We've had a lot of rain this summer and fall. (I wish we could ship some of it to Southern California where it's needed.) Anyway, lately it seems that every time I get ready to mow the lawn, it rains. This had been going on for a couple weeks and the grass was getting pretty tall. Finally mother nature strung together two sunny days in a row and things had dried out enough that I could make an attempt to mow. I did my routine checks of the little Deere 317 and found that the hydro sight tube was below its usual level. Being in a hurry, I topped it off and fired the tractor up. I thought it a bit strange that the hydro fluid was low but there were only a couple drips on the ground under the tractor and I remembered that I had spilled some Hy-Gard last year. I'd check for leaks later. More rain was forecast for later this afternoon and I had a few acres to mow. I backed the 317 outside the garage and flipped the PTO switch so I could cut down the weeds outside the garage door and started to mow. The mower deck sounded good for a moment, then the pitch of the whirring blades sounded different. "I know that sound," I thought. "Crap, I've thrown the deck belt."

I swung around the corner and drove straight into the shop. About 10 minutes later, I had the deck belt covers off. It took me a couple moments to make sense of what I saw. The center and left side spindle pulleys were about an inch lower than the one on the right side. I grabbed the center pulley and pulled. I was able to lift it up about an inch. Something's not right here.

It turned out that the sheet metal that supports these two spindles had cracked and broken. The cracks appear to be from metal fatigue. I pulled the deck off the tractor and removed all three spindles. The area of sheet metal that supports the two spindles had failed pretty spectacularly. The third one looked fine. With two spindles broken free from their mounts, it was no wonder why the belt had come off.

There's not a whole lot left of the mounts for the center and left side spindle.	This is what is left of the center mount.

The mower deck on this tractor had seen some abuse prior to becoming mine. After purchasing the tractor, I had had some trouble getting the mower to cut level. When I stripped down the deck and replaced the spindle bearings, I found that the deck was twisted. Someone had run the deck into something pretty hard. I had been planning to look for a deck shell or maybe purchase a new deck this winter. However, I still have grass (and weeds) to mow, so I was going to need to repair this deck for the time being. I had some 16 gauge sheet metal left over from another project and figured that I could make up some patch panels to repair the damage.

I took some measurements of the not broken right side spindle mount and drew up a CAD sketch of the hole layout. The bolt holes were a very close fit on the 3/8"-16 carriage bolts that hold the spindles to the deck. I measured the holes on my deck to be 0.366". The closest drill bit I had to this was a letter drill size "U" at 0.368". I measured the center hole at 3.025". I figured that it had probably been an even three inches when new, but there had been a lot of rust and corrosion since that time.

The left side mount is in pretty bad shape too.	A CAD sketch of the mounting hole layout per my measurements.

I made up a template using a piece of scrap clear plastic and checked the fit. It was a very good match to the holes in the one good mount I had left. After I got the patches installed, I started thinking about the measurements I had taken. The measurement I had come up with for the three mounting bolts was 2.1045" from the center point. This gave me a bolt circle of 4.209". I figured that on a new deck this bolt circle would actually be 4.2". I ran my idea past an engineer friend and he agreed. I drew up another CAD sketch with what I think are the stock layout for the holes; a 4.2" bolt circle and a 3" center hole. As I said, I used my first measurements and the patches came out fitting real nice, but my guess of the proper measurements are shown in the second sketch.

A revised sketch which is probably very close to the mounting holes that Deere used on the deck.	I made a Plexiglas template using the original measurements I had made. It fit quite well.

I made the first patch just large enough to cover the damaged areas and give me some area to weld to. I cut a couple "V" shape notches to make the patch easier to bend over the ledge at the front side of the center mount. I used a grinder and orbital sander to clean the metal under the patch as best as I could. I liberally coated the cleaned metal under the patch and the back side of the patch with flux and used bolts and clamps to position the patch on the deck. I was hoping that the brazing rod would get wicked between the patch and the deck and give it a little more strength than I could get from brazing around the outline of the patch alone.

Once I had one side of the patch secured, I heated a section of the patch and formed it to match the deck with some auto body hammers. I then brazed the newly bent section to secure it. I continued this process as I worked my way around the patch. I had considered securing the patch by cutting out the deck to match the patch panel and using butt welds. It would probably have looked nicer, but wouldn't have been as strong. I will have two layers of metal for much of the mount.  It also would have been more work to match the raised area around the mount. Considering the condition of the deck, my goal was to get it useable again. If I can achieve that, I will be happy.

The template matched up with the spindle housing as well. Time to cut some sheet metal.	The sheet metal has been cut and I am starting to braze and hammer form it to fit the spindle tower.

Once I got the first patch finished, I started thinking about the second patch. This side of the deck had a lot of hair-line cracks radiating out from the mounting area. The patch was going to need to be a little larger. Again, the bolt hole area was missing for one of the bolts and where the other two holes were, there wasn't much metal that wasn't cracked. I decided to cut this patch in the shape of a circle with a small flat area to fit down into the

side with the ledge. To cut the circle, I added the little table to my 4" X 8" horizontal band saw that allows it to cut in the vertical position. I don't use this table too often, but it sure came in handy for this project.

