![]() |
#31
|
|||
|
|||
Something I wrote in another thread:
Quote:
|
#32
|
|||
|
|||
Quote:
![]() |
#33
|
|||
|
|||
This post first made at Rough Cut on Diagonal in the Troubleshooting area:
A random thought . . . . what would be the effect of having too little capacitance at the power supply to the geckos? |
#34
|
|||
|
|||
A filter capacitors job in a power supply is to hold up the output voltage during the ac valleys after the rectification process. It will supply power to the load in this time frame. If the capacitance is too low, then the output voltage will droop during this time instead of being held steady. This droop is usually called ripple, as that is what it will look like when viewed on an oscilloscope. I think this ripple would have to be pretty bad to have any effect in the cut quality. (Please note this is from theoretical background and not actual experience on my part) Maybe Mike has an opinion on this one?
|
#35
|
|||
|
|||
Hmm. I'm thinking there could be two effects, one ripple, but the other is a function of demand.
Ripple would be at 2 x 60Hz (or 50Hz) so 120Hz (100Hz). For that to be directly visible at the about 4 ripples per inch shown at the top of this thread would require that you be covering the inch in 4/120th of a second, so 30 inches per sec = 1800 IPM. Not likely. If you were moving that fast, you'd probably see ripples, but based on way outrunning the cutter's material removal capability. And Nils has already told us his speeds, which aren't anywhere near that fast. Moving on to demand, then. Suppose the capacitance is low enough that when two Geckos try to draw full power simultaneously, the capacitors are completely drained and you have to rely solely on the power from the rectifier. That power would have a high degree of ripple, and could easily interact with the Geckos in a manner that caused underpowering of a motor. So the demand is toggling the power supply back and forth between well regulated, and full of ripples, cyclicly. When the power supply is ripply, the motors underperform. Seems plausible. |
#36
|
|||
|
|||
Brad,
I dont believe that the caps are getting fully drained. They will have a new charge every 8.3 uS. Also the geckos have a local filter capacitor as well for local energy storage (though smaller than the main capacitor that Nils has). They only way this would happen is if his filter capacitor is dead I believe. One easy test would be to use a multimeter (in DC voltage mode) and see if there is any voltage drop from the non loaded state to the fast cutting (loaded) state. If there is not a real difference there, then it should not be a problem. If there is significant drop or a fluctuating type event happening, then this could be part of the problem. Nils, What value of capacitor are you using in your powersupply? Last edited by domino11; Wed 07 October 2009 at 09:01.. |
#38
|
|||
|
|||
Sean is correct, the Antek Power Supply that I am using is the PS-6N56R5R12 unit.
|
#39
|
|||
|
|||
I am a little bit nervous of Antek quality and wouldn't be surprised if one of those caps were faulty.
But, I wouldn't jump to that conclusion and start stripping the power supply until we have really established how low capacitance is likely to affect our gecko/motor performance. Brad, I hear you on the ripple frequency being high. But what is the chance of the ripple frequency, combined with the step frequency, producing "beats" at a low freqency? |
#40
|
|||
|
|||
Gerald,
Question? - i would assume the ripple question would go to marius on the geckoforum? - without a scope, how would a someone even begin to test this theory? Thank you |
#41
|
|||
|
|||
I will put the question to Geckodrive.
For a practical test of the theory, the quick thought is to purposely reduce the capacitance in a power supply, but then I wonder if it could damage a gecko. . . |
#42
|
|||
|
|||
Sean, Nils
See my quick and dirty test on post #36 to try. If the voltage does not fluctuate during cutting and there is no large drop in dc voltage from loaded to unloaded states, the capacitor is probably ok. Sean is right though, a scope would be the best thing to use. Also since Nils power supply has two caps in it, both would have to have gone open to have no capacitor in the circuit. Gerald, Don't forget that eack gecko has some capacitance built in and therefore the circuit would probably still work. How well is another issue. |
#43
|
|||
|
|||
When we look at the components in a power supply, the transformer changes the amplitude of the voltage, i.e. 220VAC in becomes 24VAC out, the bridge rectifier blocks the negative portion of the AC sine wave and adds it to the positive portion, i.e., 50 cycles per second becomes 100 cycles per second, and the filter capacitor stores the AC cycles so that the power appears to be DC.
