Fine leveling adjustment means?

[S KParticipant @sk20060]

Dear All,

I wish to be able to level a plate or bar relative to gravity via a very fine adjustment with near-micrometer sensitivity. Only one axis (vertical, i.e. on one side of a bar or plate relative to another) is particularly important, and only a small range of adjustment would be necessary (a few degrees).

The leveling adjustment needs to be done to say a thousandth of an inch. A micrometer head or a very fine-pitch screw could be used. For example, I can obtain stainless-steel ball-pointed 80 or 100 threads per inch set-screws and associated brass threaded inserts that look appropriate for the task.

My question is how to arrange things to maintain rigidity. Lateral as well as vertical rigidity is needed. I can't get away with simply setting and adjusting the screw(s) on some surface as one might do with a tripod. I've thought about adding hold-down force via spring tension or pressure, but I'm not sure that would be good enough.

The easiest (and crudest) way I can think of is using a fine-pitched threaded rod or screw (seems 32-48 tpi is the maximum) with two nuts to capture the floating element that needs leveling. The other end can can be similarly captured, but is just slightly loosened during adjustment, as it mainly acts as a hinge. This is how my bubble-level is adjusted, so perhaps it's OK? But it still feels unsatisfactory compared to just turning a micrometer head or equivalent.

Is there a better way of doing this?

Thank you.

Edited By S K on 01/02/2023 19:04:17

Edited By S K on 01/02/2023 19:20:15

[S KParticipant @sk20060]

[bernard towersParticipant @bernardtowers37738]

You say that level to a thou or so can you not turn two pillars of identical length (hollow) and bolt together through these?

[S KParticipant @sk20060]

Thanks for your response Bernard, but it seems I left something out:

The leveling needs to be done relative to gravity, not merely to some other surface. So it would need to be leveled each time it's moved. That's why an easy to use screw-type adjustment is desired. I've corrected this in my original post.

Thanks again.

Edited By S K on 01/02/2023 19:20:51

Edited By S K on 01/02/2023 19:22:29

[John HaineParticipant @johnhaine32865]

Differential screws?

[S KParticipant @sk20060]

Ah, yes!

The ones I can find are only 20 tpi, though, so a quarter-turn would be 0.0125". In metric, the finest is M5x0.8, so 1/4 turn would be ~0.08". I have to dust off my trig, but it just might do if the baseline is long enough.

Edit: There are carbide insert screws that are finer-pitched. I see one that is 1/4-28, with adequate thread lengths, or 5/16-32 with marginal thread lengths. This could work, though the range of adjustment would be rather tight.

Thanks!

Edited By S K on 01/02/2023 20:02:07

[John HaineParticipant @johnhaine32865]

To be clear, I meant a screw which has different threads at each end, one screws into one part while the other screws out of the other. So if the pitches were 1mm at one end and 0.9mm at the other, 1 turn would move the parts relatively by 0.1mm. I've seen these use a lot on (electron) microscopes. The trick could be to find two standard pitches that are very close, quite possibly one metric and the other imperial so their pitch difference is very small.

[S KParticipant @sk20060]

Oh! I saw those but didn't understand the point – thank you for the explanation.

Ok, I'm seeing 1/4-20 to 1/4-28, but this is a coarser difference, and these are claimed to be intended to mate differing threads rather than the fine adjustment purpose. Anyway, I'll explore the options.

Thanks again.

 

Edited By S K on 01/02/2023 20:15:27

[Rod RenshawParticipant @rodrenshaw28584]

If you are equipped to tap and thread your own screws the ME range of threads includes sizes such as 1/4 " x 40 TPI, so one whole turn will move the screw 0.025" ( like an Imperial micrometer.) If you need finer than this then John's suggestion of differential screws is worth further investigation. Some combinations will give very fine adjustments,

I am not completely sure from your post that you have entirely understood the working of a differential system, Apologies if I am wrong.

Rod

[Frances IoMParticipant @francesiom58905]

using a LH 1mm thread on one end and a RH 26tpi on the other – assuming perfect threads would give a very fine adjustment yet the individual threads could be quite substantial and taps/dies readily available (the weight of the moveable plate should avoid backlash)

Edited By Frances IoM on 01/02/2023 20:46:33

[John HaineParticipant @johnhaine32865]

I think you need the same thread handedness at each end? You rotate the screw, it moves into one part and out of the other but different amounts.

[S KParticipant @sk20060]

Yes, I did understand: Using two slightly different same-handed threads would amount to a vernier-type adjustment mechanism.

Now I'm also thinking of a combination: A fine set-screw (e.g. the 100 tpi one) to set the distance, and an opposite-handed differential screw to lock the two surfaces together (the base and the pivoting part).

