Alternative power supplies and voltages

I've been poking around power supplies for a while and have been seeing what I can get for amperages. A lot of higher torque DC motors pull a LOT of amps for 12V models (60A stall for the ones I've found new so far, but they also top out at over 20000rpm, too), and I think "low cost gearbox" is an oxymoron. I'm waiting on testing of the strength of RepRap'ed gears before final judgement on that, though. I've been looking for either higher voltage models or alternate power supplies. Also all the brushless DC motors I'm looking at for future use start at 24V and go up to 48V. This is a little of what I've found:

ATX Power Supply: the RepRap gold standard. A "good" one can actually be rated to 25A on the +12V supply. This can be as small as 350W and cost $30 new. Unfortunately, if you want to try and split phase it (+12 and -12 to get a 24V differential) you're only going to get an amp. Not Good. These also will not like you if you attempt to go beyond their rated capacities. They'll either fail or fold back and complain at 105% normally. Also may not like being paralleled on a case by case basis (harmonics, load balancing)

Industrial power supply: costs more, no 5V or 3.3V to leech off of for different models usually. Advantage, probably stabler, but still not recommended for brush motors at least (noise = instability = fault). I've been looking at a MeanWell 600W 24V supply (25A) to use for brushless motor power. This supply is carried by a few case mod sites for a lot less than industry, lowest price being about $140.

Transformer - rectifier - capacitor: toss in a soft start resistor/contactor and you're well on your way to making your own self-contained frequency drive. I've seen a surplus 12V 10A transformer for $10 at my local surplus store. It was even an isolation transformer. This should mean that I can, carefully, parallel the inputs and outputs to generate higher amp blocks. I can series the outputs to generate 12, 24, 36, and 48V. With a slow blow or circuit breaker, they can also surge several times their rated current for short periods. Disadvantages: lower efficiency power wise, not regulated so I have to watch my overvoltage trips (although this might be easy enough to fix). Plus I need to remember what DC voltage I get when I recitfy and cap-smooth a 12VAC signal.

I believe that all current RepRaps run off of 12V in one way or another, mostly computer PSU's. Does anyone else use something different on a regular basis? My personal power interests all all over the board (due to different motor requirements). I know I've seen interest once or twice (probably Zach) in moving Mendel along a brushless DC servo system, but I doubt that will be feasible if 12V stays the standard.

... i build a power-supply with 8 parallel LM317T and shunt-resistors for balancing the differences (as in an old project from a electronic magazine).

With this i drived a laser-plotter with two 2Amps-per-phase-steppers an the supply of a RF-CO2-laser.

The power-supply is regulated and can output from 2 to 28 Volts at currents until 10 Amps (max 1,5 Amps per LM317T, so 12 Amps would be the abs. max for 8 LM's).

But with more LM's i can drive more current too.

If you want, i'l try to reenginering the wiring and the part-list ...

For supplying the electronics i simply insert dc-dc-converter-ic's in the main (as for 5 or 3,3 Volts).

Viktor

I've tried connecting two ATX power supplies to get 24V, just to see if it would work. It did, and it had no problem driving a few motors. This is probably a really bad idea for some reason I don't understand, but I have a few dozen of them lying around so I figured it wouldn't be any great loss if they went up in smoke. I only played with it for an hour or two before I put it aside.

As for brushless motors, I know a lot of model helicopters & planes use battery packs in the 12V range to drive their motors. Aside from those and the motors in hard drives and CD players (also 12V), those are the only brushless motors I've paid any attention to. I don't know what sort of torque they can produce or how much a RepRap would need, but those motors can do some damage given the right gearing.

Viktor: Yeah, I figured that may be the only way to clean up the system

emf: What you did with the ATX supplies may work fine actually. I'll have to tear into one later and see if the grounds are shared. If they are, it could have been you got lucky by not having the supplies grounded to each other. If this can be easily "broken", it would be great.

As for brushless, a lot of the RC brushless motors have very tall speed ranges, but not much torque at low end. They're also designed for 10 minute runs at speed, not continuous operation near stall. Aircraft specific motors especially have been designed for power efficiency and power output at speed. I'll have to look a bit more at the RC car ones. I was holding off from the RC market as it seemed just as expensive as the low cost industrial brushless motors I'm finding, but without the helpful hall sensors you need to run a low speed BLDC effectively. I'm still investigating them, however. I think if I attach a 12 bit absolute encoder I can tune the motion controller close enough to avoid having to do a "wake and shake" at startup. Hmmm...

The grounds on an ATX power supply are connected to the wall plug's ground, so you couldn't plug both into a grounded outlet. I took a pair of pliers and ripped the ground pin off the power cord for one of the power supplies to break that connection, then connected the 12V outputs of the first power supply on a hard disk connector to the grounds on a hard disk connector of the second one. Resist the temptation to stack the two power supplies on top of each other, since the cases are grounded.

