Step down adjustable power supply

[Michael GilliganParticipant @michaelgilligan61133]

Posted by Muzzer on 15/08/2014 22:34:19:

The LM2596 actually is a PWM buck converter, so it wouldn't actually be a bad choice for controlling a small motor. The switching frequency is 150kHz (hence the tiny inductor which is the small grey cube). The main issue here is that the max output of 3A sounds rather puny.

.

Very true, Murray … But, based on my limited experiments with them, even the 3A is optimistic without additional heatsinking … and the ebay LM2596 boards don't really lend themselves to bolt-ons.

They are great for about 1A, but I'm nervous thereafter.

MichaelG.

[Anonymous]

Many moons ago when the NatSemi single chip buck converters came out they were quite a step forward compared to the old control chips where you had to add external diodes, transistors, compensation loop components and much more. If I recall correctly they started at 50kHz switching, then 100kHz and finally 150kHz. However, I found that they never quite reached the claimed output currents. In the past I've used them a lot, but these days they're a bit long in the tooth, and are physically quite large. The slow switching speeds also mean large inductors. I have a recollection that they weren't too happy with ceramic filter capacitors either due to the low ESR.

These days I prefer the TI TPS 'Swift' family. They're smaller devices, faster switching (up to and beyond 1MHz) and the designs one can obtain from the TI Webench software actually work and provide the stated output current. So much so that I don't bother prototyping bogstandard buck converters using them as I have confidence that they're just going to work as advertised; always assuming of course that one doesn't implement a dumb PCB layout.

Regards,

Andrew

[Enough!Participant @enough]

If you get :AAQ:CA:1123″>this one, it comes with a readout display. Unsolder the trimmer and wire in a panel mounting pot. Put the whole thing in a little box for a power supply you can make with any wall-wart.

[Ian PParticipant @ianp]

Posted by Muzzer on 15/08/2014 22:34:19:

A variable voltage DC supply is not the best way of driving a brushed motor, far better for your purpose in this case is a variable pulse width controller. eBay has lots of them

The LM2596 actually is a PWM buck converter, so it wouldn't actually be a bad choice for controlling a small motor. The switching frequency is 150kHz (hence the tiny inductor which is the small grey cube). The main issue here is that the max output of 3A sounds rather puny.

On the other hand, the 8A units are rather crude in comparison – they don't even have any form of current limiting or voltage regulation. The lack of current limiting may not matter much if you can avoid wiring shorts and the speed regulation will be pretty poor when driving a brushed motor anyway, even if you had a regulated voltage. As suggested, you'd want to fit a flywheel diode for driving a motor with one of these. Even better get a proper motor speed controller….

Murray

Murray

The first paragraph of your post is a 'quote' from my earlier one. However what I should have said that for motor control a 'PWM Motor Speed Controller' was best.

A PWM motor speed controller is a different animal to the item you linked to, and to some of the others mentioned in on posts on this topic.

Pulse width modulated circuits are used for numerous purposes, DC power supplies and motor speed regulation are just two of them. The various step up, step down regulators on eBay are designed to produce a regulated DC voltage which is definitely not the optimum of driving a DC motor (maybe OK if its fixed speed and load).

Proper PWM motor speed controllers, apart from having much better speed control will also drive the motor more efficiently and allow it to produce greater torque at a lower speed, they will also incorporate protection circuitry so should not need any additional diodes.

Ian P

As has been mentioned, an electric screwdriver is not really suitable for continuous running although Mick has not given many details about what his machine is going to be used for, if its going to be filing Balsa wood or for watch making it will probably be fine.

[Ian PParticipant @ianp]

Posted by Muzzer on 15/08/2014 22:34:19:

A variable voltage DC supply is not the best way of driving a brushed motor, far better for your purpose in this case is a variable pulse width controller. eBay has lots of them

The LM2596 actually is a PWM buck converter, so it wouldn't actually be a bad choice for controlling a small motor. The switching frequency is 150kHz (hence the tiny inductor which is the small grey cube). The main issue here is that the max output of 3A sounds rather puny.

On the other hand, the 8A units are rather crude in comparison – they don't even have any form of current limiting or voltage regulation. The lack of current limiting may not matter much if you can avoid wiring shorts and the speed regulation will be pretty poor when driving a brushed motor anyway, even if you had a regulated voltage. As suggested, you'd want to fit a flywheel diode for driving a motor with one of these. Even better get a proper motor speed controller….

Murray

Murray

The first paragraph of your post is a 'quote' from my earlier one. However what I should have said that for motor control a 'PWM Motor Speed Controller' was best.

A PWM motor speed controller is a different animal to the item you linked to, and to some of the others mentioned in on posts on this topic.

Pulse width modulated circuits are used for numerous purposes, DC power supplies and motor speed regulation are just two of them. The various step up, step down regulators on eBay are designed to produce a regulated DC voltage which is definitely not the optimum of driving a DC motor (maybe OK if its fixed speed and load).

