Selecting motors for the MechMate - some history

Originally Posted by Richards

Joe,
<snip>

To determine which power supply will work best, you must decide whether you want to wire the motors Bipolar Series or half-coil. I strongly suggest that you choose half-coil to get much better response at higher speeds, but you will lose about 30% of the low-speed torque if you choose half-coil.

<snip>

Isn't that only true if you use the same voltage ?

when using half coil you have one set of parameters, when using series, you have a whole different set.

If you select a power supply for half coil, your voltage will be low and your current will be high.

if you wire the motor for series and use that power supply, it will draw half the amps, but because it is using the voltage for the half winding, it is not getting the full performance at hight speeds.

Both style motors will get the same torque up to the corner speed of the way they are powered.

(in real simple terms, corner speed is that point where any more voltage is wasted and the motor just gets hot.)

Mariss has a white paper in the files section of the geckodrives group that discusses this and how wire and select amps and voltage.

- - - - - -

to add to Joe's origional questions about power supply.

There are actually two main power supplies. one is your wall power. 220 AC volts or 110 AC volts. that powers all your tools. router and such. it also powers your DC voltage power supply.

Since the comment about using larger spindles, vacuum motors and such was mentioned, that falls to your AC or main power, not your DC power.

Since you can easily add a Solid State Relay (SSR) to power 220VAC devices there is no need to worry about that part of the power supply.

the AC needs to be able to handle the total watts. And if you use a secondary panel for all your high power AC stuff, you would need to make sure you have a circuit breaker in your main panel for those things.

The stepper motor panel does not use that much power, usually less than 1,500 watts or 15 amps at 110 VAC. your router probably uses that much, and should not be on the same breaker.

As note about steppers.....

Interestingly, a large table router will have two axes that need fast motion. the X and Y need to be able to move at high speed, with power.

The Z, and in your case the BC axes are also slow axes so could be used with a series power supply.


I agree with Mike that half coil is the way to go, and add that you do not lose torque by going that way, you do gain power at high speed.

Dave

[Richards]

Joe,
The PS-8N54 should work perfectly if the motors are wired Bipolar Series. It will be about 2X too much voltage if the motors are wired half-coil.

[joepardy]

Mike,

If I wire the motors half-coil as both you and Dave have recommended, then based on your last comment "It (PS-8N54) will be about 2X too much voltage", are you suggesting that maybe I should use a different power supply?

Or will the PS-8N54 work for both applications?

[domino11]

Originally Posted by Gerald D

Venturing into A territory was bad enough - I'll watch this B and C from the sidelines.

Gerald,
Have you done any work in the A axis? I dont remember seeing any evidence of A axis type work at the Camcraft site. If you do any A Axis stuff maybe you could show us a video?

A six axis mechmate would be awesome though

[Gerald D]

My venture into A axis was allowing it to be discussed at this forum and offering suggestions to Art.

A six axis MechMate will not be a MechMate - it will be experimental and not production.

I can see how it is exciting to try to use the basic MachMate design as the base and then do things like make a 24 inch Z axis. That is a lot of rise for the unit.

An articulating head on the end of the Z would require more height between the bottom of the car and the top of the table.

Deffinitely experimenal territory.

Dave

[joepardy]

Agreed - but the most exciting place to live - is on the edge!

[Richards]

Joe,
The PS-5N30 power supply would be the best match when motors are wired half-coil. The 5 in the part number means 500W. The 30 in the part number means 30VDC. That power supply puts out more than 15A, or about 35% more than your motors will draw, so you won't have to worry about either excessive heat or voltage droop caused by too much current being drawn through the power supply. If AnTek offers a PS-6N30 or a PS-8N30 at or near the same price, the larger power supplies would be even better. (Most power supply manufacturers would just substiture a 600VA or an 800VA toroidal transformer in place of the 500VA toroidal transformer, so the cost increase would be minimal.)

[domino11]

Gerald,
I did not intend to infer that you include 6 ,5 or 4 axis details into your plans. I just meant it would be an incredible machine.

[joepardy]

Richard,

It appears that the largest 30VDC unit is the 500W. They do go up to 1000W on the 35VDC units.

Is that an option? Or should I stay with the 30VDC.

If I need to, I will ask if they can custom build an 800W unit.

