Help with 'Speedy' valve gear and other issues...

[Roger]

Feb 19, 2021 8:51:27 GMT timb said:

Hi Roger, you have probably though of this but we always had lots of issues with pickup on 0-10V analogue signals and had to go to great lengths to ensure a smooth signal. Have you plugged anything in recently that has not been there before? Could it be that the cable screen or a screen connection has failed? Harmonics on the inverter??

I dont envy you finding the cause.

Tim

Hi Tim,
Nothing has changed on the machine for over a year. It could be an internal dodgy connection going open circuit. I've seen other frequency converters run away when the analogue signal went open circuit. I'll have to disconnect that and see what happens.
I'm also very suspicious of Windows updates changing things. The Ethernet connection has started throwing the is error, and it makes me wonder if it's a software issue.
I'm trying to get answers about that on the Warp9 forum.

[jon38r80]

+1 for being suspicious of Windows updates. Their 'bugfixes' sometimes cause more trouble than they solve.

[Roger]

I've got a pretty old version of Mach4 and I'd like to use the latest Ethernet Smooth Stepper plugin, so I think it's time to bite the bullet and upgrade to Mach4 Version 2.
Wish me luck, these things usually end up being a tortuous saga so I avoid them unless I really need to do them.

I've reconfigured the main spindle motor drive to run at 1600RPM max instead of 3000RPM which I'm never going to use. Bear in mind that the pulleys will double that figure in the highest ratio.

I've also prodded and poked this while its running and that revealed nothing.

The analog feed to the spindle drive was also disconnected and the drive runs down to a stop, so it's got nothing to do with that.

I'm inclined to think this is a software issue, hence the upgrading. The first thing they want on support Forums is to know that you're running up to date versions, so to get much help I've got to go down that road. such is life.

[Roger]

I've installed the latest version of Mach4 and the latest Ethernet Smooth Stepper plugin. That took a bit of getting to work again. First it worked, then it didn't. Then it worked but forgot what input lines were connected to the limit switches and the enable became undefined.
So it's been the usual struggle but it's back up and running again.
I did find something that didn't make sense with the Emergency Stop input. That was floating for some reason, it should have had a 24 pullup, so it's been working like that since 2008!
Anyway, I fixed that and managed to get one of the High Speed Spindles running again. It's not the quietest, the bearings have suffered from the thrashing they got, but at least it looks like I can carry on. I've clipped the High Speed Spindle to a pedestrian 12000RPM which is actually enough for most things.

I'll be watching it like a hawk for a while to see if it behaves now. It's been suggested that I turn on logging, and I'll do that if I remember when this job is finished. I'm told that the output shouldn't be changing, but the logging ought to catch it if it does, and maybe that will give a hint as to when and why it happened.

I'll be getting another batch of bearings, £600 worth, so I'll be able to repair the other spindles when they arrive.

So, it's not all gloom and doom, it can always be worse!

Back to making swarf...

[andyhigham]

Silly speed spindles usually use air bearings.
More recently turbochargers running at 100,000 to 150,000 are using ceramic ball races instead of oil floating bushes

[Roger]

I baby sat this one through without a hint of trouble, but I'm not holding my breath. I'll believe it's fixed when it's run a month without incident.

Anyway, that's the roughing operation done, I'll run the finishing operation tomorrow.


20210219_230518 by Georgia Montgomery, on Flickr

[Roger]

Feb 19, 2021 21:55:44 GMT andyhigham said:

Silly speed spindles usually use air bearings.
More recently turbochargers running at 100,000 to 150,000 are using ceramic ball races instead of oil floating bushes

It depends on what you call Silly speeds. This was my business for ten years! We used to repair the German Precise SC53H spindles that ran at 80,000RPM on greased bearings! The front bearing was a tandem pair to take the loads, with an ID of 10mm. It takes about four days of careful running in to build up to the maximum RPM. If you rush it, the balls skate on the grease before it's been worked away from the ball tracks and it destroys them. They're mighty expensive and on 16 week delivery, so you have to take your time or go out of business.

