There are a number of information sites addressing reducing spindle motor vibration, but I have not found one here. So, I thought I’d start one.
I’m in the process of a 15 year rebuild of a PCNC770 and am looking into the “fan blade balancing” solution some are talking about.
I have taken the spindle motor cover and what looks like a nylon fan off the spindle motor. I ran subjective vibration tests prior to the cover/fan removal and after. There is a very noticeable difference in the vibration amplitude at most frequencies. Running the motor without the fan is definitely an improvement.
The question becomes – why keep the fan and cover on at all when the motor end is sealed at the shaft penetration, the cover is sealed to the motor end, the cover has no side venting, the cover only vents on the end of the cover, and the fan blades are not angled to force flow direction (only circumferential)? It seems that this “cooling” design has very limited air exchange capability, while adding a great deal of imbalance/disturbance to the spindle vibration modes.
In talking with Tormach Technical Support, this “cooling” design used on 770s may be being used on 1100s still.
With the presence of fins, a large cavity for the motor to operate in, and a vertical orientation, it would seem that the fan and cover could be done away with and normal convective cooling would be adequate. In addition, the collet pullout issue would probably occur much earlier than any load condition that would require active cooling due to motor torque using the TTS system.
I would venture the advantages of the return on vibration reduction would far outweigh any benefit of a small, maybe unnecessary, cooling boost.
Comments on user experiences with fan/cover modifications as associated with spindle vibration reduction on TTS systems are appreciated.
While it’s a worthwhile endeavor, I’d think about making chips first and then see if fan vibration is still an issue. Depending on the work you do it may well be lost in the noise of everything else.
The fan is shrouded as a safety measure. That motor has enough torque to do significant damage should a finger, sleeve, bracelet, hair, etc happen to get near enough. Yes I realize that is incredibly unlikely in this scenario but the motor is not a proprietary component designed for this application, it’s just the motor that happened to be selected in this case.
As for convection cooling vs active, I’d be willing to bet you will find out very quickly why the fan is in place, should you operate the machine for any length of time without it. Speed and torque both play a role in heating of a motor and the area surrounding the motor is not that large to allow significant heat transfer to occur. I suppose if you were to run very short programs, at low speeds, with small tools, you probably would be just fine but as soon as you start taking a decent cut or switch to harder materials, your motor life is going to get short pretty quick. I’m not sure how you’ve determined that the original setup has limited air exchange ability but I can assure you, the fan moves a good amount of air and is necessary in most use cases.
As a grumpy old man I used to work with was fond of saying “Never assume your predecesors were idiots”. In other words, while the reason may not be readily apparent, there is generally a reason things are the way they are.
If vibration is your primary concern, remove the fan and shroud and add in an electric fan as most who have gone down this road have already done. Otherwise, my suggestion would be to finish the rebuild and make some chips before you make any significant modifications. That will tell you real quick if the vibrations are even an issue, and will give you an idea of the amount of heat the motor produces while actively cooled.
Good points Roy and Ian. Part of this effort is the timing and opportunity to investigate any benefits of removing the fan/cover. It has allowed us to identify harmonic nodes to avoid in spindle rpm settings with/without the fan present. Another factor for this effort is we principally work with tooling under 1/8” diameters on very small parts – small features – small cuts – small tolerances. Power is usually not an issue and monitoring of motor temp yields very low operating temps on the motor casing. We’ve seen the posts of others as Ian mentions below of others having put external decoupled cooling to the motor. It will be interesting to gather actual user experience data of cooling modifications to evaluate power/heat and removal techniques based on the machine usage parameters.
As with the Tormach motor chosen, and you are probably aware of, most industrial motors of this type allow for a small amount of air to be drawn from the motor fins through a small annular opening between the fan cover and the end of the motor cooling fins by a directional or centrifugal fan mounted to the motor shaft. But, this cooling is probably designed for the rated horsepower of the motor. Our primary interest during this “experiment” is to see if we can get by with vibration sources reduction under low torque applications, i.e., will the motor bearings be affected by reduced cooling?; will the heat rise be significant enough in our run times to have to consider balancing of motor fans?; would external decoupled cooling be even needed?
A good way of us expressing this question would be at what point does forced convection really matter and under what loads to the motor?
It will be interesting to see what other users that have made these type of modifications have experienced.
Several years ago, a guy on the CNC-Zone sold an add-on fan for the 1100 mills. It was a fan much like the ones used in PCs, was AC powered and mounted in place of the original Tormach fan with a 3D-printed housing. I installed one on my 2006 1100 and it did seem to reduce vibrations a bit. It would be pretty simple to cobble one up with a fan from Mouser or similar and if you have access to a 3D-printer.
