S Haller “Elo” midget escapement question

[Phil G4SPZ]

I’ve just acquired this clock, which matches Plate 1521 in the Tenth Edition of the Guide, but is definitely a midget and requires a new Horolovar no.44 suspension unit, as the fork and bottom block are missing too.

While this is on order, I’ve dismantled the clock to clean it. I have found that the anchor pin is bent, as is one of the pallet pins; the pallet pin nearest the camera has a distinct upward tilt. I’m sure that pallet pins should be dead straight, so with great care I propose to try to straighten it. However I’m not sure whether the anchor pin should be bent or straight - it’s conceivable that this shape is actually correct.

If not, why would anyone do such a thing? Is it my bad luck, or has every clock out there been ‘got-at’ by someone before me? I seem to get more than my fair share!



This is my first pin-pallet 400 day escapement, so any advice from more experienced practitioners would be much appreciated. Many thanks,

Phil

 

[KurtinSA]

Be very careful with the pin pallets. They are brittle. We had a thread within the past 3-4 months by a member showing how to replace them with a certain size sewing needle.

Sometimes, the anchor pin needs to be bent. Clocks with adjustable pallets should never need the pin bent, but rather proper adjustment of the pallets. For these clocks, and I assume it's the same for pin pallets, the anchor pin should be pointed straight up, with the escape wheel tooth has traveled about 1/3 the way down the impulse face of the pallet. For the pin pallet, that means that the pin will have moved off the lock face onto the sloped impulse face of the escape wheel. The clock works best if the anchor pin oscillates equally about the vertical position.

On your clock the bend is in the middle of the anchor pin. Where does the fork intersect with the anchor pin? My guess is that it is in the upper section above the bend as the lower part of the pin seems to be still perpendicular with the anchor...maybe it's not but looks sort of OK.

Kurt

 

[Phil G4SPZ]

Hello Kurt, and thanks for the quick and helpful reply.

Without the suspension and fork to hand, I can’t try the escapement out just yet, but I will not bend the anchor pin until I have reassembled everything and checked the action.

The pallet pin will hopefully respond to gentle persuasion!

Phil

 

[Wayne A]

Without seeing the rest of the clock its hard to say if its supposed to be bent or not. For example the Schatz Jum/7 pin pallet has the anchor mounted on the side of the EW so its fork pin is not centered in the anchor but its not bent, they mounted it that way. Pictures not the best angle but think you can see what I'm talking about.

 

[Wayne A]

The pallet pin will hopefully respond to gentle persuasion!

It did not look very bent from the one picture so maybe it will be ok. Some have even heated the pins to anneal them, never tried that myself but there's very little load on the pin pallet in normal use. Think most are bent from adjusting the clock and with that in mind I always put a finger on the EW when setting time on a pin pallet clock.

 

[Phil G4SPZ]

Thanks both for the helpful info.

Wayne, you make a very good point about how the pallet pins get bent. I suspect the clock had been reassembled incorrectly, as the hand setting was very tight and there wasn’t a spring washer behind the cannon pinion. “It’s been got at”...

Phil

 

[shutterbug]

If you've already ordered the unit, you might have to redo it with a Horolovar spring. Here's why.
You could also learn how to thin a spring

 

[Phil G4SPZ]

Oh, okay SB, thanks for the warning. I’ve ordered the unit from a UK seller, so we’ll see what happens when it turns up. I’ll try thinning if necessary.

You’ll never guess what I did this morning... dropped the hand nut It’s in the workshop somewhere!

 

[Wayne A]

You’ll never guess what I did this morning... dropped the hand nut It’s in the workshop somewhere!

Spent some quality time looking for tiny parts.. A very long time ago I started using a good sized white towel on the bench to stop the tiny things from bouncing as far.

Funny how the missing bit is always in the last place you look, suggest starting there!

 

[Phil G4SPZ]

I’ve already looked there! I couldn’t see it.

One of the local suppliers thinks he may have some. Several assortments are available on line, but as always the important information is missing.

For what it’s worth, the thread OD is 1.35mm (just under 1/16”) and the pitch is 0.25mm or around 100 TPI.