You can run ball bearing spindles with an oil mist, but that requires more equipment. Most are still greased, but it's not grease as you and I normally know it. Add too much, and you'll never run it in, however long you try.

The fastest ones we did were Westwind Air Bearing spindles up to 150,000RPM. Nowadays, 250,000RPM is commonplace, and some go to 350,000RPM. Remember, these are not dentists drills, these have shafts that are 15-20mm diameter! It's an absolute nightmare getting the balance good enough, and the cleanliness and alignment issues are daunting. I got out before it became untenable for us to repair the new designs.

[JonL]

What sort of kit requires bearings like that?

[Roger]

Feb 19, 2021 23:25:37 GMT JonL said:

What sort of kit requires bearings like that?

PCB drilling machines. The drive towards ever denser boards drives ever smaller hole sizes. Typical via holes are less than 0.5mm, many are much smaller. Thin flexible circuits can have 0.1mmm holes. The limiting factor for the number of holes you can drill in a minute is the maximum surface speed. You can imagine where that leads as diameters fall. You can drill three or four holes per second, it looks like they're being punched.
If you Google PCB drilling machines, you'll be amazed at what they can do, it's another world. These are the machines I designed the CNC control for when I worked for SEL in Horsham

[jma1009]

Very sorry to hear about Roger's expensive bearing failure...

At £600 pounds to repair is 'ouch'.

Going back a bit, one of the elements of Roger's design work was to establish whether the output/overflow from the gap (or in Roger's case holes) between the combining cone end and the start of the delivery cone with a lightly sprung check valve on same would increase the range of the feed water temperature and still be able to feed.

(Above that to which a standard injector will work).

I haven't seen anything yet to show that this design object has been achieved.

I have watched a number of times Roger's YouTube clips, and I have been very impressed. But it is exactly as I would expect of the cone throat sizes etc.

[dscott]

When looking for some problems the correct camera helps.
I am just about recovered from our walk down our Road during the cold time. Showing up major weaknesses
in insulation with heat escaping everywhere.
Yes our latest toy being a thermal imaging camera now the price has dropped to buyable.



Shows a motor getting very hot. We got up to late 50s yesterday but no camera.
I stopped every half hour to let it cool.
Wonderful to note that previously this heat got transferred down to the head and this is where I
thought the bearings were at fault for getting very HOT. Altered bigger holes in the base of the
motor are starting to pay off.

Time for an additional fan to be fitted and TLC within walkable distance ordered from.
Plus waterproof boxes for cable management now I have something to attach them to at the
workshop end.

David and Lily.

The camera has lots of other uses which include checking Lily for adequate clothing before
walks. A sweep of the house to check on cables getting hot!! Radiators blocked.
Pipe insulation. Cold toes! And most importantly...

Is the tea at the right temperature to drink.

[Roger]

Feb 19, 2021 23:56:49 GMT jma1009 said:

Very sorry to hear about Roger's expensive bearing failure...

At £600 pounds to repair is 'ouch'.

Going back a bit, one of the elements of Roger's design work was to establish whether the output/overflow from the gap (or in Roger's case holes) between the combining cone end and the start of the delivery cone with a lightly sprung check valve on same would increase the range of the feed water temperature and still be able to feed.

(Above that to which a standard injector will work).

I haven't seen anything yet to show that this design object has been achieved.

I have watched a number of times Roger's YouTube clips, and I have been very impressed. But it is exactly as I would expect of the cone throat sizes etc.

Hi Julian,
The £600 is for 8 pairs of bearings, so it's not quite as bad as all that. The problem with these exotic bearings is availability. If you miss a manufacturing slot, it can be 16 weeks delivery. They appear in catalogues as standard parts, but they are almost always made to order.

I don't think the high temperature performance will be that different with a delivery overflow valve. It ought to help in the extreme cases where it's just about hanging on to feeding at the highest temperature though. I haven't really explored the extreme limits of that. It might only make a few degrees of difference,  but it ought to be there. It ought to improve the thermal efficiency, since any cold being air drawn into the boiler is replacing hot water. Again, how much this represents is hard to tell, but it must make a difference, however small.

