Model Engineering Articles

Mill Quill Feed Backlash & Mill Spindle Bearing Pre-load

I have been using the 626 Mill quite happily for some time now. By chance I saw an article on a web site which described some changes made to a similar 626 type of mill. The writer also noted that the backlash present in the quill downfeed was excessive. I had the same with my mill but as long as the spring is set to lift the quill upwards then the backlash is taken care of. I decided to have a look at the problem in a bit more detail to ascertain if the backlash was inherent in the design or could it be further reduced or even totally removed.

To test for backlash ensure the clutch feed is disengaged and clamp the quill with the locking handle. By moving the RH down feed handle (used for drilling and large feed movements) the amount of play can be easily felt.

After stripping the left hand side clutch mechanism you can easily see the left bearing. This bearing is held in position as a press fit and prevented from rotating by a small keep screw. Once the screw is removed the bearing will come out by gently tapping with a wooden block. The other bearing is on the right hand side where the down feed handle is positioned. The backlash is mainly due to the mating of the quill feed gear to the quill rack teeth. In my case there was never going to be a total cure but the gear teeth could be moved into closer contact. Even then there would always be an element of backlash which I think is due to the lack of precision in either the gear or rack teeth, probably both. The picture shows the modified bearing with bush.

It was soon evident that the lefthand bearing was a sloppy fit and so this needed fixing in any case. The right hand bearing was a good sliding fit. I decided the best option was to bore out both bearings (cast iron) and fit a new phosphor bronze bushes in them. If each bush was eccentric then depending on its relative position in the cast iron bearing it could move the gear nearer to the rack.

In practice it was a relatively simple job to bore out each cast iron bearing. The phosphor bronze bushes were made to be a sliding fit in the bearings with a slight taper at one end to enable the bush to be retained whilst machining was carried out. This way the bushes would be easily pressed in and held firm to enable boring to an exact size. The bearing assembly was held in a three jaw chuck and a small distance piece (0.025") thick placed against one jaw to provide an eccentric throw. The bore was then done to a sliding fit.

With the bushes separated from the bearing they were brought to finish length and the taper end removed. This way the bushes could be placed in trial positions and later loctited once the best position was found. The right hand bearing has a small oiler which on removing was found to be completely blocked with filler! A small oil hole needs to be drilled through into the bush and the oiler refitted.

After final assembly the amount of play was re-tested. The resulting backlash is significantly reduced with only a small amount of play felt. I now also have the satisfaction of knowing each bearing is well engineered and the backlash is as minimal as possible. As a matter of interest I compared the mills modified quill play to that of my Swedish drilling machine which is a precision made machine. There was no discernible difference, very pleasing result.

Having now got the Quill gear and rack backlash or play to an absolute minimum my attention turned to the fine down feed mechanism. To measure the backlash here it is necessary to engage the clutch fully and feed downwards to ensure the drive is taken up and then reverse the feed upwards until the quill starts to move (I have a digital scale attached so this is easily seen). The amount of backlash is evident by the movement of the feed index before takeup. In my case the backlash measured 0.047" which is totally unacceptable!

Yet another case of modifying the method of engagement for the worm and wheel drive of the feed mechanism. On taking the feed mechanism out of the head it can be seen that the worm spindle is held in a cast iron body with two ball thrust bearings at each end. The bearing housing is (in my case 1.179" diameter) but since it is concentric with the worm spindle no adjustment for engagement is possible.

The remedy is relatively simple. A new bush will be made and fitted to the worm spindle housing and have eccentric adjustment. A start was made by finding a 1.25" diameter piece of mild steel bar and bore this out to an internal diameter around 15/16". Next a mandrel was made which was the same diameter of the worm spindle so that the housing could be held for turning it's 1.179" diameter down to a good sliding fit for the bore of the mild steel piece but with a slight taper at the flange end. This taper will enable the mild steel sleeve to be held firmly for further machining. At this point the cast iron housing reduction in diameter is concentric with the mandrel. The mild steel sleeve is now tapped on so that it is firmly held in position. Next the mandrel needs to be slightly offset in the chuck. The amount is not too critical but MUST NOT be so much that the ball thrust race bearing outer diameter is proud of the finished diameter. Once set up the outer diameter is machined to 1.179" to be a sliding fit in the head. Only small cuts (0.004") can be taken in reducing the finished diameter as the mandrel is not very substantial and the start of the cut is some 2.5" from the chuck with no end support.

