Screw Machine Progress ???

I have the screw machine “working” with the new 16MM screw that is more appropriate than the 20MM I was using before.

I decided to use a servo motor and gecko drive to increase speed and torque at high speeds. A servo motor should be able to accelerate faster and run faster as well as having high torque at high speeds as opposed to a stepper that has a much worse torque curve and difficulty accelerating fast.

In spite of the fact that the servo is supposed to be better in the above mentioned respects, I am having a very difficult time getting it to perform well. I think I am slowly figuring out how to get decent performance but it has been a long road of trial and error and guessing. It seems that the stepper motors drivers would just ignore impossible moves. With no feedback they would do the best they could and not get upset. Some steps might be missed in rapid accelerations. The servo has feedback from the shaft encoder and  it goes into fault mode for 1 second each time the error count exceeds a set threshold. This is great for a milling machine or other machine where position is critical at every point. For a fucking machine, this just gets in the way of progress.

So I am experimenting with processing of the joystick encoder data before sending it to the motor driver. I have made significant progress but no to the point that the machine will accelerate as rapidly as I demand.  If I could get a faster lead ballscrew I am sure the way it is working now would be adequate. I am not sure if these are easy to find. High lead screws are available but I am just not sure if this can be done in a ballscrew due to the way the balls need to recirculate.

I will keep working on the step pre-processing to get the best performance. It looks like the highest pitch for a 16MM ballscrew is 10MM which is twice the common standard of 5MM. Unfortunately it looks like they are only made by companies like Thompson and are very expensive.

My internet connection is very slow so I can not do a lot of searching. It would be great if somone with a better connection could research this matter. A low cost 16MM screw with a pitch of 10MM or greater would be ideal. There are higher pitch screws but only in much larger diameters. Also it might be possible to find a non ball nut type. I have one that is very fast but it is only 12 MM or maybe even smaller and would probably not last very long. I can experiment with it if it turns out that my step processing alone can not produce fast enough acceleration.

This entry was posted in Uncategorized. Bookmark the permalink.

8 Responses to Screw Machine Progress ???

  1. AZ_Kinkerer says:

    What length are you looking for on the ball screw??
    I’m looking around now… Found a 1610 x 400mm for $36 with tax, but it’s in Europe, and you’d need to contact them for shipping cost. So far, all I’ve found in the US start at about $100.

    I picked up a Smithy 3 in 1 Mill/Lathe/Drill Press that I’m in the midst of refurbishing and getting set up. This will give me the ability to make anything I need since my buddy’s machine shop closed up at the start of this year.
    I’ve picked up the gear and length of belt I need to change mine over from the friction drive to the cogged belt. A couple of old pulleys from one of the machines I work on look like they’ll work perfectly as idlers. 😉

    • shagmatic says:

      i got the servo motor to work quite well so i am confident that with the right screw it will be a very good machine. 10MM pitch is better than 5MM but i don’t think it is quite fast enough. i will keep working on the servo driving firmware and might get it faster. it is a little tricky because motion from the joystick needs to be intercepted and changed into controlled acceleration motion without making the machine un-responsive or “rubbery” if the acceleration is too slow the machine is not responsive and if it is too fast it will fault. the faster the screw (as long as enough torque is available) the better. i think that .5 to 1 inch pitch would probably be ideal but can not know for sure until i test it. the screw i have on order is .7″ pitch according to the description but the picture looks like something different.

  2. shagmatic says:

    I just got the servo motor working much better so maybe it would work with a 1610 screw. the 1605 is not fast enough for really rapid thrusting. it would probably be OK for short quick moves but the ramming that the servo should be able to do is lost on too fine a pitch.

    I have a 7/16″ maybe (can’t find a caliper in the house and it is pouring) acme screw with 8 starts and a 1″ pitch. this will definitely be good enough for testing. I found a few 5/8″ pitch two start ball screws on e-bay for around $65 with shipping which are long enough. Once i test with the Acme and know that the servo will work, i will probably buy one of these. Normally high pitch ball screws are very expensive. I think that a 1/2″ acme with 5 starts and a pitch of 1/2″ or would be be fine but i just don’t know how long these would last. Thompson’s web site has a calculator for all things ballscrew and it looks like a 1/2″ screw could last for something like 1000 years with the relatively low load of a fucking machine 🙂

  3. Larry says:

    Here is a screw for $22.80 US, shipping not included.

    16mm dia., 10mm pitch, 700mm length

    It may need to be cut down to size though.

