PDA

View Full Version : still cant find power supply



level head
07-21-2006, 10:30 PM
i still am looking for my cnc mill ,but with no luck does anyone have a 40-70 volt 5 amp dc regulated power supply for sale
thanks

Todd Tolhurst
07-21-2006, 10:42 PM
How's this (http://www.mpja.com/productview.asp?product=8527+PS) one?

level head
07-21-2006, 10:52 PM
i was pricing out $100 if the power supply says 8 amp output can i step that down to 5 amp

JRouche
07-21-2006, 10:59 PM
You dont step down the current. The drives will only pull what they are limited too. A bit more on the current side is not a bad thing. Just means you can drive a fan or two or similar accessories. JRouche

Also you prolly want a linear supply not regulated? I may be off here....


http://www.industrialhobbies.com/Merchant2/merchant.mvc?Screen=PROD&Store_Code=IH&Product_Code=PWR_48V_5A

Evan
07-21-2006, 11:00 PM
For steppers an unregulated supply is recommended. Regulated supplies can't provide enough instantaneous current.

Todd Tolhurst
07-21-2006, 11:02 PM
You don't need to "step down" the supply's current capacity; the load will draw whatever it requires, and no more.

Rustybolt
07-21-2006, 11:29 PM
http://www.plitron.com/ for custom toroid transformers or parts express for off the shelf. www.partsexpress

capacitor and bridge rectifier and your'e good to go.

BobWarfield
07-21-2006, 11:53 PM
These are not very hard to build. Here is some data on the one I'm building:

http://www.thewarfields.com/MTLatheCNCDrivers.htm

The page has parts info, suppliers, as well as pointers to where to get the data to design one that suits your needs. Make sure you have your steppers (or servos) before you dig in as their voltage requirements vary.

Best,

BW

RobDee
07-22-2006, 08:30 AM
First, I think you're putting the cart before the horse. Work from the stepper motor specs. back.

Secondly, you don't need a toroid, nice xformer but cheap surplus is fine. Ten to fifteen bucks.

Third, Bob, in your circuit the zener has no where to go. To start, zener diodes are low current devices. They won't do zip for the current driving a stepper over , say, 50 ma.

A one watt zener at, say, 24 volts means that the maximum current it will carry is 41 ma. I=P/E 1/24= .041 amps. I won't push a device over about 75% max so that brings us down to 30 ma.

Your zener has no where to go. If the voltage rises above the breakdown voltage of the zener it will simply blow. You need a resistor in series with the supply output. That means the supply will be severely limited and any load over the 1 watt zener power capability will be dissipated by the series resistor and the voltage will drop on the output significantly. The bottom line is you can’t use a zener for your application.

The best thing to do is design your supply to run the voltage you need or run several heat sunk three terminal regulators like the LM1084 in parallel buffered by .1 ohm resistors on each of their outputs.

To figure the rectified DC from an AC xformer secondary multiply the transformer secondary rectified and filtered voltage by 1.4 and subtract .7 volts for every diode you go thought to get to your stepper. This will get you close enough to your needed voltage.

Also a couple of 1,000 uF electrolytics in parallel is more then adequate for your application.

Hope this helps.

Rob Dee

J Tiers
07-22-2006, 09:14 AM
What has been said about the zener is pretty much true, given the assumptions.

With alternate assumptions.....

If the zener is meant to be a protection... so the voltage never goes OVER that voltage even if/when there is reverse current back from teh steppers due to over-run, etc... Most drives have a max voltage, above which they will fail catastrophically.

Then the zener setup can be useful. If it is supposed to be a "crowbar" protection against input voltages too high, it needs a fuse at least in series with input from capacitors... and probably an SCR or similar added to provide a higher current device that won't require replacement every time. The zener would be used to trigger the SCR on over-voltage.

The SCR also will positively clamp fed-back voltage from the drives (possible, depending on design) and as soon as th current drops to zero, it turns off and is ready for the next use.

But, no, the zener alone with nothing to "act against" as in your circuit will not do anything but probably fail shorted in the case of an overvoltage.

In teh picture of the thing on the capacitor, the "zener" looks like it may be an MOV. They are much tougher than a zener. BUT, they have a very imprecise breakdown voltage, and a "clamping" voltage that may be almost twice the breakdown voltage. So they arent very good at precise protection against overvoltages.

RobDee
07-22-2006, 09:37 AM
What has been said about the zener is pretty much true, given the assumptions.

With alternate assumptions.....

If the zener is meant to be a protection... so the voltage never goes OVER that voltage even if/when there is reverse current back from teh steppers due to over-run, etc... Most drives have a max voltage, above which they will fail catastrophically.


The zener can never be protection in the illustrated circuit. Any power above the device dissipation of the zener will blow it regardless of whether from the stepper or the supply.

Then the zener setup can be useful. If it is supposed to be a "crowbar" protection against input voltages too high, it needs a fuse at least in series with input from capacitors... and probably an SCR or similar added to provide a higher current device that won't require replacement every time. The zener would be used to trigger the SCR on over-voltage.

A fuse that blows from zener breakdown will still be in the ma. range and any stepper current above that range would blow the fuse regardless whether a zener was in the circuit or not.

The issue isn’t whether zener diodes can be useful in power supply applications. Triggering an SCR or TRIAC is a low current application. This is clearly not the case here. The zener is valueless in the circuit shown.