After machining the cast iron body it is removed from the mandrel, the mild steel sleeve removed and the slight taper on the body filed down. The mild steel sleeve should now slide up to the flange with no binding. The worm spindle and thrust bearings are assembled and a trial fitting is made with the sleeve rotated appropriately. When the optimum position is found where play between the worm and wheel is minimal the position of the sleeve is marked. The sleeve is now permanently fixed to the cast iron body with bearing grade adhesive as there is very little load on the joint. On final assembly back into the main head the backlash was re-tested) From an original backlash of 0.047" it now read 0.004" which is not only a vast improvement but also made the feel of the fine feed so much more precise. In my case the sleeve was adjusted to it's maximum position of eccentricity. If I had just a few thou more adjustment then it is likely all backlash could have been further reduced. Unfortunately the setting of the mandrel for eccentric turning is a bit of a guess and it is very important that it is not overdone otherwise the thrust bearings will bind in the main housing.

It is definitely well worth making these modifications as the change in performance is dramatic and the whole operation of the machine is more precise. A small amount of backlash is necessary to prevent wear and binding but as supplied the backlash was massively unacceptable. It is a pity the manufacturers did not incorporate these eccentric bushes into the design as they are simple to make and enable backlash to be minimized on assembly.

Mill Spindle Bearing Pre-load (new addition)

For some time now I have been a little concerned that the pre-load on the spindle bearings is probably not as good as it could be. After the mill has been running for about twenty minutes on full speed the spindle is only just slightly warm to the touch. This in itself is no absolute measure that the pre-load is wrong but I did expect it to be a bit hotter. Another area of possible concern was that the adjustment nut was unable to provide adequate pre-load because the top bearing was limiting the spindle movement. The only way to check things out would be to dismantle the quill and spindle assembly for a closer inspection.

Before starting this task I decided to make a tool to adjust the locking nut on the spindle. The tool has four teeth that engage in the lock-nut slots and a handle about 6" long. This handle will hopefully prove to be long enough to apply enough pressure, by hand, to judge the pre-load.

After a quick strip down of the quill I removed the lock-nut and spindle. The spindle is quite well made and has a respectable ground finish. The quill end (nearest the cutter) has two bearings but despite my efforts and no puller I decided to abandon removing them as there was always the possibility they had been fixed with bearing adhesive. One of these bearings is a substantial angular type (7202) and is backed, according to the parts diagram by a standard plain ball bearing (6007). At the top end of the spindle is a plain deep groove ball bearing (6206). It would probably make sense for this, at some time, to be replaced by an opposing angular bearing (7206), and for the lock-nut to apply the requisite pre-load as before only this time the angular bearing is a lot more suited to applying and holding axil loading. I did manage to remove this upper bearing and check its fit and relative position to the bottom bearings. I was concerned that the bearing might not be correctly seating in the quill or that the spindle was tight against the underside of the bearing and thus preventing any vertical movement of the spindle which is essential to apply pre-load. I was able to set my mind at rest and ascertain that there was sufficient clearance movement (about 0.050"). Once all parts were cleaned and re-assembled the nut-locking tool above was used to apply some pre-load. Basically, this entails tightening the nut until all clearances are removed but not so tight as to cause premature wear or rough running. Because the handle is 6" long I think hand tightening this should prevent over stressing the pre-loading force. With my previous Dore-Westbury mill the manual stated that the spindle be adjusted in a similar manner until a small amount of drag could be felt when rotating the quill but in the case of the 626 quill assembly I found it difficult to feel any discernible drag and so one has to rely on hand pressure alone. The lack of 'feel' was possibly due to the top bearing being a plain ball design and not ideal for pre-loading and unfortunately it remains a bit of a guessing game. I am reasonably convinced the best approach is for the top ball bearing to be replaced with an angular type positioned to oppose the bottom one and thus distribute pre-load more evenly however the amount of pre-load still remains a concern. Looking through a bearing suppliers list I managed to find an angular bearing (7206) of the same thickness with a 40 degree contact angle. The contact angle of the bottom angular bearing (7207) is not specified and thus yet another thing left to guess!

After everything was re-assembled I did a dry (no suds) test using a much re-sharpened 0.5", 4 flute end-mill and set the cut to 0.25" deep on some mild steel. All went well, a few noises but nothing to get worried about and the bearings did not groan under the force of cutting. This is much better than before the adjustment when such a depth of cut would have resulted in some horrible warning noises and over heated chips. I can only assume, as delivered the pre-load may have initially been done correctly but after many hours running has become less effective. This could of course be symptomatic of wear in the bearings or as a result of pre-loading a deep groove bearing which is not ideally designed for this purpose and will in all probabi;ity ease up after running. Like many things on these Eastern import machines where hand fitting is done it is very variable how well things turn out. On some inspections 'beneath the covers' one might be more than a little surprised and possibly disappointed at some of the poor workmanship hidden from view. In this respect this mill seems pretty good. As the spindle, bearings and quill are largely responsible for the performance of the machining process and resultant finish produced, it is well worth the time spent checking all this out. This now raises yet another thought, the quill fit in the head. Many adverts claim the fit is a lapped one but I fear this too is not quite as good as it could be. It is not all bad news since most operations, especially when taking heavy cuts, are only done with the quill firmly locked. For accurate boring then I think the best approach is to apply a small amount of lock pressure to the quill but just enough to still allow the quill to feed downwards. A sharp cutter and small cuts are of course appropriate for accurate work.