There isn't much motivation for setting up an injector for these extremes because it impacts on the pressure it will drop out at. The higher the water temperature you want it to work with, the larger the regulation gap. There's no way of getting around that unfortunately. I might have a play with it, just for completeness, but I don't think it's terribly useful in our climate. I know I can make it pick up at 35C, and that's probably as far as anyone needs to go.

The motivation for the valve wasn't for efficiency or high temperature working though. If those are consequences, then so much the better. Presumably these are the reasons why other injectors have two valves. My motivation was purely because there's nowhere to conveniently put the usual valve in the 8X scale injector body. If there had been, that's what I would have done. There are other solutions, but as Elon Musk says, the best component is no component! Moving the valve to the outlet solves most of the problems with scale injector design.

[dscott]

I had a friend down in Plymouth who had jobs of an exotic nature and he ended up in Specialist bearings.
He even collected the Test ones he had done from the assembly.
They specialized in Aircraft ones and the test to destruction of course. If you pay enough they will rotate at high speeds
for 14 hours at minus 10. Then heat up as the aircraft sits on some Tropical island airport at 35 degrees.
Failing is not an option... And yes they all had a roasted look but they all still worked.

Also admitting to a love of Tropical Islands in search of Steam Locomotives Of Course.
He also had a Philippine wife!!!

David. Lily is asleep.

[delaplume]

The bearing at the output end of the power turbine in a jet engine must get a wee bit hot as well I'd imagine ??..........

[Roger]

Here's the first of the mirrored Injector halves, the other is currently being roughed out. Against all expectations, I stripped one of the sick High Speed Spindles, washed out the bearings, and reassembled it with a generous squirt of slideway oil. I was expecting it to sound like a bag of nails, but it's sweet as a nut! I might modify this spindle with a couple of oiling points and run it like that instead of with grease. It's not a particularly demanding application, and I don't have to rely on flaky customers doing the maintenance.

20210220_115919 by Georgia Montgomery, on Flickr

Following on from Julian's post, I've run a few more tests with much higher temperature water to see what happens. I filled the tank with water at 50C and took the following wobbly video. I put the tripod away since I thought I'd finished with filming.

Water that hot cools down surprisingly quickly, even though it's 22C in the workshop today, so the temperature had dropped a few degrees before I got round to filming. I had to increase the regulation gap significantly to get it to work, which is hardly surprising. I think you would need to optimise the Steam Cone throat size to better match the conditions if you wanted to run at these elevated temperatures all the time. I guess that would be the case if you were designing a Backhead mounted Injector.

You can see that there's a little overflow going on, but it's still feeding happily. Maybe it will go even higher, but for the moment I think this proves the point.

Do the conventional injectors we know and love work at these temperatures? I somehow doubt it, but I stand to be corrected.

Looking at Bob Bramson's book, he states "As the feed water temperature increases, the quantity delivered decreases and at about 100F ro 110F (43.3C) most of the feed water tips out of the overflow pipe." This may well be true if the regulation gap has been set for lower temperatures, but does not appear to be true if the regulation gap has been made larger to accommodate hot water.

D.A.G Brown's book states that feed water of "40C results in  delivery temperature of around 95C, so you can start to see the reason why  these conventional injectors cannot cope with high feed water temperatures; boiling starts to take place at the point where the liguid stream should be going into the delivery cone". Why injectors with an overflow valve for the Delivery Cone should perform better isn't exactly clear. One draws in air while the other doesn't. However, you could argue that the air cools the water and stops it boiling. You could also argue that the lower pressure you get by having a valve ought to make the boiling happen sooner. My guess is that drawing in air is much worse for the performance of the injector than the effects of stopping the air getting in. I'd be really interested to have a definitive answer to this perplexing question.

In any case, neither books claim to feed water well at 48C which I demonstrate below, so I can only assume that they can't do it.