Rob Dee

Evan
07-22-2006, 09:47 AM
The zener is to protect against back emf from the drives produced by the steppers, at least it should be. The supply should be designed to take into account variation in line voltage using 132 volts AC as the maximum possible input voltage. If protection against higher inputs is required then a zener can be used together with a high capability power device to provide a shunt regulator. Putting any resistance in the output leg of the supply is a bad idea. Even with the resistance of normal wiring if the leads to the stepper drivers from the supply filter cap are more than about six inches long then additional capacitors are usually required close to the driver.

[edit] BTW, a zener such as that is recommended in the application notes by various manufacturers of drive modules. The amount of power it handles is very small, it's only meant to sink very short duration overvoltage spikes that would exceed the maximum rated voltage of the driver.

RobDee
07-22-2006, 10:26 AM
The zener is to protect against back emf from the drives produced by the steppers, at least it should be. The supply should be designed to take into account variation in line voltage using 132 volts AC as the maximum possible input voltage. If protection against higher inputs is required then a zener can be used together with a high capability power device to provide a shunt regulator.

Again, what can be done and the circuit shown are two different things entirely. We’re not assuming what can be done with a circuit component we’re evaluating a circuit with a component in it.

Putting any resistance in the output leg of the supply is a bad idea.

Exactly!
And the only way the zener in this circuit has any value is with a series resistor.

Even with the resistance of normal wiring if the leads to the stepper drivers from the supply filter cap are more than about six inches long then additional capacitors are usually required close to the driver.

Not for regulation, for noise. If you want to regulate a high current driver then you don’t use a zener. It is a low current device.

If you want to protect the back EMF then put a standard diode across it like is done for any inductive device.

[edit] BTW, a zener such as that is recommended in the application notes by various manufacturers of drive modules. The amount of power it handles is very small, it's only meant to sink very short duration overvoltage spikes that would exceed the maximum rated voltage of the driver.

Back EMF, forward EMF, doesn’t matter. The zener has a maximum power dissipation. Anything over that will blow the zener, regardless whether it is a short duration spike or a long one.

You might put a zener on the gate of a low power MOSFET circuit but it doesn’t belong anywhere near the B+ supply to the drain of an ‘N’ type MOSFET or the source of a ‘P’ type MOSFET.

The bottom line is if you want to protect the driver chip itself then use a three terminal regulator. If you want to protect an inductive device driver from back EMF then use a silicon or schottky diode across the device.

J Tiers
07-22-2006, 11:10 AM
The zener can never be protection in the illustrated circuit. Any power above the device dissipation of the zener will blow it regardless of whether from the stepper or the supply.

Then the zener setup can be useful. If it is supposed to be a "crowbar" protection against input voltages too high, it needs a fuse at least in series with input from capacitors... and probably an SCR or similar added to provide a higher current device that won't require replacement every time. The zener would be used to trigger the SCR on over-voltage.

A fuse that blows from zener breakdown will still be in the ma. range and any stepper current above that range would blow the fuse regardless whether a zener was in the circuit or not.

The issue isn’t whether zener diodes can be useful in power supply applications. Triggering an SCR or TRIAC is a low current application. This is clearly not the case here. The zener is valueless in the circuit shown.

Rob Dee


as I said.......
But, no, the zener alone with nothing to "act against" as in your circuit will not do anything but probably fail shorted in the case of an overvoltage.

Although, zeners come in different wattages, and a large wattage (10W +) zener might possibly be able to handle small back emf issues.... but not a good idea.

AND, the usual back EMF issue is "pump-up" in a PWM where the DC voltage is raised by returning energy from the load. It may be from braking, reversing, etc.

That is a DC condition, not a transient.

No regulator on the input can affect it in the slightest way.

For that, the best idea is a simple circuit to dump the power into a load resistor and draw down teh over-voltage. The Gecko guy (Mariss) showed an elegantly simple way a year or so back.

Note that "pump-up" is inherently energy-limited, as the limit is the mechanical energy in the drive motion. IT IS NOT drawn from the transformer, rectifiers, etc, so a pump-up limiter is NOT in danger from power current.

However, for that to be true, the pump-up limiter (a shunt regulator of sorts) must be set ABOVE the maximum DC output of the supply at maximum high-line conditions, which is often taken to be 132V on 120V, but really should be considered to be higher for safety (say 140V). That way it ONLY operates on pump-up.

The Mariss circuit operates ONLY on returning energy, and is NOT voltage sensitive.

JRouche
07-22-2006, 11:26 AM
LOL....Level head, did you get all that.

You listen to the ramblings of Robdee, J Tiers and Evan long enough you may get to some of the other components figured out too. Then just build your own PS. ;) JRouche

RobDee
07-22-2006, 11:56 AM
LOL....Level head, did you get all that.

You listen to the ramblings of Robdee, J Tiers and Evan long enough you may get to some of the other components figured out too. Then just build your own PS. ;) JRouche

OK, this is what I don’t get.

If a plumber, a professional who has been doing it all his life, comes into my house and says, “hey, you used 4 inch soil line when 3 inch was over kill so you're wasting money.” And “ That vent won’t work because it is to far from the basin.” Then I would listen to him. I might check out what he has to say but I would learn from that and my thinking would change.

No ego involved, no hurt feelings. I made a mistake I’ll make one less tomorrow because I listened and learned something.

Will Rogers said, “Everyone is stupid, just on different subjects.” Believe me I have a plethora of my own but electronic engineering and model prototyping are not on that list.

So what are you saying? “Then just build your own PS.” What thinking should he use to construct that supply?