Phil

 

[victor miranda]

the thread is likely 1.4mm. I have a model train engine with that size screw in it.
nuts should be available...
try here? --> Screws, Nuts, Washers, Springs

I suspect the shipping will hurt.

your photo of the pin pallet anchor looks straight to me.
the shaft and the pins should be parallel.

One set of pins in a clock I had was worn about halfway through the pins.
the pins are usually brittle. bending them is not something I'd try.

victor

 

[Phil G4SPZ]

Hi Victor, and thanks for the suggestion. M1.4 nuts are available here in the UK from model engineering suppliers for just a few pounds, so it gives me another option to follow up. I just need to double-check that the thread is indeed M1.4mm, as it definitely measures 1.35mm across the thread tips.

Phil

 

[Phil G4SPZ]

I found a hex nut to fit amongst a cheap assortment of tiny metric nuts and bolts.

I managed to snap off the bent pallet pin whilst attempting to straighten it. Yes, I know, you told me so... I will replace it with 0.3mm pivot steel which has just arrived.

Phil

 

[Wayne A]

Don't see why they hardened those pins to egg shell brittleness in the first place. Well at least those pins are a good size to work with. Sewing needles work well too, very flexible and already polished.

 

[Dells]

Oh, okay SB, thanks for the warning. I’ve ordered the unit from a UK seller, so we’ll see what happens when it turns up. I’ll try thinning if necessary.

You’ll never guess what I did this morning... dropped the hand nut It’s in the workshop somewhere!

Been there done that.
I may have one let me know the size
Dell

 

[Phil G4SPZ]

0.3mm is a bit fine for a sewing needle, but I’ve got some pivot steel ready to grind to a point and stake in. I guess the hardening is done to prolong the pins’ life, a bit like conventional pallets are hardened. I have seen pins worn half way through but still working, though they were thicker. I liken these 0.3mm pins to the thickness of a bee sting.

Phil

 

[Phil G4SPZ]

Been there done that.
I may have one let me know the size
Dell

Thanks Dell. I’ll let you know if I can’t regulate the clock once it is back together.

 

[Wayne A]

0.3mm is a bit fine for a sewing needle, but I’ve got some pivot steel ready to grind to a point and stake in. I guess the hardening is done to prolong the pins’ life, a bit like conventional pallets are hardened. I have seen pins worn half way through but still working, though they were thicker. I liken these 0.3mm pins to the thickness of a bee sting.

Phil

Needles are available smaller yet. I've used .25mm beading needles to repair Schatz Jum/7's pin pallets.
Attempted to use paino wire once but it was far to brittle once properly sized and ended up with beading needles. Those beading needles are like spring steel, very flexible.
All my pin pallet clocks are midgets and have yet to see any worn pin pallets, just broken off or completely missing ones! Was the wear on a larger clock?

 

[Phil G4SPZ]

Was the wear on a larger clock?

Thanks Wayne for the useful info on the needles.

Sorry, I didn’t mean to imply that I’d seen worn pin pallets on a 400-day clock! Yes, the wear was half-way through the pins on a very old 30-hour movement and the pins were much bigger. I chickened out of replacing them as the clock wasn’t mine and it still ran fine.

Phil

 

[Wayne A]

Yes, the wear was half-way through the pins on a very old 30-hour movement and the pins were much bigger. I chickened out of replacing them as the clock wasn’t mine and it still ran fine.

Phil
Was expecting something like that with allot more force behind the movement. Always have said its hard to improve something that's working fine but easy to make it worse! Company I worked for implemented a large preventive/predictive maintenance program, suspect about as many problems were created as prevented...

 

[Phil G4SPZ]

Well I am a firm believer in “if it ain’t broke, don’t fix it” but in some instances, with clocks, a little preventive maintenance is worthwhile.

Many of the first clocks I repaired were American, and I found them to be remarkably robust and capable of running well even when badly worn. I then tackled a 14-day Ansonia with a Brocot escapement with ruby pallets, and quickly learned that wear and incorrect escapement geometry is virtually guaranteed to stop the show. I also pay a lot of attention to things like bent or worn trundles, and have spent many a happy hour cutting and replacing trundles. The effort is well repaid, although it probably would not be worth it in a commercial context.

