Thread: Hypothetical Lapping Question

Hypothetical Lapping question for the brains department.

Been doing a little research on precision tolerances extra lately.
Mainly watching You tube videos.
And I’m finding the whole subject quite interesting and fascinating or is that my age 49 and some bits.

And ,I’m little interested in possibly having a go at the three plate method,
As original develop by Joseph Whitworth.

But ,now I will reveal my ignorance , can the three plate approach be done on say 200 mm long bar stock ?
Say 3 mm by 8 mm by 200 mm long,
On the 3mm edge.
Or is this silliness!!

Everything I’ve watched so far basically use discs of some kind.

But, if I was to say , purchase three HSS Bar lengths could the three plate idea be applied to them?
Hope that all kind of makes sense.

Cheers Matt.

In theory the shape makes absolutely no difference, the trick is in alternating the 3 plates. If they're square bits, maybe spin one around every time you change pairs, to help keep the same parts from lapping at any given time. Tom Lipton (Ox tools on Youtube) had a pretty interesting video where his granite surface plates are being recalibrated, and some of the discussions with the guy doing the job gave some interesting insight. One line sticks in my head - "As I'm fixing your plates, my tools are getting better as well."

However, 3mm wide pieces 200mm long? That sounds like a steep muddy hill to be pushing the septic tank up.... Be very difficult to keep from rolling the edges off trying to lap them together, with not much room to slide them over each other.

Better in my mind to make 3 larger plates, lapped flat, and then use one of those to lap your 3mm wide bits, along with some form of jig to stop you tipping them over all the time. But I have no experience other than a basic understanding of the theory.

Hi Simplicity, Guys,

Yes you could lap 3 mm by 200 mm pieces of material but you would need to make up jigs to ensure that all three pieces were kept square, since the jigs would also have to be lapped at the same time meaning that you now have three relatively large areas to lap together. So yes it could be done !

In my opinion very much a larger undertaking than simply getting the pieces surface ground.

I do agree with the comments made by Jekyll and Hyde.

My (limited) understanding is that to do the 3 plate method reliably and optimally requires rotating the plates 90 degrees. Therefore problematic for long thin things like straightedges.

Originally Posted by BaronJ

Yes you could lap 3 mm by 200 mm pieces of material but you would need to make up jigs to ensure that all three pieces were kept square.

That would be the hard part - keeping those surfaces absolutely square. It would be nice if they could be rotated but the key thing is that they are processed until a perfect match, one of the surfaces swapped out processed again, swapped out and so on. The end result is that the surfaces are flatted out by average. From memory, I think Connolly has the method in there.

I may be being pedantic, but my understanding of the 3 plate technique is that one plate is blued, rubbed on another and then the high spots scraped off. Clean and repeat...
Lapping involves embedding an abrasive into a lap, and then rubbing the other part on it.

Back on topic, if you wanted to try the technique out (and I would not mind doing so myself one day), there is nothing wrong with just doing a couple of small cast iron plates. I have a 4"x10" Wing surface plate - just the right size for a small part and a height gauge. I'd be pretty sure that if you scraped in 3 plates of around that size in there would be takers for the left overs as one of the big unknowns of surface plates is how flat are they? A plate freshly made with the 3 plate method would be guaranteed flat.

Michael

Quoting Post 5
“My (limited) understanding is that to do the 3 plate method reliably and optimally requires rotating the plates 90 degrees. Therefore problematic for long thin things like straightedges.”
I reckon Pippin is correct. Unless checked by 90 degree rotation three surfaces can be twisted, yet bluing would show them to be in full contact.

Really appreciate all the feedback, and yes in hindsight 3 mm wide is pushing it, especially as I’m a complete newborn to all this,
(10 mm wide ??)
The other question is ,that the bars have to be turned 180 degrees not just 90 degrees,is that a problem?

One of my reasons for this little exercise is to gain a small straight edge, also being in lock down in Melbourne (Glen Waverley) something to do.
I’m actually currently looking after my father in law who’s 90 at his place
So , it’s a project I can do with minimum gear., tho I have a vice !!