Our statements about the circuit we constructive and not ridicule, regardless or not that we had different perspectives to express. Your statement is out of sarcasm and leaves the individual making the same mistakes tomorrow he made today.

Too bad.

Evan
07-22-2006, 12:03 PM
Pacific Scientific manufactures a wide range of steppers, servos, and the associated driver modules. They also have some excellent application notes for their products. In particular, the 5210 stepper driver which is currently available from lasermotion.com for $49 each has a very detailed description of power supply design for exactly the sort of system that levelhead is contemplating.

In the application note for the Pacific Scientific Sigma 5210 driver (http://www.pacsci.com/support/documents/5210manual.pdf) they include a schematic for a recommended power supply design. You will note that a regen zener diode is called out if needed on the output leg of the supply. It is required if the power supply voltage is close to the maximum rating of the driver modules.

http://vts.bc.ca/pics/powersupply.gif

alanganes
07-22-2006, 12:12 PM
Hi,
Another option would be to connect several PC power supplies in an appropriate series-parallel combination to get the voltage and current capicity that you require. This will not necessarily result in the most compact unit and the requlation may be overkill, but these things are dirt cheap and available almost anyplace. I did this (used 2 supplies in par.)to make a 12V@30A supply to run a mobile ham radio in the house. Works great. There are more details here:

http://www.qsl.net/i0jx/pcsupply.html

and here:

http://www.antennex.com/preview/archive3/powers.htm

I'm sure there are lots of other such sites lurking about the web. Just a thought.
-AL A.

Evan
07-22-2006, 12:19 PM
As I stated earlier, a regulated supply is not recommended. Also, connecting several PC supplies in some sort of series configuration is highly unsafe and definitely not recommended. The power supply for a stepper or servo system must be very carefully designed and constructed. If not, you risk destroying all the components powered by it, an expensive mistake.

BobWarfield
07-22-2006, 01:46 PM
Wow! It's so easy to create CNCZone-style threads on HSM. First ballscrews/angular contact bearings and now zener diodes. I can't wait for the next opportunity.

FWIW, buy the little Ajax circuit board and forget the rest. It's cheap, it is proven, and it's neat. No zener there and I don't worry about it. Opinions have varied all over the map on the value of that zener. Plenty of experts weighed in on both sides. If you don't have your own strong opinion, you're probably better off with the Ajax board anyway.

Mariss F. himself weighed in against the zener, at which point I quit worrying about it too. Even if you like the zener idea, it was never more than a minor filigree at best. The important thing is the basic power supply is easy to put together, even more so with one of the little Ajax boards. Incidentally, Tormach uses a similar little board, but they want a lot more money for theirs. PMDX also make a real nice board.

Best,

BW

PS Wait a minute, what zener diode are you guys all talking about? I don't see a zener in the diagram on my page (any more)? <G>

RobDee
07-22-2006, 02:11 PM
Pacific Scientific manufactures a wide range of steppers, servos, and the associated driver modules. They also have some excellent application notes for their products. In particular, the 5210 stepper driver which is currently available from lasermotion.com for $49 each has a very detailed description of power supply design for exactly the sort of system that levelhead is contemplating.

In the application note for the Pacific Scientific Sigma 5210 driver (http://www.pacsci.com/support/documents/5210manual.pdf) they include a schematic for a recommended power supply design. You will note that a regen zener diode is called out if needed on the output leg of the supply. It is required if the power supply voltage is close to the maximum rating of the driver modules.

http://vts.bc.ca/pics/powersupply.gif



Evan,
Thank you. I have looked over the design of the driver chip and I understand their thinking.

I stand corrected but I do have some strong caveats.

First, I remember building SMPS circuits in the late 80’s and early 90’s. We were all scratching our heads back then as the blown power supplies came rolling back in! It wasn’t until later on that we took a hard look at what was blowing these things up and we saw step up shoot through and spikes that we never saw before.

I looked at their chip (according to the data sheets this was designed around 93) and they used shear power to overcome design flaws. I’m sure this circuit works fine with the zener clamp but the problem is that they should never have put the B+ power drive pin and the chip voltage pin together on the chip. The circuit to the ‘H’ bridge should have been tightly regulated and the high side MOSFET gates driven from a charge pump configuration in the manner that the Intersil HIP4080 series chips do ‘H’ bridges. Not only would this have resolved the regen. Problem but would have allowed running steppers above their 44 volt limit. In their case the easiest path was to put a 5 watt zener on the power supply. If you noticed they went into the designing of the supply and stressed the maximum output to the caps. That’s because that zener can never limit a high current supply that is over 44 volts.

Today there are a lot better solutions for stepper design. I use embedded uC chips either alone or running the L293E chips or something close to it from ST Micro.

A couple of cautions for people running their configuration. Make sure you haven’t blown the zener and it’s not just sitting there open. If you do you’ll blow their chip. Make sure your power supply doesn’t run to close to the zener voltage and finally put a MOV (metal oxide varistor) with a couple of caps on the primary side of the xformer to cut down power line spikes.

Rob Dee

Mcgyver
07-22-2006, 02:20 PM
level, I built mine for $50 and I am an electronics ignoramus.