RC motors: makes sense. At least for the plane motors, I'm sure you pay a lot for super-lightweight motors and they might assume they'll have air from the propellers cooling them.

Dodgy removing the earth lead from a PC power supply. They usually have Y caps in the mains filter which will cause the case to float at half mains until you ground it. On this side of the pond that is 240/2 = 120V, so not fun to touch!

---

[www.hydraraptor.blogspot.com]

mmm... lunchtime

It would work for testing, but yes I'd not want to touch that system as "production" either. If the grounds can be separated (possibly), then we're safe. I know this is possible if they're using a full bridge topology switching power supply. This would also transformer isolate one side. Anyone know where we can take a look at schematics?

I'm almost tempted to look into building my own modular switcher, but I have neither the time nor money for that right now. From app notes, I think it would be possible (possibly simple and definitely not easy) to make a parallel/serial switching PSU in 500W 12V modular form, so you can stack for volts or parallel for amps.

Hm... I'm trying to electrocute our users, Viktor is trying to blind them, we've already got the burning part working pretty well. Sharp blades, anyone?

I came across this schematic for a computer power supply a while back. Maybe it will mean more to you than it did to me :-)
[www.pavouk.org]

... here i have some images of my '20-years-old-but-still-working-prototype' 2-28V/12Amp-powersupply and the second-generation-pcb for 8xLM317T.

With more LM's parallel i can add the currents - i think with a good heatsink there's {N}x1,3Amps without limit - as much N's, as you insert LM317T and shunts ...

The other images are from my 2xIMT901-finestep-motordrivers and from an MC3479P-prototype ...

Viktor

emf,
That PSU looks like all you need to do to get the same current on the -12V rail as +12V is to fit a bigger cap and make sure the inductor is the same rating. Not sure what the extra diode is for but I expect you can ditch it to make it fully symmetrical.

---

[www.hydraraptor.blogspot.com]

Sharp blades belong to nophead so far. I think a small milling bit in a 30krpm rotary tool works just fine. Or for even more fun, a fiber reinforced cutoff wheel.

There's a lot of discrete parts in that PSU. Figures. Nophead's idea may work, and I think we'd all love it if it does. Anyone have an ATX they want to crack open and look at? I have access to several AT supplies here, so I might take a look inside of them. They should be simpler but related.

To build a stackable system from the schematic would require decoupling the secondary side grounds (not too hard depending on layout) and figuring out what to do with the two feedback lines that violate the isolation barrier (um.... um.... stumped on that one). Otherwise, it may still be good to steal the magnetics and work on that digital PSU design.

I've seen a PSU put together from a guy at OpenServo. Similar to yours, Viktor, but he used one voltage reference and built FET voltage followers to add on capacity. Toasty. Hmmm... power supply and plastic processing in one unit! Who needs Nichrome?

SOI,
I don't think the feedback crosses the isolation, the mains is only on one side of each transformer.

---

[www.hydraraptor.blogspot.com]

Hi SOI,

... my PSU inserted in the laserplotter supported the interface-electronics (PIC16??-CPU), front-controls and IMT901-driver-boards, two steppers with 1Amps-per-phase (in chopper-mode) and a RF-CO2-laser with 5 to 6 Amps at continuous running.

The temperature over some hours continuous working was not more then 40

Hello everyone.

This is an interesting sort of discussion to see on the forums. As always there's a lot of ways to do any particular job.

My particular slant: is a regulated 12 / 24V supply actually mandatory?

Consider what's being powered:
1) Electronics with their own 5V (or 3.3V) regulators
2) Stepper motors
3) Heating element
4) DC motors?

The only thing that doesn't have its own independent controls are the stepper motors. If these were controlled in some manner (e.g. had a constant current driver) then there is no need for precise regulation.

One could simply use a generic mains supply transformer (of suitable rating), a bridge rectifier and some suitable capacitors. There will be some ripple but not a huge amount...

This could also be used to generate split supplies e.g. +/- 12V which means that things could be driven with half bridges instead of H bridges. At this point it becomes much more palatable to 'roll-your-own' bipolar stepper drivers as you only need half the parts.

On the switched mode front, I have tinkered with these before and while you can certainly build one in the garage (a 200W sinewave inverter on veroboard!) they need quite a lot of know - how and an oscilloscope. Plus a lot of potential RepRap people won't be electronics heads and may not be too happy about playing with 350V DC rails (230V * sqrt(2)). With good reason.

