Two motors on one drive?

[giggler]

Hi,

I have two small'ish Nema 23 step motors, single stack, 4 wire configuration, 1A rated, at 5V. I want to drive an X axis with two motors, same model and specification.

The single drive I have is a leadshine, 2M415, 1.5A max drive, powered by 24V DC. Microstepping, currently at 64steps but might reduce since lead is very small to probably 16steps.

The machine is to inspect PCB's so there is no tool and it is very light, no real loading, lead screw, gantry design.

I can see only two routes forward as I cannot add an additional drive,

1). Parallel them up on the single drive and set the current to 1.5A output, enough to drive both motors at 75% rated current, full 24V.


2). Serial them with the single drive, current set at 1A output, accepting the drive voltage loss of basically half, i.e. 12V.


What would you do? Which wiring would you choose and why?

I feel more pushed to the parallel option, obviously if one motor dies, the other single motor will be over driven but in that situation a lot worse things will happen since one leadscrew will turn and the other not so I doubt this situation will be prolonged (i.e. Bad noise, oh crap, panic and E-Stop pushed).

Regards,

dc

[wildwestpat]

Looks as if your motors are high inductance types from the 1amp 5 volt rating. Your drive is of the current modulated type so it will sort out the current shaping (PWM) correctly for series connections of the two motors. For parallel IMHO you would not set the current as higher value as steppers run hot. You need to find out if your drives can be set for reduced current when not stepping - if they can then you could increase the current setting a bit but not to twice the current as you cannot be sure how the current will be divided. If you have the means of monitoring the current pulses in each motor winding then do so - lead lengths and lead routes should be similar for both motor to drive connections and to the power supply. In general parallel connection is to be preferred as it reduces the inductance seen by the driver and hence the speed with which the current can be made to rise in the time allowed by the pulse width modulated drive. Or put another way parallel connection is good for more torque and a bit more stepping speed.

There is a mechanical problem in that the drives need to be synchronised - this has two components - first the motor to axis drives should be mechanically connected at a stationary full step - secondly there is with all stepper motors an error in angular displacement with step command and this means that the two motors will tend to tussle for the position at each step and this might accentuate any resonance problems also there will be variations in the drive screws to consider.

Do I understand you correctly you are proposing to connect two lead screws one screw and one motor on opposite sides of a common gantry. If so the gantry will stall unless the screws and motor can free wheel when failed.

A simpler option might be to replace the motor with a more capable one.

Regards - Pat

[giggler]

Hi Pat,

Thanks for your reply.

The drives can't produce 2A continuous (double motor rating), so I figured 1.5A split approximately between the two might work. However from your comments, I can reduce that down to around 1A which matches my motor ratings.

First you suggest series, and then make a statement about parallel connection being prefered. I'm still a little confused about the best approach.

I forgot about cable lengths, yes I should attempt to make them the same length, that means relocating the control box but I guess that's life.

You have the right idea about the two sided driven gantry design, however it's not a motor issue. I mean, one motor happily powers the gantry on 1A setting, but the torsional affect of one side pushing and pulling on the other means that the axis moves trapezoid, or at an slight angle.

The reason for duel drive was to solve this issue without adding weight and renforcement to the X and Y axis's, that would also mean expensive material purchase of new metal, design time and machining of axis sides.

I know some commerical routers that use rack and pinion drive; have duel drive on the X axis, but I'm using fine pitch leadscrews and brass nuts. It's all pretty small and light stuff, but I want more accuracy. 8mm leadscrews, 1mm pitch on all axis. Presently I am having to over drive to position and pull back so the axis is straight and almost true (I guess you'd call that a sort of backlash). That's the only reason for the additional driven side.

The 12mm linear rails are all good and smooth but do flex a little over length because they are only end fixed and not supported all the way which is an issue because they tend to grip a little at intermittent points as they flex. Again I figured driving both sides would solve this as it seems to when I disconnect the leadscrew and move by hand, the same flexing and gripping occurs when drive by 1 hand; both hands and the movement is smooth and free.

Regards,

dc

[wildwestpat]

Sorry I did not explain. Putting the windings of each motor in series increases the inductance and that limits the maximum speed / torque in the same way as wiring an eight lead motor where parallel connection of the windings will give more high end torque particularly with PWM drivers. Putting the motors in series solves the current sharing but at the expense of torque at the higher stepping rates. That is why I suggested trying to keep lead lengths and resistances the same for each drive if you connect the motors in parallel.