The opposite view of the previous picture. The notches make it easier to bend the patch to fit the spindle tower.	For the left patch, I started with a circle with one flat side. The flat will get bent down to align with the ledge to the front of the raised spindle mount.

I made the circle seven inches in diameter and set it up on a piece of pine board backing on my milling machine. I marked the center of the disk and drilled the size "U" bolt holes. I had looked up the size of a clearance hole for a 3/8" bolt and found that size "W" is often used. Size "W" is 0.386". I had measured 0.366" on the rusty holes on the right side of the deck. As my friend mentioned, it appears that the design considerations were to minimize any play in the spindle housing to deck junction if the bolts were to loosen. With the size "U" bolt holes, the bolts are a very snug fit.

Next I turned my attention to cutting the 3.025" hole for the center of the spindle housing. Since I don't have an offset boring head, I made do with a fly cutter with an angled bit. The fly cutter works well, but setting the bit to cut a precise hole involves some trial and error. It took a few test cuts on some scrap before I was able to cut the hole diameter I needed. Once the bit was set correctly, cutting the center hole went quickly.

Drilling the letter size "U" (0.368") holes in the patch.	Here I am using a fly cutter to cut the hole for the spindle housing.

With the second patch made, it was back to brazing. I used the same method as before and tacked one side, then heated and bent the sheet metal to the proper shape. Again I used a lot of flux between the patch and deck and fed brazing rod into the weld until I saw that the bronze had been wicked between the parts. Once I finished brazing the bottom side of the deck, I turned it over and started on the top side. I clamped the broken fragments into their respective positions and brazed them in place. Since the spindle housings sit on the top side of the deck, my goal was to build up the top side so that the two cracked mounts were at the same height as the undamaged one. I have always had a belt alignment problem with this deck due to the metal under the spindle housings being bent and sagging a bit. Now I had a chance to improve the alignment by making sure that the height of the deck was uniform. There were lots of little pieces of broken sheet metal to clean and braze into position. After I took the picture below of the deck top side, I spent another couple of hours brazing, grinding, and leveling the two mounting areas.

Both patch panels have been brazed in place.	On the top side of the deck, I brazed in all the broken pieces. I did this so that the height of the spindle housing would be the same as the not broken right side.

When I finished brazing, both sides of the mounting areas were ground and sanded flat and painted with a couple coats of primer. If my repairs hold up, I will strip the deck down this winter and get a few more coats of paint on it to try and keep the rust at bay until I can find a nice deck shell to replace this one. A wider deck would be ideal but they're even less common than the 46 and 48 inch decks, but who knows, I may get lucky.

When I was talking to my friend about the size of the bolt circle, he mentioned measuring the spindle housing as another reference in addition to measuring the holes in the deck. I have two different styles of housings. One with slots instead of bolt holes and one with square holes to capture the carriage bolts that Deere uses to hold the spindle housings to the deck. After measuring the housing with the square holes, I am pretty confident that the bolt circle is 4.2".

Primed and ready for the spindle installation.	Comparing the two different styles of spindle housings.

The last time I had the spindle housings out of the deck was November 2010 when I replaced the spindle bearings and added power steering to the 317. Four years later, all nine of the spindle housing nuts and bolts came apart easily. The main reason for this was that I used an anti-seize compound on all of the threads. I had never used anti-seize on anything but spark plugs in aluminum cylinder heads until we moved from California to Virginia in 1986. The mechanics in the Maryland dealership I worked for used it for all nuts and bolts that were exposed to weather and swore by it. With road salt being used for winter snows and lots of rain the rest of the year, there was a need for using it. In California we didn't have much of a rust and corrosion problem. I started using it for my projects at home and now wouldn't be without it. Last year, I found a anti-seize product for use in marine environments. I haven't been using it for long enough to know how it holds up after a few years, but everything I have used it on has come apart easily. I am hoping that it does an even better job than the non-marine stuff I have been using for decades.

I lubed up all of the nuts and bolts and installed the spindle housings. I sharpened the three blades and installed them. The blades are getting pretty rounded on the ends. I think it's about time to buy some new ones. I installed the belt and was real happy that the idler pulley alignment with the center blade pulley was better than it's ever been. This should make my belt last a little longer.  The deck was now ready to be installed back on the tractor. However, there was one last thing to check.

The spindles are installed.	The blades have been sharpened and installed. (I really could use some new blades.)

Last Tuesday when I thought I was going to mow, I had checked all of the fluids before starting the tractor. While this is my usual practice, finding the hydro fluid a little low in the sight tube was not a usual occurrence. I topped of the level with Hy-Gard low viscosity fluid and checked under the tractor for leaks. There were a couple fresh drips on the ground and the bottom of the transmission had some oily dirt on it, but it didn't appear to be a big leak. I figured that I would pressure wash the tractor and find the leak after I mowed. As it turns out, it is a good thing that the mower deck had trouble and that I couldn't mow. The leak only happened when the hydro was under power and it was a pretty good sized leak. 

