KB58's Power Drawbar - Electric

[kb58]

This can be adapted to any Bridgeport-style mill with a drawbar nut up top.

I use my mill just enough to be bugged by having to constantly loosening the drawbar, to the point that sometimes I don't even when I should. One thing that makes it tougher is that I raised the mill to a more convenient working height, but it raised the drawbar height such that I had to stand on my toes. The project was started in February 2024 and took one month. For this reason, ignore the posting dates below since they're being copied over en masse.

Looking at power drawbar ("PD") kits, the obvious answer was PM's own unit, but being a hypocritical cheapskate, I decided to make my own. After watching many YT videos where everyone uses the Harbor Freight butterfly impact wrench, I went to get one, only to discover that HF has apparently discontinued the tool. That started the wheels turning, could I do something different, and what exactly did I want?

Nearly all homebuilt PDs use compressed air, with many people leave their compressor on all the time. Thing is, my California Air Tool compressor did something once that broke all trust - one day when it was turned, the motor stalled(!), getting very warm very quickly. If it hadn't been shut off; it may well have caught fire because it wasn't drawing enough current to pop the breaker, so it's never left unattended. Between that and not wanting to bother waiting for it to build pressure just to use the PD once or twice, thoughts began to wander. Why hasn't anyone designed a PD using an electric impact wrench? Looking up the specs for what's needed, it appears that only 50-100 ft-lb is needed to tighten an R8 adaptor. Nearly all corded impact wrenches generate far more than this, so much so that I worried that it could either strip the threads or even shear the drawbar itself. Attention then turned to cordless impact wrenches, and they, too, brag about how much torque they can generate. Looking at the lower end of the range though, I came across a DeWalt unit that generates up to 150 ft-lbs, and decided to give it a go. More posts to follow

 

[kb58]

A common issue on Bridgeport-style mills is accessing the draw bar nut, located at the top of the machine at the end of an approximately 20" long bolt that retains the tool holder. It must be accessed every time a tool holder is changed out. For example, an R8 adapter drill chuck and R8 end mill holder are swapped back and forth all the time. On some machines, unscrewing the draw bar nut is made difficult due to the height. Also, some R8 adapters get stuck in the spindle bore, necessitating tapping the top of the draw bar nut with a hammer to dislodge it, and again, access can be difficult. As a result, some people leave the drill chuck installed and put the end mill in it instead of a proper collet-type holder. Drill chucks are not designed to be side loaded and cannot grip the end mill properly, and can slip.

There are many YT videos about making power draw bars, nearly always air-operated, incorporating a small butterfly-type impact wrench. The one that everyone uses is from Harbor Freight but they apparently no longer carry them. While other vendors carry them, I thought, "why build something like everyone else when I can try something different - you know, more difficult?" Kidding aside, not everyone has an air compressor, and having to start it up and wait to build pressure can be annoying in a home shop. Another reason to go this way is that cordless tools are becoming so good that they're negating the need for air tools and compressors. The question then becomes: How to design a power draw bar using a cordless impact wrench instead of an air tool? It's not a combination I've seen before, so it has novelty going for it, giving people more choices.

So, the idea is to fabricate a holder for a battery-operated impact wrench, one that can slide up and down and rotate left and right. The trick is creating a holder and linkage that extends down to within each reach to operate engagement, speed, and direction. The added twist is the aforementioned "tapping with a hammer" requirement." If the impact wrench on its own can't shock the R8 adapter to pop free, the wrench has to be kept from moving upward so that it doesn't rise up as the draw bar unscrews. Said another way, the impact wrench may have to literally push the R8 adapter out of the spindle. Much sketching will ensue, and the impact wrench modeled in CAD. It remains to be seen if actuation will be all mechanical, electrical, or a mix of the two.

 

[kb58]

The cordless* impact wrench arrived, a Dewalt DCF890, chosen because it only produces 150 ft-lb of torque. Drawbars are designed to be tightened by hand, so it's unlikely that anything more than about 100 ft-lb is ever used, or necessary. Most impact wrenches are 1/2" units producing crazy amounts of torque. and with the draw bar being about 20" long, I'd fear for its life.

First test was loosening the draw bar to see if the impulses would drive out the R8 adapter. No, it didn't, not by itself, but it took surprisingly little additional down force to pop it free - good. Now to work out the mechanics of how to rig it.

Searching for an online 3D CAD model of the impact wrench turned up nothing, but the task was simplified after realizing that there are only three critical part locations: the impact socket, the direction button, and the variable speed trigger.

* I considered a corded wrench, but they aren't much less expensive, they produce too much torque, and I already have Dewalt batteries.

 

[kb58]

Looking a bit like a 6-shooter revolver, here are the socket, trigger, and direction button relative to each other. All CAD work is with Alibre, though a handheld 3D scanner would have been a HUGE help... oh well.

