Latest (Cool) Work Project

The Incident

My department at work is responsible for some test equipment, among other things. One of these pieces of equipment uses a 62kW electrically operated steam generator to test... some products that we're developing. This piece of equipment lives in another building across the street from my office.

One day, my boss is across the street, checking up on this machine. It just seemed like a typical day until he sent me an email titled "Houston, we have a problem!". It had this picture attached...

What you're looking at there was a bank of six 40 amp solid state relays, in groups of three. Each group controlled the three legs of the 480 volt power going to two of the boilers four heating elements.

The Investigation

Once we finished laughing, we had to figure out what happened.

Upon inspection,  we discovered that one of the legs of one of the heating elements had shorted to its sheath (and ground), and immediately sucked more current than the poor relay could safely handle. Poof - all the magic smoke got out. (Probably more like BANG!)

How could this have happened? Wasn't there overload protection for the relays?

Nope. Unfortunately, the contractor that designed and built this piece of equipment for us didn't use the correct size fuses to protect the relays. The fuses that were upstream of these relays were sized for the conductors running to the relays.

The Solution

In addition to the obvious issue of the barbecued electrical enclosure, there had been recurring hiccups with this machine that had been bugging us. This was the perfect opportunity to hopefully resolve some of those too.

First things first: Add over-current protection to everything...

New SSR enclosure

The 50 amp fuses that were upstream of the SSRs (Solid State Relays) were replaced with 40 amp fuses. These will act as a last line of defense for the SSRs.

Circuit breakers, sized for the conductors and heating elements, were placed between the SSRs and the associated heating elements. This meant that each SSR fed a pair of 3-pole breakers that were tied together with Edison combs. These circuit breakers act as the first line of defense for the SSRs, and as the over-current protection for the conductors to the heating elements.

All of these upgrades (and the ones that I haven't described yet), meant that the old enclosure just wasn't going to be large enough. The old, charred enclosure was removed, and a new 24" x 30" x 8" NEMA1 enclosure was mounted on the adjacent wall.

Power was fed from the new 40 amp fuses, out of the boiler through six 8 gauge conductors to the new SSRs. After running through its various components in the new enclosure, it went back to the boiler via six 12 gauge and six 10 gauge conductors (three for each heating element). All of these conductors were initially run through 1 1/4" EMT (more about that later).

Next Up: Monitor the crap out of this thing...

In this new wall-mounted enclosure, I also added 12 current transformers. Each current transformer measures the current drawn by each leg of each heating element, and sends a 4-20 mA signal to a PLC (Programmable Logic Controller) in another enclosure. These signals allow us to see, in real time, the condition of each heating element.

PLC enclosure

If the current begins to drop on one leg, we'll know that the resistance is going up, and it's about to burn out. If the current begins climbing, we'll know that the resistance is going down, and the element is either about to short between two legs, or short to ground (both really bad), and that it needs to be replaced.

The internal temperature of this enclosure (and the PLC enclosure) is also being measured via a pair of thermocouples. This information allows the PLC to control fan and vortex tube cooler in the large enclosure to maintain the temperature within the operating limits of the SSRs. If for some reason the temperature climbs above the safe operating temperature, the PLC will kill power to the SSRs, to allow them to cool down.

This PLC (a Ethernet Click by Automation Direct) monitors all of these inputs, along with some others that allow for more pinpointed preventative maintenance of the boiler and the test equipment.

A 6 inch touchscreen HMI allows operators to see whats going with the machine, and change settings on the fly.

A side note about Automation Direct: I've been using their hardware for about 6 or 7 years now. While I've always preferred their stuff over the other guys (I'm looking at you, Rockwell), their Click series PLCs, and cMore Micro HMIs are absolutely amazing. They're so easy to work with, and super inexpensive. That's a great combination, because you can put them in just about everything.

Finally: I mean really monitor it...

Since this PLC is on the network, and I'm a sucker for brute-forcing my way through problems, I decided it would be cool to write an application that would communicate with the PLC and display information back at my office.

<Nerd Alert!>

What I ended up doing was writing an HTA application that uses the winsock networking component (OSWINSCK.dll) to create a TCP connection to the PLC, and poll it via MODBUS TCP\IP.

For anyone who is both interested in programming, and is a cheapskate, HTAs are amazing! HTA stands for Hypertext Application. Its basically a web page, that runs locally on your computer. But, since HTAs aren't technically a web page, they're given access to all kinds of cool things on your computer. You can learn more about HTAs here, among other places.

In my case, I used vbscript (I know, I know: I really need to start using javascript) to use the File System Object to read and write txt, csv, and htm files on my computer; and the Winsock Object to open a TCP socket and communicate with the PLC.

I had to create modbus related functions to poll the PLC, and to interpret its responses. Figuring out how to decode a 32-bit floating point value was... surprisingly easy, once I wrapped my head around how it was encoded.