I was pretty certain that I could silver solder the crack and close it up, but the question was how long it would stay sealed. The tube is attached to the bottom of the transmission and extends upward to form a loop. When the tractor is running, the loop vibrates and there is no upper mount to dampen the vibration. I didn't think that my silver solder would hold forever, but I was hoping that it would hold long enough to get the yard mowed before our party on Sunday.

I found out where the Hy-Gard was going. A hair-line crack 2/3 the way around the cooler pipe.	It was hard to see the crack in person. It was harder to get it to show in a picture.

I started mowing around noon. It usually takes about four and a half hours to get all of the lawns and the trails done. The deck was working well aside from leaving a thin stripe of long grass where the rounded ends of the blades didn't cut. I stopped every 15 minutes or so to check the cooler pipe for signs of leaks. No leaks for almost two hours. I finished up on the back yard and started mowing in front of the house. After a few minutes of mowing, I decided to check the cooler tube again. This time there was a drip hanging off the pipe. I'd better get the tractor back to the shop. After I covered the 50 yards back to the shop, the drip had turned into a tiny stream. I had caught the crack early enough that I didn't lose more than a couple ounces of fluid.

Low heat from the Oxy-Acetylene torch, lots of flux and some silver solder sealed up the crack, but will it hold up?	I removed the adapter, cleaned the threads and added some Teflon tape to help prevent future leaks.

I drained the Hy-Gard and pulled off the cooler pipe again. I would order a new pipe on Saturday. I still wanted to get the lawn mowed, so I decided to try brazing the pipe. I used a grinder and a small disk sander to remove as much of the silver solder as I could from around the crack. I coated the pipe in flux and heated the pipe just enough for the brazing rod to flow and follow gravity downward. With the crack facing down I added enough rod that some drips started to form on the bottom of the pipe.  I was trying to get as thick of a coating over the crack as I could.

Using a 1-1/8" wrench to tighten up the adapter.	A 1" wrench is used to tighten up the cooler pipe fittings.

After the pipe had cooled a bit, I plugged one end and pressurized the pipe using a rubber tipped air chuck. I dipped the pipe in some water to look for bubbles. No bubbles were found, so hopefully that means no leaks. I had been thinking that I wished that there was some way I could secure the pipe so it didn't vibrate quite so much. There may well be a way to secure the upper looped end of the pipe, but that would mean pulling the fender pan off the tractor to get a good look at the possibilities. That would take some time and I didn't have a lot of time to spend on the tractor right now.  I decided that I could add a gusset to the bend where the crack had formed and this might be enough to buy me a few hours of mowing time.

Back to the drawing board. This time I will braze with a bronze alloy and hope for the best.	I let the braze drip a bit and made sure that gravity guided the drips on top of the crack

I cut a small triangular piece of sheet metal and ground it to the approximate shape of the bend. I covered the side that would be facing the pipe with flux and tinned that side with braze. I did the same to the area of the pipe that the gusset would attach to. With low heat, I then brazed the gusset to the pipe.  I reattached the cooler pipe to the transmission and ran the Hy-Gard that I had drained through some paper filters to make sure that I hadn't picked up any dirt.  I had used the last of my jug of fresh Hy-Gard when I had filled the transmission on my first attempt to seal the leak. Now all I had was what I had drained before removing the pipe for the second time. I wasn't going to have enough to fill the trans to the proper level and I wasn't going to mow being low on fluid. I would hit the Deere dealer tomorrow and get another gallon of Hy-Gard low viscosity when I ordered the new cooler pipe. 

To reduce the amount of movement at the point where the tube cracked, I brazed in a gusset.	Once again, I installed the cooler tube. It's a little too dark to try it tonight but I am hoping to be able to finish mowing the lawn tomorrow.

Saturday morning we took a ride up to the Deere dealer. I bought a new belt for the deck, some Hy-Gard, and ordered the cooler pipe. The parts counter guy said that the pipe should be in on Wednesday with their stock order. I had no desire to pay extra to get the part sooner. If I couldn't get the yard finished today, I didn't really care if I mowed on Monday or even next Friday.

It took about a half hour to install the new belt and fill the hydro. Time to see if it works. Again, I stopped to check for leaks often. I found none. I got the front yard finished and started on the trails. I was a little hesitant about mowing the trails. There are lots of pot holes, tree roots, and rocks to contend with.  I would also be a good distance from the shop if I developed a leak. However, the grand kids enjoy walking down to the creek and the weeds were getting pretty tall. I took a chance and mowed the trails. Fortunately the pipe held and even hitting the big ruts didn't cause the crack to open up again. After I get the new pipe installed, I will keep this old one as a spare. As long as I keep it around, the chances are that I will never need it.

© Fager October 18, 2014