Power feed for a benchtop mill

There are lots of vendors supplying their variations of mills based on standard components produced, and usually assembled, in China. These are excellent value machines, and the few I have used I have found to be quite capable.

The power feed solutions available for the X-axis are all pretty much the same. It looks to me like a power feed design, from many decades ago, has been replicated and then very basic solutions used to adapt the fitting of that design to other mills.

The original product looks like it was designed for a large floor standing Boxford knee mill. This is not ideal for a benchtop mill and none of the attachment methods that I have seen are quite right for my requirements. Therefore, I have designed my own bracket. 

I have a Warco WM18 mill but my design, or a slight variation of it, probably applies to similar mills marketed by Chester in the UK and Precision Matthews in the US among other vendors.

Position

The main requirement I was trying to solve with my solution was to mount the power feed so that it does not foul the bench or the chip tray of the mill, while also not requiring too much bench real-estate.

The fitment for benchtop mills provided by all the versions that I could find use a horizontal bracket held by a rather disappointing looking pair of bolts clamping that to the mill table. The power feed sticks out a long way, usually to the left of the mill table. I did not have enough space for that arrangement.

I briefly thought about designing and building a solution from scratch, using a stepper motor, but the dog clutch needed to instantly stop the feed made this unlikely to be significantly better than the off the shelf motor with the clutch and controls built-in. I therefore concentrated on a better method to mount a readily available power feed.

For inspiration, I looked at the way a few others had mounted the feed and concluded that I would prefer to replacing the left handwheel and have the power feed motor horizontal, but facing the front.

Typical advert on eBay

Having made that decision, I bought the standard kit intended to hang from the right handwheel of a Boxford mill.

Mounting

My design fits to the left hand end with the bronze gear between the motor and the table. This is the opposite of where it would fit on a Boxford mill.

3D printed mockup

I designed and 3D printed a couple of possible brackets to bolt on in place of the original end bracket. The result was a bit too large to mill out from a block of aluminium so I opted to weld together some 15mm thick steel plate.

This mounts using the original M6 threaded holes and the holes for the 6mm dowel pins.

My mill has a DRO scale mounted on the left hand side. I adjusted the bracket to maximise the movement of the table, but my design does reduce the table travel by about 50mm. I could probably gain a further 15mm if I moved the DRO scale, but for the moment, I don't need that travel.

Gears

The main bronze gear is supplied bored for a 16mm shaft. The newer WM18 mills, such as mine, have a 17mm shaft with a 5mm key.

I found it fairly easy to mount that gear in the lathe and enlarge the bore. I also pressed in a 5mm keyway.

At a later time during prototyping I cut 15mm off the length of the gear.

I also cross drilled a 5mm hole and threaded for an M6 grub screw.

I bought and used some 17mm ID shims to help position the gear, but these are probably redundant with the grub screw.

Bracket

As mentioned, the larger bracket is made from 15mm and a small length of 6mm mild steel plate. The end mounting plate which the motor housing is bolted to, is made form 6mm steel plate.

Inevitably, the design evolved slightly as the prototype progressed. 

I originally had a bearing in the design, but I have eventually opted for a ball oiler instead. That is in keeping with the rest of the mill. I also had to mill a clearance slot for a stepped radius on the bronze gear, that was just a fraction larger than I had initially allowed for.

On the largest section I drilled the mounting holes on the mill, before assembly. Also before assembly, I faced one side on the lathe and bored out the hole for the shaft.

My welding leaves a bit to be desired, but it's more than strong enough and I've been able to clean up most of it.

As welding inevitably pulls the assembly out of alignment, I've used the mill to get to the final dimensions.

Alignment

The last job was making sure that the motor axis aligned correctly with the shaft.

I made a bush to fit with a 17mm ID for the shaft and a 24mm OD to fit the motor mount spindle.

Sketch of the bush

With the bush in place on the shaft and the end plate attached to the motor mount, I held the motor in place on the shaft and square to the bracket while carefully transfer punching the mounting holes.

Ready To Go

With the holes drilled and tapped and everything assembled it was time to test.

It worked well, first time.

I'm cleaning it up and painting it yellow to match Warco's colour scheme.

I've also 3D printed the cosmetic covers to match.

I'm pleased with the end result.

==

For anyone interested in making their own, here are the drawings:

Hardware:

4x M6x25mm - to mount the bracket to the table.

2x 6x18mm stainless steel dowel pins - to align the bracket with the table.

4x M6x20mm - to mount the motor housing to the end plate.