</Nerd Alert!>

When this all comes together, it allows us to see what the machine is doing, from a quarter mile away, without leaving our chairs! The data on the display updates every .25 - 2 seconds (depending on network congestion). It can also be logged to a csv file for later viewing in excel, where trends can be seen by graphing the data.

It's super cool, and I've been super jazzed to work on it. As of me writing this, I'm still polishing the user interface on my application. Hopefully it'll be wrapped up in the next week or two.

Lesson Learned

Finished installation

All of the conductors between the boiler and the SSR enclosure were initally run through a single 1 1/4" EMT conduit. I selected this size conduit because that was the size of the knockout in the boiler, and because the fill ratio was allowable by the NEC.

Once the machine was up and running, I discovered that this conduit was hot enough you couldn't keep your hand on it, and roughly 160°F on the inside.

It turns out that once a raceway has over three current carrying conductors, the ampacity has to be derated. In this case, I was way over!

The conductors ended up being split into three conduits: the original 1 1/4" for the six 8 gauge feeds, and then two 3/4" conduits for each of the six 10 and 12 gauge returns. This configuration still required derating, but everything fell in line with the circuit protection already installed and current requirements of the system. 

Crisis averted, and lesson learned.

Thanks for reading!

Turning brake drums

This job really threw the brakes on all the fun!

A friend of mine contacted me a few days ago about turning a couple brake drums off of an equipment trailer. They're big honking things - 14" in diameter and every bit of 40 lbs each.

Tailstock adapter seen here.

My Logan only swings about 10", so it was out. Luckily, I have access to a 16" lathe in my shop at work. Unfortunately, the drums' geometry didn't allow for them to fit in the lathe without some specialized tooling - which I was not in the mood to make.

What I ended up doing, was making an extension for the tailstock quill. It has a male MT4 taper on one end, and a female MT4 socket on the other. It pushes the gauge line of the tailstock out about 4 inches or so.

For anyone that's wondering, the tapers on this adapter were cut with the compound slide on this lathe. I chucked up a drill chuck with a MT4 taper, and trammed the compound against it until I had less than .0005 variation along the length of the taper. It ended up working a lot better than I thought it would!

A shot of setup. That's an 8" chuck for a sense of scale.

This adapter allowed me to chuck the drum on the hubcap bore portion, and support the other side with a bullnose center in the seal bore. The extension gave me enough space between the tailstock and the drum for the carriage to move, but not enough room to do one complete pass inside the drum - doh!

I ended up having to cut about 2 inches worth to the finished diameter. Once it was clean, I'd move the boring bar out farther and cut the remaining 1 1/2 or so. I was able to blend the two cuts pretty well. Because of the protrusion of the boring bar (≈5.5 L/D), I clamped a square aluminum bar to the back of the boring bar. This is a trick I learned from a now retired, never-touched-a-cnc, machinist. This trick works wonders at getting a little extra usable length out of a boring bar without scapping your workpiece.

After all the setup and figurin', the boring operation itself went fairly smoothly... once I got past the learning curve of turning a brake drum. The thing is effectively a giant bell - so it resonates like a mother if you even look at it wrong.

In the middle of a cut. Note the Shop-Vac
sucking up the rust and asbestos dust.

To reduce chatter, I wrapped some rubber sheeting around the drum, and slowed the lathe down to as slow as it would turn - 45 rpm. The slow speed worked wonders at reducing chatter, but it also meant that each cut was unbearably slow! (approximately 2 inches long at .011/rev at 45 rpm = about 4 minutes per pass)

In interesting observation I made while cutting these parts: They cut a whole lot better when they're cold! They were about 35°F when they were first put on the lathe (stored in my truck). When I'd bring one in and chuck it up, I could get away with running it at 90 rpm. Once it started to approach room temperature - I'd have to slow it back down. I don't know what the mechanism is that drives that (if anyone out there knows, please let me know!), but it was interesting.

The surface finish turned out pretty good considering the precarious setup, the material (unidentified iron, based on the chips), and the cutting speed (160 fpm with carbide).

Now that its all said and done, I'm not sure if I'd do another brake drum - at least not for anything short of about 35 bucks a pop.  I have between 4 and 5 hours in these four brake drums. If another set ever comes up, they should go a little bit smoother now that I know whats up.

One about to come off the lathe.

Hopefully, if anyone out there is contemplating do anything like this, you'll know what not to do... or to run away.

Thanks for reading!

Bringing a 10" Jet Bandsaw back to life: Part 2

See part one here!

Making the Motor Happy

Disassembled motor

In the last post, we were waiting for replacement parts to arrive. Well, they've arrived! 

Before we continue, to the left you'll see a better picture of the disassembled motor. The staining from the rust that had brought this motor to it's knees can be seen on both the rotor and stator. You can also see the old bearings stuck on both the shaft and in the end cap.

These bearings were on there! The one on the shaft came off with some light tapping on the shaft while supporting the bearing in a vise. The other bearing was a little more stubborn. A little heat was introduced to the end cap using a heat gun, and a little "persuasion" on the bearing with a punch got it out in no time.

While the new bearings went in without incident, I wasn't happy with the way the motor felt. There was too much drag coming from somewhere - even with the new bearings. I discovered during the second or third reassembly (more about that later) that this increased drag didn't rear its head until the end caps of the motor were fully tightened. The culprit: too much axial preload on the bearings. My solution: take the rotor to the lathe and take approximately .005 off of one of the bearing locating shoulders.

As a side note, the lathe hiding in this picture is a Logan model 200, circa 1946. It belonged to my grandfather for longer than I've been around. He had it shipped back to Maryland from Lompoc, CA when they moved back in the early 1970s. A complete cleaning, repainting, and bull gear replacement is in it's future.

This extra few thousandths of clearance made all the difference in the world. Final assembly with the mounting bracket, driving sheave, cooling fan and fan shroud were next. Easy peasy.

I'd like to share a valuable lesson i learned the hard way during this process: Note the orientation of the stator and rotor during disassembly.  Some motors (like this one, for example) have identical ends on the stator housing. If the rotor goes in backwards, it'll spin backwards. 😖

Tire Replacement

Naked wheels

I ordered polyurethane tires off of Ebay, and they turned out coming from a company called Polybelt. Polybely was kind enough to send detailed instructions and an installation tool (the little dowel, nail, plastic tube thing in the picture to the left) with the replacement tires.

The instructions were pretty straightforward, and are as follows (condensed version):

1. Remove old tires.
2. Clean wheels.
3. Warm up new tires in soapy water.
4. Put new tires on wheels using clamp and included tool.
5. Let new tires cool down and dry.
6. Enjoy

Waaa?

So, after removing the last old tire and cleaning the wheels, I slinked off to the kitchen with my urethane tires in hand (after the boss went to sleep, of course - I didn't want to try to explain why I was cooking rubber on her stove).

In a small saucepan, mix a small amount of dish soap into to a pot of water. Add two polyurethane bandsaw tires for flavor. Once everything is heated to approximately 120°F, remove the pot from the stove and hope no one is watching while you carry the pot back to the shop.

That orange is snazzy!

The instructions recommended clamping the wheel to your work surface while installing the new belt. I can't imagine attempting to get this thing onto a loose wheel. That being said, the clamp and the included tool made installing the new tires much easier than I though it would be.

I don't have any pictures that go with this next bit, but I feel it's my duty to share it - and this seems like as good a place as any...

When the drive wheel was going in, I realized two things...

One: With the tire on the bottom wheel, it doesn't clear the housing of the saw any more. It took some rather heavy duty flexing of the housing to get that thing back in there. I'm sure the trunion didn't mind taking one for the team. 😉

Two: The drive belt I ordered (see the first installment of this series) is too short. Like 4 or 5 inches short (At least the profile is correct). I'm not really sure what happened while I was measuring the old one - maybe something shiny caught my eye, or it could have involved my nemesis: Math. Anyway, the old belt is on there now, and doing surprisingly well. I'll eventually measure it again and order a correct replacement. Probably after the old one breaks.

New Blades

I ordered two Timberwolf brand blades for this saw directly from a company called Sulffolk Machinery. I decided to try a 6 TPI Raker style blade, both in 1/4" and a 1/2". This pitch and style should match the type of cutting that I can foresee doing on a saw like this, and the widths are at the upper and lower limit for the saw. (Actually, while the published minimum blade thickness is 1/8", multiple reviews stated that the guide bearing design won't allow for blade smaller than 1/4").

Putting the rip fence rail back on

While putting the 1/2" blade in, I discovered something goofy about this saw - the rip fence guide has to come off to change blades! I was not amused!

With the 1/2" blade installed, I began playing with the blade tracking, the lower wheel alignment, and the guide bearings. Once I was happy with the way everything was running, I decided it was time to see what this little saw could do - and what a better way then by resawing some 3" thick white oak.

Now, I know that doesn't sound like too much, but keep in mind, that's eating up 75% of the height capacity of this saw.

A few test passes

Overall, I was happy with the results.While it isn't going to set any speed records, it was capable of doing the job. It cut at roughly 40 inches per minute. The cut was straight (parallel to the rip fence), and square to the table.

I later swapped blades, and did some contouring work (a small oak heart for she who must be obeyed, as a peace offering). It cranked right along on that thinner material (roughly 3/16" thick).

Wrap-Up

So, this just leaves building a table insert, ordering the proper drive belt, and making some sawdust! This has been a fun distraction from other not-so-fun projects around the homestead. I'm looking forward to using this saw in the future to resaw some dimensional lumber to make casement molding and french cleats for the shop. Hopefully I'll be ready to start putting up trim by spring.

Thanks for following along!