Without a capacitor, the output would be just a series of sine waves going from about 0V to (SQRT(2) X output voltage). In other words, the output would basically be a positive AC signal. With a capacitor, the voltage is smoothed and filtered. The filtering refers to the fact that a properly sized capacitor removes all of the AC. That's why the standard formula is (80,000 X Amps) / Volts = Capacitance in microfarads. With a properly sized capacitor, there will be very little ripple. With an undersized capacitor, the current being drawn through the capacitor will empty the capacitor faster than the transformer and bridge can fill it, and the stepper motor will get less voltage than desired. Stepper motors rarely all pull the maximum current at the same time, so using a capacitor that is at least 75% of the computed capacitance works well for me. Be sure to select capacitors that have a VOLTAGE rating higher than the voltage of the transformer. The AC voltage is measure using RMS but the capacitor has to deal with PEAK voltage, which is SQRT(2) X RMS voltage or 1.414 X higher than the RMS voltage. When I use a 25VAC transformer to make a 35VDC power supply (remember 25VAC X 1.414 = about 35VDC), I use 50V capacitors or even 63V capacitors if they are readily available. Now the real question is whether low capacitance can cause rough cuts. The answer is a definite "Maybe". When two axes are in motion at the same time and the stepper drivers are micro-stepping, one axis could push harder than the other axis which could cause a problem; however, I get that exact same pattern on my Shopbot PRT-Alpha which has Alpha motors and Alpha stepper drivers. Those drivers are supposed to be perfectly matched to the motors, meaning that the motors would never be starved for current or voltage. My guess is that the rough cut is primarily due to factors other than low capacitance. |
#44
|
|||
|
|||
...and Mike, I certainly would agree with your summary comment :
"My guess is that the rough cut is primarily due to factors other than low capacitance" |
#45
|
|||
|
|||
The reply from Geckodrive:
" The capacitor is there to reduce ripple voltage and keep it nice and consistent. If the capacitor is inadequate you will run the risk of multiple things happening, including but not limited to: 1.) Unpredictable motor and drive heating due to fluctuating voltages that are out of the range for the motor (based on inductance). 2.) Stalling of the motor because it is being asked to do too much with too little voltage on the low end of the ripple. You will get some pretty miserable performance but it will be immediately noticeable. It is not something that will sneak up on you and occasionally throw in a missed cut but will simply not work depending on the voltage required from the motor. " |
#46
|
|||
|
|||
A little Capacitor sizing question
I am going to run 4 PK299-F4.5A (parallel) motors with a 50V Antek power supply PS-10N50R12 - 50VDC 1000W Power Supply. It comes with 2 10000uF capacitor where the calc from Mike's spreadsheet [80000*(4*A)/V] is 39845uF. Which one should I observe? |
#47
|
|||
|
|||
Because you are running the motors with a parallel wiring connection, I'm assuming that you're going to drive the motors as hard as you can. You can START with the stock configuration, then, with an oscilloscope, watch the voltage as you run the motors. When ripple gets to be much above 5%, add another capacitor in parallel. Keep adding capacitors until the ripple stays within about 5%.
I've run four high-current motors on the test-bench with 11,000 uF with good results. Speed and load determine how much current the motors are actually going to draw. The higher the speed and the greater the load, the greater the current draw. Because a CNC router is expected to handle a wide variety of speeds and loads, it's really hard to tell you exactly how much capacitance you'll need. So, start with what you have and then add more as necessary. |
#48
|
|||
|
|||
Thanks Mike, I am a fish out of water with the electrical stuff, That is why I took circuit theory in my final semester as a senior ME student, sat next to all of the freshman EE's. I just don't want to make a stupid mistake.
|
#49
|
|||
|
|||
Resistor size,
Hi Mike,
Sorry yoou seen to have made a error in your calculations 3. You could add an LED; however, an LED uses Current to work, about 10mA. To get 10mA at 70V would require a 700 ohm resistor at 7W! You would want to use at least a 25W resistor to keep the heat down and a 100W resistor would be much better. in fact it would require a 7000 ohm resistor(R=V/I , R= 70 / 0.010A) at 0.7 W (W=V x I , W= 70 x 0.010) which becomes realistic, a 1 watt resistor would be fine. But i agree there is no need for a LED, bleed resistors should be looked at with sceptisum they are in use all the time the power supply is operating and will eventually fail DON'T trust them,(they also reduce the output of the power supply and reduce the effectiveness of the capacitor) use a resistor to bleed the capacitor and multi meter to check then short them out (most capacitor failures can be traced to excessive current flow). Hope this clears things up, Andrew |
#50
|
|||
|
|||
Andrew,
Thanks for catching that error. I had the decimal point in the wrong place. A 7000 ohm 1W resistor would work and it would be within specs, but it would still be too hot to touch. A 5W resistor would be the minimum size that I would use. (I don't like heat in electronic circuits because I never want to repair anything that I design.) When I've built that type of indicator circuit on other projects, I've used a zener diode to clamp the voltage at 5V and then used that output to drive a transistor. The transistor drives the LED. Even an inefficient transistor would have a beta of 10:1, so you would only need 1mA of current flowing through the zener. A 47K to 70K 1/2W resistor would work along with a 1W zener (I tend to over spec wattage to keep the heat down). |
