0 - 90+ volt variable DC power supply

[webbch]

Everyone needs a variable power supply for bench testing of your motors right? Well, a lot of the motors out there seem to run on 12 & 24V, which is no problem, but what about the ones that take 90V? Must you make/buy a separate power supply for each one?

My goal right now is to run some 24V brushed DC motors (Yaskawa Minertia R-series servo motors). The data sheet (for the 50 oz-in one) specs the stall current at ~28 amps IIRC, which would dictate my overall power rating I would think. However, if some higher voltage (i.e 90 V) DC motors come along at a later date, it would nice to be able to run them off this same supply as well.

I've considered going the PWM route, with 90VDC as the rail voltage, but have concerns about such a large square wave pulse if used with lower voltage motors, even if the duty cycle provided the appropriate "RMS" voltage. And if I used a 24 V rail voltage, then it wouldn't be capable of driving a 90 V DC motor I wouldn't think.

I was thinking of making a linear power supply to have the capability of running any range of voltages between ~0 and 90+volts DC. Anybody do this before? The vast majority of one's I've seen out there go ~1.2-36 VDC (probably for good reason, I'm just not experienced enough know why)

After trying to find some multi-tapped transformers that would provide the appropriate levels of AC, I started thinking about using a variac to supply 0-120 VAC, followed by a diode bridge rectifier of appropriate maximum rating, and some fairly large filter caps dictated by my max voltage (rectified 120VAC). This would provide an unregulated source of DC. I haven't thought through how to regulate anything above 36V yet, but.....

I was considering using an LM317 voltage regulator (~1.2V - 36V IIRC) that would be switched into the circuit for cases where I was needing voltages below 36V, to provide much better regulation without a lot of fuss. I haven't checked the current capacity of the LM317 yet, so it may not be up to the task. Perhaps if I lower the current requirement to ~3-5 amps @ 24V (since I won't be using this to power the system, just want to bench test the RPM/volt spec on the encoders, probably under minimal load), this would become more practical?

I haven't thought through cases higher than 36V yet -- would the unregulated supply have a LOT of voltage droop under load at, say, 90V?

Again, I haven't considered all the details of this design yet, I was just wondering if you folks tend to draw a line between 0-24 VDC and the "higher-voltage" motors and get different power supplies for these two cases.

Chad

[H500]

A linear reg is too inefficient. Why not use a switching supply? A discarded PC supply will have most of the parts, if you're ambitious.

[jerber]

The lm5007 is a 80V regulator, www.national.com/pf/LM/LM5007.html, but it's only 700 mA.
Maybe you can put a power transistor over it. I know you can do this with the 78xx series.

I also saw somewhere on www.headwize.com (or another forum about tube amps) some guys using a regular lm317 with lots of diodes between the adjust pin and the ground.
Every voltage drop over each diode equals the raise in voltage of the output.
But again this is a couple of 100's of mA max.

You also could make a giant batterypack out 1.5 V NiCd Cells.
60 of them will give you 90 Volt in steps of 1.5 V.
(And 60 switches, more then airwolf )

[Evodyne]

I'd forget about switching in the regulator-just add a second, regulated output (or outputs) and leave the high voltage unregulated. Really, if you think about it, the only real purpose of the regulated is to feed electronics. So what, maybe 5 to 24 VDC at one amp?

I'm doing the variac (10 amp) through a suitable bridge and 400 VDC caps. The goal is to be able to test various servo motors I come across. The one thing I read was that the ouput of the variacs isn't isolated from the line and that can present a hazard for electrocution. The advice was to use a 1:1 isolation transformer ahead of the variac.

Lance

[eman5oh]

If you use a vaiac make for sure that it is a isolated type variac or you will let the magic smoke out of some expensive electroics. I speak from experiance, smoked a stepper driver and a motherboard in the computer.

[webbch]

Really, if you think about it, the only real purpose of the regulated is to feed electronics

I was actually thinking of using it to feed the motors, but maybe that isn't necessary.

If you use a vaiac make for sure that it is a isolated type variac or you will let the magic smoke out of some expensive electroics. I speak from experiance, smoked a stepper driver and a motherboard in the computer.

I don't see how the isolated variac will save any of your electronics. I see it as a safety net between the variac and the wall outlet. I was originally thinking the isolation transformer went between the wall outlet and the variac, but perhaps it goes between the variac and the electronics??

A linear reg is too inefficient. Why not use a switching supply? A discarded PC supply will have most of the parts, if you're ambitious.

Maybe I'm not creative enough, but the PC power supplies I have seen have a 5V and 12V output -- perhaps you're suggesting a boost converter on the end of one of these. I don't know much about boost converters, but an interesting suggestion. However, I'm trying to avoid making this a bigger project than it needs be. I've never built a SMPS before, and although it would be interesting, perhaps another project for another day -- this (Linear supply) is more likely to be quick and easy and entirely sufficient, with the added benefit of not needing to worry quite so much about the effect that braking on the motor might have on the electronics. Since it will only be used for bench testing where the critical aspect is variable voltage to encompass a very wide range of motor testing, I thought a SMPS to be somewhat overkill. Yeah, it's inefficient perhaps, but it's not like I'll be using a voltage divider circuit to vary my voltage, nor will the supply be run continuously for long periods of time, negating any efficiency savings IMO. I'll save the SMPS for times when I want just a constant voltage -- or a source where I can interface a PWM (see my concern about PWM in this case in my first post).