With 400-day clocks, I rapidly discovered that there are no quick fixes, and everything has to be exactly spot-on if reliable running is to be expected. Whilst challenging, and often frustrating, this emphasises the need for careful workmanship, on which - as a mere retired hobbyist - I can afford to devote the time required, purely as a learning exercise.

Phil

 

[Phil G4SPZ]

Update. I made a replacement pin pallet by grinding a point on the end of a piece of 0.3mm pivot steel and pressing it firmly into the hole in the anchor. Worked perfectly. The anchor pin is supposed to be bent - I established this by assembling just the EW and the anchor between the plates and trying the action. The clock now runs fine with a new suspension unit and gives around 340 degrees total rotation.

I also solved the lost hand nut problem. Tension for hand setting was originally provided by the hand nut which was not fully drilled through, so designed to ‘bind’ on the threaded end of the centre arbor. There is no spring washer. An ordinary M1.4 nut fitted okay, but kept coming loose. I squashed the nut (very gently) using parallel-jaw pliers. After about five attempts I got a nut that was just tight enough on the thread to stay put and allow hand setting in both directions.

In case anyone else finds themselves working on an “Elo” midget, I’ve discovered two errors in the 10th edition of the Horolovar “Guide”. The clock matches Plate 1521, but the description matches Plate 1522A. Also the Guide states that the beat rate is 10 BPM, but I can assure you that the clock is keeping time at 12 BPM.

Phil

 

[Phil G4SPZ]

...the Guide states that the beat rate is 10 BPM, but I can assure you that the clock is keeping time at 12 BPM...

Sorry, I was talking rubbish. The Guide is correct, the beat rate is indeed 10 BPM. The clock was not keeping time at 12 BPM and was soon found to be gaining at a rate of about 12 minutes per hour. Shutterbug in post #7 was absolutely right, the spring as supplied measured 0.05mm thick against the Horolovar spec of 0.046mm. Not only that, the top block was too thick to fit in the slot, and one of the screws had a stripped thread! Rather stupidly, and having seen no reply from the supplier regarding a refund, I decided that, with nothing more to lose, I would file the top block down to fit, and teach myself how to thin a suspension spring.

Taking the established rule-of-thumb which says for every 0.0001" (0.0025mm) of thickness sanded off, you lose 4 minutes per hour, I calculated that I needed to remove 0.0075mm with a target thickness of 0.0425mm. Sadly, I over-did it a bit, and although the thinned spring measures correctly on my micrometer, the clock now runs about 9 minutes per hour slow. So rather than mess about any longer, I'll now order the genuine Horolovar suspension unit and put the whole episode down to experience.

I discovered after a week that the supplier had offered me a full refund after all - unfortunately his e-mail went into my 'spam' folder and I didn't see it until I'd done the modifications.

At least my replaced pin pallet is still working perfectly!

Phil

 

[Phil G4SPZ]

I do apologise for keeping on asking questions, but I have now fitted a brand new suspension unit 44 supplied by Cousins UK and the clock is still racing. The suspension spring measures 0.046mm as it should, yet the pendulum is rotating at over 13 beats per minute when it should be 10 BPM.

Is it normal to have to thin suspension springs routinely? I’m beginning to wonder if this clock has been fitted with the wrong pendulum bob, but it all seems to match.



Phil

 

[KurtinSA]

Could be the pendulum is the wrong weight. Looks like no one has offered what the weight of their pendulum is. What does yours weight? I don't have a midget like this so I can't help.

Kurt

 

[Phil G4SPZ]

Thanks. I’ll weigh it later and report back. It might be interesting for me to add some small lead weights and measure the effect.

 

[KurtinSA]

You said the suspension spring was 0.046mm which is 0.0018" but the guide calls out 0.0022". I guess you arrived 0.0018" by some of your previous testing and using the rule of thumb. Clearly 0.0022" would only make things worse. Generally, I don't trust what's in the guide anymore and prefer to do some side testing on a little rig I created before I make final cuts on my suspension springs. You are still quite far off. I presume that you have the pendulum adjusted for the middle range...that way if you find a spring that gets you close, you have some adjustment left in the pendulum.