As John Haine will surmise, this is to level a pendulum that is hanging on knife-edges.

Thinking through some math: My machinist's level (6" Starrett 98) has a purported resolution of 0.005" per foot. Over a 6" baseline, that would target 0.0025" as an accuracy limit at the adjustment thread. At 100 tpi, that would be a quarter-turn, which seems practical. If the adjustment error is half that resolution, then the error from vertical at the bottom of a 24" pendulum would be 0.005" (or up to twice that – not sure how Starrett is quoting the accuracy). It's hard to say if that's really good enough, but it doesn't sound horrible, and in any event it's all I would get from the Starrett.

Edited By S K on 01/02/2023 21:09:48

Edited By S K on 01/02/2023 21:10:18

[Martin KyteParticipant @martinkyte99762]

The MK3 Quorn uses differential threads for the rocking arm adjustment by having a thread within a thread arrangement. A threaded adjuster fits in a threaded hole and has a knurled collar to turn it. Another threaded rod passes through the centre of the adjuster but with a slightly smaller TPI. If the central threaded rod is constrained and the hollow adjuster turned then if it is screwed in the central rod will be screwed out the movement being the difference of the pitches. All threads are right handed.
So if the central thread was constrained in a lower plate and the hollow adjuster threaded into an upper plate the separation could be adjusted to a fine degree. 3 such adjusters would give you what you require.
Alternatively you could just purchase a theodolite tri bracket, a device that has been doing the job for years.

regards Martin

Edited By Martin Kyte on 01/02/2023 21:16:28

[Robert Atkinson 2Participant @robertatkinson2]

The adjustable mounts used in optical systems use micrometer heads with either a ball end or using a flat end sitting on a ball bearing. The adjustable part is held down to the fixed base using relatively strong springs. The pivot being another ball sitting in machined Vs or a pairs of short hardened rods e.g. needle rollers. So for a bar you would have the micrometer on the center line at one end and a pair of ball and roller bearings near the edges at the other end. To stop angular movement the V or rods on one side are at 90 degrees to those on the other side.

If you want really fine movement and / or low profile use the micrometer to push a ball bearing sideways along a ramp between the two halves. Again using Vs or pairs of rods to keep the ball in place.

Robert G8RPI

[duncan webster 1Participant @duncanwebster1]

Posted by S K on 01/02/2023 20:04:27:

Oh! I saw those but didn't understand the point – thank you for the explanation.

Ok, I'm seeing 1/4-20 to 1/4-28, but this is a coarser difference, and these are claimed to be intended to mate differing threads rather than the fine adjustment purpose. Anyway, I'll explore the options.

Thanks again.

Edited By S K on 01/02/2023 20:15:27

1/20 – 1/28 is 1/70, how fine can you want it.

[Peter Cook 6Participant @petercook6]

Something like

With an M4 (0.7) thread at the top and an M5 (0.8) thread at the bottom, each full turn of the thumbwheel would lift or lower the top bar 0.1mm 0.004". So one quarter turn would give you the 0.001" adjustment

Edited By Peter Cook 6 on 01/02/2023 22:23:51

[S KParticipant @sk20060]

Thank you. Yes, I could use a micrometer head instead of the ball-tipped fine set screw. And in combination with a spring-loaded bolt, the net result should be easier to use than messing with two counter-acting screws.

At the pivot end, I think a single rod sitting in book-matched V's should do, no? I suppose a spring-loaded bolt at that end would help, too.

[HopperParticipant @hopper]

Posted by S K on 01/02/2023 21:04:54:

As John Haine will surmise, this is to level a pendulum that is hanging on knife-edges.

If that is the case, then surely clockmakers have done this very thing many times before and have established models that could be copied?

Re screw threads, you say you want an accuracy/resolution of one thou. An ME 40TPI thread will give you the same as a micrometer, which can read down to one tenth of a thou. Making your own version would allow for larger diameter dials that would make such resolution even easier.

[Pete RimmerParticipant @peterimmer30576]

What is wrong with just a fine thread screw? Differential screws are all well and good but they have a woefully small range of adjustment. A fine thread screw say 10mm x 1mm pitch will give a range of adjustment which is the effective length of the screw itself and easily adjust to better than the 1 thou you require. In fact, I use a sensitive level on parts supported on 1mm pitch thread screw jacks and find no difficulty adjusting to tenths with a 4" long spanner on the hex head of the jack.

[S KParticipant @sk20060]

Hopper: Knife-edge pivots for pendulums have been quite rare. Only the finest master time-keeping clocks have used them, and even then not often. They are just too sensitive and problematic for general use. But I'm not really trying to make a clock, I'm just making a "precision pendulum" in a fashion simple enough to implement using my meager tools. I've mostly completed the pendulum itself, and I'll start a separate thread once I get just a little further with the mount.