Thanks Mike. We are familiar with those types of designs and have very large resources in terms of manufacturing capability to do the same. Our primary emphasis is on seeing if we can reduce all vibration inducing sources without having to do optional retrofits or add-ons. We also use some very small spindle machines that have no auxiliary cooling and are trying to determine if the same principal can be applied to the 770 size machines with very low torque production requirements. It’s more of an interest in users that have completely deleted their fan/cover or have gone through a balancing effort – how far can we go in torque/power/time without affecting windings, bearings, etc of the motor. The 770 we have has been highly modified for precision without getting away from the TTS tooling concept. As an example, we have 0.00005” DRO scales on the axes. Precision/repeatability/surface finish impacts are the focus on very small features/parts over the entire table working area. We have also learned to not push the 770 over about 0.2 horsepower in order to not risk tool holder pullout.
While it probably doesn’t matter given the size of tooling you’re using, your last statement makes me think your drawbar is very loose and needs adjustment. My 1100 (prior to upgrading to the BT30 spindle) could absolutely utilize more than .2 hp without pull out. And there are plenty of documented cases of 4-6 cu in/m removal rates with various larger tools out there that would indicate the same.
All that said, I’d be surprised if you found many, if any users that have removed the fan without replacing it with something else and can give you first hand experience. Your axes scales alone are so far beyond what is reasonably expected from these machines that I suspect you are in totally uncharted waters. For that matter, I’d be surprised if the mechanics of the machine allow you to achieve anywhere near that level of precision and accuracy. While sub .001” is absolutely achievable, most comments that I’ve read, and my own personal experience, is that a .003”-.01” total tolerance is where your expectations should be in a production environment. Less than that generally requires careful walking in of final dimensions on a per piece basis. The dovetail ways and gibs force a compromise between rigidity of each axis and lost motion.
All THAT said, if you’re actually able to achieve sub tenth precision, be sure to post about it. You would be the first I’ve ever seen at that level for sure, so it would be terribly interesting to learn how you got there.
Ian, we are actually in uncharted waters! This is a special machine modified with the help of Tormach in the 2010’s for very small parts. Our horsepower limit is self imposed based on performance using the Tormach Power Draw Bar system with TTS. We have been round and round working on that issue with Tormach, and yes, you can definitely achieve higher material removal rates with manual draw bar torquing (or just avoid the issue and use BT or similar spindles/tooling). Our emphasis is on the investigation of potentially determining where the motor operational limit parameters are if there is no forced cooling (convective, conductive, auxiliary, looped/controlled systems, or other cabinet changes to the standard installation). And you are correct in that special techniques of using and adjusting the stepper/bearing/gib positioning capabilities must be employed and are not suitable for use across the full horsepower range of the machine for general machining. Again, thanks for the input. Just trying to share to see if any others have tried the “no cooling” approach. We may be experimenting on our own it seems.
Probably a centrifugal fan, so it may not look like it does much but it does. Chuck the fan up on a shaft and see how much it wobbles. Vibes should be somewhat isolated by the belt. I would focus on a spindle rebuild, make sure all surfaces are getting lubed to prevent sticktion, snug up the gibs and tram the column last. New spindle bearings with good lube makes a huge difference. Wash and dry new bearings before lubing. Really need to pop the table off to make sure oil is getting through. That really effects backlash settings.
Still want to know why axis scaling can’t be adjusted.
Doesn’t the carbide attached to the spindle, chewing into the metal make significantly more vibration than a plastic fan on a motor connected to the spindle by a rubber belt?
Maybe/maybe not. How to tell without data from a PSD, RMS calc, then look at node data under different configs material/cutter/speeds&feeds. Quite a test matrix. Haven’t seen any of that on the 770 anywhere. What would be the evaluation criteria — surface finish, tool life, material removal rates? Why not just eliminate as many of the frequency distributions to minimize the system resonance nodes or at least characterize them to avoid operating there. We already have data with the fan installed which is indicating avoiding spindle speed in the 7500 to 8500 rpm range due to system resonance (for our machine under our config with balanced tooling– not to be taken as a general rule for a Tormach PCNC770). I cannot validate your assumption of the tool cutting spectrum vibration amplitude being greater than the spindle/pulley/belt spectrums may/may not be correct. But in terms of the system cannot be the the criteria for decision. The real parameter to look at is the system spectral density. Who knows – maybe we can find a clocking angle of the fan relative to the spindle shaft that acts as a damper to the nodes, or maybe we’ll just develop a dynamic variable balancing fan for the system. Fun stuff! Thanks for your input.