number 5, given the sequence of change wheels. The set of imperial thread TPI seems fairly comprehensive, as does the metric, except for some of the metric pitches below 1mm. May be it was assumed that one wouldn't be screw cutting small diameter metric threads? How big is the lathe?

Regards,

Andrew

Jason: Thanks for posting the pictures. As far as I can see the circuit is a pukka capacitor discharge ignition system. It's a pretty elegant piece of design; clearly done by somebody who knows his volts and amps.

Here's an annotated copy of one of Jason's pictures, identifying the various parts:

From the left, the input filter capacitors provide some filtering for the incoming DC supply. I assume that the LED indicates when firing occurs. One of the resistors marked 'Rs' is probably associated with the LED as a current limit. Together the transistor and transformer make what I assume is a blocking oscillator. This uses an auxilliary winding on the transformer to provide the positive feedback necessary to create an oscillator. One of the resistors 'Rs' will be to initially kick start the oscillator when power is applied. The transformer will have a third, step up, winding to provide a high voltage AC output. This is rectfied by the diode and feeds the high voltage capacitor with DC. I assume that the small yellow filter capacitor near the high voltage capacitor is to control RF emissions. The coil is obvious, and the SCR (silicon controlled rectifier) is triggered to complete the circuit, discharging the high voltage capacitor into the coil. I assume that the SCR is a logic level device, and can be driven direct from the hall effect sensor. The resistor and capacitor 'RC- gate' are to provide a pull down on the SCR gate to ensure no mis-triggering, or triggering on noise, of the SCR. All in all a pretty neat circuit.

Here's an interesting link detailing the pros and cons of CDI. **LINK**

Regards,

Andrew

Edited By Andrew Johnston on 28/07/2012 22:18:40

Same as Jason here. I don't differentiate between roughing and finishing as far as the tool is concerned. I use mainly insert tooling. Standard depth of cut for roughing is 0.1" to 0.2" off the radius, for finishing more like 0.01" to 0.04" off the radius. I will, however, normally reduce the feedrate for finishing cuts, from anywhere from 4-12 thou/rev for roughing down to 2-4 thou/rev for finishing.

Regards,

Andrew

The cunning plan is now coming to fruition. After my unsatisfactory conversation with the UK distributor a further internet search revealed a metrology company in St. Albans that advertised all the Haimer kit, for purchase online. Their prices were much more reasonable, basically Euros converted to £, give or take a few percent. A couple of days after ordering I get an email to say everything was out of stock, will be available at the end of July, do I want to cancel or wait. I wasn't entirely surprised; I had wondered how a company operating out of a small industrial unit could afford to hold stock of all the items shown on the website. These included high end slip gauges and hardness testers; not cheap items! Anyway, I said I'd wait, and then had second thoughts, as Haimer had a note on their website saying that prices would be going up at the beginning of August. After a 'phone call to the company I got an email stating that the price would not go up; the price I'd been quoted would be the price I paid.

I got all the items ordered, minus one additional probe, this week. So far I've only had time to unpack and look at the items, not had a chance to try them yet. First impressions are that the build quality is excellent; really nice bits of kit. Nicely packaged, complete with short manuals and a calibration certificate.

I am now looking forward to sending an email to the UK distributor 'explaining' why they didn't get the business!

Regards,

Andrew

Power drawbar and tool setter have finally arrived, compressor bought, delivered and tested, in the comfort of the hall; just need to move it into the workshop. Pneumatic fittings, filter/regulator/lubricator and piping bought, mostly from J&L, but I used Ebay for the 1/4" NPT fitting for the drawbar itself. Extra electrical sockets and cable bought so I can add some more mains sockets on the CNC mill where needed. Just need to get a UK mains to US mains adaptor to fit the PSU that came with the drawbar.

I've done a quick test with the tool height setter to make sure it works. Now I need the time to sit down and use it properly, with a master tool, and try filling a tool table. Need to find time to fit the drawbar too!

