Evaluate, Repair and Adjust a Second-Hand Micrometer

Good tools are expensive. Good precision tools are very expensive. The first precision tool for most people is a dial or electronic caliper. A reasonably inexpensive caliper may indicate to a precision of 0.01mm (0.0005") and have a claimed accuracy of 0.03mm (0.0015"). The best do only a little better than this, and the worst do much worse.

Next on the slippery slope to precision measurement is often a micrometer. They have a smaller inherent uncertainty than calipers, in part because they satisfy Abbe's Principle, properly used have greater immunity to the influence of operator technique, and generally have an order of magnitude greater precision and accuracy. This all comes at a price.

Many students and hobbyists can't afford top quality measuring tools new, and many can't afford even mediocre tools new. Unfortunately, you get at most what you pay for. Cheap tools don't have the repeatability or precision to do good work, and are often unpleasant to use. Used good to top quality tools can often be had used for less money than an unusable cheapo costs new.

Unfortunately, used precision tools often come with problems, some of which are repairable, while others make the tool scrap. If a measuring tool can not reproducibly measure features to the appropriate precision, it needs repair or should be disposed of, as incorrect measurements may be more trouble than no measurement at all.

There are various opinions about the quality of different makes of tool and methods of repair and maintenance, but for the most part, these are like arguments over pizza topping. The quality makes all do the job well and if the tool meets the requirements for accuracy, precision, and repeatability, that is what matters.

Let's examine the evaluation, repair, and adjustment of a second hand Starrett model 436, one inch range, 0.001" resolution caliper. The tool was bought at a yard sale for US$1. The current (2016) closest equivalent tool from Starrett lists for about US$140 (the 436.1XFL) new. In prime used condition, this tool might have cost about US$40 or more, but this is not in prime condition.

First, lets look at the parts of an outside micrometer caliper, or mike for short, and how it operates:

The feature is measured between the face of the anvil and the face of the spindle. The space between the measuring faces is changed by rotating the thimble. The portion of the spindle that is inside the barrel includes screw threads, and as the thimble is rotated, the spindle and thimble advance or retract. The frame holds the parts in the appropriate configuration.

In use, the spindle is advanced until the measuring faces contact the feature to be measured, and the graduations on the barrel and graduations on the thimble are combined to produce the measure. There are a number of different schemes, depending on the unit the micrometer is designed to measure (metric or US conventional/Imperial) and the precision to which it can be read, and several manufacturers have excellent instructions and videos available for the types they manufacture.

The sample here is an Imperial unit (inch) with a precision of one one-thousandth of an inch (0.001). The spindle has a thread pitch of 40TPI, meaning that the screw will advance 1/40" for each revolution, and each revolution represents 25/1000". There are 25 uniformly spaced marks around the circumference of the thimble that allow measurements to be read to 1/1000".

The micrometer under evaluation is old and beat and rusty. Why would I spend even one thin dollar to buy it? Honestly, because I knew that even if it was irreparable, I can still use it for parts. On the other hand, I had good reason to think it can be repaired, though a few things make it a definite borderline case.

When inspecting the tool, two things come in to play: visual inspection and tactile inspection.

Visual inspection starts with overall appearance. Has it been hit with a hammer? Had it been dropped? Is the frame bent? Is the thimble round or has it been damaged? Are the measuring faces undamaged? Are parts missing?

Visually, this micrometer has a few issues, good and bad. The paint is worn, but there are no dents or dings that imply that the tool was dropped. The damaged ratchet speeder is not making for a warm, fuzzy feeling, but the thimble itself appears undamaged, so it isn't a big worry. The speeder sticks out and they are relatively vulnerable. The manufacturers sell replacements.

The indicating edge of the thimble isn't dented and is concentric to the barrel, so the thimble and spindle end are probably not bent. Though the measuring faces can't be brought into contact due to rust, they are not visibly misaligned. If I could bring them together, I would look for alignment of the spindle with the anvil and see if the faces meet plane-to-plane. Any gap or unevenness in the meet is a negative. Any indication that the frame is bent is an instant rejection.

The big concern is the rust. There is a little on the faces, spotting on the spindle surface, and the spindle is frozen, making the tactile examination simple. The rust on the thimble doesn't matter as long as there isn't heavy pitting.