Relevant equations for interested parties:

VdcMAX = Vac * sqrt(2) - Vrec
VdcMIN = Vac * sqrt(2) * (100% - regaultion) - Vrec

Vac = transformer output voltage
sqrt(2) = the ratio of the peak of a sine wave to its RMS value
regulation = a transformer property. The transformer output voltage falls as more current is required. 10% is a usuallly good guess.
Vrec = rectifier voltage drop. Ranges from 0.2V (for a single fancy diode) to 2V (for a heavily loaded diode bridge).

In general, a 12V AC transformer will give you 15 - 16V max, falling to 13-14V under load.

I took the cover off a 200W AT pSU I have sitting around. Single sided (as expected) and several safety/noise parts have been removed and replaced with jumpers. Found out that the mounting screws are the only connection of ground to case.

I'm working my own electronics design, with the motors and probably the heater going on a separate power bus than the electronics. The mains transformer is a consideration for me, as I know where to get some 12VAC isolation (I think from meter testing) transformers surplus, but it's not a final decision. I'm also in the US, so unless I build a universal power supply, it'll only be 170V rails

What bugs me about the current RR power system is that I've heard that DC brush motors do not play well with switching power supplies. The noise from the commutator sparks plays havoc with the loop stability. This will become more of an issue if closed loop is implemented at some point with DC gearmotors. A transformer supply for the motors is one option. I'd like to switch over to sensored brushless DC motors instead, but most of those I find that I'd consider buying run on 24-48VDC. We'd also get longer life, too. A transformer power supply may be more expensive, larger, and less efficient, but it'll be simpler and provide true mains isolation. I know it can also surge current well, too. Plus no problems with noise collapsing your voltage control loop.

I did peek at RC aircraft motors. A lot of the small outrunners are getting 7000RPM and pulling 8A while swinging a 9x6 prop. I don't want to think of stall. However, it is feasible to wind your own. I don't have data on RC ground motors, but they may use the same (well, more like standard inrunners that can go 30k to 40krpm). I'd definitely need encoder feedback to get this to work right, though. I'd not have the advantage of carefully aligned hall sensors in phasing.

Edited 1 time(s). Last edit at 10/04/2007 11:53PM by SOI Sentinel.

SOI Sentinel Wrote:
-------------------------------------------------------
>
> What bugs me about the current RR power system is
> that I've heard that DC brush motors do not play
> well with switching power supplies. The noise
> from the commutator sparks plays havoc with the
> loop stability. This will become more of an issue
> if closed loop is implemented at some point with
> DC gearmotors.
>
I haven't had a bit of trouble with this. Mind, I'm running the axes with Solarbotics gearmotors not unlike the ones used on the Mk II extruder (different gear ratio, of course). A simple polyester capacitor across the power lines to the motor from the 754410 seems to do the job.

I'm also running the whole rig off of an ATX power supply with no hassle.

My motors draw much less than an amp each, mind. With larger brushed DC motors I have no doubt that there would be big trouble.

That is exactly what I'm expecting. Plus, as RepRap adds new head capabilities (coughzachandhislaserscough ) I suspect we may start seeing random failures from voltage glitches if run on DC brush systems. Could always split the difference, low cost 7.5 deg/step steppers run in high pole count brushless DC mode with the right controller and an encoder. Thoughts for the future, be it quiet and clean or noisy and dirty.

I had severe noise problems from the GM3 and that was while using its own separate floating PSU. I managed to cure it completely with four caps and two ferrite inductors, details in the builders blog. I think this should be added to the official design, at least the three cap version.

110V switch mode PSUs I have seen still have a 340V rail, they use a voltage doubler on the input. Two diodes and two caps instead of a bridge and one cap. Better ones, with active PFC, step it up even higher with the PFC circuit.

Edited 2 time(s). Last edit at 10/05/2007 09:59AM by nophead.

---

[www.hydraraptor.blogspot.com]

yeah, i like your design, Nophead. i'd like to talk to you a bit more about it and get it added to the new dc motor controller board i'm working on. maybe send me a PM or email?

The suppressor needs to be at the motor end of the wire. Perhaps a tiny PCB is called for but two of the caps have to be soldered to the can at one end anyway. It is very easy just to make it on veroboard.

Edited 1 time(s). Last edit at 10/05/2007 01:16PM by nophead.

---

[www.hydraraptor.blogspot.com]

I think I'm going to have to track down my local ACME rep. We've used them before at work as a good but not excessively expensive transformer supplier. Their 24VAC transformer line goes up to 1kVA. 40A continuous is definitely capable of support anything I think we'll see and this shouldn't exceed your standard recommended single phase loading or outlet loading. They're even 120/240V input. It's also an isolation transformer. With a microprocessor watching the power supply, we could actually PWM a high side FET on the 33V DC bus to let us regulate the power. Does this sound workable?