  4. Meviane says:

    Hi from France !

    Thanks a lot for all the infos available here.
    The fucking machine based on a ballscrew is a very nice idea, here is some complementary info about the design.

    ¤ All ball screws are provided with a nut mounted on them, it needs to be removed with a rigid cylinder, which needs to have very close dimensions (I’m talking about Internal Bore), it will allow to retain all the balls contained into the nut.

    ¤ All ball screws needs to be machined with a lathe, but the tracks are highly accurate, so the steel is highly hardened. In order to avoid any breaking of lathe’s cutting tools, it’s often necessary to cover the screw with a clothe with stringent sealing; in order to use an angle grinder, thus allowing to remove the hardened steel. To avoid any contamination of balls with particles; the cleaning stage after machining needs to be perfect.

    ¤ The servomotors can be easily ordered at DMM servomotors, located in Canada.

    They are providing 2 families : the DYN2 and DYN4. DYN4 can be powered directly with 230 Volts AC, they can run up to 5,000 rpm with movements obtained in less than 10 milliseconds. But the power is close to 3/4 or 1 HorsePower, with a NEMA34 framesize. It’s too much power with serious risk of death.

    The DYN2 servos seems much adapted in that case of use, the framesize are rated with 4 holes in a square, dimensions from 40 to 60 millimeters, the upper range is sized in NEMA34. But they requires a separate DC rectifier (the tension can be used between 24 to 75 volts, typical 60 Volts DC), and also an independent driver.

    Both families can easily be commanded via an USB cable, through a Windows PC, or a programmable microcontroller like Arduinos. Decelleration/Acceleration can be easily set up, with angular position or speed or torque commands.

    A ballscrew connected via a linear guide should provide a “profiled, narrow” fucking machine, easier to integrate in a seat/bed/dedicated furniture. But the power used by servos must be kept into consideration, because the efficiencies of ballscrew is really high : the force applied onto the sextoy can reach several hundreds of Newtons, here is a simple equation :

    For a ballscrew of 16 mm O.D., with a lead of 10 mm, by using an efficiency of 0,95 (95% yield), if you pick the DMM’s DYN2 servo model no. 410-DST-A6TS1 (price $133) coupled with the drive model no. DYN2-T1A6S-00 (price $138), this 100 watts motor will deliver a torque of 0,318 N.m at 3,000 rpm, the resulting force is 180 Newtons (or 18 daN), an equivalent of mass of 39 kilograms, so it’s close to the motion of a young female.

    Here the specs :
    Servo :
    Drive :

    This servo can reach 5,000 rpm, but at a torque close to zero. Please keep in mind that DMM present the maximum torque of 0,955 N.m, obtained at a speed close to 3,000 rpm, then the screw will create a thrust of …570 Newtons (57 daN…).

    Here is the DMM products infos :

    It will be a pleasure to follow this thread, personnaly I will order some parts quickly for 2 prototypes, using the 1610 screws, to be mounted onto 6002-type ends bearings, coupled onto a linear guide.

    I’m aiming a max. stroke of 10 to 12″

    Wishing you the best results !

    • shagmatic says:

      i did some work on the screw machine firmware and got it working pretty well with my servo motor but not as good as i want. i need to spend some more time when i can to tune the motor driver and possibly find some way to improve acceleration in the shagmatic firmware without adversely affecting the performance. it might just be that the brush type servos can not be made to perform well enough. only time and experimenting will tell. i have not been able to connect a scope to the test pin of the gecko servo driver in order to tune it. so, i can only tune it by “feel’ i am hesitant to spend a lot of time on the firmware if the issue is really something that can be fixed with driver tuning. i even called gecko support and was on the phone for a while be “we” could not get the scope to read the diagnostic output. will try this again when i have time. if anyone has done this successfully it would be good to know what they did.

      • Meviane says:


        The best way to achieve such improvements with acc/decel features is to “simply” connect the servo via an USB port, a dedicated software helps DMM users to finely adjust any settings; once it has been determined, a dedicated microcontroller can send a series of orders via a “string output” or a variable DC signal or a series of phase’s quadrature.

        Here is a nice example of settings on a DMM Dyn4 servo, rated at 750 watts :

        Your website & datas has given to me the idea to make some orders for a build within weeks. Please consider making a video like RuTube or others platform about your research, it can help you to find help/resources/collaboration.

        Best regards.

Leave a Reply

Your email address will not be published. Required fields are marked *