Microwave transformer: free
10 Gage magnet wire from electrical motor repair guy: $12
2 Beer can sized 30,000 uf capacitors from surplus store, $20
rectifier: $5
sundries, hook ups, fuse, switch cardstock and fiberglass resin, $12.

and to complete the cliche, accomplishment of DIY'ing yet another part the cnc machine, priceless. liked it so much i wrapped another secondary winding to give my low voltage (yeah if i was better organized i could have tapped the first, but hey, I'm a beginner)

basically you scrounge an old microwave transformer, remove the secondaries and shunt and rewind to get the voltage you need. I knew nothing about it when i started but there are lots of online resources and its not hard. if you want to have a go i'll find some links

alanganes
07-22-2006, 02:44 PM
As I stated earlier, a regulated supply is not recommended. Also, connecting several PC supplies in some sort of series configuration is highly unsafe and definitely not recommended. The power supply for a stepper or servo system must be very carefully designed and constructed. If not, you risk destroying all the components powered by it, an expensive mistake.


I am curious as to why a regulated PS would be a problem. Please explain. it seems to me that the function of a power supply is to supply power, at the specified voltage with sufficent current. Protection from back-EMF, spikes, noise, etc. is a seperate issue. Typically, if a system requires X Volts +/- some %, at Y Amps, the PS need only to hold its output within those values under all expected operating conditions. If it can achieve that without regulation, that's great. But I cannot see how a regulated PS would cause a problem. Particularly with stepper or servo systems where, in general, good voltage regulation is an advantage.

How is connecting power supplies in series highly unsafe? It is done in lots of places for lots of reasons. Again, please elaborate.

Please understand that I am not trying to be a wise guy, just asking for clarification.

Evan
07-22-2006, 03:19 PM
A regulated supply cannot supply the surge current that an unregulated supply can. The regulating elements in the supply act as a resistance in the supply lead, as they must, with the exception of a shunt regulator. With a regulated supply there is a limit to the amount of capacitance that may follow the regulator as too much capacitance can result in the regulator shutting down on startup through current foldback because of the high charging current of the capacitance.

In order to combine the outputs of two PC power supplies the safety ground of one of the supplies will have to be removed. The grounds of a PC power supply are connected to the case and the safety ground. In order to connect two supplies in series one supply must be reworked so that the case is floating at the output potential of the first supply. This poses several hazards. It would be very easy to short the case of the floating supply to ground.

Even if the boards are isolated from the enclosures a component failure could place line voltage on the outputs, especially the "ground", and with no safety ground it would present an extreme shock hazard as well as destroy all equipment powered by the supplies. There is a very good reason that the supplies have the grounds connected to the case and to the safety ground.

JRouche
07-22-2006, 03:37 PM
So what are you saying? “Then just build your own PS.” What thinking should he use to construct that supply?



RobDee, see that little face lookin thingy ;) after my statement. Kinda hard to make out but it is a "wink". Which I interpreted as joking or tongue in cheek type expression. I could have used :-þ or :P or wink with a tongue ;p or actually tongue in cheek ;-^) Dunno, crazy thing this forum communication thing.

No, I dont think the original poster wanted to build his own PS, he didnt seem to imply that. I think he was looking to buy one. Even talked about a certain dollar amount.

You all obviously know what you are talkin about, the originator of this thread may not so you basically hijacked his thread.

As for learning new concepts I'm all for it. JRouche :)

RobDee
07-22-2006, 03:58 PM
A regulated supply cannot supply the surge current that an unregulated supply can. The regulating elements in the supply act as a resistance in the supply lead, as they must, with the exception of a shunt regulator. With a regulated supply there is a limit to the amount of capacitance that may follow the regulator as too much capacitance can result in the regulator shutting down on startup through current foldback because of the high charging current of the capacitance.

In order to combine the outputs of two PC power supplies the safety ground of one of the supplies will have to be removed. The grounds of a PC power supply are connected to the case and the safety ground. In order to connect two supplies in series one supply must be reworked so that the case is floating at the output potential of the first supply. This poses several hazards. It would be very easy to short the case of the floating supply to ground.

Even if the boards are isolated from the enclosures a component failure could place line voltage on the outputs, especially the "ground", and with no safety ground it would present an extreme shock hazard as well as destroy all equipment powered by the supplies. There is a very good reason that the supplies have the grounds connected to the case and to the safety ground.

Actually there is some new thinking on this issue.

You can run two or three 5 amp regulators in parallel with a low value resistance off each leg and very little capacitance on the output (I’d say 1 to 4.7 uF tantalum) and have very little voltage drop, quick recovery and small footprint.

Let’s say you wanted 24 volts at 3 amps with three regulators. Even using a plain vanilla LM317 regulator and three regulators I can get very good results.

Steppers are very slow devices in electronic terms and this is in our favor.

To keep the regulators from ‘fighting’ with each other I’ll put a .1 ohm resistor on each output.

So if we have an 18 volt xformer we get 32.5 volts after the bridge in a full wave bridge configuration. (18 * 1.414 – 1.4 = 32.54). This gives us 8.54 volts overhead (32.54 – 24 = 8.54) which is plenty for our application.


Each leg carries 1 amp (3 legs / 3 amps = 1)
With a .1 ohm resistor on each leg the drop on each resistor at 1 amp is .1 volts (1 * .1 = .1) Without large caps on the output and our 8 plus voltage over head our output voltage stays very stable and within a tenth of a volt of our 24 volts.

Power dissipated by each device is 8.54 watts (8.54 volts * 1 amp = 8.54 watts) well within specs.

Great regulation from the LM317, low ripple without large caps and 85 cents each.