 

[kb58]

Spent most of the day designing a top and bottom adapter to hold the wrench, sandwiching it so that it doesn't need modification (making replacement easy). The first version of the 3D-printed nose piece fit perfect, but the top piece needs another revision. Pics when they're done.

 

[kb58]

Here are the nose and top adapters for the impact wrench. I haven't decided whether to use them as-is, screwed to aluminum plate, or try machining them out of aluminum - will probably stick with this design. This is just the first step; there's also how to activate the direction button and the variable-speed trigger.

Oh and why are the parts different colors? I learned a $20 lesson about 3D filament. I didn't realize that the filaments I bought from Bambu Lab were refills, meaning that they don't include a reel. "I can make do with it as-is", I thought. That idea was challenged after cutting the last strap holding the bundle together, when it rapidly spun into a messy ball. "I can make do with it as-is", I said again, and had to stand there for two hours as it fabricated another part - ironically, one half of a reusable reel. It got to 90% done and I figured I could let it finish unmonitored. Nope, it used more filament than expected, kinked the filament, stalled the extruder, and caused the "100% finished with no problems" part to end up missing its upper portion.

 

[kb58]

Work continues on the CAD design. I set out to model the entire assembly in CAD, complete with moving parts to determine ratios of various parts. Unfortunately, CAD systems have issues with how part constraints are applied. As far as I can tell there is no guide for what to do in what order. It'll fail part way through the process, but if you start over and apply the exact same constraints - but in a different order - it may work. I have most but not all of the model working, with a few parts not playing nice right now but should be okay by tomorrow.

 

[kb58]

The idea:
The key parts are in place and dynamically constrained to replicate actual part motion (trigger, direction button, and socket). (The part that looks like a disc brake represents the top of the mill.) The next steps are tricky, creating a linkage that:
1. Selects impact wrench direction
2. Engages the draw bar nut
3. Applies torque as necessary

What seems like the best way - but possibly difficult to do - is to first move the activation lever left or right to set loosening or tightening, then move the lever downward and have the socket start spinning right as it contacts the draw bar nut. The thinking is that even though the socket is made to easily engage the nut, there's still a fair chance that it can get hung up. Having it just start spinning when it comes in contact with the nut will hopefully assist it to engage, but not allow it to skip over the top of the nut. This may require a fine offset adjustment to set the exact point where power is applied during nut engagement.

As mentioned before, some R8 tool holders stick in the bore, while others fall straight out. For the sticky ones, a mallet is needed to tap the top of the draw bar nut to pop the R8 piece loose. This can be a problem with a power draw bar unit blocking the nut, as there's then no easy way to tap it with a hammer.

The math says that it'll work:
The draw bar on my Precision Matthews PM-935 is threaded approximately 23 mm, and at idle, the power draw bar socket sits 2 mm clear of the nut. Full engagement of the socket on the nut moves the impact wrench downward 17 mm, so as the drawbar unscrews, it starts pushing the impact socket upward. Because the impact socket sits at rest 2 mm clear of the draw bar nut, it can only be driven upward a total of 17 mm + 2 mm = 19 mm. Thing is, the draw bar still has four more millimeters before it's fully unscrewed, when it runs into the power draw bar's idle stop position, where it won't go any higher. Therefore, as the unscrewing continues, it's now pushing downward on the R8 adapter, and 4 mm is plenty to pop the R8 adapter free.

Or at least that's the theory.

 

[kb58]

The puzzle is becoming more challenging, currently stuck at the "how am I going to pull this off" stage.

At first I thought that as the impact wrench descended onto the nut, it could contact a fixed point on the mill head that presses the trigger. The problem is that as the nut unscrews, it pushes the wrench back upwards, so that solution is self-defeating, cutting off power before the nut is completely unscrewed. The idea was to pull a lever forward or back to set the "tighten" or "loosen" direction, then pull the lever down to engage the nut and apply power. The problem is that the wrench moves downward relative to the mill to engage the nut, while the trigger has to move upward relative to the wrench.

One way is to have a pull cable, like a motorcycle brake or throttle setup for motor power. Another way would be to somehow lock the lever in the down position until the nut is fully unscrewed. There are endless complicated ways to do it - but doing so elegantly and simply, is hard.

 

[kb58]

I tend to over-think things, and in this case it's the assumption that R8 tool holders can get seriously stuck. The question is: how stuck? I have two R8 holders; one falls out, releasing easily, while the other only pops free after the draw bar nut is tapped with a hammer. Well I just tried the impact wrench on the sticky one and it comes free with surprisingly little down force on the impact wrench. I'm going to make the (convenient) assumption that this sticky one is typical, because it greatly simplifies the design, allowing the trigger to contact a fixed point on the mill head, so pulling the wrench downward engages both the nut and applies power, eliminating a separate control.

Bushings, bar stock, and suitable springs are on the way (although I've seen in many projects that designers end up trying a few different springs).