I notice that the guide shows 0.0018" springs for plates 1522 and 1522A. Would your clock be closer to those plates?

Kurt

 

[Wayne A]

That pendulum looks to have solid brass balls so not likely to have been altered. Already have ran across ones with lead in shells being altered, lots of fun sorting that!
Don't have a similar clock to compare but evidently you need a much thinner spring. That suspension unit a horolovar?

 

[victor miranda]

this post is going to be me sharing my conclusions.
the Horolovar book appears consistant when you use horolovar suspension springs.
when I read that Horolovar sells springs that perform as a 023 and that the spring itself is not that exact dimension...
I ordered the size I needed and sometimes did a little sanding.

this is not a comment of what suspension spring may or may not work, rather, attempting to fix this clock
with an explanation for what may be both being asked and why it is asked.

I have no tools that can measure a suspension spring accurately, the thickness is one of two
dimensions that can affect spring performance. so one does end up evaluating the spring in hand
no matter what dimensions it may have.

your clock may have a pendulum that didn't get evaluated into the book.

all that to say, were I in your place, I'd order a thinner spring (I think, from your description.)

victor

 

[Phil G4SPZ]

Kurt, Wayne, Victor, thank you all for your interest and helpful input.

To clarify, my clock exactly matches the appearance of plate 1521 (10th Edition of the Guide). However, the description of plate 1521 is, I believe, wrong as it is described as a miniature clock taking 0.0022" suspension unit 43. My clock is definitely a midget, and unit 43 would be much too long. However, plate 1522A looks nothing like my clock, but has what appears to be the correct description for my clock - 4 ball midget with pin pallet escapement taking 0.0018" suspension unit 44. So my conclusion was that the plate images and descriptions have simply been reversed between 1521 and 1522A. And what of plate 1522B? It has no description at all...

Wayne, the suspension unit is as supplied by Cousins UK and is not a Horolovar branded item, although it is claimed to be the direct equivalent of Horolovar unit 44. I have checked the thickness of the suspension spring with a micrometer, and it measures 0.045mm as close as I can read it. I took the precaution of also ordering a pack of three genuine Horolovar branded 0.0018" (0.046mm) suspension springs. Using the same micrometer, these measure 0.050mm...! I do recognise that the compliance of a spring is a function not only of its thickness, but also of its width and of the material from which it is made. Hence a 0.05mm thick Horolovar spring may actually be more compliant than an 0.045mm thick spring of another make.

Victor, the 0.0018" spring is the thinnest that Cousins sell in the UK, so I can't order a thinner one! And Kurt, my clock looks nothing like plates 1522 or 1522A; it is exactly like plate 1521, no doubt about it!

I have several options. I know I can sand the springs I have, to thin them down to the point where the clock can be regulated. I don't mind doing this, but I had hoped to be able to fit-and-forget this new suspension unit without having the chore of repeatedly sanding, measuring, trying... I read in a previous Forum post that someone used the technique of suspending the pendulum outside the clock and sanding the suspension till the correct period was achieved, in my case 6 seconds per swing or 10 BPM, which will save me some time so I'll copy that idea.

Thanks again everyone for your advice,

Phil

 

[KurtinSA]

Phil -

That was me regarding swing tests outside the clock. I'm doing that right now as a matter of fact. I bought a Becker at auction and the bottom block broke off during shipping. The guide says the spring is 0.004". I tried that thickness...it's too slow for 8 beats in 60 seconds. I have no option but to pull out a 0.0045" and begin sanding, reattaching the top block and rehanging my pendulum after a number of sandpaper strokes. I was about 2 seconds slow for the 8 beats and even moving the weights as far in as possible only got me close...not a workable situation. So, a-sanding I will go!!

Kurt

 

[Phil G4SPZ]

Oh thanks Kurt, very helpful advice.

Phil

 

[Phil G4SPZ]

Well, it's now at long last working and keeping good time. In my (limited) experience, the smaller the 400-day clock, the more difficult it is to fix... this one has proved a right awkward cuss. Not helped of course by the Guide contradicting itself about the spring thickness.