My other concern is with the way each step is translated into physical movement and the way that movement has an affect on the other side of your gantry. If your machine is a light build then the two sides of the gantry are coupled by a structure that can flex a little bit and this has to absorb the differences due to motor non linearities and those of the screw - we are talking differences between the two sides. This might induce resonances if you are unlucky. However since there is no cutting loads to consider you should be OK but if the inspection head is a camera then it may need a short period between steps to allow the structure to settle. This used to be a problem with the old optical flicker comparators for PCB work.

OK I missed the reason for wanting to drive the other side was to keep the gantry legs in step. A simple way would be to use a timing belt with a pulley on each lead-screw. This would keep the screws in step - this is a common configuration to solve the crabbing of the gantry. The motor or motors could then be added to the timing belt path if this simplifies the layout of the modification.

Hope this clears up any ambiguity in my response to your question.

Regards - Pat

[giggler]

Thanks again Pat, that clears things up in my mind.

I have considered the timing belt option and will think about it more at the drawing level this week. I have three issues with this and my current design:

1). the leadscrew is nearly maxed out with the limit sensor positions so I'd probably only have 10mm at best to play with, adding a pully - doable.

2). I already have the metal and motor to add if needed (-new leadscrew) but not the pully's and belt. Also I am going to struggle getting a large enough belt and would probably have to go for some kind of joined belt. Something, I've not really considered in much detail.

3). I did a quick look at the drawings with a rule and I may not be able to run the belt far enough away from other parts of the machine to not interfere.

The head is a usb high-res. colour cmos camera with auto-adjusting lense. I don't at present intend to move the machine, taking pictures and accept there will probably be some settle time, as it is the auto-adjusting feature will probably take a few mS to get focus.

I guess now I know that the axis will be a little more flexible than I want, next time I design an X/Y table of similar dimensions and build, I'll factor the possibility in at the start I live and learn

Regards,

dc

[wildwestpat]

Here is a link to a belt supplier 3mm pitch 9mm wide. http://www.motionco.co.uk/timing-bel...-25_38_39.html There are some others that are smaller and when I am back home I will try and find the link. It is a good idea to have some form of belt tension adjustment so perhaps you could use a pair of ball-races on fixed pivots to divert the belt away from any fixed parts as well as providing tension control - just a thought!

I always think it is a fool who makes the same mistake every time. It sounds as if you have learnt a lot with this build and mark2 will be better and then on to mark3 if you get bitten by the CNC bug.

Regards - pat

[wildwestpat]

If you are going to compare the "new" PCB portion with a known good PCB image then in my experience the registration does not have to be 100% as the eye is drawn to the mismatch. I have instigated the use of an optical version on production checking in the 60's where there was a problem with the colour codes on resistors not being that easy to read. These components were miniaturised but large by to-days standards.

Regards - Pat

[wildwestpat]

Found the links this morning for small toothed belts motionco_Pulleys & Belts_Grooved Round Belt Pulleys_Buy from Stock also has a belt pulley centre calculator. Another source of the slightly larger pitch of 3mm is : Zapp Automation Ltd - Stepper motors & Stepper Motor Drivers - Servo motors and Servo Motor Drivers - Ballscrews & linear motion products..

Another way of keeping the legs of your gantry in step is to use a pulley system. (At one end of the track mounted on each side one pulley with their axis in the "Z" plane at the other extreme of the track two pulleys also with axis in the "Z" plane. A cord is then wrapped round the pulleys such that a single loop covers both sides and crosses over to the other side at the end with two pulleys per side. Simply anchor the legs to the appropriate cord to ensure the legs move together. For a miniature cord I would suggest trying the carbon fibre used in GRP layups as this does not stretch yet is very flexible and since the cord is anchored in two places there is no need for an endless loop/band. This sort of arrangement was common on the large drawing boards in the days when large scale drawings were the only way of designing small mechanisms before being over taken by CAD.

Good luck with the build - Regards - Pat

[harryn]

Thank you for asking this question - it has been on my mind as well. The reason for me is simple economics of the cost of 2 x drivers for 2 x motors on an axis.

I have a gecko G213v which can drive 7 amps, and 2 motors rated at 3.5 amps x 12mH inductance.

It would be great if I can put these two motors in parallel and drive them from one driver. (saves a money)

I understand the aspects of being out of alignment for each side of the gantry, but I think I can deal with that part during the setup.