Tactile inspection would include, if the spindle was free, running the tool through the entire range and noting any non-uniform drag or places where it sticks. The spindle lock and force control device are not key, as they can be repaired easily, and are options.

For the price, for me, this was a definite buy.

The ring for the lock turned freely in the unlock direction and caught the lock pin in the lock direction. That, with the pattern of rust on the surface and little pitting on the spindle, led me to believe that the spindle was rusted in at the exposed end, but the rust wasn't into the threads, so, if nothing else, I could cut the frame and spindle and use the thimble end as an adjustable lathe stop.

The ratchet speeder can be replaced if I want, and, in fact, I have a couple of spares from other damaged mikes.

If the spindle had been pitted or there had been visible damage to major parts, I would not even spend a nickel on it. I would not recommend a tool in this condition as a first tool, even at the price, due to the need for significant repair, and the uncertainty about whether it can be repaired at all. The same tool in better condition, but cosmetically poor, commonly goes for less than US$10.

If the spindle moved through the full range without major hangups, and there weren't major dings in the measuring surface, I would recommend it, with the qualification that it will take complete disassembly and will need to be checked over the full measuring range.

If the micrometer operated at all, the first step would be to unthread the spindle from the frame. Due to the rust freeze, the thimble was removed first.

The method varies from make to make, but most Starrett, and makes modeled on them, have the thimble fit to a precision tapered section on the spindle end. Other makes, such as Brown and Sharpe and Tumico, used other schemes on their older tools.

The thimble is a light friction fit on the taper, which holds it centered and straight, and positions it along the axis of the tool. The thimble can be mounted in any rotational position, so cutting the threads on the spindle and body is easier in manufacture than if they needed to be cut to an alignment. Most makes do something analogous. The thimble is held on by a screw through the end. In this case, the screw is part of the speeder assembly, but micrometers without the speeder and use a headed screw, most often slotted.

That made the first step removing the speeder body. This was done using a pin in the end of the body and soft jaw pliers to turn it. If the speeder hadn't been damaged, I would have used a different soft jaw plier on the outside of the speeder body. It is a standard right hand thread.

It shouldn't need any excessive force, whether a speeder or a headed screw. You can hold the thimble with your fingers. If it requires enough force that the screw might strip or the head break off, then it gets interesting. This can occur even on micrometers in near perfect condition, as over time, even a small bit of moisture or organics getting into the threads will cause them to bind. Since these only get rotated when taking the unit apart, they don't get kept free. In that case, the solution is penetrating oil, maybe a little gentle heat, and gently working the screw. Even if it won't rotate, trying will help work in the oil and, eventually, it will probably come free. Work it both ways (righty-tighty and lefty-loosey)

Once the screw is out, removal of the thimble is easy. Some people put the screw back in about one turn shy of tight, then rap the screw with a hammer while holding the thimble in the free hand or a soft vise fixture. I prefer not to, as, even though it should be a light, sharp rap, speeders and screw heads can be damaged, and sometimes even the threads. I use a pin punch (the punch has a flat end and parallel sides. No taper) to reach the bottom of the screw hole and rap the punch instead. A 3mm (1/8") punch was a good fit here, and I made a fixture to hold the thimble by boring a hole about the same size as its outside through a piece of soft wood and sawing it in half through the hole diameter. One sharp, light rap with a 200g (8oz) hammer while supporting the micrometer frame with my free hand to prevent it dropping.

Next, the wear adjuster nut came off. There was no meaningful rust on the adjustment threads or the exposed spindle threads, making it worthwhile to continue.

This is mostly time. Just time. Oh, and penetrating oil.

Penetrating oils do a great job for this type of work. Rust removal chemicals (derusters, rust converters, naval jelly, acids, lye, etc) and electrolysis should NOT be used.They are much more likely to change the dimensions by removing excess material or growing a surface than a simple penetrating oil. Some of the "magic" penetrating solutions have corrosive properties and can damage precision surfaces. The goal is to not ruin the precision of the tool.

I supported the frame in small vise with some paper to soften the grip (the vise was snug, not tight) so that the spindle was straight up and down.

Then, a drop or two of penetrating oil at the top to get things started. Gravity and capillary action did the work. As the oil works in, another drop. No rush. No oil goes anywhere else. We don't want to trap air or, worse, water in there between two seals made of oil, do we?

When oil showed up at the locking ring, I added a few drops there, and not long after, it worked through the last bit.