4x M6x12mm - to mount the end plate to the bracket.

2x M6x10mm - to mount the front and end cosmetic covers.

2x M6x45mm - to mount the rear cosmetic cover.

2x 5x10mm roll pins - to align the motor housing.

Download:

PDF drawings and 3D print parts (zip)

==

Update on the mill and lathe

Having used the mill a bit more, I have made a couple of adjustments and completed some bits I had planned for the lathe.

Guards

On both the mill and the lathe I've added some simple clear guards to replace the cumbersome factory supplied ones.

These are magnetic, so I can easily move them about to suit where I'm working. To me, it nearly always feels like guards that come with machines have been designed by someone who has never worked at a machine that is even close to the one the guards are fitted to.

Obviously I don't want chips heading towards my face, but I wear glasses or a face shield, so the guard is not the primary protection in that area. I find that more hot chips hit my hands, which are by necessity closer to the machine and largely unprotected. Most commonly, this is while using the handwheels. My solutions sort that out.

R8 Spindle

Since I've had the WM18 mill, I have been disappointed with the poor engagement of the keyway in my R8 tools. Never having used this before, I did not know what to expect, but I decided it should be better. The collet chuck would engage but not very positively, the Jacobs chuck arbour would fit at any angle!

Over the weekend, I decided to take some time to check that there was not a fault. It did not take me very long to find the issue.

I now know that the key is a simple grub screw and it was nearly all the way out the wrong side! I added a little medium strength thread lock and put it back in, to a depth that felt right. It now works as I imagined it should.

==

First chips from the new lathe and mill

My new lathe and mill arrived a few weeks ago. Between then and now I've been going through the various processes of setting them up. Cleaning, oil, levelling, tramming etc. and adding some digital scales to the lathe.

For the first jobs, I've turned some threaded inserts and milled a couple of slots.

==

Fast printing PETG filament

[Updated: 5 Aug 2025 Bambu Lab are predicting stock back in towards the end of August.]

In June 2025 Bambu Lab were out of stock of their PETG-HF (High Flow) filament, worldwide. Because of that, I was forced to try other makes. I should point out, that apart from the odd specialist filament, like TPU, I now almost exclusively print using PETG. I mainly print functional parts and they are often for use inside cars, where the heat resistance of PETG is needed to avoid warping.

As mentioned in an earlier article, I wanted to be able to use an alternative filament as a direct replacement for the Bambu Labs version, without the need to change any settings on my Bambu Lab H2D.

I'll start with the benchmark, the Bambu Labs filament:

==

Bambu Lab PETG-HF

• Print settings: Automatic defaults.
• Colours tested: Black, Grey, Red and Yellow.
• Print Result: Good. Negligible stringing, few blobs, tidy seams.
• Surface: Satin to matt.
• Spool AMS compatibility: Good (OD 200mm, ID 55mm, Width 67mm)
• Number of kg spools used: >10.
• Print fails: None.
• Eco: Very good, uses refills.
• Cost: ££££

==

Filament Choices

The main criteria for selecting the following filaments have been the cost at any point in time, and that they get substantially good reviews on Amazon.

If a spool is on special offer, I have taken advantage of that, as my next filament to try. 

These are in the order I tried them:

==

Sunlu Rapid PETG

• Print settings: Same as Bambu Lab PETG-HF.
• Colours tested: Black.
• Print Result: Usually good. Negligible stringing, few blobs. Developed very untidy seams before the nozzle clogged.
• Surface: Satin to matt.
• Spool AMS compatibility: Good. (OD 195mm, ID 63mm, Width 59mm)
• Number of kg spools used: 6.
• Print fails: 1x clogged nozzle.
• Eco: Poor. Reusable plastic spool but no refills available.
• Cost: ££

==

Tinmorry Rapid PETG-eco

• Print settings: Same as Bambu Lab PETG-HF.
• Colours tested: Black and Light Grey (too pale).
• Print Result: Usually good. No stringing, few blobs. Usually tidy seams but one spool of light grey was a little ragged.
• Surface: Satin to gloss.
• Spool AMS compatibility: Good (OD 197mm, ID 53mm, Width 56mm)
• Number of kg spools used: >10.
• Print fails: 1x clogged nozzle.
• Eco: Poor. Disposable plastic spool.
• Cost: ££