All of the above reveals some of the shortcomings of the 626 mill and variations in supply. Does this matter? For the hobby user I don't think so. If one is aware of the shortcomings you can still produce very accurate work but the 626 is never going to be a high volume accurate production machine. On the other hand it has a lot of advantages over such a typical production mill, such as a used Bridgeport. Unless one is very fortunate the used mill will have a fair degree of wear in slides and bearings, the latter being extremely expensive and difficult to replace. The 626 has a basic and crude bearing arrangement but as these wear they are easily and cheaply replaced. For me it is all about getting the best out of my machines and where possible enhancing or eliminating design shortfalls and cost cutting. The 626 is a very capable milling machine but a new machine straight out of the box may need some minor adjustment to give good performance and some extra enhancement work in order to achieve the optimum performance it is capable of.

The replacement 7206 angular bearing arrived, standard grade and not expensive. After fitting the new bearing in the quill assembly comes the dreaded adjusting the pre-load. As mentioned above It seems that applying pre-load is a bit of a guessing game, The objective of the pre-load is to remove all axial clearances and apply a small additional load to maintain this condition. All things considered, probably the easiest method to apply a measured and repeatable pre-load is to adopt the following procedure. First, the assembled spindle in the quill is tightened using the lock nut (without the locking ring) with the tool and a reasonable amount of hand pressure until no vertical movement can be felt. The spindle is revolved to ensure bedding in and further tightened, the position of the locknut is released and the same done only this time the nut is tightened (with the locking ring) by applying medium hand pressure so that the nut is tight. This now offers a bit more confidence to judge the pre-load from light to heavy. The objective is to apply a pre-load that is enough but not excessive and using this method, whilst crude, does enable a reasonable guess to be made. In practice there is very little difference in the amount the lock nut is turned between very tight and medium tight and almost impossible to discern. The saving grace is using the lock tool which has only a short handle, 6" long, but the area gripped is probably only about 4" from the centre of the spindle. This will apply a moderate pre-load torque and unless you are very heavy-handed and strong it is unlikely the bearing will be over-stressed. Whilst not ideal and very basic this is all that can be done.

The real test comes after assembly and the trial cutting. I repeated the same tests described above and with the same cutter. The difference was quite noticeable in that the cutter seemed to proceed effortlessly and chips rolled off with a clean shiny appearance. there were no unexpected noises and everything seemed much smoother than before. I decided to try depths of cut which I had previously considered impossible and everything seemed to cut well. I do not intend to use such heavy cuts in normal use as this is expecting too much, but this change of bearing does appear to be a sensible modification with real improvement. Best of all it costs very little and takes about an hour to do.

Quill Backlash - eliminating radial movement (new addition)

Now I thought I had finished making mods to the mill but was milling some metal today and I thought about the problems with the quill and its fit. Mine is not what I would describe as good close-honed fit but on trying to twist the bottom part of the quill I discovered there was some radial play. The substantial detent-block at the front which fits in the head-slot is supposed to limit this but as with all things Chinese they prefer to leave a little clearance to save having to fit them which takes time and skill. This is yet another probable cause of vibration and ultimately wear on the splines unless the quill is firmly locked each time. Re-fitting the detent block is an easy modification and quickly made.

The detent block is removed and a 0.5" slot about 0.1" deep (correction) is milled on one front side edge (I chose the LH side). Into this slot I fitted a flat piece of brass and secured it with a 4BA csk. screw. The assembled block was then placed in the mill-vice and the brass surface milled until it was a good close sliding fit in the head (it was probably only about 0.004 - 0.006" higher than the original). The brass filler piece was then removed and a piece of paper placed under it and re-assembled. A quick touch of surface lapping and it was now a very close fit with no perceptible lateral movement even under strain. The down-feed spring was unable to freely lift the quill but I don't think that matters as I rarely use the mill for drilling and it is hardly sensitive anyway. With the fine-feed clutch engaged, feeding now felt much smoother and the quill rack and pinion has lost a lot of the play (this was previously modified using eccentric bushes to close the gap). The end result is the machine feels a whole lot better. Any future wear can easily be made-up with another shim of paper underneath. Initially I had thought about some kind of adjustable gib arrangement but the simple brass pad works better and as long as it fits and moves smoothly it should transform the feel and operation of the mill under cutting loads and remove another source of vibration.

Return to Home Page or next to Next Article ... or related Buying a new mill or related Minor Modifications or related Mill PowerFeed