What I had no idea about, is whether or not there would be problems with one motor somehow slipping into a different micro step (angle) vs the other motor. In this case, there is not an easy way to mechanically tie the two motors, so it is either make this work, or buy 2 drivers.

I wouldn't care, but I have to do this for several axis.

Thanks

Harry

[wildwestpat]

Harry - you do not say why you want to do this. Giggler's question turned out to be keeping the gantry from crabbing and a second motor and lead screw is not the only option or necessarily the best.

The problems with connecting two motors to a single drive are as stated earlier in this thread. Much will depend on circumstances but if the aim is to increase drive torque or speed then separate drivers is the way to go if you can tolerate the motor step non linearity and the drive positional error differences.

Regards - Pat

[harryn]

Harry - you do not say why you want to do this. Giggler's question turned out to be keeping the gantry from crabbing and a second motor and lead screw is not the only option or necessarily the best.

The problems with connecting two motors to a single drive are as stated earlier in this thread. Much will depend on circumstances but if the aim is to increase drive torque or speed then separate drivers is the way to go if you can tolerate the motor step non linearity and the drive positional error differences.

Regards - Pat

Hi Pat, Thanks for the reply.

The router I am attempting will have 2 motors on the X, 2 on the Y, and one on the Z. The reason, is that based on my testing and calculations, this is what is needed for power and to keep things aligned.

I can also save considerably by eliminating gears, belts and jack shafts. Frankly, I am also building this with hand tools, so there is not a good way for me to be confident in precision mounting and aligning parts all over the place to each other I could be wrong, but my calculations indicate that a belt needs to be about 1/2 wide per ft long in order to not stretch under what I consider normal cnc use. I don't mean stretch out in the conventional sense, I mean springiness. That can get wide pretty fast.

A gecko driver cost about $ 150 / each, so if one driver could be used on two motors, that saves another $ 150 each x 2 axis.

In the case of nema 23 motors, the inductance is pretty low - 2 ish mH, so putting them in series might make sense. (assuming it is technically a good idea)

I am also playing with a nema 34 motor that has 1200 mH but needs only 3.5 amps, so in that case, perhaps putting them in parallel might be better, since the G203vs can put out 7 amps with an 80 V DC supply.

So in answer to your question, this build is a hobby, so economics are a factor. That being said, the nema 34 motor is nearly $ 200 / each, so if I make an error and burn this stuff out, then it is a huge false economy.

[wildwestpat]

Hi Harry

You aim to use two screws per axis i.e. one each side of sole sort of gantry or table. Yes this will give you some increase in torque. However you need to consider how a stepper motor works because this can introduce some spring into the system. The stepper motor tries to step as commanded but the exact angular movement is dependent upon the linearity of the motor as well as the torque required to keep that position. The later error increases with static load as well as any cutting forces on that axis. Try this with a motor and you will feel the springiness as you try and force the rotor off station.
Apply too much force and the stepper motor cannot progress to the next position. This is why it is necessary to ensure that there is always sufficient motor torque to prevent stalling but prior to an actual stall the angular position becomes less precise than expected due to the magnetic springiness inherent in the motor action. Adding microsteps makes the situation a bit more complicated as the rotation becomes smoother but the importance of low inductance coils becomes ever more important.

The setting up of axis can be simplified by adopting a structured build. by that I mean testing the axis without the drives attached. use a simple lever on the shaft that will connect to the motor and measure the static torque necessary to move the axis. Keep adjusting the alignment until the required torque is a minimum. Sounds tedious but is quick to do if you just use finger pressure to gauge the force necessary on that lever.

As to the costs these are also important for commercial builds but for slightly different reasons! I am learning this having been forced to retire - now time rich and cash poor! Have you splashed out on any hard war? If you have post the size of the machine you are building - materials to be cut - accuracy you are looking for both in terms of real accuracy and that of repeatability as these are different in a subtle way which impacts greatly on the cost.

As to the series parallel argument this is down to sharing the current in the parallel connection and driving sufficient current in the series case. The inductance is a measure of the resistance to alternating current. Each step is modulated by most electronics by varying the pulse width and this equates to raising the frequency.

You could consider simple analogue system of driving such as Linstepper. Look at their web site Creator of the Linistepper stepper motor driver

I am surprised about your view on the bounce of a toothed belt as these are used in many CNC designs including axis drive but the belts are tensioned.

Regards - Pat