Member postings for Andrew Johnston

As previously stated the value for the addendum is arbitrary. Gear teeth can have a multitude of forms depending upon arbitrary parameters. The Victorians attempted to standardise gear teeth so that gears from different manufacturers would fit together and also allow for standardised machines and tooling. It's no different than Whitworth threads - there is no theoretical basis, it was simply an amalgam of common practice at the time. Similar to threads there are preferred values of DP and Mod for which cutters are readily available.

Much of the design of gears in Victorian times was done graphically, with supporting equations, rather than algebraically. Search for odontics, which substitutes a circular arc for the involute. The "button" method of making a cutter is nothing new. Much of the original work on standardising gears was done by Brown and Sharpe in the US.

As mentioned, it only gets worse for other gear types. A lot of the parameters for straight bevel gears were specified by the Gleason company in the 19th century. It's no surprise they sold machines to cut straight bevel gears. But there are competing standards, such as the octoidal tooth form. Spiral bevel gears have at least three different tooth forms created by Gleason, Klingelnberg and Oerlikon.

Some years ago, when I was exhibiting at a ME show, I got talking to someone who worked for the Mercedes high-performance transmission division. I started quizzing him on gears, but he simply said, we specify what we need and leave the details to the gear manufacturer. We're only scratching the surface of gear design in this thread.

Andrew

The PCD is fundamental to the design of spur gears. If the gears had no teeth, but were simply cylinders, then the speed increase, or decrease, will be set by the diameters of the PCD. The DP system is defined in imperial while Module is defined in metric. The PCD is the circle from which the gear is designed. In the DP system the addendum (part of the tooth above the PCD) is 1/DP while the dedendum (part of the tooth below the PCD) is also 1/DP, in theory. In practice the bottom of the tooth space needs clearance so the dedendum is increased. Common factors are 1.125, 1,157 and for gears where one has a small tooth count 1.25, although any value can be used as seen fit by the designer. The common values give a total tooth depth of 2.125, 2.157 or 2.25 respectively all divided by DP. The tooth depth (in inches) aka cut depth is usually marked on DP cutters as it saves knowing which clearance value was used.

In the DP system the OD of the gear blank is (N+2)/DP where N is the number of teeth. Simple algebraic manipulation will show that the equation is equivalent to the PCD plus twice the addendum.

Spur gears are simple compared to other types of gear. Gawd help us if we ever get onto other types of gear!

Andrew

I don't think it matters if the disc is at an angle. As the tool rotates it doesn't move vertically or horizontally relative to the disc, and so it's height above the work doesn't change either. Suppose the disc was canted over by 45° and the tool was perpendicular to the disc. Now take the disc away and we have the tool in the position it would be in a conventional flycutter, ie, at an angle to the axis.

The important constraint, apart from the spindle axis being perpendicular to the table, is that the axis of the flycutter is parallel to the spindle axis. Concentric is good too, but not essential.

Andrew

That's a weird way of looking at gears. The base circle is the circle from which the involute is developed. Depending upon the gear specification the base circle can be larger than the root circle. Which leads to the question of what should the tooth form be between the base circle and root circle where the involute does not exist?

I don't need to suck it and see, my granny already taught me to suck eggs.

Andrew

Working again for me. Seems they ran out of money, so I've just made a fairly substantial donation. It's by far and away the most useful of the forums I am on.

Andrew

I don't generally use flycutters for flat surfaces, but I use a homemade one for curved surfaces:

I've never thought about angles; I just grind as per a lathe tool, in all cases, and it seems to work fine.

Andrew

Darn it, summat else to put on the list of things I don't need, but that would be nice to have. After all if you don't have it you certainly can't use it.

When I started making gears I spent some time faffing about with gear verniers:

I didn't find them easy to use and the results were indifferent. Now I just calculate the theoretical values and machine to those.

Andrew

You can't measure CP with calipers. Circular pitch is defined as the length of an arc on the pitch circle diameter. A caliper measures between two points on a line, which is also not likely to intersect the PCD at the point of measurement. Using a caliper to measure across multiple teeth doesn't give CP either. There are published tables that give theoretical values measured across multiple teeth to get pitch and backlash, but they don't give CP.

Andrew

Looks more like a cylindrical grinder.

Andrew

Don't confuse quantity with quality.

Andrew

Worked fine this morning, but now seems to be emulating a dodo.

Andrew

For me it's easier to use standard ballnose cutters and let the CAD/CAM do the grunt work. These prototype 10 tooth, 6DP, pinions have undercut as part of the CAD model:

Andrew

Thanks for clearing that up, I understand one DP or Mod, but any number of teeth, per cutter. I originally interpreted the post as one cutter for all DP or Mod, which is why i got confused.

In theory it should be able to cut small numbers of teeth as well. The undercut arises automatically. In some of the 1930s books I've got on gear cutting there are formulae for calculating whether a hob will undercut or not, even when not needed.

Andrew

Not sure I understand that? Do you mean it'll cut gears of different tooth count for a given DP or Mod value, or that one cutter will work for multiple values of DP or Mod?

Andrew

Which is nothing to do with respect, but simply a lead in to egregious nonsense. Had Chris had the wit to read my profile before posting it would have answered his question about engineering. We're left with the thought that amateur has several meanings, but amatuerish only has one meaning.

