View Full Version : NEW MEMBER: Hi from South Africa, a little help please

18-03-2014, 08:21 PM
Hi there all CNC'rs,
Coming from a technical background due to my father, I was always interested in automatic/computer control for machines.
My interest is actually the electronic part of the projects.
I have semi-finished a table top cnc, to be used as an engraver in my friend's jewellery shop.
The driver boards were home made using a 4070 Ex-Or ic, 4027 Flip-Flop and L298 H-bridge driver
I have tested movements with 4 wire stepper motors running at 12 volt/0.3A, recovered from old plotters.
The results look ok but now I have to move on!
In South Africa I can get hold of Nema17-3.7V/1.68Amp stepper motors.
I am planning to use a beefed-up 5 volt supply but with current limit, set at approx 1.2 amp.
My question, if some-one can perhaps help is the following;
Will the L298 H-bridge be able to except a supply voltage of 5 volt/1.2Amp and drive the 3.7 volt stepper.

I don't want to buy the low voltage steppers and discover that it does not work.

Thank you forum group in advance


Lee Roberts
20-03-2014, 05:54 PM
Hi Danie,

Just to let you know I've moved your post to a new thread for you, hopfully others will be along shortly to help and say hi to you. Welcome to the forum !


20-03-2014, 06:22 PM
Something that I didn't find obvious when I first started looking at stepper motors is that their voltage rating is actually pretty much irrelevant. That's why, for example, my machine uses a 68V power supply to drive nominal 4V or so motors! What you are looking for is rapid motion and adequate torque, and in practice this comes from current. You are aiming to push something like the rated current through the motor during the whole of a drive pulse, and given their inductance that means a high voltage via a current-limited driver. If you look at the specs of a typical stepper driver, you will see that it has the ability to set max drive current but output voltage isn't mentioned. My concern with what you have built using a 5V power supply is that you won't have the current drive capability, that is, the ability to drive the rated current through the motor during the whole of a short pulse. You will be able to generate enough current to hold the motor in position but not enough to step it fast under load. I've never looked for published stepper driver designs but maybe that would be a source of more design info?

[edit] Just been taking a look at the L298 datasheet. Looks as if it can accept something like a 36V supply, and at least a couple of amps of motor current. That would probably be fine for the Nema17 motors in your machine. My 3D printer uses (from memory) Nema17 running off 12V but that's a pretty undemanding application in terms of speed and torque. The L298 actually has two supply voltage pins, one for the logic circuitry and one for the switched motor supply, so it's designed to work in that mode. (Hope I haven't misunderstood your original question and just told you a lot of things that you know already...)

20-03-2014, 11:43 PM
Neale beat me to it, but he's said pretty much what I was going to say. Those NEMA 17 motors have precious little torque anyway so before buying anything you might want to do the calcs to see if they are suitable.

21-03-2014, 12:49 AM
Dito Every thing Neale and Irving said but Also the L298 36V is Maximum and in reality if you push them much over 30V they don't like it and won't last very long, that's why most use 24V with these chips and play safe.! . . Even then they can go pop if you so much as look at them in a funny way.!!

21-03-2014, 06:59 AM
Dito Every thing Neale and Irving said but Also the L298 36V is Maximum and in reality if you push them much over 30V they don't like it and won't last very long, that's why most use 24V with these chips and play safe.! . . Even then they can go pop if you so much as look at them in a funny way.!!

Jazz is right; the early L298 were 36v and had a habit of going pop and were overtaken by better devices using MOSFETs, the L298 is a bipolar device. However the later ones were updated to 50v so were quite happy on 36v. In their day (mid 1990's) they were good little drivers, indeed I used to run them at 42v, but check the datasheet for your actual device/manufacturer and not a generic datasheet, not all manufacturers updated their devices as newer more integrated chips came along.

However Dan, without wanting to be overcritical, you need to learn how to read the nuances of a datasheet.

As already said, one of the main issues with the L298 is that it's bipolar; the sink and source saturation voltage of the output transistors totals 3.2v max @ 1A so on a 5v motor supply the steppers will be lucky to see 2v - you'll get no useful work out of them. You need at least 12 - 24v to be useful and then you need to limit the current using a sense resistor and chopper circuit (or use the L297 driver).

The other main issue is heatsinking. You get about 5v drop @ 2A on each bridge. For two windings that's 20W of heat generated in the chip to lose through a heatsink. Since the junction to case thermal resistance is 3℃/W you need a heatsink of better than 1.6℃/W to stop it frying. That's a fairly big heatsink, about 300mm square of 3mm alloy plate; I think the ones I used to use @ 1.4A were 120mm x 80mm with 20mm deep fins and they got pretty hot to touch, about 60℃.