Since the advent of SMPS (switch mode power supplies) designers are changing their thinking from the days of brute force, low freq. expensive capacitors.

level head
07-22-2006, 03:58 PM
hey ,if it is as easy to build a p.s from a micrwave than you can give me those links .



my motors are 2.3 volts at 5.5 amps .I read earlier that my power supply voltage has to betweeen 5 and 20 times the motor voltage .So to get this straight the 2.3v times 20 is 22.3v then i found a wiring diagram for my gecko 202 at cnc zone and it says the p.s should be between 24 - 80 volt at 5 amps since my motors are 5.5 will that make a difference .And is this wirring diagram a accurate diagram for my motors.

http://www.cnczone.com/forums/attachment.php?attachmentid=19401

Mcgyver
07-22-2006, 04:25 PM
level, if you're already on cnczone, that's where i got the info from. here's some threads. mostly follow what bubba has to say. last link fills in some of the questions i had and photos of mine in progress.

http://www.cnczone.com/forums/showthread.php?t=7656&page=1&pp=15&highlight=microwave+power+supply

http://www.cnczone.com/forums/showthread.php?t=5658&highlight=microwave+power+supply

http://www.cnczone.com/forums/showthread.php?t=14446&page=1&pp=15&highlight=microwave+shunts

there was a good discussion here on what to seal the new windings with (they'll vibrate like crazy otherwise and wear off their enamel coating) I ended up using fibreglass resin. I completely sealed one side of the transformer with duct tape (handy mans secret weapon), hung it from a wire with the open side up. poured the liquid resin in the open side and let it sit.

make sure your boards can handle whatever voltage you make. The reason why you use such a high voltage is that the drivers pulse or chop the power to the motor - the power is on at a much higher voltage than the motor's rated for but for a small perioed of time. otherwise they've very little holding power.

your call if you want to make one, I just wanted to point out that they aren't hard or expensive to make and that way you can get the exact voltage you want. For me, I'm trying to learn more about electronics so want to diy

Now if only i got get those dang controller cards working properly....

level head
07-23-2006, 01:24 AM
please help me i just need to kow the voltage and amp for my p.s to fit my stepper motors by the way i am using gecko 202
this is what i got at another form

I have 3 stepper motors at 2.3 v 5.5 amp .will this power supply work with it


http://cgi.ebay.com/50VDC-10A-Power-Supply-CNC-Servo-Step-Motor-Amp_W0QQitemZ270010494398QQihZ017QQcategoryZ78189Q QssPageNameZWDVWQQrdZ1QQcmdZViewItem






It will depend on the drives you are going to use. Some drives can only
handle 30v other like the geckos can go to 80v.
10Amps may be on the shy side for all three of your steppers, think you should be looking at 2/3 your total current which would be around 11A but you could possibly scrape by.

if you could tell me what i need it will be very help full because in a wiring diagram it said 5 amps so please send me in the right direction

thank -you

J Tiers
07-23-2006, 01:27 AM
The AnTek supply looks a little "loose" but might work for Gecko. The large capacitance will minimize regeneration voltage increases, but at 50V falling to 45V it might be a little lower, IIRC the Gecko likes more than that, aren't they rated for 75V?

In any case it probably would function OK. Maybe not the best. The Geckos I think have a current limit up around 5 or 7A, so several of them might overload that if set for max current.


I looked at the Pac Sci data sheet for the drivers as linked above by Bob W.....

Ya gotta love the stuff that has a nice low power supply voltage that you need to get close to for best power, but can't go over it ever...... In this case 44V.

Anyhow, according to the data sheet, the current draw has a limit. Therefore, there is no special need to worry over peak current demands. Per teh data sheet, as soon as current reaches the limit, it is shut off and allowed to decay. So presumably max current is the sum of all the current limits as noted in the data sheet. And the limit is settable, so you have some control.

I do NOT think there is a big issue with regulated supplies.... Should work great as long as the supply can provide teh required current output.

The motor inductance will limit the rate of rise of current, as noted in the data where the figure shows the triangle-wave motor current waveform. So arcane "transient current" requirements are probably not in the picture.

If any issue is anticipated, a larger output capacitor can be provided to allow the regulated supply a chance to "catch up". A small value resistor can be used to isolate the output cap enough to avoid overcurrent and stability issues with a linear regulator.

They do require a local 4700 uF capacitor, to ensure stable operation. But they ALLOW a larger capacitor if required.

They mention regeneration (although AFTER the index..... Eh? almost missed it...) their figuring is OK, and you may note that they mention larger zener ratings of 5W. The duty cycle is low, so that can be OK if there is sufficient heatsinking of the zener.

BUT, there are other ways to handle it if you don't like the zener.

If there are several stepper drivers, and they are known to not all be regenerating at one time, then some will act as a load to keep down teh voltage due to regen energy from another.

A larger total capacitance will reduce the voltage increase with any given amount of regen energy. That alone can prevent over-voltage.

A bleeder resistor can be put in place, which will drain energy and help keep voltage down. But it is an energy waste, and should be a last resort.

A dumper circuit can be set up to sense reverse current, or excess voltage, and connect a load resistor temporarily to dissipate the excess energy.

As far as regulated supplies with low output capacitances..... NOT a good idea for a regenerating load. In this case the maker REQUIRES a certain minimum capacitance, but others might not.

UNLESS the regulated supply can SINK current as well as supply it, the regeneration energy will raise the voltage by an amount related to the returned energy by energy in joules = 1/2C*V^2, where C is the capacitance, and V is the voltage.... they give a calculation which you can use....

