Member postings for S K

True. But since 3/4 of a 2" compensating sleeve (as calculated) would be inside the bob anyway, those time delays would have to be lived with. To improve that situation, an Invar sleeve could be used inside the bob, but I don't have the material for that right now and I'm not sure I want to go that far. I need to keep things simple until I have something working.

More progress:

I successfully drilled and reamed the bob - very carefully. That was a bit nerve-wracking.

I did make a 0.5" brass sleeve (3/8" OD) as a compensator, supporting the bob from the bottom. This was per the calculation assuming that the TC of the rod is 1.2 ppm. This will inevitably be wrong, and possibly quite wrong, but I might as well add it since it was trivial to do and it's the best guess I have.

I also drilled a 3mm hole in the bottom of the rod and created a stainless-steel pin (with a very slight taper) to fit into it. This will be the rest for the compensator or bob. It's easy to put in and pull out if/when I change the compensator.

Now I need to clean up and wax or lacquer the bob and the other parts, and the physical part of the pendulum should be about finished.

Compensation below the bob, i.e. with a compensating sleeve, only became practical after suitable low-TC materials such as Invar became available. Otherwise, since common metals all have roughly the same TC (~15-20-ish ppm), the below-the-bob extension would have to be very long, i.e. a quite significant percentage of the above-bob rod length. Grid-irons are fairly extended in length for this reason, too.

After some thought, I'll just include the ~0.5" compensator below the bob anyway, since it's easy enough to incorporate. If I'm going to throw a dart at compensation (because I don't know the actual TC), I might as well aim for the bulls-eye. But I'll think about how to conveniently change it.

Yes, Duncan, I've felt worried about those sorts of potential variations, and commented on how easy it would be to make matters worse earlier in this thread. Unfortunately, the traceable certificate I received for this metal did not quote the TC.

So with this contingency in mind, how would you recommend cutting the bob? I suppose all the way through at 1/4" (the rod's diameter) is good enough for now, as I don't have an easy way to cut a flat-bottomed hole 1.5" deep at the moment anyway.

I did just experiment drilling a sacrificial rod 3" deep to see if it would come out straight when reamed 0.001" oversized, and thankfully it did, so I'm ready to go.

Also, I'm not too stressed about the stability of Invar, infamous as it may be here, since my expectation is that carbon fiber is just as unstable if not more so. Quartz might be the real go-to for both stability and low TC.

I could also go with electronic temperature compensation, as many here are. I already have a high resolution thermometer up and running on an Arduino.

Edit: If supporting the 3" bob from the bottom, the computed TC of Invar would need to be 0.9 ppm for zero net TC. Lower than that would require drilling into the bottom, and higher would require stacking it on an external compensator below the bob. So I think that's a reasonably safe starting configuration, and any compensation for higher than 0.9 ppm can be external.

Edited By S K on 25/08/2023 20:41:57

Actually, Joseph, I still don't understand. What about suspending the bob at its midpoint makes anything at all difficult?

Joseph, I know that's not what you said, I was just taking the argument to a limit. And I know that your scheme can be made to work, too. I just think it's overly complicated since you have to worry about both TC's - those of the bob and of the compensating tube. By leaving the support at 50%, one only needs concern for the compensating tube's length and TC.

But is my proposal sound and mathematically correct, or wrong somehow?

Edited By S K on 25/08/2023 19:54:49

I'm sorry, but there's some fundamental difference in strategies that I'm not following, at least per my understanding of your proposal.

You are placing the adjusting nut at the center of the bob, and the bob's rest location above that (i.e., 75% from the bottom of the bob). I feel this is wrong. In fact, if your bob and your compensating tube is made of the same material (e.g. both 360 brass, as I am proposing), then your scheme will not address the temperature coefficient of the rod at all! It will just leave the bob with its center at the rating nut regardless of the temperature. If it's not the same material, you are just complicating matters, since by supporting the bob at other than its mid-point, you have to consider its TC as well as the TC of the compensating tube.

What I'm proposing: First, keeping in mind that I am making an arbitrary-length, arbitrary-period pendulum: I don't need a rating nut, but merely a support (e.g. a pin), and there's no point in leaving unused rod below that. Hence, in my case, that support can and likely would be merely at its bottom.

Next, I think we can agree that, if we had a perfect rod, then the ideal place to rest a non-zero TC bob is at its center point (50% from the bottom). That way, the bob's own expansion or contraction is a wash, as then the position of its center of mass along the rod would not change.