I received the second suspension unit from Cousins and fitted it without checking, or tried to. The open-ended fork was too tight and resisted my attempts to widen it, so I fitted another with a closed end. The new spring was 2mm too long, and despite apparently being of the correct thickness, 0.0018" or 0.046mm, it still had to be sanded in order to achieve 10 BPM and ended up about 0.043mm. Thanks to Kurt for the tip about doing the swing test out of the clock between sandings. I found this a quick and straightforward method. I clamped the top block in a swivelling vice with the bob hanging over the edge of the bench.

Unlike most of the 400-day clocks I've attended to, this one seems to have plenty of power at the escapement and tended to flutter. I had to raise the fork a fraction above the Guide setting to tame it, but it now seems to be behaving itself, so I've treated myself to a glass of something amber by way of celebration..!

Sometimes, the anchor pin needs to be bent. Where does the fork intersect with the anchor pin? My guess is that it is in the upper section above the bend as the lower part of the pin seems to be still perpendicular with the anchor...

Kurt, you were spot-on. The bend is necessary to keep the pin rocking equally either side of the vertical.

Phil

 

[Wayne A]

Phil
Nicely done. Just have to sand suspension springs every now and then.

High powered 400 days can have a very infrequent flutter, I'd keep an eye on the clock for a couple of weeks. Unless very closely watched flutter can be missed and end up offsetting flutter with rate. If a new clock goes fast on me I always check the rate externally as a way to see if its fast or fluttered.

Wayne

 

[Phil G4SPZ]

Nicely done... High powered 400 days can have a very infrequent flutter, I'd keep an eye on the clock for a couple of weeks. Unless very closely watched flutter can be missed...

Thanks, Wayne. I have another miniature that does this, but only when fully wound-up, so I wind it about three-quarters full and it runs perfectly.

This Haller has kept going all night, the pendulum rotation has grown to 270 degrees and it has lost a few minutes, so I’m tentatively confident that it’s going to be okay once I’ve finished regulating it. It is fully wound.

Phil

 

[Wayne A]

I have another miniature that does this, but only when fully wound-up, so I wind it about three-quarters full and it runs perfectly.

Phil, thats manageable for sure. I don't fully wind any of my clocks as well because I'm trying to keep the going spring in its linear range. Just always striving get my clocks to keep the best time they can.

Wayne

 

[Phil G4SPZ]

After losing steadily for most of the day, the clock is suddenly 5 minutes fast, so it is obviously fluttering, although it never does it while I’m watching! I can either (a) raise the fork a fraction of a millimetre, or (b) unwind the mainspring a bit to reduce the power. I’m going to try (b) first!

Phil

 

[Wayne A]

Once your into the linear range of the Going letting down more is not going to matter. Probably have to raise the fork also helps to keep any fork to pin gap to a minimum.

 

[Phil G4SPZ]

Thanks for all the advice on this, Wayne.

I didn’t unwind the mainspring as the click is covered by the dial pan on this clock and I couldn’t be bothered to dismantle it again, so I raised the fork. I watched it for half an hour... and spotted one flutter!

I’m not sure why this is happening, although the very thin suspension spring doesn’t offer a lot of damping to the anchor, and the fork tends to ‘bounce’ as the EW tooth drops onto the exit pallet pin. I replaced that pin with pivot steel, which may be more springy than a hard pin. The escapement action is otherwise smooth and there is good lock, and the total rotation is around 300 degrees.

Overnight, the clock has lost one minute - I’ve deliberately rated it to run slightly slow - so it may have settled down.

What I’ll try next if necessary is to increase the pallet depth slightly by lowering the back pallet arbor pivot hole, which is in the top bracket. There is a bit of adjustment available in the bracket fixing screws.