I make no special claims for the performance of this design. Any Injector that prevents air being drawn in to both gaps ought to have similar characteristics. You certainly could set up an Injector to run up to these temperatures, but it would probably be tiresome to have to regulate the water manually to get it to run dry at more usual temperatures.


20210220_144634 by Georgia Montgomery, on Flickr

[Roger]

Here are the completed halves, with the burrs removed and degreased ready for Silver Soldering.

The water and steam channels are being protected with Tippex because I can't easily remove any Silver Solder that gets in there.

The other holes have been left open because I'll put rings of Silver Solder inside those. I don't want to add any Silver Solder manually unless I have to. I ought to be able to supply enough in the various places to fill the gaps.

20210220_210320 by Georgia Montgomery, on Flickr

I made the mistake on the prototype of just putting the rivets in and forming the heads. That was a bad idea, because it moved and it was difficult to get the halves registered accurately again.

This time I'm using all four M1.4 bolts to hold it while I do the riveting.

20210220_215433 by Georgia Montgomery, on Flickr

I mixed up some thin flux and painted that on the faces before joining them. There ought to be enough space for the water to be driven out when I heat it, but I might put it in the oven for an hour just to make sure.

20210220_220013 by Georgia Montgomery, on Flickr


Quite a lot of Silver Solder ended up on the outside on the prototype, so I'll cover the bulk of it with Tippex if I remember!

Anyway, so far so good, I'm pretty pleased with them. The challenge is to finish them without making a mistake.

[92220]

They look brilliant Roger. Superb bit of machining!!!................or are they secretly, castings?...........just joking of course!!!!

Bob.

[Roger]

Feb 21, 2021 10:19:45 GMT 92220 said:

They look brilliant Roger. Superb bit of machining!!!................or are they secretly, castings?...........just joking of course!!!!

Bob.

Hi Bob,
I'm glad you like them, but the credit must go to the clever sods who write the software. I've been on the other end of it, writing code for a PCB CAD product and also the control system for a PCB drilling machine. Believe me, using it is a doddle compared to writing it.
Hopefully they will indeed look like castings, I'll leave the witness of the joint to give that impression. Hopefully they won't leak!

[Roger]

Here I've covered the parts I want to stay clear of Silver Solder with Tippex. I've removed the M1.4 bolts because they're not needed and there's no point in wasting them. I've wrapped two short lengths of Silver Solder wire through each of the bolt holes so that when it melts there's a good chance of it getting into the join and flowing inwards. I need the two halves to be totally Steam and water tight, and that includes under where the lugs are. I've also bent up a thick piece of wire and threaded that into the water outlet holes and there's also a 'U' shaped piece in the overflow hole.

20210221_093933 by Georgia Montgomery, on Flickr

Almost ready for heating, flux needs to be added to the bare metal parts.

20210221_094428 by Georgia Montgomery, on Flickr

I must have had to do this on the prototype, but I'd forgotten. Anyway, not all places showed signs of Silver Solder, so I decided to turn it each way and add a little to each of them for good measure.

20210221_095431 by Georgia Montgomery, on Flickr



20210221_095858 by Georgia Montgomery, on Flickr



20210221_100129 by Georgia Montgomery, on Flickr



20210221_100938 by Georgia Montgomery, on Flickr



20210221_101148 by Georgia Montgomery, on Flickr



20210221_104238 by Georgia Montgomery, on Flickr

Everything looked fine except for this place which hasn't got any Silver Solder in the gap.

20210221_104429 by Georgia Montgomery, on Flickr

So that went back for another go...

20210221_105209 by Georgia Montgomery, on Flickr

... which sorted that out.

20210221_112347 by Georgia Montgomery, on Flickr

I've cleaned these up with wire brushes and a needle file on the joint all the way around to make certain that I haven't missed anywhere.


20210221_122606 by Georgia Montgomery, on Flickr

So far that looks ok.

20210221_122629 by Georgia Montgomery, on Flickr

[delaplume]

Hi Roger,

I think we are all holding our collective breathe on your behalf in the hope that all is well internally....Fingers crossed !!