| Author |
Message |
    
Mike Richards
Registered
Username: Richards
Post Number: 91
Registered: 05-2006
| | Posted on Thursday, March 29, 2007 - 04:09 am: | 
|
I built up an extra power supply from a 12A, 24V standard TRIAD transformer (Part No. F-226U). It is being used to drive four 3-amp PK296B2A-SG3.6 Oriental Motor stepper motors via four Gecko G203 stepper drivers. It also has a 11,000mF capacitor.
This transformer runs hot - 160-degrees F. or 60-degrees C. The figures show that it is a 288VA transformer (24V X 12A). Using Gecko's formula for four G203 running 3A motors, 4 X 3A X 0.667 = 8A, I should be able to use an 8A (200VA) power supply. I figured that using something 50% larger would help keep things cool.
Until now, I've only used a Toroidal transformer with my Gecko test bench. It is an AVEL Y236801 rated at 500VA. It has never even been warm to the touch. Perhaps the toroidal is just a lot more efficient, but I was really surprized to see the amount of heat this Triad generates. The G203s, the G100 and the stepper motors aren't even warm, but the transformer is too hot to touch.
It looks like it's time to order another A236801. |
Gerald_D
Registered
Username: Gerald_d
Post Number: 1187
Registered: 11-2005
| | Posted on Thursday, March 29, 2007 - 08:10 am: |
|
Mike, I have started to question the use of that 0.667 factor that is so automatically applied....
If you have 4 drives set at 3 Amp each, and they are all being pulsed at very low speed, they will draw 12 Amps together. Your DC voltage is 34V. Those 4 drives want to consume 34x12= 408VA. If you could put an ampmeter on the transformer, you should be reading 400÷24= 17A (ac) which means your 12A transformer is way overloaded and getting hot. The 500VA toroid would smile at the 408VA load. (this rough calc assumes transformer/drive efficiencies of 100% - the real situation is a bit worse)
That 0.667 factor cannot apply where all motors run slowly simultaneously. The factor will only work for applications where some motors are standing still most of the time. Or where the motors turn very fast under low load. |
Mike Richards
Registered
Username: Richards
Post Number: 92
Registered: 05-2006
| | Posted on Thursday, March 29, 2007 - 09:14 am: |
|
Gerald, I think you're right. I duplicated the test using the larger toroid transformer and had no heat at all. Later today (it's currently 1:15 a.m. in this part of the world), as a cross-check, I'll reconnect the 12A transformer but only run two drives, which should require about 200VA.
As a general rule, I could probably assume that, on average, 2-1/2 axes would be active on my machine - depending on whether the 2 x-axis motors were active. I do very little work that requires active use of the Z-axis. It usually plunges at the start of a cut and then 24 or 48 or 96 inches later, the z-axis pulls up again. But having a larger than necessary toroid transformer is cheap enough to give peace of mind. |
Gerald_D
Registered
Username: Gerald_d
Post Number: 1188
Registered: 11-2005
| | Posted on Thursday, March 29, 2007 - 09:52 am: |
|
Mike, the definitive answer will come from an ampmeter (or is it ammeter?)...AC before and after the transformer, DC after the rectifier. I know it is a bind to wire in a DC "current"meter, and that smoke may be released, but a clamp/tong meter is easy for the AC. Do you have a "tongtester" (our popular name for it) to check that 17 Amp that I think you have? |
Mike Richards
Registered
Username: Richards
Post Number: 93
Registered: 05-2006
| | Posted on Thursday, March 29, 2007 - 03:04 pm: |
|
Gerald, I dug out my Greenlee CM-600 Clamp-on Meter, but got some strange readings. They were much lower than expected. Until I can find another meter to cross check the readings, I'll assume that they are in error.
However, here are the readings:
1.1, 1.6 with 1 motor on, 3 motors on standby
1.4, 2.2 with 2 motors on, 2 motors on standby
1.3, 2.7 with 3 motors on, 1 motor on standby
1.5, 3.6 with 4 motors on
The first number is the current reading of AC Hot (Line) going to the transformer. The second number is one of the AC leads going to the bridge rectifier.
The meter only has two ranges to read current through the clamp: 0-200A and 200A-600A. I'm guessing that the meter can't read these small currents correctly, so I'll see if I can borrow a meter that has a range somewhere around 0-20A.