Proper PWM motor speed controllers, apart from having much better speed control will also drive the motor more efficiently and allow it to produce greater torque at a lower speed, they will also incorporate protection circuitry so should not need any additional diodes.

Ian P

As has been mentioned, an electric screwdriver is not really suitable for continuous running although Mick has not given many details about what his machine is going to be used for, if its going to be filing Balsa wood or for watch making it will probably be fine.

[michael howarth 1Participant @michaelhowarth1]

Ian …..The filing machine that it is to be used in is only a small affair which uses needle files. I have made it mainly to assist in making valve gear parts for Gauge 1 and similar scale (1/2 – 1" models.

Mick

How did that smiley get there??? Can't seem to get rid of it.

Edited By mick H on 16/08/2014 19:42:36

Edited By mick H on 16/08/2014 19:43:32

[Ian PParticipant @ianp]

Mick

The self inserting smiley is one of the more charming features of this website's editor. Basically it gets the hump with certain combinations of characters!

Knowing what the motor is now going to be driving I would say that an ex-screwdriver should be OK. I would still recommend driving it with a DC motor speed controller like the one I suggested in an earlier reply.

Ian P

[MuzzerParticipant @muzzer]

Ian, I'm not quite clear what you believe a PWM motor controller to be other than a voltage regulator, given that we aren't talking about any form of speed or position feedback. Without such feedback, the torque-speed characteristic inherent to the motor is fairly poor and can't easily be improved beyond voltage regulation. It'll be fine for driving a filing machine.

BTW, the datasheet I linked to is simply the PWM controller IC that is used in the original ebay purchase. For a brushed DC motor, you can't really drive or control it any more efficiently than this. The controller is already using PWM drive and there is also a flywheel diode in there, as well as various protection functions.

Murray

[Neil WyattModerator @neilwyatt]

PWM gives better torque/speed characteristics, as you get lots of tiny pulses of full torque.That's why electric model railways run far smoother with PWM than the old 'rheostats'.

Neil

[Michael GilliganParticipant @michaelgilligan61133]

Posted by Neil Wyatt on 17/08/2014 09:57:14:

PWM gives better torque/speed characteristics, as you get lots of tiny pulses of full torque.That's why electric model railways run far smoother with PWM than the old 'rheostats'.

.

Neil,

This begs an interesting question [which I have been pondering for a while] …

It is, as you point out, an established fact that motors run well on a PWM supply … but, that was well-demonstrated when the pulse repetition frequency was comparatively low [circa 100Hz ?]

Now that repetition frequencies [for general purpose power supplies] are up in the KiloHertz range; does this make them less effective at driving a motor ??

I suspect that the Inductive nature of the motor will prevent it from responding to these very rapid pulses …

There is probably an optimum pulse repetion frequency for any particular motor; but I don't know how to calculate it.

Any thoughts ?

MichaelG.

[Neil WyattModerator @neilwyatt]

A good question. The carrier frequency for my inverter can be set between 0.75 and 15 KHz. The default is 15Khz, they note this causes more heating, but that torque needs to be derated to 75% at 0.75KHz.

As to calculating the optimum it will depend on duty. I would assume that you would need to know the inductance of the motor and a place to start could be the frequency when the reactance rises to a comparable level to the DC resistance of the windings?

Neil

[Michael GilliganParticipant @michaelgilligan61133]

Posted by Neil Wyatt on 17/08/2014 11:00:09:

The carrier frequency for my inverter can be set between 0.75 and 15 KHz. The default is 15Khz, they note this causes more heating, but that torque needs to be derated to 75% at 0.75KHz.

.

Neil,

That's an interesting observation … but it's probably worth noting that the inverter is synthesizing a lower frequency [approximation to a] sinewave; so the situation is not quite the same as "DC" PWM.

All very interesting stuff, this.

MichaelG.

[Neil WyattModerator @neilwyatt]

Michael,

I understand the reason they call it a 'carrier frequency' is that the sine waves are generated by pwm – in effect the inverter is three 'class D' amplifiers 120 degrees out of phase.

Neil

[Michael GilliganParticipant @michaelgilligan61133]

Posted by Neil Wyatt on 17/08/2014 12:33:59:

Michael,

I understand the reason they call it a 'carrier frequency' is that the sine waves are generated by pwm – in effect the inverter is three 'class D' amplifiers 120 degrees out of phase.

Neil

.

Neil,

Assuming that things have not changed out of all recognition since I was involved in Vibration Testing; I think you will find that the "sinewave" is synthesized by adding a set of squarewaves [Fourier synthesis] … which is not really Pulse Width Modulation.