#51
|
|||
|
|||
Mike, sorry for resurfacing a two year old comment, but it fits almost my question, just wanted to be sure:
Im building a 34VAC*18A power supply, to get 48VDC, and I already have 2x 10000uF 100v caps, is 20000uF enough for this supply or I need to add more capacitance? It also has 2 extra secondaries (9 and 4, to get 12 and 5vdc) but these are simple. Thanks! Quote:
|
#52
|
|||
|
|||
Pablo,
80000 X 18A / 48VDC = 30,000 uF. I doubt that you'll be pulling 18A, so 20,000uF might be adequate, if not, just add another 10,000uF capacitor in parallel with the other two. Usually only the 48VDC is unregulated. The 5VDC supply would normally be regulated and perhaps the 12VDC, depending on what you're driving. The most simple way to regulate a 5V or 12V is to use a 7805 or 7812 regulator. Depending on the case style, and the heat sink used, they handle about 1A of current; however, they need about 3V extra. So, a 5VDC supply would start with an 8V or 9V unregulated DC source. The 7805 would regulate that higher source to 5VDC. A 12VDC supply would start with 15V or so. TTL type circuits (chips) like a well regulated voltage source. They don't work too well if the voltage source is not regulated. The specs on the 78xx regulators say that they can have a source voltage up to about 35VDC. That is true, but remember that the wasted volts are turned into heat. Even dropping 3 or 4 volts across a regulator will cause significant heat. You could use the 12V unregulated as the source for the 5V regulated line. I've done that many times. If you're just driving proximity sensors and relays, the 12V doesn't have to be regulated. If you're using the 12VDC as the voltage source for a PMDX-122, it doesn't have to be regulated. The PMDX-122 has a built-in 7805 regulator to change that 12V down to 5VDC. |
#53
|
|||
|
|||
I had a gut feeling that I should had ordered one more 10000ufx100v cap when I ordered them together with the proximity sensors, I think I can find 10000ufx63v locally, that should work too because I'm using just 48vdc.
For the 5 and 12v I have separate windings for each so I will just put a rectifier bridge, a capacitor on each side of the 78xx regulators and have both supplies regulated. I have a lot of LM350 lying around, with a couple of resistors they do the same job as specific regulators with the bonus that they can handle 3A Thanks!! |
#54
|
|||
|
|||
Pablo,
Remember that a diode bridge will waste about 1.4V (two diode "drops") and that many regulators require a 3V drop (minimum), so that means that the 5V line must start out with 5V + 3V + 1.4V = 9.4V. To get 9.4VDC, your AC winding would need to be at least 6.6VAC before rectification. (I tried using 6.3VAC in one project to get 5VDC, but I had to replace those transformers with 8VAC units.) The 12V circuit would require 12V + 3V + 1.4V = 16.4V minimum. That would require at least 11.6VAC before rectification. |
#55
|
|||
|
|||
Hi again Mike, is it normal for the transformer (EI not toroidal) to reach 60C (140F) measured with a digital thermometer?
what can cause such a high temperature? Keep in mind its only driving 2 geckos set at 1A each (10k resistor) and nothing else. BTW, its summer here, around 30C (86F) room temperature. Thanks! |
#56
|
|||
|
|||
My EI transformer runs less than 50C at any weather condition under full operation condition. BTW, my EI transformer is rewind from a 450VA iron.
As you only run 2A max. I suspect poor winding or too thin secondary wires. |
#57
|
|||
|
|||
The 'standard' transformers that I've used with stepper motors run hot. I never measured the temperature, but they were too hot to touch. All of those transformers were the Triad brand, which is well known and reputable. The toroidal transformers that I now use exclusively all run at room temperature or just a few degrees hotter. None of them ever get hot to the touch.
Today, I ran a test on some steppers to see how hot they would get when I used a 47VDC power supply to power some PK296B2A-SG3.6 motors (which are rated at 1.5 mH inductance and 3A). I knew that I could use more than 39VDC with those motors because their current carrying capacity was underrated by 33% by the factory. They got toasty warm - about 45-degrees C - well under the 80-degree C rise that is permissible. However, the transformer never got above 72F in a room that was 70F. Even the Gecko G202 stepper drivers that I was using stayed under 80F. I don't know why standard transformers run so much hotter than toroidal transformers. Many of the transformer web sites claim that toroidal transformers are much more efficient than standard transformers, but they don't explain why a standard transformer rated at 6A runs at a temperature that will blister your fingertips while a toroidal transformer also rated at 6A runs at room temperature under the same load. |
#58
|
|||
|
|||
Higher iron losses & hysteresis.
Even with these factors, 72C is abnormal... Maybe the factory did not underrate the tranny by 33%... Copper price had since sky-rocketeng in the past years & still rocketting... Anything can happen... |
#59
|
|||
|
|||
68000uF capacitors
Hi would it be ok to use 68000uF capacitors as I might be able to get some verry cheap ? Thanks
|
#60
|
|||
|
|||
Yes you can, it only takes more time to discharge after you powers off. Just make sure the rated voltage of the capacitor is 10V or 20V more than the voltage of your power supply, just to be safe.
|
![]() |
|
|