While the oil was soaking through, light rotation of the spindle helps, like with any stuck screw. Even if it doesn't move, the force can help open up microscopic gaps that let capillary action do it thing. No tools for this. Just fingers.

It wasn't needed here, but, if the penetrating oil doesn't get through in a few days, a little heat can help. The oil is flammable, so flames are out, but any ignition safe heat source that can be used to quickly warm the barrel an frame where the spindle go through up to maybe 80 or 100 degrees (175 to 212F) will help open up gaps. The frame and thimble will grow from the heat sooner than the spindle.

Once the oil worked through and the spindle, begrudgingly, began to free up, I ran it in, not out. Running it in provides the shortest path for getting at least some of the rust out, without dragging it through the threads that give the tool its precision. After several minutes working the threads back and forth, the spindle advanced until it just about met the anvil, and I was able to clean off a lot of the rust. No abrasive, just lint free wipes and brown paper.

While waiting, the inside of the thimble was cleaned. Bamboo skewers and cotton swabs do most of the work here, with a little oil. It is a scrape and wipe job. Abrasive paper can be used, but it is awkward to get in, and getting rid of the residue can be difficult. The outside was cleaned with steel wool and oil.

Once the spindle was substantially free and as much rust was off as practical, little light machine oil was applied to the exposed spindle and the and then the spindle was unthreaded from the frame. As it was removed, there was no binding or rough spots. If there had been, then those are the places to stop, put a little oil on the threads, and run it back in. The key thing is minimal force. It should come out freely. Once out, it was cleaned to remove the last of the rust using the same lint free cloth and brown paper technique. The threads were clean, so not a big deal.

If there had been rust on the threads, then that would have been the end of the job. Rusty threads make it impractical to repair, as it will never operate smoothly and accurately.

Once the spindle was removed, the lock was removed. On this model, it slides straight out of away from the frame. The keyway at the front of the frame. The assemble comes out as a unit, and may require a little finesse. The lock has a spring cut in it so it has friction to hold it in the frame slot. I generally use a small blade type screwdriver as a wedge behind the lock ring. The target is straight out without cocking the key in the keyway. A little care prevents loss of the lock roller.

Once removed, the parts were cleaned with a brass brush and mineral spirits.

The lock was adjusted and its parts reassembled. The parts should stay together on their own. Assembly can be difficult, as the lock should be sprung open just enough that it has a light friction fit in the ring, and the roller may take a bit of finesse to insert because it is held between the ring and the lock by the spring force of the ring. Do not operate the lock when the spindle is not through it. If this has ever happened, the lock will be sprung small, and the micrometer will never operate correctly. The lock can be opened up using a small blade screwdriver to carefully spread the gap. Test fit the lock assembly over the spindle. It should slide freely up the spindle. A minimum amount of light machine oil is appropriate. Not even a drop. I apply it with a sewing needle. If it is dripping wet with oil, the oil attracts and holds dirt.

Cleaning the inside of the frame is a repeat of cleaning the inside of the thimble, only in a tighter place. Mineral spirits to get oil and residue off and light oil, cotton swabs, and bamboo skewers to remove residue, grime, and rust. Abrasives should not be used here. The alignment and size of the bore influences the accuracy and repeatability of the tool. Cleaning of the anvil was done at this time.

The anvil had a small ding in the edge that raised a burr on the working face, so the burr was dressed with a fine, dead flat Arkansas stone until the burr was gone and the face flat. I check this using an optical flat. Carbide faces don't have this problem, but, instead, they chip. In either case, this is not desirable. Any damage to the face limits the utility of the tool and can affect the accuracy. This will need to be lapped later, and the spindle face needs it as well. Lapping the faces is really something that can be done most accurately and economically by a professional, but this takes money, not time, so it waits until all other issues are addressed.

After thorough cleaning and drying of all parts, the spindle was test fit to check the thread fit. If it didn't go in dead freely, then the fingers for the wear adjustment could be gently spread using a small screwdriver in the gaps. This must be done carefully, as if they are not spread evenly, the spindle may not be properly aligned. In this case, the spindle was free.

The adjuster ring was put back on- it has a taper in the threads, so goes on the right way much more readily than the other- and adjusted a fraction of a turn at a time until there was no detectable play in the spindle. It turned freely through the entire range with no play, so the adjustment was good and thread wear is not an issue.