I also see lighting transformers (with circuit breakers on both sides of the transformer available) that do 12VAC and split 12/24VAC. Also isolation.

(Addendum)

Also of note, anyone want 80A of 12V DC bus available from a single PSU?
[www.newegg.com]

$100

(Addendum 2)

Also found a place selling industrial PSUs in the US. I know the Mean Well units are rebranded by well known companies in my industry (Lutze in particular), so they're stable. Cheap too, although not as much as your standard PC power supply. [www.trcelectronics.com]

Edited 2 time(s). Last edit at 10/06/2007 12:21AM by SOI Sentinel.

Rock, While they need to be tested, It's now apparent that most PC PSUs can be series and physically stacked safely if the case ground is severed.

[www.procooling.com]

Old article, but it should still hold steady.

I'm starting with the old AT PSUs I have laying about to verify this first. Many modern ATX PSUs have multiple 12V rails that can be tied together (parallel, not series). The rails usually average 15A each, although better (or perhaps less honest?) PSUs can push 20A or more. 30+A from a $20 PSU, or 80A from a quality $100 PSU. Not bad, although I'd start stacking $20 units first.

Now I need a motor and driver board...

i have a 12v 1 amp power supply i want to up grade it and turn to higher ,about 12 amp something what should i do? is there a way to make it?

Your best bet is probably to simply buy an ATX powersupply.

Please be very carefull with switch mode supplies at all times avoid running them with covers removed. There is a lot of power stored in eletrolitic cappacitors faults can occur.

About 10-12 years ago a friend of mine was working on a switch mode power supply when an electrolitic capacitor failed the aluminium caseing flew off and hit his temple by his right eye it hit with such force it damaged his retina causing a split. This means he lost depth perception in the center of his visual field.

Ever since he explained what happened to him I have taken extra care with all switch mode devices ensuring that some kind of protection is always between me and the DUT (device under test) / repair. Safety goggles could be another route.

Switch mode supplies ~ inverters all contain a lot of raw kinetic energy.
The covers are not just there to protect you from flyng capacitors they are also there to shield the RFI and the high voltages.

As NOPHEAD mentioned the voltages are also very high and have a tendancy to bite you after the power has been removed, as it remains stored in the electrolitic capacitors.

200-300 volts DC will split the skin open.
During the first year (1971) of my appreticship one of the guys in my group picked up his valve TRF reciver chasis, it was off but fully charged it split the palm of his hand open.

I don't meant to scare any one from experimenting / hacking just be very carefull.

1/. Cover the device at switch on.
2/. Ensure the primary caps are discharged before handling PSU boards.

Another saftey rule is the "one hand rule" :-

"One hand rule. When working on energized circuits one should always keep one hand behind your back or in your pocket. Many times a person will carelessly ground himself with the hand he is not using the increases the chance that he will complete an unintentional circuit. To avoid this we suggest that the one hand rule be followed at all times."

I tend to think of this as the Left hand rule and put my left hand in my pocket, with your heart being on the left I just think it must be a little safer.

Ref.. [www.helium.com]

Edited 2 time(s). Last edit at 06/15/2009 06:42PM by BodgeIt.

---

Bodge It [reprap.org]
=======================================

---

BIQ Sanguinololu SD LCD board	BIQ Stepcon	BIQ Opto Endstop
BIQ Heater Block PCB	BIQ Extruder Peek clamp replacement	BIQ Huxley Seedling
BIQ Sanguinololu mounting	BIQ standalone Sanguinololu or Ramps mounting	Print It Stick It Cut it

My rep strap: [repstrapbertha.blogspot.com]

Buy the bits from B&Q pipestrap [diyrepstrap.blogspot.com]
How to Build a Darwin without any Rep Rap Parts [repstrapdarwin.blogspot.com]
Web Site [www.takeaway3dtech.com]

As someone who've been hit by 50kV from a CRT/flyback cicruit while it was powered off, I can attest both to the effectiveness of the left hand rule, and the cautious attitude towards any high-power circuit, even when it's off. Even though 50kV from a CRT isn't really this dangerous (still, sounds scary - nontechnical people usually go like "ooooh, you must've been lucky you're still alive!") and there's more risk of self-inflicted mechanical injury due to muscle reflexes in the hand, rather than electrocution, it's damn painful. Keep in mind, filter capacitors in a computer PSU can easily be much more deadly than that.

For added protection, use a pair of rubber-coated garden gloves (cheap) or a pair of professional electrician's gloves (expensive, but also more comfortable). They won't hold up to more than a few hundred volts, but will do fine in contact with low-voltage live circuity.