Cheers Matt.

Hi Matt,

Download and have a ganders at this book from

https://archive.org/details/Foundati...anicalAccuracy

Its by Moore & Wright, and goes into detail about achieving a flat surface.

Originally Posted by BaronJ

Hi Matt,

Download and have a ganders at this book from

https://archive.org/details/Foundati...anicalAccuracy

Its by Moore & Wright, and goes into detail about achieving a flat surface.

Thanks Baron,

Will do, cheers Matt.

Originally Posted by Neil317

Unless checked by 90 degree rotation three surfaces can be twisted, yet bluing would show them to be in full contact.

Doesn't matter. First you do 1 and 2 so that you get reasonable spotting, then you do 3 to spot with 2. Lastly you do 1 against 3. Because 1 and 3 will have high spots in the same places, knocking off the high spots will bring them closer to flat. When all three spot each other equally, they are all flat. Rotating is nice but not essential. Straight edges can be done this way too. (To put it a way that can be visualised, if 1 is concave then scraping 2 to it will make 2 convex. Scraping 3 to 2 will make 3 concave. Scraping 1 with 3 will be concave to concave.)

Michael

Originally Posted by Michael G

Doesn't matter. First you do 1 and 2 so that you get reasonable spotting, then you do 3 to spot with 2. Lastly you do 1 against 3. Because 1 and 3 will have high spots in the same places, knocking off the high spots will bring them closer to flat. When all three spot each other equally, they are all flat. Rotating is nice but not essential. Straight edges can be done this way too. (To put it a way that can be visualised, if 1 is concave then scraping 2 to it will make 2 convex. Scraping 3 to 2 will make 3 concave. Scraping 1 with 3 will be concave to concave.)

Michael

That's all fine and good, but remember we're not talking about scraping. We're talking about lapping alone, where you don't have the ability to pick and choose what to remove. The action is guided by the surface underneath it. A high corner will create a (smaller) low in the plate above it, and create a diagonally opposed low of roughly matching size in both plates (assuming square plates). Eventually it should cancel out anyway, but flicking one plate around 90 degrees (or rotating them together rather than oscillating) partway through lapping it to another is going to help get the surfaces averaged out faster.

While 3 plates is the minimum to end up with a flat surface, I believe that in the case of lapping, more plates actually results in better overall flatness (but probably at an overkill level for most applications)

As per post 5, I'm talking about scraping. You can not lap surface plates together, because all that grit that you have pushed into the surface will stay there and ruin your measuring gear.

Michael

I agree with Michael that concavity/convexity would be obvious and correctable using the 3 plate method on long narrow straight edges,
But if one surface is twisted before scraping then I think you would end up with three equally twisted yet neither concave or convex surfaces.
Try this: cut 3 pieces of aluminium sheet, about 100 x 20 x 1.0mm. Twist all three about 15 degrees.
Stack them together, they will sit with good surface contact with each other. Rotating one 180 degrees makes no difference, but with 90 degree rotation the ”error” becomes obvious.
Neil

Originally Posted by Neil317

I agree with Michael that concavity/convexity would be obvious and correctable using the 3 plate method on long narrow straight edges,
But if one surface is twisted before scraping then I think you would end up with three equally twisted yet neither concave or convex surfaces.
Try this: cut 3 pieces of aluminium sheet, about 100 x 20 x 1.0mm. Twist all three about 15 degrees.
Stack them together, they will sit with good surface contact with each other. Rotating one 180 degrees makes no difference, but with 90 degree rotation the ”error” becomes obvious.
Neil

Given the one dimensional motion the lapping would take with such thin plates then you are unlikely to get such a twist. developing.

Originally Posted by Neil317

Try this: cut 3 pieces of aluminium sheet, about 100 x 20 x 1.0mm. Twist all three about 15 degrees.
Stack them together, they will sit with good surface contact with each other. Rotating one 180 degrees makes no difference, but with 90 degree rotation the ”error” becomes obvious.

Yeah but...

If the plates are scraped together, only two will twist one way. The other will be mirror image.

Michael