==

Eryone Hyper Speed PETG

• Print settings: Same as Bambu Lab PETG-HF.
• Colours tested: Black and Grey (too dark).
• Print Result: Very Good. No stringing, no blobs, tidy seams.
• Surface: Gloss.
• Spool AMS compatibility: Cardboard spool, not recommended. I use a simple workaround.
• Number of kg spools used: 3.
• Print fails: None.
• Eco: Good. Cardboard spool.
• Cost: ££

==

Deeplee Rapid PETG

• Print settings: Same as Bambu Lab PETG-HF.
• Colours tested: Black and Grey.
• Print Result: Good. Negligible stringing, no blobs, tidy seams.
• Surface: Satin to gloss.
• Spool AMS compatibility: Cardboard spool with strengthened edges. I use a simple workaround.
• Number of kg spools used: <1.
• Print fails: None.
• Eco: Good. Cardboard spool.
• Cost: ££

==

Subjective Views

The results above are my opinions after a relatively limited number of hours printing with each filament. This is not an organised review or test, this is just a reminder to myself of how well filaments have performed for me.

I have printed everything at Bambu Labs defaults for speed and temperature. I have not attempted to establish the fastest speed a filament can print at. I want to be able to print reliably, repeatably and with minimal tinkering.

My only changes have been to affect the strength appropriate to the component I am printing. Things like wall thicknesses and infill densities.

AMS Compatibility

Being easy to use in the Bambu Lab AMS 2 Pro and AMS HT is important for me. In my results above, I am referring to those two AMS's. I have no access to any other AMS's to try them with.

Cardboard Spools

Cardboard spools are not recommended by Bambu Labs for use in their AMS's. As the cardboard wears on the rollers, the dust formed is not conducive to the operation of the AMS over time. My solution is to use an overlay spool that sandwiches the entire cardboard spool. The running edge is then plastic. As long as the cardboard spool has an overall thickness of 60mm or less and an overall diameter of 200mm or less, the result easily fits in the AMS.

Conclusion

All the filaments that I have tried, so far, are good and can be used as direct alternatives to Bambu Lab PETG-HF.

==

This page is updated from time to time when I try new filament.

==

Spool desiccant bag and NFC holder

Recently I was forced to buy another make of filament. My original plan was to always use Bambu Lab filament, so I could benefit from the automatic filament recognition and pre-calculated settings. What I hadn't factored in is that Bambu Labs could be out of stock.

Based on some reading, Sunlu PETG Rapid appears to be a very close match to Bambu PETG HF.

I wanted to add an NFC tag holder to the filament spool. I've already printed out a desiccant holder for my Bambu spools, so my idea was to design one with a place to fit the NFC tags. A quick look on Maker World and it appears that nearly everyone has that idea. 

There are a number of good models on Maker World but, as is typical, my requirements are not quite met by the pre-existing designs. I'd like to thank everyone else that has done similar designs, they have no doubt inspired elements of my version:

My specification:

• Fit a range of filament spool sizes from different brands.
• Fit a 20g (0.75oz) desiccant sachet.
• Position the NFC tag in the same position it would be on a standard Bambu Lab spool.

I prefer silica gel sachets rather than loose desiccant. A 20g bag is very similar in size to the double bags supplied with the Bambu Lab H2D, AMS 2 and the AMS HT.

My design of hub fits spool thicknesses from 50mm (2") up to 75mm (3") and central hub diameters from 50mm (2") up to about 70mm (2-3/4").  Sunlu spools are 59mm thick with a central hub diameter of 63mm. Bambu spools are 67mm thick at the outer edge with a 55mm hub diameter. I have also used this with Tinmorry spools which have a lip round the inner hub. In that case, the holder fits and works well, although not very elegantly.

The NFC tag is at the same radius it is on a Bambu spool. A tag from an empty Bambu Lab spool can be cut at the bend and it clips in effortlessly. I've modelled in a central undercut to help extract the tag, should that ever be necessary.

As a bonus, this is the first time I've created a thread to 3D print using FreeCAD, so I've learnt a new technique.

Using the Part Design workbench in FreeCAD 1.0, I created a simple 85 deg., thread profile and used the helix function. The depth of the thread is 2.5mm and is offset into the surrounding cylinder by 0.3mm, inside and out, to create the clearance. The helix uses a pitch of 4.5mm. The screw worked perfectly first time. It's quite a loose fit and I could easily reduce the clearance if needed on other models.

==

Download:

Spool hub for desiccant with NFC tag STEP and 3mf files (zip)

https://makerworld.com/en/models/1541990-spool-desiccant-bag-and-nfc-holder

==