I'm not so sure its entirely fair...
David Barnett
...to drag Rob's post from another forum to this one without his consent, but I'll express and reiterate my concerns from this different framework. If Rob or the list moms would prefer I remove this unsolicited (by Rob) reply, I'll respectfully absterse it from this discussion. I'm not criticizing you, Tom, for posting this -- it's an interesting opinion Rob offers. I'm just not sure of forum etiquette and protocol in this case.
This is all interesting material and I'm in agreement with most of it. I'm even sympathetic to the considerable challenges of making affordable diamond honing plates. Where I still have reservations involves the final maximum allowable deviation from flat, especially regarding flattening the backs of broad edge tools. I'm willing to accept my shortcoming in understanding the factors and explanations for how 0.005" maximum flatness deviation can be overcome by production or end-user technique and will endeavor to keep an open mind.
"1) Tolerances are not a target - they are the gate that acceptable products must pass through. Think of them as a passing grade... when you take a test, you're always shooting for 100%, not just a pass. Parts too, range from pass to 100%."
This sounds all right at first until you realize that setting the passing grade of 0.005" for overall flatness sets the bar pretty low, at least too low for me. I understand the concepts involved but am not so sure I understand why we should expect lowered standards from 0.001" or even 0.002" all the way down to 0.005" to accomodate production challenges to achieve an attractive price point.
It reminds me that beginning musicians need well-made instruments to learn upon, and that poorer made instruments can discourage and defeat beginners before they've had a fair chance to experience an adequate instrument and learning experience. If 0.001" earns an 'A', and 0.002" gets a 'B', it doesn't speak well for 0.005", at least not to me.
"2) Flatness tolerances - when objects are machined and ground, they undergo tremendous forces. Metal parts can flex, magnetic chucks and tables can hold a slightly curved piece flat, and tools can wear from the start of an operation to another. Even expansion of metal parts due to heating can cause deviation. Holding a part can also flex it - it will flex back upon release - after machining or grinding. All of these things can work to take a part out of tolerance. However - when a surface is out of tolerance, it's rarely a very localised defect (like a ridge, or a bump) - it's usually a very gradual change over a larger distance."
Although there's nothing here I would disagree with -- every bit of it is true -- it addresses only production challenges and difficulties that need to be overcome in machining a flat honing plate for use with diamond. In other words, none of these obstacles should be left unresolved in the final product, nor can they serve as excuses for lack of final flatness.
"3) Location of variation - where a part has variability can be more important that what they variability is. Does the variation occur where it will affect performance? Is the tolerance on a functional aspect, or a cosmetic aspect? Looking at lapping plates, is a plate out of tolerance within .010 of the edge really defective? (answer is yes - because it's out of tolerance, but practically it makes no difference)."
There are different to think about localization. In one view, the anomaly is small in relation to the total surface area of the substrate, either as a pit or a protrusion, a pit being preferable for honing purposes. If the bump is anywhere near the middle, this demands correction. If at a corner, say, one could presumably avoid it with considerable ease. The bump could also be corrected (flattened) with considerable ease, although this shouldn't be left for the purchaser to have to do on his or her own. It's similar to anvils that have a localized bump where it's out of the way of where things get flattened. In some cases, the bump even gets used as a feature. Prominences and dishes involving broader percentages of total surface are can be more complicated and deleterious.
Unlike anvils, with hones, however, these issues are genereally undesireable. If a higher ridge occurs along an edge, this can dig into and scratch the back of a blade during flattening. It can also raise the angle creating a long angled surface on the tool back where flatness is desired. But even if the aberration is gradual, concave or otherwise, what may not too badly interfere with bevel sharpening, is apt to cause flattening problems, especially on wider blades in 2" widths or greater. Having experienced such problem stones which were out of flatness well below the Lee Valley allowance of 0.005", I'd have to consider anything above 0.001" to 0.002" unusably aberrant and in need of correction.
Furthermore, I'd say the above offered example, where Rob postulates an out-of-tolerance 0.010" deviation, saying "practically it makes no difference", isn't something I can so easily rationalize into agreement. In my opinion, this is not so.
Lastly, under Rob's point 4), he explains:
"Metals in solid form are in a state of equilibrium, but there may still be internal stresses that express themselves over time. Grey iron castings do this - especially if you change the surafce. Think of it as a block of ice with embedded stretched elastic bands frozen in it...stable, until you shave the surface, causing a band to snap (it's just an analogy - not a model!)."
Rob's right, metals move and cast iron benefits from letting in season and settle so the casting stresses can sort themselves out. Some Chinese woodworking machines could do with more homogenous iron and better casting schedules to minimize jointer bed torquing and tablesaws that depart considerably from flatness, although this occurs less now than before more stringent quality and production inspection was implented.
Now Bill Tindall and I avoid much of this distortion by using continually cast grey iron plate, such as Meehanite, which is far more stable when it cools, is more homogenous in its grain structure and settles in a very controlled manner owing to its composition, method of manufacture and manufactured shape. What I'm trying to convey is that iron isn't always just iron. There are broad differences not only in composition but in realization of final product. So thinking of cast iron "as a block of ice with embedded stretched elastic bands frozen in it...stable, until you shave the surface, causing a band to snap" isn't nearly so applicable here. And as Rob said, "it's just an analogy - not a model!"
And as Rob also says, "Sudden shocks, changes in temperature can also cause movement." Couldn't be more true, and applicable both in the case of plain cast iron and continuously cast iron. Machinists and metrologist must deal with this constantly. What's more, it's generally more true for iron than for mild steel, which can be more stable and relaxed that cast irons.
So while I'm obviously still resistant to sharpening plates being called flat with an accepted tolerance of 0.005", I welcome efforts to convince me otherwise. Won't likely be easy and probably unlikely, but I'm willing to listen.