Now, having nullified the bob's TC (by supporting it at its center), we want to compensate for the rod's non-zero TC. I compute that a 2.05" long compensating tube is needed. This would nominally support the bob at its center, and for my 3" tall bob, it would extend 0.505" below the bottom of the bob.

But if that compensating tube is the same material as the bob itself (as I propose), then supporting it at its center is redundant, since the 1.5" of the compensating tube within the bob would expand or contract exactly as much as the lower 1.5" of the bob itself. Therefore, I can just support the bob - from its bottom - with a shorter 0.505" tube of the same material.

Does this make sense?

Edit: I think this also means that, for a 35" long Invar rod, a 360 brass bob that is 2 x 2.05" tall, supported from the bottom, would not need any special compensation - the bob would move its CG up and down by the appropriate amount on its own. And also that my 3" bob would be appropriately sized for an Invar rod that's ~35/2.05*1.5=25.6".

Also, the math is just a bit more complicated, as the rod has its own CG and, as it changes length, its CG will shift too. So the net CG of the pendulum would still change slightly with temperature unless one works harder at it.

Edited By S K on 25/08/2023 19:41:56

The profile will not be free-standing, and so its stiffness is not very critical. Rather, the profile will be firmly bolted to a solid wood plank, and then to a wall. I'd expect that assembly to be quite stiff, solid and flex-free.

The rod is 36", and with a bit more subtracted here and there in the supports, it will be a little under a "seconds" pendulum. The arbitrary net length will be rated computationally.

I don't understand supporting the bob 2/3 up from the bottom. That would leave a net increase in temperature sensitivity vs. supporting it in the middle. My rough math (may easily be wrong; someone please correct if so):

TC of Invar: 1.2 ppm. Times 35 net inches = 4.2E-5 inches expansion/contraction per degree.

TC of 360 brass: 20.5 ppm. To counteract 4.2E-5" per degree, I'd need the mid-point of the bob to be supported by 2.05" of 360 brass below it. As the 3" tall bob itself is 360 brass and the mid-point is already 1.5" off the bottom, I need an additional 0.55" of 360 brass below the bob. Therefore, I think I can support the bob from below (not the midpoint) via 0.55" additional brass. Does this sound right?

Edited By S K on 25/08/2023 14:29:37

This project isn't intended to generate the best pendulum ever, just hopefully a pretty good one. In part it's a beginner's exercise in machining, and so some design compromises are accepted in that light. I've read Matthys and as much of the other literature as I've found.

I'm close to finishing a bob that is a 3" diameter by 3" thick brass cylinder, weighing about 6 lbs. I wish I could make an oblate spheroid, etc., but that's beyond me and my equipment for now. The bracket seems appropriate for the pendulum's weight, but I'll consider adding a cross-brace for added horizontal stiffness. I've also contemplated silver soldering it in addition to the bolts.

The V is 90 degrees simply because that was the cutter I had, but I also allowed for using the flat portion too, albeit with the bob a bit closer to the profile. I polished the riding surfaces. The cross rod shown in the V is 1/4" stainless, ground to a polish.

I'm planning to bolt the aluminum profile to a 1+ inch thick plank of wood, and then to a wall. Yes, 2" from the bob to the profile isn't a lot of distance, but it will have to do. The profile includes channels which may reduce aerodynamic efficiency, too, but I could cover the channels near the bob.

My next dilemma is whether and how to attempt temperature compensation, and I've hesitated trying to drill a straight and accurate hole all the way through the bob (on my sad, inadequate hardware). The rod is Invar, and I'd need to add a bit over 2" of 360 brass under the bob. I'd like to think of a way to trim the compensation in place.

I've also been thinking of making a new hinge assembly, this time out of Invar too, as my current design will contribute to temperature errors.

No, I'm not painting it! 😄

This is the kind of thing I'd like to finish. I'm not fully satisfied with the sanding. Maybe a bit more tooth?

I think I'll try the suggested wax.

Thank you!

More progress:

I've finished the pendulum's bracket. It's designed to fit 8020 aluminum profile (80x20mm), and the bolt positions were spaced accordingly, with the two supports being 60mm apart. The throat is about 3.25" to the V.

The top of the bracket is intended to allow a flat-placed pendulum cross-rod, with or without the clamp. Of course the V can also be used with or without the clamp. Flatter is better, but the clamp is currently set with a very slight down-angle, capturing the 1/4" cross-rod (better than ejecting it). Taller, well-tuned clamp supports would be needed to clamp a "flat" cross-rod.