Your inverter [Jaguar Cub?] will be differently rated at different carrier frequencies because the shape of the "sinewave" becomes more, or less, accurate. … The more jagged approximation has a residue of "square" rising edges, which [by Fourier analysis] create very high frequencies.

You have probably blown a tweeter or two in your time, by overloading an amplifier to the point of severe clipping … it's a similar effect.

MichaelG.

… Sorry, this is getting some way off topic; so best to stop there.

[Neil WyattModerator @neilwyatt]

> Assuming that things have not changed out of all recognition

An assumption I would not make!

I think that change is exactly why a modern inverter can fit in a box little larger than a packet of tea.

Neil

[Michael GilliganParticipant @michaelgilligan61133]

Fair enough, Neil

Yes … I see they do it differently in the inverters

[Thanks for making me look that up … learn something new.]

… But what's happening there is still not particularly relevant to the type of simple PWM controller typically used for a "model railway" type DC motor [which is where we came in]

 

PWM gives better torque/speed characteristics, as you get lots of tiny pulses of full torque.That's why electric model railways run far smoother with PWM than the old 'rheostats'.

 

Neil.

MichaelG.

 

Sorry about the white stripe … I can't change it

Edited By Michael Gilligan on 17/08/2014 15:36:09

Edited By Michael Gilligan on 17/08/2014 15:37:09

[Neil WyattModerator @neilwyatt]

My assumption is that the effects of winding inductance would be similar (apparently in broad terms inductance is greater with smaller motors, so the effects in small DC motors will be more marked).

The traces near the bottom of this page are interesting – and relevant!

**LINK**

At high frequency the inductance of the motor smooths the PWM into 'virtual DC'.

Neil

[Michael GilliganParticipant @michaelgilligan61133]

Posted by Neil Wyatt on 17/08/2014 15:49:40:

The traces near the bottom of this page are interesting – and relevant!

**LINK**

At high frequency the inductance of the motor smooths the PWM into 'virtual DC'.

Neil

.

Thanks Neil

That link looks useful

Your closing comment is, of course, why I raised the matter in the first place.

… 'virtual DC' probably does not give the desireable pulses of torque that make simple PWM controllers so effective on small motors.

MichaelG.

 

 

Edited By Michael Gilligan on 17/08/2014 16:11:22

[Neil WyattModerator @neilwyatt]

I must admit my assumption was that at too high a frequency the motor would show too great a reactance and not pass enough current.

The peak power at the low frequency is 32 watts, at the high frequency just 13, and a much briefer spike.

It's interesting he interprets 500Hz as being too low a frequency, yet roughly integrating the two curves by eye, the motor in the lower trace is using much less power and producing no back EMF between pulses suggesting it is running much slower?

Can I be bothered to replicate the experiment….

Neil

[Michael GilliganParticipant @michaelgilligan61133]

Posted by Neil Wyatt on 17/08/2014 16:31:37:

Can I be bothered to replicate the experiment….

Neil

.

Go on … you owe it to your readers

MichaelG.

Edited By Michael Gilligan on 17/08/2014 16:56:23

[Anonymous]

Having been involved in the design of a number of inverters, albeit of rather higher power than being discussed here (~100kW), I can say that the outputs are definitely PWM, not Fourier synthesis.

Of course the output stages for driving a DC motor and an induction motor are different, as are the timing of the PWM. For a given setting for a DC motor the PWM value stays constant and hence the pulse width stays constant from one cycle to the next. For an induction motor operating at a constant frequency the PWM value changes on every cycle in order to synthesis the sine wave.

Quantization noise on DACs usually manifests itself as a noise like signal. The level of the noise is determined both by the number of levels, ie,the resolution of the DAC, and by the sample rate compared to the signal being synthesised. As part of my Ph.D I derived a similar analysis for quantization errors in time sampled signals, ie, PWM, that showed that the error spectrum is essentially constant amplitude noise.

It is interesting to note that the standard use of a DAC to output a series of pulses, the width of which is determined by the sampling period, means that output signals cannot be accurately reconstructed up to the Nyquist limit.

I'll leave it as an exercise for the reader to work out why.

Regards,

Andrew

Hint: Think sinc

[Michael GilliganParticipant @michaelgilligan61133]

This thread has certainly moved on a bit, from Mick wanting a way to drive his salvaged Black & Decker Screwdriver motor.

Hope you found an answer somewhere amongst it all, Mick.

MichaelG.

[Anonymous]

That's progress for you – Andrew

[Ian S CParticipant @iansc]

If you get the right battery drill(it will be higher voltage than the screwdriver), it might have a speed controller built in. I find drills with U/S batteries at garage sales for a dollar or two.

Ian S C

[michael howarth 1Participant @michaelhowarth1]

MichaelG…yes, still working on it but just got back from a lengthy trip to see a v. sick relative….and before that my multimeter ammeter in the appropriate range went on the blink…so a bit of catching up to do.

Mick

[Author]

Posts