Phil

 

[victor miranda]

I claim to be an old hand at fighting flutter.
raising the fork a little ... often works, and at the cost of making beat critical
and often simply stopping the clock.
the other possible solution, if easy, is to move the anchor pivot a little closer to the escape wheel.
Give that a try after you set the fork to 'normal.' This is more suited to a deadfall type escape
because it is attempting to get better locks. it has a similar effect on pin pallet types.
the next 'fix' is to tighten the anchor pivot. think "no slop" as exactly the correct amount of slop.
the escape wheel pivot may need the same attention.
the last solutions are effectively reworking the escape wheel...

victor

 

[whatgoesaround]

If you are getting sufficient overswing, wouldn't it be easier to just raise the fork again? Adjusting the pallets is something I only use as a last resort, unless they have been obviously moved before. They were set at the factory and unless moved are generally in optimum position as it is.

 

[Phil G4SPZ]

If you are getting sufficient overswing, wouldn't it be easier to just raise the fork again?

Yes, but the fork is already almost at the upper tip of the anchor pin! Can’t go much higher...

Over the past 20 hours the clock doesn’t seem to have fluttered at all. I plan to let it run unattended for a couple of weeks and check the timekeeping at the end of ghe period.

I really appreciate everyone’s advice, thanks all.

 

[victor miranda]

hohohohohoho.
install a longer anchor pin!

lift the fork a little. ...very little... and point the fork down a bit.

victor

 

[Phil G4SPZ]

...point the fork down a bit...

That’s a good point. I’ve thought long and hard about the interaction between the pin, the fork and the proportion of the suspension spring between top block and fork. Angling the fork up or down adds a new variable and another interesting dimension! I see from the Guide that some clocks are supposed to have the fork at an angle, up or down, although most including mine are shown level.

As things stand, the clock seems to be running very well. Total rotation is 315 degrees with equal overswings of 90 degrees, keeping good time and no apparent fluttering. I have posted a short clip of the fork’s motion here:


Phil

 

[victor miranda]

In my testing or frustration... I can tell you, you can do a lot of angling and the clock will run.
and flutter...

with 400 day clocks, you can angle some, think 10 degrees is possible, and the clock will run.
for the most part, the only problem is the fork may need to be opened a little.

victor

 

[Wayne A]

Running out of fork adjustment room can be a problem, I've done the angle down trick and also flipping the fork depending if it was a top of the block fork. There is always deepening the locks but on a pin pallet I don't like to do that because is seems to reduce the accuracy of the clock and you end up with a more of a recoil escapement the further down the tooth you go.

 

[victor miranda]

hi Wayne A,
I think I understand. on a personal kind of response,
I think I was advocating a little of deepening the locks.
more as a way to say there is a perfect place for the anchor/EW and
the slight error of being shallow on the locks will get more far more flutter than
that same error being a little deep of perfect.

that said... you have me curious about the 'more recoil' comment.
I had not noticed that. (caveats here... in no way disagreeing... had not made that connection that way.)
and I have to ask you to expand a bit? more recoil is from the fork, uh, lets call it vibrating after the drop?

oooof if that is right... I have to spend some time adjusting an anchor and watching the result.
interesting.

victor

 

[Wayne A]

you have me curious about the 'more recoil' comment.
I had not noticed that. (caveats here... in no way disagreeing... had not made that connection that way.)
and I have to ask you to expand a bit?

Victor, the edge of the escape wheel teeth on pin pallets are angled back a few degrees, so the further you drop the lock the further it has to rebound up the hill to release.
I'm always trying to get my clocks as accurate as I can, its a thing I do so in this pursuit I also find ways to make them less accurate!

Wayne

 

[Phil G4SPZ]

I get the impression that the thinner the suspension, the more prone to fluttering the escapement will be. It seems that flutter is prevented by having a shorter length of spring between top block and fork, which damps the fork’s movement.

When I thinned the suspension, I thinned it all the way up. I’ve read of others who thin from the fork downwards, which in the light of my experience could be a better option.

Anyway, for what it’s worth, the clock is currently keeping perfect time, so hopefully this can be the end of the matter! Thanks for everyone’s helpful input.

Phil

 

[KurtinSA]

I’ve read of others who thin from the fork downwards, which in the light of my experience could be a better option.

What I've "read" from John Hubby is that is not the right thing to do. The critical part of the suspension spring is the small section between the top block and the fork. That's where the thinning really needs to take place. It's important to thin the whole spring, but holding on to the bottom block and stroking away from it is the way to go.

Kurt