The spindle was removed and the lock ring reinserted. The same spring slit that makes it awkward to remove makes insertion a tender job. Be sure the key is aligned to go in the slot, the assembly is lined with the center of the frame, and press it in. It may take a little force. Watch through the bore for alignment. It probably won't line up right, but get it close. Do not turn the ring, or the lock may compress making spindle insertion impossible.

When the lock is in place, reinsert the spindle carefully. A minimal drop of oil on the outer surface near the lead end is all of the lube it takes. Watch through the lower end of the bore since will probably hang up entering the lock. A little finagling of the lock assembly will generally get it in, only to need a little more to get back into the last bit of the frame bore. I line the lock up on really annoying ones using an old spindle with a taper ground at the lead end. This is a precision grind and polished smooth, not right off a bench grinder.

A small drop of oil on the threads and run the spindle in. Don't bring the measuring faces together yet, because before we can put on the thimble and zero the tool, we must set the barrel.

If you are lucky, you got an adjusting wrench with the micrometer, or, like me, have the appropriate one. The wrench has a pin that goes into the adjusting hole in the barrel, allowing the barrel to be rotated against a fair bit of friction to fine tune the zero.

I didn't use the wrench until I was sure it wouldn't be broken. Older mics often have barrels that won't turn easily, or are even rusted solid. In support of my suspicion there might be a problem was the condition of the adjustment hole. It had been damaged. Note that manufacturers have used many, many schemes for adjustment over the years, though this is the one that became most common by the second half of the 1900's. So, back to the vise and a different sized hole in a split block of wood.

Once the barrel was lightly held, I tried to rotate the frame in it. The barrel turned in the holder. A little more pressure and a little working back and forth, things began to loosen up. Too much freedom isn't good since this friction needs to hold the final adjustment. It needs to be able to be moved without being so tight the wrench pin will break.

Once the barrel was free enough to use the wrench, I rotated it so the index line (the longitudinal line the 25/1000 marks extend from) was where I want it.

After insuring the measuring faces are clean, the spindle was run in to just contact the anvil. Then, a light wipe of oil on the spindle end taper, wipe it off to leave a bare minimum film, and place the thimble on lining its zero mark with the index line. Light axial pressure to hold it in place while running the screw in the end to lock it. Once the screw is finger snug, the spindle is backed out a turn and the screw snugged up. Now the tool can be tried out. If the thimble is straight and concentric, no rubbing is going on, and lines close to zero when the spindle is brought in to the anvil, then it is backed off and the crew given a final tightening. It isn't a big screw. Tight enough doesn't need a lot of torque.

At this point, final adjustment is done. With the spindle in contact with the anvil (again, after cleaning the faces) at the appropriate measuring force, the barrel is tweaked with the adjustment wrench until the index line matches the zero on the thimble exactly.

Calibration is where we establish a relation between the reading of the tool and an actual measurement. We don't adjust anything. That is done. Now we want to know: how good is it?

With a micrometer, the base point is the zero. when we bring the spindle in contact with the anvil, it should read zero. If all other characteristics are good (thread lead, flat and parallel faces, etc), this is the only point that need be regularly checked. The questions are: Does it zero? If we open it up and re-close, does it still? In this case, the answer is: on each of ten repetitions of the process at a temperature of 22 degrees (+/-1 degree), the unit zeroed to within the width of the index line. This is good.

For many situations this is more all that is needed. Since the measuring faces are less than perfect, and I have no history of the device, I also checked several other factors. Using a set of gauge blocks, sized over the range of the micrometer and with size differences that bring the spindle to different rotations, I checked for parallelism of the faces and appropriate thread lead.

To oversimplify, and without propagating all of the uncertainties involved, the error was less than 3/10000" for all test points, and the cumulative error was about 2/10000. This is pretty much consistent with the faces being a little out of parallel, and the thread being a little worn at one end of the range. Until the faces are lapped in, this is suitable for use at the lathe, but not for final check to less than 1/1000". There are more comprehensive checks that could be done, but the errors that have shown already don't warrant going there. This tool was intended to be used as a demonstration piece, and that it is, to some extent, usable is a benefit. Many of the "economy" micrometers fail pretty much every inspection point out of the box, and, even in need of work on the measuring faces, this is more accurate and more repeatable than a number that have come across my bench.