One thing is certain; this particular transformer is not efficient. The voltage going into the transformer is 120.1VAC and the voltage coming out is 25.6VAC or a 4.69:1 ratio, which shows that the transformer has the correct number of windings. (Dividing 115V by 4.69 would give 24.5V.) If the transformer were 100% efficient, then 4.69A being pulled from the secondary side would cause 1A of current to flow through the primary side. As you can see from the numbers, we're really getting about 2:1 (if the readings from the meter are linear - which I tend to doubt).
As far as heat goes, with one motor on and three motors on standby, I read 122-degrees F. after 1/2 hour. With two motors on, I read 135-degress F. after another 15 minutes. With three motors on, I got 157-degrees F. (69 C.). The amount of heat produced also verifies that the transformer is not efficient.
If I've done the math right with the GeckoMotion program, the motors were running at 4000 steps per second. With 20-tooth spur gears and 3.6:1 gearbox, that would equate to 1.75-ips, meaning that the motor is still in the high torque range where it should draw a lot of current. |
Gerald_D
Registered
Username: Gerald_d
Post Number: 1190
Registered: 11-2005
| | Posted on Thursday, March 29, 2007 - 03:23 pm: |
|
Some very strange readings indeed...are you sure the tongs are closing properly? Does the output voltage stay consistent with load?
Anyway, there is no doubt that toroids are the better transformer, but I didn't think the non-toroids are so bad! (What does one call a non-toroid?) |
Mike Richards
Registered
Username: Richards
Post Number: 94
Registered: 05-2006
| | Posted on Thursday, March 29, 2007 - 06:27 pm: |
|
I just finished making the same AC Current tests with a new Fluke T5-600 meter. Its accuracy is supposed to be 0.1 amp with an error +/- 3%. It's not ideal for measuring low current, but the only meter available at the moment that I can use to cross-check the results.
(The first number is the 120VAC Hot lead, the second number is one of the transformer secondary leads.)
This time the Fluke T5-600 read:
0.8, 0.7 - all motors at auto current reduction
0.8, 1.3 - 1 motor
1.0, 1.8 - 2 motors
1.1, 2.7 - 3 motors
1.2, 3.3 - 4 motors
The Greenlee CM-600 read:
0.9, 0.8 - all motors at auto current reduction
1.0, 1.6 - 1 motor
1.3, 2.1 - 2 motors
1.4, 2.9 - 3 motors
1.5, 3.5 - 4 motors
The original readings with the Greenlee CM-600:
1.1, 1.6 - 1 motor
1.4, 2.2 - 2 motors
1.3, 2.7 - 3 motors
1.5, 3.6 - 4 motors
Even though the numbers are not what I expected, they are consistant. The Greenlee reads a few tenths higher, but its clamps have a distance from inside to inside of 1.3-inches at the point where the measurement is made. The Fluke has a slot rather than a clamp and the sides of the slot are 0.5-inches apart, with a reading area clearly marked on the slots. The accuracy of the Greenlee is 1.5% +0.5A, which probably accounts for the slightly higher readings.
With both meters agreeing, and with LOWER readings (about 0.1A) with the motors spinning 4X faster, I have to conclude that these steppers pull a lot less current than I expected. I believe the test is valid because the steppers reach a temperature of about 125 to 135-degrees F. after prolonged use and the G203 reaches a temperature of 85 to 90 degrees F. when bolted to the top of a 3x3x1.5 inch aluminum heat sink.
To complete the test, I hooked up my Fluke 77 meter in line with the + voltage from the capacitor to the DIN rail star + voltage connection to read the DC current.
0.54A - all motors on standby
1.53A - 1 motor
2.44A - 2 motors
3.50A - 3 motors
4.49A - 4 motors
With all motors off, the power supply voltage is at 33.2VDC with 0.081VAC ripple. With four motors running the power supply voltage is at 31.2VDC with 0.556VAC ripple. That shows that the capacitor is large enough for the job. (I'm always happy to have less than 10% ripple when I use an unregulated power supply.)
The pulse generator is a G100. The stepper drivers are G203s. The steppers are Oriental Motors PK296B2A-SG3.6 steppers wired as unipoloar (half-coil) with 33K current set resistors which should allow the steppers to run at just under 3A.
The GeckoMotion instructions for the final series of tests are:
GPO 1000 0000 0000 0000
GPO 0000 0000 0000 1000
Rx05
Mx3
Ax1000
Vx4000
Ry05
My3
Ay1000
Vy4000
Rz05
Mz3
Az1000
Vz4000
Ra05
Ma3
Aa1000
Va4000
x+250000
y+250000
z+250000
a+250000
x-250000
y-250000
z-250000
a-250000
You'll notice that each motor is started independantly of the other motors so that no motor will be affected by the velocity of the other motors. A final P* instruction showed that all motors had returned to position 0 - and that's not bad after traveling 500,000 steps. |
Gerald_D
Registered
Username: Gerald_d
Post Number: 1192
Registered: 11-2005
| | Posted on Thursday, March 29, 2007 - 07:00 pm: |
|
Mike, great work! It will take a while to digest ALL those results.