So perhaps voltage regulation isn't quite as critical as I originally thought. If I go the all-linear route, I don't think there will be any electronics on the board that will need to be powered from 5V, eliminating the need to figure out how to obtain 5V from the variac, or use a separate ac adapter.

Chad

[Manuel Bernal]

Now I'm using for my L297-L298 based stepper controller, a PC power supply. The 5VDC for electronics, and the 12VDC connected in series with two very cheap ($3.00 USD) transformers, these are used for home halogen lamps, they are 127VAC to 12VAC 50 watts , of course each has a bridge rectifier and a filter capacitor at the output. This way I have 12, 24 and 36 VDC, and I´m sure that you can do this up to 4 times more. May be this is not an ortodox solution, but it works very well with my 3.6 V 1,5 Amp, 200 steps/rev stepper motors.

[Torchhead]

It may be overkill for what you want since it also features soft start, short circuit protection and other features but checkout the specs on our new ACM-01 module. It will handle 18V to 100VDC (regulated) at up to 2KW in a package a fraction the size of a variac. It needs an isolation transformer and controls the primary AC voltage. It's intended market is the big machine retro fits where big motors and lots-o-power are needed.
http://www.CanCNC.com. Contact me off list for more details and pricing.

[webbch]

Posted by Manuel Bernal: This way I have 12, 24 and 36 VDC, and I´m sure that you can do this up to 4 times more. May be this is not an ortodox solution, but it works very well with my 3.6 V 1,5 Amp, 200 steps/rev stepper motors.

My purpose is not to have several discrete voltages, but a continuously variable voltage output between 0 and 90V. Sorry if I didn't make that clear. Many times one (particularly those not extremely well versed in the art of electronics) wants to slowly be able to "ramp-up" the voltage to avoid a smoke-out of the electrical components

Torchead,

Yeah, that would probably be overkill for my application -- With my ebay and digikey shopping, I'm currently looking at about $80-120 tops for the variac/rectifier setup. Size is not a huge concern with the variac.

Chad

[OCNC]

I recently purchased two power supplies on Ebay. One is a 24v 6amp switcher from General Signal (Solas) for $11.00 and the other is a variable linear supply from Raytheon (Sorensen) that goes from 0-80v , 2.0a for $40.00. Not the capacity that you're looking for but Ebay seems to be a good source for this kind of gear. I also have a Woodmaster planer in the shop that uses a 90vDC gear motor for the feed drive. It has a continuously variable speed conrol that works quite well but I don't know the motors current draw off hand.

[Torchhead]

The supply is continuously variable from approx 18 to 100VDC at 20A primary current using a single pot but does not go all the way to zero. The auto transfomer is the cheapest way to go but just be aware that it has no isolation from the AC line so it's like putting a bridge across the AC. The DC - side could, with the wrong phasing end up being hot in relation to true ground (the part your body is at!). You just need to be very mindful of the way you connect the Variac and use the safety ground on the chassis so you don't light yourself up.

I use a big variac to simulate varying line conditions on the bench. My scope chassis has ground floating through an isolation transformer BUT my plugin current probe (I found out after the sparks) is referenced to AC safety ground and grounded my scope when it's plugged in. I sent 10 ^ 6 AMPS through my logic circuit (for 100 usec) that had it's ground referenced to the AC hot side. Fortunately it wasn't through me.

Even when you know what you are doing and being careful it's possible to have a lapse in focus.

You also need to be aware that conventional linear regulated power supplies don't handle inductive loads (like motors) very well and the pass components can be easily destroyed by the inductive spikes or back EMF. Switchers on the other hand tend to get unstable if they happen to be driving a PWM motor control unless they have added caps on the output to deal with the highly no linear current loads.

It's not easy to supply high current and high volts in a 0 - 90 VDC supply. There are circuits available to build PWM DC motor controls and several sold commercially and many show up on Ebay. I have one on my 120 Reliance Servo's on my big mill (80% converted to CNC). They just rectify and PWM the AC line but there is no isolation and yu have to be careful if you plan on using tested equipment!

[gar]

051005-0928 EST USA

webbch:

I suggest that you use a Sola constant voltage transformer to feed your Variac. As well as voltage regulation this will also provide electrical isolation from the ac line so you can ground one side of your bridge rectifier.

A Variac will not give you fine resolution, but is a cheap way to get reasonable power and continuous adjustment from zero to max.

You do not need an output filter capacitor. In this case at 120 v ac rms sine wave input your average dc output from a full-wave bridge is (0.636/0.707)*120 = 108 v minus two diode drops. A half-wave rectifier would give you 54 v minus one diode drop. If you have an output capacitor that is large enough that ripple is small, then the dc output voltage with said input is 1.414*120 = 170 v. A half-wave rectifier and adequate capacitor will give the same voltage as the full-wave bridge rectifier, but one less diode drop.

For your lower voltage motors I would suggest a step down transformer on the Variac output.

.