I am a professional engineer, but largely self-taught as a machinist. That doesn't stop me wanting to learn and try new things, and at least aspire to professional standards. If Chris is happy stuck in his back shed that's up to him, but it's not for me.

As is often the case the comment on power factor completely missed the point. The question was about which motor to turn on first. I think that the OP made an assumption that the spindle motor would be taking less current than the pump motor as it starts off load. However, since the power factor of an induction motor off load is poor the spindle motor phase currents may be higher than expected, and possibly more than the pump motor at full load. The VFD controls current, so that's the important parameter.

Andrew

A) As per Dave, select the drive according to the input VA. In star each phase winding is connected from a phase to a star point. The star point is nominally equivalent to the neutral but is not connected externally. In delta there is no star point, windings are connected phase to phase.

B) Select the drive accoriding to the next size up from VA of both motors.

C) When I looked recently I couldn't find any 1-phase 240VAC in to 3-phgase 415VAC VFDs offered by the industrial manufacturers. There are some "digital" drives advertised as 240V in, 415V out, but it's not clear what they are. There's no industrial demand for such a unit. If you get 240v in and out and run the motor in star then the torque will be down to about 60%. If the motor can be connected in delta then the phase currents will be the same as for 415V/star, so the torque will stay the same.

D) It does matter. At no load the power factor of an induction motor is appalling, around 0.15-0.20, so significant phase currents will flow even though the real power is low.

Andrew

It doesn't, it's the sine of 14.5, not 45. I suspect that 14.5 was chosen as a PA not as a result of experiments that showed it had the characteristics posited by Brian but because the sine was a round value, at a time when calculators were unheard of and many people wouldn't have known how to use trigonometric tables.

Mathematically a rack does have a PCD, it just happens to be infinite.

Andrew

This thread has brought back some memories. I was never seduced by the matchbox radio, digital watch or calculator kits, although a number of my school friends built them, with varying success. The Sinclair approach to QA was to let the customer do it. In the 6th form I had a Sinclair Oxford calculator, which was ok until the LED display failed.

I remember the Acorn Atom and Sinclair MK14, along with a myriad of other offerings. I think Acorn had an office in Market Square in Cambridge while Sinclair had an office nearby in King's Parade. About that time I had a college set in King's Parade, above Campkins Cameras and looking out over college. At this time I bought a Tangerine computer kit as it was more 'professional' in the sense that it had a series of Eurocard boards such as memory and disk controllers which fitted into a 19" rack with a backplane. Tangerine were based in Ely, where I visited when I couldn't get the processor board to work correctly. It turned out to be a duff IC socket. Oddly their chief designer became my manager for a short while some years later when I was working in the mad world of motor racing.

A friend of mine bought the ZX80, ZX81, then a Spectrum and finally the QL. The first three were innovative designs albeit with some faults. If you knocked the memory pack on the back of the Spectrum it momentarily lost power along with anything in the memory. The QL was a disaster, the original manual had most of the pages missing, ie, not yet written. Nobody in their right mind would trust business records to a mini cassette player. It's ironic that Sinclair failed in his bid to go up market with the QL at the same time as Acorn failed to go down market with the Electron. My friend visited the Sinclair "repair centre" to get his ZX81 fixed; it was a man in his shed in a back garden somewhere in Cambridge.

I think Clive Sinclair used to live in the second house in Madingley Road in Cambridge. it's a large house built in a lovely honey coloured stone backing onto college playing fields. It's now offices.

I never had a HP calculator so didn't really get into RPN, although I did have a brief fling with Forth, which is a similar stack based programming language. In my first job after leaving academia we managed to convince some of the secretaries that reverse polish was a technique so advanced that it didn't appear in the Karma Sutra.

The punch up in the Baron of Beef was mentioned in a retrospective in The Telegraph today; the other protagonist was Chris Curry, a founder of ARM. In the late 80s I was a regular drinker in the Baron of Beef as it was round the corner from the above mentioned company. In those days it was a proper spit 'n' sawdust pub.

We used a QL in the above company as it contained a 68000 processor, which was the family I was using in the design of active noise control systems. So the QL allowed the software creator to do basic testing before I'd got the hardware finished.

I watched Micro Men a good while back and really enjoyed it. No doubt there's some exaggeration, but I suspect the story line is basically true. it's really annoying that it never seems to be repeated.

Man, i must be getting old what with all these trips down memory lane.

Andrew

I suspect that the fact that:

sin(14.5° ) = 0.250

is no coincidence regarding the use of 14.5° as a pressure angle.

A hob does not generate an exact involute, it produces an approximation consisting of a number of facets. Of course in many cases the tooth form created is perfectly satisfactory. But it interesting to note that in the professional world hobbing is often regarded as a roughing operation. Look at the literature for a geared head lathe and great play is made of the fact that the gears are shaved and/or ground as finishing operations.

Andrew

Edited By Andrew Johnston on 17/09/2021 20:21:54

The helix was cut on a CNC mill for historical reasons, but everything else (hob and parts) were machined on a manual lathe and a big old horizontal mill that cost me £175.

Andrew