Don't let this put you off Dan, I applaud your desire to experiment and try to achieve things with limited resources but you have a lot of learning to do before you're ready to spend money

21-03-2014, 09:45 AM
If you are going to replace the power supply to give a bit more grunt, then it's generally best to avoid switch-mode psu's. Use a simple linear design - traditional transformer/rectifier/capacitor. What you need is the peak current capability that the big smoothing cap gives (which may be over the steady current rating of the transformer) rather than the stabilised output of the switch-mode psu which can't handle pulse current demands very well. Watch output voltage, though, as combination of low-load output plus back emf of motors can make output transistors go pop.
Not many people build their own electronics - most people seem to stick to the mechanical bits - so good luck with it.

22-03-2014, 06:37 AM
Thanks Lee for the moving of the post and thanks to Neale, irving and Jazzcnc for information.
Yes I realise that I am still on a learning curve but with the response received I am willing.
Up to now it was a tremendous experience, making all the components for a cnc engraver.
Not only the making of components but a human interface that it created. Free advise, free scrap aluminium and free assistance.

The volt/ohms value of the steppers are so low so I thought that keeping the supply low will assist in limiting the wattage. I presume the stepper values are kept low to keep inductance low.
What seems to come out of the discussion is that higher voltage, ordinary transformer power supply and current limiting will be the answer. (thanks Neale for the SMPS tip)
My power supply is not built yet so any voltage is possible. I am going to start off with an existing, 20Volt/5Amp, CC-CV, bench power supply.
My home made driver PCB does not have built-in current limiting, hmmm, I should incorporate it in the next version. I do have fly-back diodes on the motor leads for the back emf. BTW I am using PCB SPRINT Layout software, available from Abacus in Germany.
Again thanks for all the replies, much appreciated.
I will post the progress.

22-03-2014, 05:35 PM
Here is a picture of my L297/L298 drivers from a few years back, see the size of heatsinks for 1.4A operation


Sorry Dan, but you're on a hiding to nothing, a current limited supply simply won't work. With the L298 you must use the sense resistors/chopper circuit on each winding. A stepper motor works by phasing current through 2 windings. At any point in time current is flowing in none, one or both windings... to get constant torque and accurate stepping you have to current limit on a per winding basis.

The next 3 pics shows the voltage across the sense resistor (a proxy for the current in the winding) at different step rates for a fairly high inductance motor running on 36v supply and 1.4A. The inductance of the winding is what limits the rate of current change so you have to drive the winding hard with a high voltage to get enough current flow to get useful torque at any useful step rate, but then you have to limit the current to avoid frying the winding. If memory serves, these are at 75, 112 and 133rpm. You can see the initial slow current build up followed by the chopper action. In the last pic the current limit is never reached - above this speed the stepper will generate less and less torque (in fact by 300rpm it hasn't enough torque to overcome its own friction and stalls). This motor should ideally be running on 100v to be useful above 100rpm, and this is a 1.5Nm NEMA23 motor with a 1.4A/2.5v winding.


23-03-2014, 08:59 AM

Good to hear from you, hope you are well.

Thanks again for the information and pictures. The heat sinks are quite big I see.
As I am experimenting I don't mind using a lot of aluminium at this stage. (junkbox full !)

I took notice of the fact that current limiting should happen on a per winding basis.
On My circuit board I have allowed for the 2 current limiting resistors for the L298.
It's not clear in your picture whether your boards have them fitted or not.
Will the L298 perform the necessary current limiting with those resistors or should there be an extra current limiting circuit.?

23-03-2014, 06:19 PM
The resistors are on the boards but they don't do current limiting, they are sense resistors that provide a voltage proportional to the winding current, typically 0.5v per amp. This is used to drive comparators that turn off the drive for a short period of time, the so-called 'chopper' action. This is done most easily with the L297 companion chip or a dual op-amp, a dual flipflop and a clock pulse e.g. 555 timer. There are several examples on the web.

There are other ways to do the necessary current limiting but not with the L298 chip.

(thanks for the PM btw, appreciated).

25-03-2014, 07:38 PM
Hi Irving,

Thanks for the above info. The penny has dropped !
I am working on a PCB using the L297 and L298, which is quite inexpensive in south Africa.
The size of the desktop engraver is only 400mm x 400mm with a Z movement of around 50mm.
I will first do a lot of testing and measuring before final set-up.
The next engraver is already planned. For that I will use more conventional drivers, maybe from the UK.
(the Rand Pound exchange rate is now around 18 / 1.)
Your remarks and positive information is very much appreciated.