So the smaller the total supply capacitance, the higher the voltage may be "pumped". In this case, if you expect significant regeneration, you WANT a brute force approach of more capacitor, OR you need to provide a "dumper" to get rid of the energy, which provides a way for the supply to sink current if it cannot inherently do that.

The dumper may be a zener, or it may be something else. But if it is a zener, you need to determine roughly how much average regeneration energy you expect at worst case, and provide at least that much zener power dissipation.... Otherwise you will be fine until you stop all three axes at once and blow it.

John Stevenson
07-23-2006, 04:32 AM
That wiring diagram posted by Evan is incorrect in that there is no smoothing cap present in the circuit.

Go to the Gecko site at http://www.geckodrives.com and have a look at the diagram there under support.

This is the classic no frill power supply needed for a CNC supply.

In it he explains what voltage you need to aim for although this is biast on the 80 volts needed for the Gecko drive obviously,

Here's a similar circuit but with fixed value caps for a certain driver.

http://www.arceurotrade.co.uk/projects/PSU/images/PSU1.gif



Rule of thumb is it's 20 times your motor voltage but about 10% lower then your drivers max rating.
Bear in mind that an AC supply will increase by 1.414 when rectified and smoothed so a 50 volt AC transformer will output about 71 volts DC when smoothed.

Amps is based on 2/3 of the total motor amperage as it never runs all three motors flat out and even so you will be hard pressed to even see the max motor amperage.

There is also a formula for working the cap size on on the same page.


It's not rocket science and many hundreds of power supplies have been built by people looking at Marris's pages.
This isn't theory, it's practical, it's been done may hundreds of times and it work.

I must have done over 50 by now, all work, none have blown up or failed, what more can you say ??
.

RobDee
07-23-2006, 09:17 AM
J. Tiers,
I would not use a regulated supply with the chip expressed here but frankly I would not use the chip.

I can see its attractiveness from the simplicity stand point but it is an old design with severe limitations and while the zener is the fastest easiest solution it is not good engineering practice. As I said before the zener has no place to go once the voltage reaches it and a stepper with enough inertia will blow even a 5 watt rated part along with the chip.

My hunch is that the zener was an after thought stop gap measure from excessive field failures. I’m not blaming them, look at it from their standpoint, they would have to reengineer a completed chip when they already had one that works well enough in most applications.

However, when I come up with the same problem I reengineer the circuit once the original stock has been depleted . In their case they have had almost fifteen years to do that.

Rob Dee

J Tiers
07-23-2006, 09:58 AM
Bear in mind that an AC supply will increase by 1.414 when rectified and smoothed so a 50 volt AC transformer will output about 71 volts DC when smoothed.

True, in that that (less rectifier drop) is the MAXIMUM voltage. That is the no-load voltage.

But it falls rapidly under load. It bottoms out at a voltage that is dependent on the load current and the series resistance of transformer etc. So you can't count on the power supply voltage being at that. The 50V transformer might , depending on load, give anything from about 70V down to maybe 55V.


Yes the Pac Sci "chip" (actually it is probably a hybrid circuit) is not the newest... in fact I don't think it is made any more.

But, the principles of stepper operation are the same with any controller. The regeneration voltage issue is one that really can't be avoided with any design, it can only be dealt with.

And, a zener CAN handle it, as long as the average dissipation isn't over the zener wattage. The requirement is really just on temperature. If you don't go above the max temp, it will last indefinitely, no matter what theoretical power you present to it.

In practical terms, that means it can handle well over its rated power so long as that doesn't continue for long enough to heat up the zener too much. That is why they specify a zener, and you will see a section devoted to determining exactly what you require for power ratings in teh data sheet.

sure, there are better ways, but that way can and does work if applied right.

As far as the original question, if you build a supply, 48V transformers are not too rare. That voltage should give a maximum high-line voltage of about 73V, which IIRC is under tha max voltage for the Gecko. Under load it will be below that. Other multiples of 12V are also not rare, and for instance, 36V gets you about 50V nominal.

You do need to get the right current level. The formulas etc in teh Pac Sci data sheet near the end, appear to be appropriate for determining the current and various necessary other ratings.

In teh US, a company out east, Signal Transformer co, has a wide variety of standard UL recognized transformers, and they specialize in small volumes and single pieces.

Also, some can be obtained from Mouser electronics, or Digi-Key, for those in the US.

Once you have the transformer, the appropriate fuses, rectifiers, etc are also easily obtained from Mouser, Digi-key, etc.

While you CAN wind your own transformers, and I often do wind prototypes at work (for SMPS) I don't recommend it for the inexperienced.

The surplus places also may have appropriate transformers. Surplus center, etc.

nheng
07-23-2006, 10:10 AM
Several problems with zeners as clamps are speed (a turtle to today's hardware but perhaps not a problem in motor drives) and impedance. The impedance of a random 3W, 82 volt zener is over 100 ohms. Hit it with any current and you are way beyond 82 volts ... rendering it all but useless. Den

Evan
07-23-2006, 10:18 AM
That wiring diagram posted by Evan is incorrect in that there is no smoothing cap present in the circuit.

Yes there are. They are placed in close proximity to each of the driver modules.

J Tiers
07-23-2006, 10:34 AM
Several problems with zeners as clamps are speed (a turtle to today's hardware but perhaps not a problem in motor drives) and impedance. The impedance of a random 3W, 82 volt zener is over 100 ohms. Hit it with any current and you are way beyond 82 volts ... rendering it all but useless. Den

Actually zener (avalanche) diodes are quite fast.... it depends more on the driving circuit impedance, and the capacitance of the junction.

