Sow's Ear

I was bothered by the poor results I was getting from my MD A4 trapezoidal screw machine, so since I have had it, I have looked at what can be done to improve it.

Now I know it is an ‘entry level’ machine and to be fair it has taught me a lot about what to do, and more pertinently, what not to do when constructing a machine.

Let’s have a look at the problems one by one :-

First, the top of the gantry was a single piece of 3030 extrusion, offering little resistance to sideways forces skewing the parallelogram. The simple fix there was to put in a plate to brace the top, with plenty of T-nut and screws to fix it down as solidly as possible.Attachment 23136Attachment 23137

The second and more major problem is that all the screws were floating (at BOTH ends) as they were just in ball races pressed into the gantry sides. Any side float limitation was abdicated to the stepper motor. Not good ! So I pulled the motor end of the screw into the bearing and put in a thrust bearing and fixed it with a collar with a grub screw.Attachment 23138 All 3 screws received the same treatment and the accuracy and repeatability of the machine was much better.

I have changed the positioning of the home proximity sensors after crushing one by jogging the machine. The machine supplier wanted £17 for a replacement. (You can get the same one for £2 on eBay).

My machine operates in ‘portrait’ mode, so my gantry moves back and fro on the Y axis. I was having problems with noises from rapid jogging (any body that has seen me will know I have problems with jogging, or any sort of exercise). When I took the table off to have a look, I unscrewed the mounting for the nut and it sprang upwards by about 2mm. I fixed it back with a shim to bring it on centre with the screw. That fault was just sloppy construction !!

The last fix I have done is to stiffen the Z axis. I noticed that holes were being drilled oval, so I had a look at where the movement was. I found that with fairly light pressure I could move the top of the Kress towards the top plate of the Z axis, so obviously cutting forces would be moving the cutter. The problem was that the spindle mount was bolted onto a piece of 3030 and the Z axis rails were just bolted in from underneath.Attachment 23140 This was a weak design with no triangular bracing. My solution was to create a ‘box’ using two spindle mounts and two 6mm ali plates, so I could attach it to the top mount for the rails and make a more solid structure.Attachment 23139
I bought two mounts from China for £14 each. Chosen because the sides weren’t necked in and I had a flat surface to drill and tap M6. The other advantage was that the mount put the spindle 10mm nearer, so less leverage. So far, results are encouraging with cleaner cuts, no snapped cutters yet and less vibration.

Other rigidity issues have become noticeable, but I don’t think it is worth while changing all the unsupported rails. It would cost too much and my pennies would be better saved for a new machine.


There are some features I have added, an air blower nozzle, dust extraction and LED lighting.

There is one goody I have saved for last.

One day the Z axis fell apart. It is basically constructed of two pieces of 30120 profile bolted together at right angles with linear bearings pressed into it. The lower bearings on the vertical just fell out. They were being held in by the plastic profile end cap ! I put in a couple of angle brackets to hold them as an emergency measure and really ought to do a better engineered solution.Attachment 23141

This machine will never be a ‘Silk Purse’ but now it is less of a Pig’s Ear.

Nice write up Cropwell. Some worthwhile improvements to squeeze out some performance and reliability.

I have since done some testing, in the form of a project, two plates cut out of ali. Once I got the feeds and speeds right, it cut well, using a 2mm single flute open architecture cutter from Sorotec. Just blowing the chips away with air, no coolant.

The Chinese cutters I tried either chattered horribly and produced a poor finish to the cut or just snapped.

The finish is done with a dremel steel brush at a slow speed, 0.4mm DOC pocket and WD40.

Attachment 23206