Here it is bolted to an 8020 aluminum profile, with a weird perspective lying down on the chessboard. M3 bolt heads are small enough to slip into 2020 T-channels, allowing bolting things from the back. M5 bolts are used with T-nuts to clamp the bracket to the profile.

I chose 8020 profile because it should allow mounting sensors and coils wherever I please, with easy adjustment. With this bracket, a 3" diameter bob will have nearly 2" clear spacing to the profile.

I'm pleased with how it came out. 🙂

Hello forum members,

What are your favorite methods of preparing (sanding, etc.) and finishing brass for a long-lasting bright finish without tarnishing and staining?

Thank you.

Ah, your magnet is oriented vertically, with the coils then applying a twisting torque in the direction of the motion. Now it makes more sense to me that you put it at the top of the rod (and the Helmholtz arrangement that suits it). Unfortunately, I can't easily do the same.

In answer to your second point, I don't know. It's clearly not an optimal solution, but I've assumed that the pair will mostly net out as a single magnet. The assembly certainly appears to when I play around with it.

I could retreat back to a single magnet and a single coil, but my instinctive preference is for a symmetrical arrangement.

Also: I've almost finished the pendulum's support bracket. I now have to decide whether to cut V's or leave the bracket flat on top (i.e., for lower friction during settling). Someone, maybe John, had suggested the latter.

If flat, I'd add clamps, and locating the pendulum in one axis would need to be done manually (no V for it to fall into).

Any comments about this choice?

(I guess I could make both V's and clamps.)

Edited By S K on 23/08/2023 23:32:15

I'm close to having to decide where I want to support the bob. There are three main possibilities: support it from the top (easiest, but worst temperature effects), support it from the middle by a compensator tube (requiring a larger hole from the bottom too), or support it from the bottom.

If both the bob and temperature compensation support is the same material (e.g. brass), then there should be no need to support the bob in the middle. Instead, it's exactly equivalent to support the bob from the bottom and stack it on top of a shorter compensator, correct?

Great observation! Neat!

What does this position repeatability say about time resolution? After all, slowly incrementing a flag across the sensor only differs in velocity compared to swinging it across. I haven't calculated it, but doesn't one micrometer of travel correspond to a few ms in time, and isn't that around what most of us have been seeing?

A very small update:

I've decided on electromagnetic impulsing using Helmholtz coils. Since my rod was not hollow, I made a magnet holder capable of holding two magnets on opposite sides of the Invar rod. These are 1/4" diameter by 1/4" long grade 54 neodymium magnets in a NS-NS arrangement. They essentially forming one larger magnet with the 1/4" diameter rod in the middle acting as a bridge. The magnets grip the rod strongly, but the assembly can be positioned easily:

I noted that the magnets form a surprisingly strong compass, and wonder if that might apply a twisting force to the rod large enough to influence the pendulum? I think I'll close my eyes to that possibility.

Edited By S K on 22/08/2023 01:16:50

Very nice indeed! 🙂

Oh, I see, you have a GPS disciplined oscillator? Nice! Can you provide a link to it? Thank you.

The TDC won't be in range most of the time (nearly all of the time, actually) if the two are asynchronous. They would have to be synchronized to within 500 ns or 8 ms for the device to work, depending on the mode. But I see you plan to delay the 1PPS signal. I'm curious about how you have done that or plan to do it? Up to a second or two is a long time, and jitter in that delay would likely show up too when you are trying to measure to 35-55ps!

And I meant "hard or impractical" as in "basically impossible" in that situation (though there's always some way!).

It's very easy if you are pulsing the opto's input yourself, as SOD was (see the above post). But it's hard or impractical if it's installed in a pendulum, as you note. And the data sheet doesn't discuss that parameter. On the other hand, for a pendulum as a whole, just calculate the RMS variation in the period.

Yes, I know about TDC's. I designed an 8-channel, 16 hit per channel one with 0.5ns resolution in a former life. But my query was due to the rather short range of that device relative to the period of a pendulum, with a maximum of 8ms in mode 2. If the two clocks (reference and pendulum) are asynchronous to each other, as they normally would be, then the difference between the two would not be in range most of the time. That's why I was wondering what use it would be put to.

Edited By S K on 21/08/2023 15:55:02