For this applicatin the speed is, however, totally irrelevant, since the rate of change of voltage is dependent on the circuit capacitances which will, if properly sized, hold it way down.


And, that zener impedance is current dependent. it goes down with current, until it reaches a minimum value. The minimum value is not always given.

A 1N3001 10W zener at 68V rating has a nominal zener impedance of 18 ohms at nominal 65 mA current. At 1 mA it's nominal impedance is 600 ohms. You see it is a curve......for which they give you two points. The minimum for higher currents is not given.

Evan
07-23-2006, 11:12 AM
Several problems with zeners as clamps are speed (a turtle to today's hardware but perhaps not a problem in motor drives) and impedance

Use a close relative of the zener, a TVS diode. They have response times in the picoseconds and will withstand peak powers in the kilowatts with clamping curves that are cliff shaped, with maximum currents in the hundreds of amps. The total dissipation is a product of time and power with steady state values of 5 to hundreds of watts depending on type. Typical rating for a 5 killowatt peak unit is a 1 kilowatt for 10 milliseconds and 100 watts for .1 seconds. They are also cheap.

John Stevenson
07-23-2006, 11:13 AM
Yes there are. They are placed in close proximity to each of the driver modules.

Not on that circuit you quoted.

So you have three or 4 large caps to do the job of one ?

.

Millman
07-23-2006, 11:25 AM
Now if you guys would have stuck to the manual machines.....as God intended...you wouldn't be having these problems. Builds up your arm muscles as an added benefit. :)

J Tiers
07-23-2006, 11:34 AM
Use a close relative of the zener, a TVS diode. They have response times in the picoseconds and will withstand peak powers in the kilowatts with clamping curves that are cliff shaped, with maximum currents in the hundreds of amps. The total dissipation is a product of time and power with steady state values of 5 to hundreds of watts depending on type. Typical rating for a 5 killowatt peak unit is a 1 kilowatt for 10 milliseconds and 100 watts for .1 seconds. They are also cheap.

A good plan.

Their one problem is that they have rather loose specs for breakdown voltage, and usually have a much higher clamp voltage than breakdown voltage, relative to a Zener (avalanche) diode.

So they are not very good for a situation needing a closely limited voltage.

But they are MADE for transient limiting, and within their limitations they work very well.

However, they basically ARE a zener, optimized for power dissipation and clamp speed. The zener speed is comparable, and I use one or the other depending on the requirements of precision limiting traded against raw pulse power handling.

Evan
07-23-2006, 11:37 AM
John, they are shown in the schematic.

http://vts.bc.ca/pics/powersupply2.gif

Placing individual caps at the modules is more effective (and required) as it reduces effects from resistance and inductance in longer wiring to a single cap located further away.

Evan
07-23-2006, 11:59 AM
Their one problem is that they have rather loose specs for breakdown voltage, and usually have a much higher clamp voltage than breakdown voltage, relative to a Zener (avalanche) diode.
Not according to the data sheets.

http://vts.bc.ca/pics/tvsspec.gif

I've used these for many years and many types clamp very precisely with values within millivolts of spec and within less than a volt or two of the avalanche voltage in the current ranges of interest. In this application they aren't going to see currents anywhere near Ippmax.

[edit]
As for looseness in the breakdown voltage they are usually very consistent within a batch and if on the low side can be adjusted with a regular diode in series to raise it.

John Stevenson
07-23-2006, 12:41 PM
John, they are shown in the schematic.

http://vts.bc.ca/pics/powersupply2.gif

Placing individual caps at the modules is more effective (and required) as it reduces effects from resistance and inductance in longer wiring to a single cap located further away.

This isn't the diagram you first quoted.
Remember we have plenty of newbies on this board who take everything the more senior members say literally and leaving much needed caps out is a problem.

You only need caps at the drivers IF your run of cable is over 12" from caps to driver [ recommended ]

By using one large central cap you can handle any back EMF from the other drives better.
This way with a safety margin built in on max voltage the whole thread on zenners is not needed as you won't have to dump any excess voltage.

A lot of what is advised is overkill for safety reasons or CYA.
I have a big mill, size of a Bridgy Series II running at 72 volts into Gecko 210's driving those big horrible type 42 motors and the main transformer is 4 feet away from the cabinet, two of the drives have 470 uf caps direct on the motor leads, two don't.
Been in production now about a year and a half from conversion and no problems other than silly ones like keyboards packing in.

If you ever want to try anything on the limit then those 42's as fitted to the Bridgy's will guarantee to loose steps on acceleration and pump out loads of back EMF on braking.

.

Evan
07-23-2006, 12:55 PM
Yes it is the diagram I first quoted. I posted only the part relevant to the discussion at hand and it should be obvious looking at the part I posted of the schematic that it isn't the entire schematic. The people here, new or not, aren't idiots and I also made it clear the information was to be found at the link I posted to the data sheet where the entire schematic is found.

According to application notes from various manufacturers for various driver boards and modules the filter cap(s) should be no more than six inches away. If the main cap is further than that then smaller caps should be placed as close as possible to the driver board.

By placing individual caps at the boards they also serve as decoupling caps which helps prevent ripple from the load of one driver affecting the others.

J Tiers
07-23-2006, 01:31 PM
If you look at the 1.5KE75, by 10A peak they are pushing 10V high in voltage. That could be an issue, and is a comparable voltage change to that of the regeneration energy itself in a typical circuit.

