This is the American No 20 Universal Saw Bench, USB, which I recently imported to Oregon from Massachusetts. I think I must have wanted one pretty bad. Not disappointed yet.

Saw at work. Advantages of a tilting table can be having the weight work for rather than against the operator.


Below I am just getting into clean up and some spare parts. Starting to think about casting models for the appliances the saw is without; primarily, the quadrant gauge.


The saw has a sliding table. It is supported at left in slotted and self-lubricating slide which keeps the table perfectly aligned. To the right, it rests on machined surfaces front and back. In the center, three rollers bear in the center. I am playing with the adjustment of these. I believe the intent is to slightly bear weight without being high enough to create an appreciable rocking. A pin at the front left corner locks the table.

The sliding table is mounted on a nicely fitted casting which is adjustable perpendicular to the blade for dado blade widths.


This hand wheel rotates the large disc and paddle which carry the two arbors. The rotation creates blade elevation with the highest point at the center of the arc over the table. Lower blade elevations can be used forward or aft depending on what is needed for the work at hand.

The purpose for having a choice of blades, 14″ in this case, is to easily go between ripping and crosscutting or from sawing to dadoing. Saw requires a dado arbor extension which I do not have. This will require some research. I think the extensions shown in the catalog cuts may have been intended for the older style of dado head with inner and outer flanges and spur cutters sandwiching a pair of straight knives.

The saw is capable of carrying one 20″ blade. I don’t think I will need to try that though.

The fence is of the rack and pinion variety, secured by tapered pins. This saw has the 2″ extension that allows for 24″ ripping. An option is listed for an “extended fence”. The fence here is longer than the tilting fence shown in the catalogs. I am not sure if this is the option I have or if it refers to the additional width capacity.


This is from behind. This is an original motor mount. I will likely need to alter the plates to suit the motors I have to choose from as I think it unlikely to find the correct frame size for this. I will be getting a paper drive pulley made of diameter sufficient to make 2900 rpm.

The dust chute exhausts the blade cavity which is very nicely isolated from all the bearings and the rotational gearing. A great design. A large door accesses the blades under the sliding table.

Apologies for the dark pictures. As I proceed with this saw I am sure I will be adding more to look at.

Hey, better light, whoa!

So here we are getting to going on getting the saw going again. This is a nice solid little GE motor that I figured looked pretty tidy back there. I had to make an adapter plate the accommodate the frame size differences. I actually had an old motor that would have fit but it goes too fast. Better to save it for a shaper or something.

The pulley was made by Paper Pulley in Columbia Tennessee. Tell them I sent you. I sized it to get 2900 rpm which was the speed I figured from the catalog. I can’t handle big saws at high speeds. Some guys swear by it but they usually have Direct motor drive so they have to like the wrong speed. It was clear from the relationship of prior frame holes that the new holes needed to on the same shaft side axis. This required the countersinking of the bolts using the original holes. I centered the new to keep the shaft center at the same place as the original.

I’ll be back with the belt stuff and we’ll get into the arbors and see what’s going on with them.

I’m back. Now we are measuring for the belt length. These are the two saw arbors on their paddle. In this position neither saw will be above the table. The rope is threaded the course of the belt around the idles and over the drive pulley and knotted. So now I will rotate the paddle and watch the movement of the idler/tensioner to make sure the belt not rub on itself or anything else.

This is the paddle dialed around to bring a saw into full height above the table. As rotated I watched the idler movement to make sure my presumed belt length was correct.

With a 6″ drive a 109″ belt gives nice clearance as the arbor paddle travels its orbit. Here a 3/8″ shim adds to the center shaft distance off the base. I believe this is real close to what the original frame would have delivered. The large drive pulley develops a nice wrap for additional adhesion.

I made up a temporary test belt out of some sections of old leather belting from my drums of belting. I want a continuous belt on this machine. So once I am done getting this cleaned up and squared away I will come up with a belt to glue up in place. I figure I will glue the splice between the arbors.

Arbor 1 is running real nice and true. Arbor 2 I needed to take a shim away on the outside bearing. I am running it a bit to see about heat. The inside cap cannot be removed with the yoke on the paddle. I may try again shifting the pulley over a little. I would like to avoid removing the yoke at this time.

Arbors 1 and 2. The riving knife mounts onto a polished surface. On the upper right is an oiler for the revolving disc. The upper door hinge has the repair common to saws of this kind. Guys tilt the table and forget that the door is open. In this position, both arbors run at the same time.

To get the rust and crud out of the taper pin holes for the fence I wrapped some sand paper around the pin and twisted with oil, coarse then fine. When clean the pin fits with no movement.

This cleaning the threads for the fence when used for coving. Visible here is also the optional extension for ripping 2″ beyond original specification.

There are two pin holes on the rolling table for the fence.

It is desirable to have the fence below when ripping bevels.

I am testing the cut here. Scanting. Enjoying the micrometer fence. The blade is at the front of its arc here. Nice for close work. When at the back of the arc there is no riving knife. The modification of the fence appears to be a good idea after using it. It is longer than the stock fence.

Arbor 2 is running nicely. I added a blade and started checking the alignment. This picture shows a 14″ and a 12″ blade. So far, I am very impressed by the precision of the saws. Looking forward to getting the endless belt even though it runs fine with three joints in the belt!

Here we see the sliding table pulled back. And opened up for dado work. I see potential for tapering here. The table is tapped this side also to swing the fence in an arc.