But to your point: Some have used an orthogonal pivot with higher damping to mount the main pivot, such that the pendulum adjusts the orthogonal axis on its own due to gravity. Before I start cutting absurdly-expensive brass (wow, what a shock!), I'll give that a bit more thought, but it's more complex to make.

Pete: Nothing wrong with a fine-pitch screw. It's all the rest, to achieve a rigid setup, that I've inquired about.

[SillyOldDufferModerator @sillyoldduffer]

Posted by Hopper on 01/02/2023 22:42:53:

Posted by S K on 01/02/2023 21:04:54:

As John Haine will surmise, this is to level a pendulum that is hanging on knife-edges.

 

If that is the case, then surely clockmakers have done this very thing many times before and have established models that could be copied?

…

It has been done before, but my books mention it mostly in terms of disadvantages, and don't give any practical constructional details. From memory, the bearing is prone to wear and easily disturbed.

The closest analogue I know of is the old-fashioned Chemical Balance of the beam type, in which the fulcrum is an Agate edge. My school had four, studied briefly because the history of accurate weighing is important in chemistry, but we weren't taught how to use them in anger – chemical technique had moved on. Agate is harder than steel, and I don't know what the edge rested on – maybe a 'V' in another hard gemstone. Each scale lived in a front-opening glass case, I think levelled by four ordinary screw feet; my memory could be faulty!

I don't think chemical balance construction helps SK much because the scale's fulcrum edge was only put in contact with a lever whilst actually weighing. Most of the time it was lifted off. As there are over 30,000,000 seconds in a year, I don't know how long a sharp edge would last swinging a heavy pendulum bob.

On the subject of levelling, the support can be pretty crude as long as it's "good enough". A mechanically sophisticated way of levelling my clock has been on the TO DO list forever, but this is how it's done at the moment:

Though this abomination is levelled within the accuracy of a Wixey, about 0.2°, balancing a pendulum clock on a stack of coins isn't Horological best practice! I only get away with it because my bob is tiny and swings well below the clock's centre-of-gravity, close to the base.

Dave

 

Edited By SillyOldDuffer on 02/02/2023 10:39:03

[HopperParticipant @hopper]

Dave, not familiar with those coins but judging from the copper, that must be the budget model leveller?

Good point about wear on knife edges under repeated to-ing and fro-ing of pendulums. . I wonder if using something like hardened and ground round bar such as linear slide rails, would be more durable yet still minimal friction? Or is the knife edge essential for the latter point.

[Martin KyteParticipant @martinkyte99762]

I do wonder if on a clock with active impulse compensation reducing frictional losses is actually helpful. In order to exert control over the pendulum you actually need a certain amount of amplitude decay as impulsing can only work to increase amplitude. If you imagine the limiting condition where there are no losses there is nothing for the amplitude control to work against.

Very small impulses require driving currents with hard to achieve signal to noise ratios so the frictional losses so long as they are not huge and are constant help to keep a decent signal to noise ratio on the impulsing cuts.

I recently designed a Cryostat to maintain liquid ethane at a constant temperature. The device was a copper cup in a bath of liquid nitrogen with insulation between the two. Two copper ‘ears’ of 2mm wire were attached between the cup and the nitrogen to ensure a reasonable cooling of the cup. The cup has a heater and a temperature sensor and the whole is driven by a PWM PID controller. Without the extra losses provided by the ears the controller had little ‘command’ of the ethane temperature.

I can understand the desire to minimise friction in a totally mechanical timepiece but with impulse compensation I don’t think it’s relevant and may well be counter productive.

regards Martin

[S KParticipant @sk20060]

Posted by Hopper on 02/02/2023 10:57:46:

I wonder if using something like hardened and ground round bar such as linear slide rails, would be more durable yet still minimal friction? Or is the knife edge essential for the latter point.

Yes, the knife edge is essential, but for a different reason: A round pivot would result in the hinge of the pendulum rocking back and forth too, disturbing its performance. Of course, even sharp knife edges have a non-zero radius, and changing knives could result in changes in performance as well.

 

Edited By S K on 02/02/2023 13:24:41

[S KParticipant @sk20060]

Posted by Martin Kyte on 02/02/2023 12:29:17:

I do wonder if on a clock with active impulse compensation reducing frictional losses is actually helpful. In order to exert control over the pendulum you actually need a certain amount of amplitude decay as impulsing can only work to increase amplitude.

You don't have to apply an impulse each swing, you can wait for a certain amount of decay and then impulse. The classic Hipp toggle does this, for example.

In another thread, there's been a bit of a debate about whether it's better to impulse with each swing or more occasionally. I'm not sure it's been completely settled.

[Author]

Posts