The very low current readings overall is the big surprise. That 12A rated tranformer was never asked to deliver more than 4A, and still it was getting hot. Bottom line on that issue is to stick with toroids.
But, the other thing uncovered here is that the 3 Amp drives only consume about 1.1 Amp each ??? This doesn't make any sense. I am wondering what the DC current readings would be if the voltage was doubled - theoretically it shouldn't change, but will it?
Scratching my head after a postprandial glass of wine..... |
Gerald_D
Registered
Username: Gerald_d
Post Number: 1193
Registered: 11-2005
| | Posted on Thursday, March 29, 2007 - 07:08 pm: |
|
Might be a good time to visit this post again:
http://finance.groups.yahoo.com/group/geckodrive/message/567
"7) Load the test motor with the paper towel folded over several times
until it is only 1/2" wide. It makes a great brake-shoe. For bigger
motors you will need the channel-lock pliers to apply sufficient
pressure. You may also want to wet it a little to keep it from
smoking or catching on fire when you do.
8) Slowly load the motor while watching the DC ammeter as you do.
Note the reading the instant the motor stalls. Note the power supply
voltage as well at stall if it is unregulated. This may need to be
repeated several times until you get the hang of it."
Looks like a pliers is needed to get the current draw up to near the limit...... |
Mike Richards
Registered
Username: Richards
Post Number: 95
Registered: 05-2006
| | Posted on Friday, March 30, 2007 - 01:54 am: |
|
"Postprandial!" Today is the first time I encountered that word. (It's a good word with a good definition.)
As to the loading of the motors, my motors are very lightly loaded, since they're just sitting on the test bench with flywheels attached to their shafts. It's clear that my test environment is less than ideal when compared to the real world where a stepper is used to do something. Maybe it's time I actually built a machine instead of testing things. I'll bet that if I looked hard enough I might find some plans that would get me started. I think a know a guy in Africa who said that he was going to offer something like that to anyone interested in building a machine. . . |
Gerald_D
Registered
Username: Gerald_d
Post Number: 1194
Registered: 11-2005
| | Posted on Friday, March 30, 2007 - 07:20 am: |
|
From Africa you say? Surely you meant somewhere else more, ahem, "advanced"? I mean, those guys probably still use feet and inches...??
Mike, have we got it wrong....does a stepper motor not use max current when it is holding a static position with no load? The G203V are said to drop current to 50% while stationary - yet you see those 4 motors drawing only 0.54 Amp together. Can they hold a position while only drawing such a small current?
Another test if you will indulge...
All motors static, or just one motor coupled up and static,
No G100 connected, ie. no pulses to driver,
DC ammeter in supply to drive,
Check what the current reading does if the motor shaft is loaded, ie. put a little torque on the shaft to move it out of its detent position. See if the current climbs.....
This weekend I'll be taking my tongtester home and experimenting a bit....if you havn't beaten me to it!  |
Mike Richards
Registered
Username: Richards
Post Number: 96
Registered: 05-2006
| | Posted on Friday, March 30, 2007 - 02:44 pm: |
|
Gerald, we've opened a can of worms. I wired up the X-axis motor with the leads of my Fluke 77 in series with the ground wire going to pin 1 on the G203 and measured 137.4mA to 154.0mA, about 1/10th of what I expected to see because the module has a 33K current set resistor to give just less than 3A. (Because the readings were so low, I checked them with the 300mA range on the meter instead of the 10A range.) With an advertised reduced current mode of 50%, I expected to see 1.5A. So, I removed the current set resistor (which is done for 7A motors). The current changed to .46A, when it should have climbed to 3.5A.
With the resistor back in place and a 5-inch pulley attached to the shaft of the motor so that I could 'feel' the shaft turn from its standby position to the next 'detent', I slowly turned the shaft as I watched the meter. At the very most, I saw a 15-mA increase in current just before the shaft slipped to the next position. I expected to see the current increase at that point, but I expected to see a huge increase instead of the 10% increase.