[Richards]

Joe,
The PS-5N30 should work perfectly. I use a toroidal transformer that outputs 50% of the current of the PS-5N30 with four PK296B2A-SG3.6 motors without any problems. I would ask AnTek if they can/will substitute a 600VA or an 800VA toroidal transformer, but if they can't (because of physical size limitations, etc.), then the PS-5N30 will still work just fine. (I'm very conservative when it comes to power supplies, so I tend to over-specify the power supply whenever I have a choice. Going larger is not always good. It costs more and it is bulkier. If the smaller power supply works, then nothing would be gained from going larger. Even if all your motors were just creeping along at minimal speed, where they draw the most current, the PS-5N30 is rated to handle 100% of the combined amp requirements .)

[Gerald D]

Originally Posted by domino11

Gerald,
I did not intend to infer that you include 6 ,5 or 4 axis details into your plans. I just meant it would be an incredible machine.

. . . checks dictionary for "incredible" . . . . beyond belief or understanding . . . . that's sounds right

[joepardy]

Mike,

I just sent them an email and asked. Awaiting a reply. I would like to go a little over - in the event that I need to change the motors on the BC axis - or if at sometime I want to add an A axis. You just never know the future.

Again, THANKS for the recommendation!

I would like to point out Mikes post of earlier today

quote : Current requirements would be ((4 X 3A) + (2 X 2A)) * 0.66 = about 11A MINIMUM. And 11A X 30VDC = about 350VA. I would select a 500VA power supply to give some margin.
end quote


As you may or may not know, it is not possible to use any more power than the motors can take. Using his calculations 350VA is the limit. getting anything larger is like putting an extra 20 gallon gas tank in the trunk of your car. not connecting it to anything, just dragging it around wherever you go.

using 30 volts and a 500 watt transformer is already 43% over sized.

It's not magic, it's math.

Dave

[Richards]

Dave,
The point of concern for me is the 0.66 multiplier. I believe that that multiplier 'computes' the average current draw of a stepper motor, not the worse case current draw. Because stepper motors used in CNC routers are expected to perform flawlessly over a wide range of speeds, I like to assume that every CNC router will spend a good deal of its time with all motors creeping along at less than 50 RPM (V-Carving, 3D, etc), so I always make sure that the power supply can handle 100% current draw from all the motors at the same time. I also assume that every shop is like mine - without air conditioning - and that every power company lets the power droop 10% when everybody else is running their air conditioners. That means that I have to plan for real world conditions where stepper motors might draw 100% of their rated current and the power company might only give me 90% of the power that I need. Bottom line - oversized power supplies.

On the other hand, in my other life as a process control computer designer, just the opposite was true. Those computers were designed to ramp the steppers up to 1,000 RPM as fast as possible, run a certain number of steps and then ramp down as fast as possible. The Kodak-S series photo printers were housed in comfortable air conditioned rooms - because photo paper is very sensitive to heat. Voltage regulation had to be within 0.5% otherwise the color temperature of the printing lamp would change. Because of the controlled conditions and the specific function required of the stepper motors, I could get away with a much smaller power supply.

This last summer there was a rash of heat related problems on another forum where an upgrade kit was offered allowing the owners of older model machines an easy upgrade path to Gecko controlled stepper motors. The idea to move to Gecko stepper drivers was good, but the implementation of that idea caused some grief. The Gecko G202 stepper drivers were mounted to a PCB board instead of mounting them to a substantial heat sink. The 'reduced current mode' was jumpered OFF, causing the stepper drivers to furnish full current when in standby. The power supply was a switcher without a large filter capacitor. In addition, the power supply was set to deliver about 50% of the voltage that the motors could handle. After the initial kinks were worked out, the upgrade worked pretty good - until summertime temperatures hit 100 F. The usual 'fix' for that problem was to open the controller case and then to use the biggest fan possible to circulate air past those stepper drivers. Although I disagreed with their design philosophy from the beginning, I'm NOT pointing fingers at them. Their upgrade helped a lot of people get a remarkable increase in speed and cut quality at a reasonable cost. In my opinion, the problem was that they read the documents, did the math and then built the controller - without being aware of some of the 'gotchas'. The motors used in those machines are the same as the PK296A1A-SG3.6, which draw 1A when wired Bipolar Series. The Gecko G202 documentation states that heat sinks MUST be used with motors drawing 6A or more. Almost anyone reading that documentation would assume that the G202 would never need a heat sink when used with a 1A motor, so a heat sink was not used. Well, the problem is, the controller has provisions for up to six G202 stepper drivers, all mounted very close together. In addition, when a G202 is jumpered to turn 'reduced current mode' OFF, those stepper drivers run A LOT hotter than normal. Put all of that in a closed case and you end up with a lot of heat - and stepper drivers that shut down until they cool off.

My advice to Joe may be a little too conservative, but I can guarantee that if he uses the recommended power supply, that his stepper drivers and his stepper motors will work, even on the hottest summer days. Bottom line - I'm an old white haired guy who really gets tired of fixing things that spending an extra $10 or $20 could have prevented. Over the last thirty years, I've driven more than 1,800,000 miles (60,000 miles per year - or more), installing process control computers and fixing competitors' machines. Many of those miles were driven to fix a competitor's machine - mostly because the designer went by the book and assumed that the machine would live in a perfect environment and that the power company could be trusted. I'm really not bragging about my machines; I just had an excellent mentor. George, my mentor, before opening his own business, designed electronics for the military. He often said that, as a designer of military electronics, he knew that if and when the devices he designed ever had to be used, that many lives would depend on his ability to design things better than the average engineer. That philosophy worked for him, and, so far, it is working for me.

Things are really no different in our CNC router world when we try to determine how fast we can spin a cutter and how fast we can push it through the wood. We start with the math (chip load calculator) and then add a large dose of seat-of-the-pants experience. Before long, each of us becomes one of the go-to-guys that the newbies seek out when they can't figure out why their cuts look and feel terrible when compared to ours. Usually, that's when we start enjoying our work because we know that there will be few surprises. That's really my goal when I suggest electrical components. Most of the time I have the recommended components right in front of me. And most of the time, I've run those components harder than they would ever be run in the field. That gives me peace of mind - knowing that there will be few surprises.

[Gerald D]

Edit to add: I cross-posted with Mike above - havn't read his whole post yet, but the first sentence tells me we are on the same wave-length.

Dave, it is not that simple. The 0.66 is not a mathematical number, it is a general rule of thumb for 3-axis routers. Nobody has ventured to give the rule of thumb for 5-axis. Even the 4-axis (indexer) blokes who don't current reduce on the one axis are pushing it. (The 0.66 seems to be for the current reducers on all 3 or 4 motors).

[joepardy]

Again, THANKS to Dave, Mike & Gerald for information. I contacted Antek and they informed me that they can produce a PS-6N32 and a PS-8N32. I have not received a price yet - but I am sure that they will be reasonable. With regards to the potential for problems when "heat" is an issue, I would like to ultimately install it in a factory environment - without air conditioning. Summertime temperaturs in this facility routinely reach 100 degrees.

Originally Posted by Richards

Dave,
The point of concern for me is the 0.66 multiplier. I believe that that multiplier 'computes' the average current draw of a stepper motor, not the worse case current draw. Because stepper motors used in CNC routers are expected to perform flawlessly over a wide range of speeds, I like to assume that every CNC router will spend a good deal of its time with all motors creeping along at less than 50 RPM (V-Carving, 3D, etc), so I always make sure that the power supply can handle 100% current draw from all the motors at the same time. I also assume that every shop is like mine - without air conditioning - and that every power company lets the power droop 10% when everybody else is running their air conditioners. That means that I have to plan for real world conditions where stepper motors might draw 100% of their rated current and the power company might only give me 90% of the power that I need. Bottom line - oversized power supplies.

<snip>

I can see where you are coming from. the Gecko is a PWM chopper style driver. most of us think linear and not digital. Mariss has been designing stepper drivers for over 30 years and has a failure rate of less than 1% He has been making the Gecko no for over 6 years, so when he says 66% is enough, it means not just worst case, but WORST CASE. Mariss is very conservative in his numbers. his failure rate and success show he is pretty good about such things.

Joe listed his motors as :
(4) PK296A2A-SG (3 amps x 4 motors = 12 amps)
(2) PK264A2A-SG (2 amps x 2 motors = 4 amps)

for the 16 amps listed, and using 30 volts, we can calculate 16x30 = 480 watts. That is full amps as if the motors could all run at full amps all the time.

Although it is not possible for that to happen due to the nature of PWM and choppers and motor inductance, one would be fully covered with a 500 watt power supply.

Taking into account that Mariss is quite conservative in all his calculations, you can use much less that the 66% is recommends. His 66% is already well over sized for his drivers.

Just as a note, Mariss has said he uses a rule of thumb of taking the full design and cutting the voltage back to 80% and the amperage back to 70%. If you look at the components, the Geckos use components rated for 100 volts and 10 amps. He lists 80 volts and 70 amps following the 80/70 rule of thumb.

So, the 66% is not some out of the air number, nor is it the result of an over optimistic engineers adding up all the specs. it is a real world, worst case number. If you have the opportunity to actually test a power supply on a bunch of steppers to find out the real power number, you will not believe how low it really is.

I think the bottom line is to address Joe's question. I think Mike nailed it with the recomendation to use single coil wiring, and I think that when Mariss, the manufacturer, offers a calculation, you are extreemly safe with it. We use all of his other calculations for his product, why change what he recommends ?

Dave

ps: As for your comments about heat in a sealed enclosure, I cannot agree more, but at the current levels of the Gecko's the heat generated is in low single digit watts per driver. Makes me concerned, but since others have not had a problem, I remain (mostly) silent on that topic.

[Richards]

Dave,
I see that you've asked Mariss for a definitive answer on the Geckdrive forum. I was going to do the same thing. Somewhere I remember reading that the 0.66 multiplier was for the older round motors and that the newer square motors, being a better design, could pull more current. That post, if it even exists, might have been made by someone other than Mariss. At any rate, let's hope that Mariss helps us out on this.

I did a search on Geckodrives for the exact reason. I really like to get Mariss to answer such questions. he does the whole 9 yards and actual testing and such. That way, there is no question.

But, this industry is so dynamic that if someone were to have problems, we would all hear about it pretty quickly. (look at the rabbit problems)

My first unit was a HUGE transformer, something like 35 amps. the next was 20 amps, then the last one was 7 amps, and on 4 steppers of 3.5 amps, I have never encountered a problem.

As for the geckodrives yahoo site...

the last post was in Oct 2005, from some guy named Dave

Mariss's resonse, post # 6893

Re: Power supply selection ??

Square motors deliver more power than a round motor of the same
dimensions. Consequently they draw more from the supply. My anectodal
experience is the supply current should equal the motor's unipolar or
parallel nameplate rating. I will be running tests when time allows
to nail down the multiplier and back it up with actual data.

Mariss

[larry1larry]

Hi Guys,
This is my first post after reading everything on the forum about stepper motors.
My brain is fried for the moment,so I will make some unqualified statements.
If confused,should I just use a 4 wire motor?
Different wireing schemes result in half voltages or current or whatever.Perhaps the variable is how big a power supply you build.
I read many times in posts of the power supply pushing current to the motors. I thought power supplies provided current to the motors and the motors drew current from the supply,with the gecko's limiting the current.
The geared oriental motors have less torque than the un geared motor?
For the moment my brain is wired bi-polar series,100volts/20amps/80c.
After cooldown things may be clear.
BTW great forum,Thanks
Larry

[Richards]

Larry,
Keep reading, just keep reading.

[smreish]

Mike....funny how you mention reading...just keep reading. I was soldering the harness connections on my motor's last night while the kids were busy watching Nemo....just keep swimming...
The kitchen table project has so many parallels

- Almost ready to put power to the box for a test run. Waiting for paint to dry so I can reassemble the cabinet.

Sean

[larry1larry]

Information overload.
Sorry my brain is still overloaded but I will try to make some assumptions.
From what I have read here,.66 is OK as all 4 steppers are never fully on at one time.the others act as a current dump.
I like the low current and sealed cabinet.For higher current you could cut a hole in the box and mount the Gecko's to an external heat sink.
Larry

[Richards]

Larry,

Although 66% is generally accepted as enough current, I always go 100%. Power supplies are cheap compared to the material that might be wasted if the motors can't do their job.

As far as heat goes, the G203v doesn't generate much heat. A slightly oversized power supply doesn't generate much heat. I run G203v all the time on my test bench. They've always stayed at, or very near, room temperature. The various toroidal transformers that I use also stay at or very near room temperature.

As to whether a device 'pulls' current from the power supply or whether the power supply 'pushes' current through a device is kind of a loaded question. Sometimes it's easier if we think of a power supply as a device that can produce pressure (voltage) and volume (current). If you've ever played with water, you know that the higher the pressure, the more volume - in a fixed amount of time. So, a power supply can 'push' electricity out to various devices, as long as the various devices can't allow passage of more current than the power supply is able to produce. That brings us to the device. If we use the water analogy again and think of each device as a certain size of hose, it is easy to visualize what is actually happening. Lets say that device A has a 1/4-inch hose and device B has a 3/4-inch hose. Using that example, it is easy to understand that device B will allow more current to pass at a given pressure than device A allows to pass. The same thing is happening with power supplies and electrical components. The power supply is able to push out electricity at a certain pressure as long as the various devices connected to that power supply allow less current to pass than the power supply is able to supply. So, to bring this long analogy to an end, if a power supply can supply 10A of current and all of the devices connected to that power supply can only allow 8A of current to pass, then the power supply can be considered adequate to handle the job. On the other hand, if the power supply could only handle 6A of current and all of the devices connected to that power supply can allow 8A of current to pass, then the power supply would be considered inadequate to handle the job (just like running too many sprinklers off of one water line causes pressure to fall to the point that the sprinklers can't do their job).

[Gerald D]

Mike, an early New Year's resolution of mine is to do some measurements of these lovely vague concepts we spend hours talking about. . . . .

1. Put a logging amp-meter on the output of the power supply driving motors which are doing various types of cutting (3D, cabinet doors, etc.). This will tell us if the 0.66 factor is okay.

2. Take a variable transformer powering a recommended motor via a g202 and g203V and log motor temperature versus voltage. This will tell which voltage is the limit for the motor.

But, if Santa has been kind while I am away on holiday, there will be lovely graphs and test reports waiting in my stocking when I get back.

[Richards]

I've received several emails over the last few weeks from CNC builders (not those who've built a MechMate - as far as I know) who are getting really slow cut speeds with the PK296A1A-SGxx motors. The main complaint is that the motors start to lose steps at around 2 to 2-1/2 inches per second.

We've got to realize that the PK296 motor is basically a 300 oz*in motor. It was never designed to push a gantry that weighs hundreds of pounds. With proper ramping, it can do much more than would be expected, but it is a small motor. However, to be fair, it is the only practical motor that is affordable. The Oriental Motor AS-alpha series motors cost about five times as much. The PK299 motors with a belt-drive added to them are big and bulky and expensive.

Sometimes we concentrate on the wrong numbers when we select a particular motor. For instance, the PK296A2A-SG7.2 motor, with a 30-tooth spur gear, could move an axis about 14.7 inches per second if we only looked at a pulse rate of 45,000 pulses per second and multiplied the per-step resolution of .000327 by 45,000. What we sometimes ignore is that full torque for that motor is a little less than 6 inches per second. If we're using the PK296A1A-SG7.2 motor, then full torque is about 4 inches per second.

What I'm trying to say is that the emails I'm receiving make me think that too many people are expecting too much from those little motors. I normally run my Shopbot PRT-Alpha, which has 600 oz*in motors with a 3:1 belt-drive, which gives me 1,800 oz*in, at 6-ips when I cut MDF and plywood at a depth of 1/4-inch to 3/8-inch per pass. That's with a 3-hp spindle. If I put those same motors on a MechMate, I'm sure that I could run faster per pass without deflection, but my experience makes me think that I would never achieve the speeds that the math would make me assume were possible.

Expect less than the raw numbers would indicated. With proper ramping, proper cutting methods, and sharp cutters, you will learn to get the most from your machine - including reasonable speeds.

[Charlie]

600oz*in X 3.1beltdrive = 1800oz*in.
Then
300oz*in X 7.2 = 2160oz*in ?
I think you said that the max load on the gearbox is 400oz*in, is that true? I couldnt find it anywhere yet. So is that why your not calculating the same?
I cant yet figure out why those people are looseing steps. Isnt the motor on easy street behind reduction gears? Sure I understand looseing speed with reduction but looseing torque too?? Unless your saying it is a bad rampping setting. Meaning not the motors fault. Im confused. Could you help a guy out?

[Richards]

Charlie,
The gearbox used with the Oriental Motor's PK296AxA-SGxx motor is rated at very low torque - compared to the mechanical multiplication that you would expect the drive and gearbox to have. I don't know of anyone who has actually broken a gearbox by trying to push too hard; but, I have received many e-mails from users who have had the motor lose steps. From that, I assume that something inside the gearbox might be binding to the point that the gearbox starts to seize - and the motor loses steps.

I had a long telephone conversation with tech support at Oriental Motor, about 1-1/2 years ago, when we discussed this very issue. The end result of that telephone call was that the maximum torqure for the PK296AxA-SG7.2 motor is about 45 lb*in, or about 720 oz*in. The maximum torque for the PK296AxA-SG3.6 motor is about 20 lb*in, or about 320 oz*in.

Those torque ratings, combined with the maximum allowable speed from those gearboxes place limits on how hard and how fast those gearboxes can be driven.

(Oriental Motor publishes torque charts for all of its products, including the PK296AxA-SGxx motors. Just go to the motor's specifications and then click on "More Specificatons" to get the torque charts.)

[revved_up]

Ok, I am about to commit to the build and start buying hardware etc. I have read this thread 3 or more times and agree with many of the posters before me that this can be quite confusing and thankyou Gerald for updating your postings and keeping everything current. Is it a possibility that someone that has tested different motor control combinations could possibly post a spreadsheet to make it easier to compare and list pros and cons. Maybe if possible leave it open for people to add their specs of what they have working on their machines in the future so everything could be in one consise location. I know this may be asking for alot for somebody to take the time for this as time is my most precious asset but it would help all of us with less experience to make good decisions.