Regards,

Andrew

Jason: Neat units, but they seem remarkably cheap for what I understand a capacitor discharge ignition to be. Normally a capacitor is charged to high voltage (usually 400-600V) and then discharged through an ignition coil in the normal configuration. They're doing pretty well to include a boost converter at that price! May be it just charges a capacitor up to the battery voltage?

Regards,

Andrew

So is that because you do not know, or are you merely teasing the rest of us to see if we know?

Suppose we want to machine a worm to mate with a given gear. Clearly the worm is a thread and so can be screw cut in the conventional way, if we knew what the 'thread' pitch was. Working in DP, the gear has N teeth on its pitch circle diameter (PCD), where N is a positive integer. If we 'unwrap' the pitch circle we get N teeth along a straight line, where the distance from one tooth to the next is the pitch. However, the length of the line is PI times the PCD, which will be an irrational number. Hence the pitch will also be an irrational number. Using a gear pair that removes the effect of PI (at least to a good approximation) simplifies the remainder of the gear train calculation by taking out the irrational factor.

Regards,

Andrew

On edit: Of course if both the worm and gear are under one's control you can cut the worm at a rational pitch and adjust the DP of the gear to suit.

Edited By Andrew Johnston on 27/07/2012 13:13:17

Jeff: Errr, if the capacitor goes short then the circuit will not work at all. Operating the points with a short across them will have.........no effect! If, however, the capacitor goes open circuit then I agree that the circuit will continue to operate. Removing the capacitor changes the way the circuit works. So your conclusion that the capacitor is not critical for basic operation is not correct, as removing the capacitor changes the basic operation of the circuit even if the end result (a spark) is the same. There, that's another hair split.

Generally capacitors do not fail short, at least not for long, with the exception of tantalum capacitors, which nearly always fail short. I hate 'em!

Regards,

Andrew

Simple enough, use a ratio of 22/7, or as it is implemented on my lathe, 88/56, give or take a factor of 2. The ratio 88/56 equates to PI/2 within 0.04%. The ratio 355/113 is a better approximation to PI, but possibly a bit less practical in terms of gears.

Regards,

Andrew

Jason: If I make it to Stratfield I'll be sure to bring my oscilloscope and high voltage probe; which is almost bigger than the 'scope!

I agree that the energy in the coil at the point at which the contacts open is (1/2)*L*I**2. Once the contacts open we have a series resonant circuit, so all of the energy will transfer to the capacitor, and then back to the inductor and so on. The energy stored in the capacitor is (1/2)*C*V**2. In this sense the value of the capacitance determines the peak voltage swing, which is what we're after. Here I disagree with Russell; the stored energy does not cause the spark, it's the induced voltage in the secondary that creates a spark. Once the spark is formed and a current flows, then the stored energy is transferred to the spark, and that energy is what ignites the air/fuel ratio. Think Van der Graff generators, these produce pretty impressive sparks by generating high voltages, but very little energy is involved. There are two different parameters at work, one the voltage needed to breakdown the spark gap in the first place, and second the amount of energy needed to ignite the mixture once a plasma has formed in the spark gap. As a first approximation they are not related.

I agree that the capacitor reduces, or eliminates, sparking on the contact points, but it has a second, equally important, role. Since the capacitor has prevented contact sparking by stopping the rapid collapse of the magnetic field in the inductor we need another way to generate a high voltage to cause the spark gap to breakdown. The capacitor provides this by resonanting with the inductor, creating peak voltages that can be many times the supply voltage. Think here Tesla coils where very high voltages are induced on a secondary winding by using two loosely coupled tuned circuits to generate voltages much higher than might be supposed by just looking at the supply voltage and turns ratio.

Of course all bets are off once we have a secondary winding and spark gap. When time permits I'll try replacing the inductor by a transformer in the simulation.

Regards,

Andrew