That new little wheel is one of the three “anti-friction” rollers that takes a little of the weight off the slides. The grooved track has two wicks in it for lubricating. When closed, there is no gap between the tables but just enough clearance to slide.

I finally got the thing off its pallet. A little tricky to pick up. Sure runs smooth on the floor. It’s still in a temporary spot. That’s Prudent the little white helper.

One thing about the fence is that it takes away the space I have been accustomed to placing my work. Plus, frequent changes of width of wide variance requires re pinning the fence. I guess that’s why it’s not a “Variety Saw”. It’s a saw best for set ups with some duration, as in a production setting.

Instead of getting a new belt from the saddle shop I am using some old from my barrels of old belting. I have never skived before. Doing is learning. Using a spent planer knife with gloves is working for me. With the 30″ knife I have the leverage to flex it a hair enough to get a hollow scrape. Those curved scrapers also work good for high spots. There is a machine for this, I know. This is the lap I will need to glue up on the machine. One thing is that the skive needs to be the same direction as the others in the belt. That way I can run all skives point forward on the belt. The “Treatise on Leather Belting” has been a good resource for information. That where I understood that I could clean and re-skive old belting. I had always figured it would be too oil soaked. We will soon find out how it goes.

Here I have the belt in place on the machine. Between the two arbors is a nice place to make the joint. the belt is tacked down at each end. I am dry fitting the joint to make sure it falls right. The lap point needs to run towards the pulleys and the grain is on the pulley side. My skive is the same direction as the old skives in this salvaged belt.

This is what we want. Straight. One thing I would do different next time is to study the belting better at the start. I skived one end before I measured what I needed which then put me close to an old skive. Didn’t really want to start over. And I don’t think it will be  problem anyway.

While the coats of cement dry, the surfaces need to stay separated.

After the third coat of cement dried, I laid the joint together. It seemed logical to hammer from left to right (direction of movement) up the middle and then towards the edges.

I clamped from contact side leading edge counter to the direction of rotation. So I will let this sit for a day or two before I release the idler pulley and let her rip.

After a couple days I unclamped it and let ‘er rip. Well, it wasn’t silent like my expectation. But it is much quieter. Still has a little slap. I think my joint is a little rigid. Maybe it just needs some use.

The idlers were sounding dry so I doubted the new grease was getting though. I pulled the cups off and cleaned the port in the axles for the idlers. It’s slick willy now.

Next it’s time for the belt guard. Time for some sawing.

The Delta 12-14 is long gone and the new daily user is in place beside its other Clement product from Rochester.

Three horsepower gets me through 3″ material just fine but if I ever replace the motor it will be with a five horse motor. The endless belt would not stay together. So I have to use a clipper jointed 4 ply vulcanized canvas belt which works great. I do find that I am often back and forth between working near and far. Sometimes to get rid of the riving knife, sometimes for working in close, sometimes ripping plywood — so the blade position is not static and that requires extra attention. I will eventually get rid of the fluorescent light as it makes the blade less apparent for some reason.

The bearings hold oil for quite a while and seem to always run cool. The idlers will chirp up after an hour or so with a different tune. Blade position when low will run both spindles which lessens power. But not a problem really because the most power is required when at the high arc position which is when the unused arbor is off the belt. At low heights, they are both running.

There is a bonus fanning effect for blowing saw dust out the chute from the second arbor saw. When at full height, in a three inch cut, a chip collector duct would be helpful to keep the chute clear.

The rolling table does collect dust and does need to be blown out and kept clean underneath for smooth operation.

Here is the motor cover. I believe it is fairly close to the original. And extra belts. I may try the endless belt again once I learn what I am doing wrong. I kinda think the application requires a two ply leather belt on account the stresses caused by the short idler drive.

It’s several months later. I have set the motor cover aside. Sometimes things are missing for a reason. It’s easier to keep things going and tidy without it. Still no complaints with the saw.  Very smooth cut. Accurate.

Without a sleeve or dado arbor, a 1/2″ dado will fit the arbor without riding the threads. In this shot, you can see the separation of the tables for the dado. In this case, 3/8″. An important note is that the tilt is designed to keep the relationship of right table to saw the same through the range of table tilt. this does not apply with a dado set. Also, one really doesn’t want to forget they have a dado on the other arbor when using the saw on the other arbor. For safety, leave the gap to suit the widest tooling.

Another thing, when oiling, have the fence out of the way when dialing the arbors around to get to the oil plugs.

This is my new fence. It is off an older model of American USB. The knob details and the lack of alignment adjustment differs from my original fence. This is a short tall fence. The face tilts. Very similar to the Clement jointer fences. For certain beveling moves with more bearing on the fence than the table. The pins are larger but on the same centers. My machinist made them all the same and true for a perfect fit, both fences. Four degrees taper.  This one has a slight lead away from the saw. I don’t know if that was intentional or not. But at some point, the maker decided to make the fences adjustable.

This is the feature that allows the fence to be adjusted. I believe an innovation of the Model 20.

A nice view of the fence and the 2″ extension. Have to add to the scale, but gives 24 inches.

Today I made 30 table legs. I decided to try sawing them rather than using the jointer. I found the handle on the sliding table useful to lean against while advancing the table.

Video Show and Tell/Part 1

Video Show and Tell/Part 2



Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Google photo

You are commenting using your Google account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s

%d bloggers like this: