I thought I would tell of a recent weekend, and give everyone something to laugh about.  It’s funny once the sting wears off.

image About a month ago I was requested to produce a prototype Invar level strip bracket.  I figured “simple job, it’ll be fun”. Well…sometimes things just go wrong. Sometimes everything goes wrong.

The Invar strip is a special barcode that gets scanned by a laser, and a computer calculates horizontal range and vertical difference, without the human eye judgment factor.  The company was applying the barcode tape to a very thin price of aluminum, to reduce cost and weight.  The bracket’s job was to provide precise vertical reverence to locations at both bottom and sides, in relation to the actual strip label.

What was required

I mocked up different concepts, and after creating numerous designs that created mounting away from the digital strip, the company decided on a  different approach.  While different individuals wanted differing uses, the common goal was no frills, simple light, and fast.  We got everything except the fast part.

Then they notified me they wanted it the morning of the third day!

I decided that we could mill the bracket in 2 parts, and weld it together.  Then go back with a final pass in critical areas to true it.  Some of you are starting to laugh.


I started an assembly, and derived components from a base sketch containing all the parameters I needed, as well as some basic geometry.  Then fed the strip specifics to the design.  I sent it to a weldment, and  applied chamfers to the rear and side contacting edges to get some good penetration on the welds.


Since they wanted to fasten through the digital strip (and had already fixed the tape to the aluminum) I needed to get an accurate idea of where the barcode was in the design, I scanned it in, and scaled it to the measured values.  I set the fastener holes using the bolted connection generator, so that they would not interfere with the digital strip divisions.

Since I had no time to order materials, I went out to a construction yard, and got 1/4” thick square blanks of cold rolled mild steel.  Most of the blanks available had a slot in it, so I figured I’d run the first one out of the slotted one, and use it as a lightening hole.  This was fine except my fastener locations had to be modified.

Day 1

I rough cut the blanks down with a grinder, since the inserts in my favorite endmill are 13 bucks each.  Then rough milled the bracket pieces.  Once completed I was happy; those parts were beautiful.  I pulled out the MIG and decided that the prototype could be run with flux core, and save the gas for the final production model.  I cleaned the parts with what I thought was a clean rag, and good solvent.  Then braced the parts together, set the current and feed, and started going.

It seems that there was something on that rag, and it was NOT happy with the arrangement.  (now everyone is laughing)  The slag splatter went EVERYWHERE, and I REALLY had to pour it on to get the weld to take.  Needless to say, the quality was poor, and had to be reground in areas and welded again.  When I got done it was a mess.  The flux and foreign material fight put slag all over my finished areas, including bouncing through the lightening hole into the front slot.


Hey, No Problem…I have to final machine anyway. Right?  …r-i-i-i-i-i-i-ght.

The finish truing was intended for any warpage to the Channel part, and to square the base to the channel slot.  I did not intend to machine every last cut face.  This was bad.

So I started again, taking off about 0.001” over all the bottom mount areas.  After all the tortured welding, the warpage was a bit much, but I had left plenty of room for that.  I squared off the strip channel and mount bottom, and cleaned up the channel rear side. While on the back, I decided to cut out as much of the overkill weld as I could, since the weld was deep, I did not need much of the fillet weld material.  r-i-i-i-i-i-i-ght.

It started looking presentable.  It was time to fit check the bracket.  So as I took the thing out of the mill vise……that’s when I heard it.  Yeah, you know the sound.  The tiniest ‘tink’ sound.  Nah, can’t be the weld.  As the vise loosened, the base began to twist perpendicular to the axis of the channel.


What in the H$%# was on that rag? Why didn’t I normalize the weld? I was tired.  I mean tired, and it was 2 AM.

Day 2

There was no time to start a new part. It was due tomorrow.

The prototype was begging, no more.  I grabbed the grinder, and started cutting out the weld.  I cleaned the thing with a VERY clean rag, and poured the filler in again.  This time warpage or not, she was gonna get it.  I tacked the welds evenly, and ran 3 passes over it real slow.  That weld was solid this time.

I stared at it.  The slag splatter was substantially less, but the fillet weld got into the base tangs a bit, so I cleaned up the weld bead just a bit. Then I checked the channel.  twist.  You couldn’t see it, but when the dial indicator hit it, the twist was there.  So I re-cut the channel, and squared the base.

When it was complete, the thicknesses were reduced substantially, but this payed off as everyone liked the outcome weight.  I clamped the strip and bracket together, set the holes, and cut the strip to prescribed length.

The holes.  Everything went fine, but I should have turned the fastener heads down a bit. The countersink cuts ended up a bit large.   I couldn’t tell because I was dealing with reference measurements that I had mic’d 3 times.  I knew the fasteners were going dead center, but the oil cover I wrapped the part in blinded me from seeing the relationship to the strip.

It’s Like flying on instruments. The fog is so thick that you can’t see anything, but you still hold to those instruments, and you’ll put her right down on the runway.  The only problem is you can’t see that the runway is smaller than your 747.

It wasn’t a problem.  Thank God! There was about 0.010” clearance at the edges of the countersinks.


They liked it.  The next version will adopt a few different procedures.  All welding will be done before the machining processes.  The part is too small for welding after the cuts.

  • Clean the parts with a CLEAN rag
  • We weld the blanks FIRST – groove weld all 4 sides!
  • Normalize the weld afterward
  • Then our rough and final pass milling.

Sure a solid 1” bar stock would have been better, but I couldn’t get the bar issued before the deadline.  Maybe next time it’ll be easier using solid stock.