Finally, with the meter reconfigured at the 10A range, I started spinning the motor with the G100 attached. At slower speeds, the current rose to 530mA. At faster speeds (Rx06, Ax100, Vx10000), the current rose to 960mA. With a gloved hand and fast RPMs, I increased friction to the pulley and was able to see a little more than 2.5A current before the motor stalled. At lower RPMs (Rx06, Ax100, Vx100), the motor went from a maximum of 1A to stall with very little increase in friction. I repeated the tests several times with almost no deviation.
At this point I started wondering whether the G203 was faulty, so I re-ran the tests with a G202 stepper driver. This time, at rest I read 260mA with the motor in reduced current mode. With the motor running, I read 530mA, which was identical to the G203 when the motor was running at the same speed.
In about five hours when it is 9:30 in California, I'll try calling Gecko support. If Mariss is there, maybe he can give me some hints. |
Gerald_D
Registered
Username: Gerald_d
Post Number: 1195
Registered: 11-2005
| | Posted on Friday, March 30, 2007 - 04:39 pm: |
|
When I got home from work this afternoon, I confirmed what you were seeing, but then I moved the ammeter onto one of the motor coils. The currents there measure correctly.
It started to make a little sense then.... The Gecko is a constant current driver - loading/unloading the motor cannot change the current output. Then I tried to measure the voltage of the Gecko output while loading/unloading a motor, but got around 200V readings and realised my equipment was not up to the digital waveforms involved.
You found max current draw at a certain combination of speed and torque, and from my mechanical experience, that is how it should be...
Power = Speed X Torque. At zero speed, or zero torque, the power equals zero, except for heat losses. Since the power supply output is a fixed voltage, only its current output can vary to match the Power needed by the motor and drive.
While the motor is still, it needs very low power (only for heat losses) and the constant current device feeds it the designated Amps, but at some very low "voltage" I presume. On the input side of the device (drive) that same power is seen as a low current because the voltage is held high by the capacitor.
It must all be to do with the balance of Power (VA) before and after the Gecko. Before the Gecko there is a constant voltage, after the Gecko there is a constant current. ????
Anyway, I think we have proved the theory that a stepper draws its max current at standstill to be false? |
Mike Richards
Registered
Username: Richards
Post Number: 97
Registered: 05-2006
| | Posted on Friday, March 30, 2007 - 05:09 pm: |
|
Gerald, you've figured it out. I got a reading of 1A at standby when I connected the meter in series with one of the Phase B wires. When I disconnect the current set resistor, the current jumped to 2.6A. Both figures are still about 30% less than I expected (based on the 50% standby current), but much better than before. Edited: Since I'm only reading one of the two coils, it looks like everything is working properly.
The odd thing about this entire experiment is that when I stalled the motor by squeezing on the 5" pulley with a leather glove on my hand, the friction 'seemed' about right. I've run a similar test using a spring-scale to weigh fish and a nylon cord. Those tests showed torque figures right where they should have been.
So, end result = Mariss is still brilliant, Gerald is much better at trouble-shooting that I, and I learned something useful. |
Gerald_D
Registered
Username: Gerald_d
Post Number: 1196
Registered: 11-2005
| | Posted on Friday, March 30, 2007 - 05:53 pm: |
|
Mike, it is not a contest! You are a far better diplomat than me. (You can say things that I can't). I am the one that led you up the garden path of sticking an ammeter in there so I have foot part of the bill on this one.
I used a G202 and got the full 7.27 Amps in one coil with no resistor, and 4.70Amp with a 93k resistor. When jumpered for standby reduction, the respective currents were 2.35 and 1.47 Amp (The G202 drops to 33% for standby). So there is still something odd with your G203 readings? |
Mike Richards
Registered
Username: Richards
Post Number: 98
Registered: 05-2006
| | Posted on Friday, March 30, 2007 - 07:20 pm: |
|
Gerald, here's a response from Mariss that he posted on the Yahoo group's gecko forum:
"140mA is just about what I would expect during standby. You didn't
mention if the stopped motor had about half its holding torque when
you turned it by hand.
During standby:
At the beginning of each 20kHz switching period the G203V switches the
supply voltage across the winding. The current thru the winding
increases until it reaches the set value. The winding is then shorted
for the remainder of the switching period.
Current thru the winding rises rapidly when the supply voltage is
across the winding and decays slowly when the winding is shorted. The
power supply delivers the rising current, the supply current is zero
while the winding is shorted.
For you to read 146mA means the supply current is 1.5A at a 10% duty
cycle (5uS on, 45uS shorted).
Mariss
That explains how it all works and seems very reasonable to me. |
|
Power Supplies