You don't even need it, maybe, as it won't hold pump-up any closer than a larger capacitor will.


No problem for many applications, typically transient spike limiting, but if you need a closely held max voltage for surges, you are far better off with a different approach, or some way to avoid the need.

I agree with J.S. .....I also am highly in favor of the large capacitor solution. It works, it actually saves energy, and in general, another motor will pull the energy that the first one dumps.

Improperly applied zeners are worse than no limiter, in many ways. They don't work, and yet they tend to make a person THINK they are safe from regen surges when they are not.

Stay 20 - 25% under the max voltage of the device, and use a larger than required capacitor. Should usually keep problems away except in really odd cases.

Is it optimumized? No, but that isn't required. What IS required is a working system that does not give trouble.

Evan
07-23-2006, 04:02 PM
I agree with J.S. .....I also am highly in favor of the large capacitor solution.

I as well. It's just a matter of deciding how much.

http://vts.bc.ca/pics/caps1.jpg

RobDee
07-23-2006, 04:24 PM
Yes the Pac Sci "chip" (actually it is probably a hybrid circuit) is not the newest... in fact I don't think it is made any more.

But, the principles of stepper operation are the same with any controller. The regeneration voltage issue is one that really can't be avoided with any design, it can only be dealt with.



Thisis exactly my point about spliting the control circuit from the mosfets. You don't have the regen problem to contend with. Mosfets can run pretty high voltage or you can use IGBT's.

I'll never see it as anything less then a poor design.

John Stevenson
07-23-2006, 05:49 PM
I as well. It's just a matter of deciding how much.

http://vts.bc.ca/pics/caps1.jpg

The formula that Mariss has on page 12 of his support page gives a very good idea of what size to use given the total amp load and voltage.

The rule of thumb is amps multiplied by 80,000 divided by the output voltage, result in Microfarads and working voltage needs to be 25% plus of the output voltage.

.

wmgeorge
07-23-2006, 07:52 PM
I'm using the same steppers amp draw as yours, I did an over kill. My PS is 18 amps at 36 volts DC, running Gecko 202 drivers, works fine. I went a little larger in case I wanted to add a 4 axis. I purchased a donut type transformer off eBay, Radio Shack 25 amp bridge rectifier for $5 and I had to buy 3 30,000 MFD caps to get one, so I you need one... I'll make you a deal!!



easy to build a p.s from a micrwave than you can give me those links .



my motors are 2.3 volts at 5.5 amps .I read earlier that my power supply voltage has to betweeen 5 and 20 times the motor voltage .So to get this straight the 2.3v times 20 is 22.3v then i found a wiring diagram for my gecko 202 at cnc zone and it says the p.s should be between 24 - 80 volt at 5 amps since my motors are 5.5 will that make a difference .And is this wirring diagram a accurate diagram for my motors. http://www.cnczone.com/forums/attachment.php?attachmentid=19401[/QUOTE]

JRouche
07-23-2006, 09:25 PM
The rule of thumb is amps multiplied by 80,000 divided by the output voltage, result in Microfarads and working voltage needs to be 25% plus of the output voltage..

Pretty much the standard. I have researched many schematics and formulas and they all revolve around this basic concept. JRouche

Evan
07-23-2006, 09:27 PM
I like the supply to be as stiff as possible so will use more capacitance than recommended. By using a soft start resistor and a time relay the inrush current to the caps can be limited so it isn't a problem. Likewise for a bleeder resistor and relay for fast bleed down. I also like to put a latching relay on the AC input side so if the power kicks out it doesn't turn back on when the power does.

John Stevenson
07-24-2006, 02:20 AM
I'm using the same steppers amp draw as yours, I did an over kill. My PS is 18 amps at 36 volts DC, running Gecko 202 drivers, works fine.

my motors are 2.3 volts at 5.5 amps .I read earlier that my power supply voltage has to betweeen 5 and 20 times the motor voltage .So to get this straight the 2.3v times 20 is 22.3v then i found a wiring diagram for my gecko 202 at cnc zone and it says the p.s should be between 24 - 80 volt at 5 amps since my motors are 5.5 will that make a difference .And is this wirring diagram a accurate diagram for my motors. http://www.cnczone.com/forums/attachment.php?attachmentid=19401[/quote]


Check your maths, 2.3v times 20 is 46 volts, not 22.3v and your amperage needs to be around 11 to 12 amps if you are running 3 motors or 15 if you are running a 4th axis.

Sorry can't see the wiring diagram as although I'm logged on CNCZone says I'm not.

Sophiedoc
07-24-2006, 09:53 AM
I have a Bridgeport clone to which I would like to add at least an x axis drive or Cnc control.No one recently at least has discussed the coupling mechanism.Are the motors coupled with cnc control connected the same was as the x axis electical feed with large brackets etc.Sorry for being naieve about this but I would at least like to know how this works.Used to have electronics as a hobby and understand that part of the equation pretty well.If I convert to Cnc what size motors for x and y are needed for this machine with an 8x 29" table?

BobWarfield
07-27-2006, 05:59 PM
I hate to resuscitate this rather exciting exchange, but I did note that Mariss has decided to incorporate a TVS directly into his updated driver design:

http://finance.groups.yahoo.com/group/geckodrive/message/9048

His products are extremely well made and he will test the thing to destruction to make sure it works properly. At any rate, it won't be long before the debate about whether to zener the back-emf is taken care of for us by the drivers, at least if you are using Geckos.

Best,

BW