About Stepper Motors

[_q12x_]

Actually there is another forum discussion until reaching this cct here. It was an entire evolution of it.
This cct is working ok but in certain parameters. It is limited to power and heat over the Tr's.
Is working fine in the 5V range for 2 of my stepper motors that I tested with it. But if Im bringing up the voltage, to 10V, I get 140dgrC on one Tr and about 60dgrC on a few others. At 5V, everything stays at a constant 50dgrC which I call it normal and safe.
One of the test motor is a Nema17 17HS4401. The other one is a scrapped stepper. Both work fine at 5V and even at 10V, no one was heating. Over 5V, especially at 10V, only the driver's Tr's were cooked to (almost)death. I stopped the power after smelling hot metal in the air.
I am also receiving a couple of L298N IC's. They are coming in the mail, not arrived yet. But I dont expect much difference from them, compared to what I have here. It would be nice to be better ! I hope so.
- My question:
- How to make this cct more powerful? To drive more power through the coils of the stepper motor? (and not blow up it's Tr's!)
I already have 2 answers to this question in my mind, but I dont want to influence your answer. I expect to be pleasantly surprised. Haha.

Thank you !

 

[_q12x_]

Mister @marconi , good morning to England. A few words about my cct here?

 

[Julie.]

You have a fundamental problem with using BJTs like that, in that the switch off time is longer than the switch on time.

As the lower BJT only needs the voltage to go from 0 to 0.7v or so, whilst the upper one needs the voltage to swing down much further.

So when you drive one transistor on, the other which is on via the inverter will remain on for some time before switching off, resulting in both transistors on in series across the supply for a brief time.

I assume it is the lower one which gets too hot BTW.

You either need a delay between switching one off and the other on, or use FETs/MOSFETS which as long as you drive the gates correctly will avoid this.

EDIT:

You could of course just replace the whole driver circuit with a dual H bridge chip, which control this internally, something like the DRV8841, this is good for diving circ 2.5A in the coils up to 24V or so, which will allow a 5V circuit to drive a higher voltage on the motor, or the one you suggest.

 

[marconi]

Mister @marconi , good morning to England. A few words about my cct here?

I will have a ponder and respond as soon as I can. What is it you want to drive which requires more mechanical power?

 

[_q12x_]

What is it you want to drive which requires more mechanical power?

A stepper motor NEMA 17

 

[_q12x_]

Here is the conclusion of this cct:

 

[marconi]

I have studied what you presented and here is a quick response.

1. The 2 phase bipolar stepper motor is being driven by the circuit in 'wave mode'. This means that only one of the two coils is energised at a time. The angular steps of rotation are 90 degrees.

2. The use of the CD4017 provides a state sequence 1000, 0100, 0010, 0001. This 4 bit code controls the two H bridges to energise the stepper motor as I have shown in the attached electrical art.

3. You could produce more torque by a factor of sqrt(2) ie 1.414 if you had a switching sequence which energised two coils at a time - see the link below and the section on 'Full Step two phase on' mode. Steps are still 90 degrees.

4. If you used the mode mentioned in (3) you could stick with the present coil currents and yet produce more torque. Would this be helpful? Of course the transistors need heat sinks etcetera.

5. I would use a H bridge ic designed for the purpose rather than make my own as you have done. If I had to make my own I would use paired npn and pnp power transistors but preferably complementary power mosfets - in both cases to avoid the inverters.

6. The use of a 4017 decade counter limits your ability to generate switching waveforms to control the H bridges. I would use binary counters with decoding to switch transistors by combinational logic or an eeprom.

Have a read through this:

Stepper Motor Basics - https://www.orientalmotor.com/stepper-motors/technology/stepper-motor-basics.html#:~:text=In%20the%20wave%20drive%20method,only%20one%20phase%20is%20energized.

 

[_q12x_]

Thank you @marconi

 

[_q12x_]

Here is the version that I tested already and is NOT working with the current logic cct and the IRFZ44N mosfets. I believe it should have worked with what I have here already. No?

 

[marconi]

There are two versions of this enhancement mode mosfet. The IRFZ44N which requires a gate to source voltage of at least 3V to start to turn on and 10V to be fully conducting. Your drive circuit does not achieve this for the upper two transistors.

There is also a version suited to being driven by 5V logic - the IRLZ44N - which turns on at about 3V and is fully conducting with 5V between gate and source.

But there is another problem with this circuit which is that the source of the upper transistors is not directly connected to 0V. So these source pins will be at a higher potential than 0V because there is the coil between their source pin and ground. And the coil is switched by the upper mosfet to a positive rail which can have a voltage higher than 5V. Your 4017 has a maximum output of about 4 to 5V and thus is unable to elevate the upper mosfet gate potential sufficiently high to turn the IRFZ44Ns on fully. The potential on these gates may not even exceed the minimum gate source voltage of 3V required to begin to turn them on.

I will say more tomorrow. A bit rushed but I hope you get the idea. Fortunately there are solutions.

 

[_q12x_]

You explained this one very well ! I completely understand it ! Well done.
I only have the IRFZ44N and not the IRLZ44N. So a 10V for the full gate opening we will require.
I believe it is used a Bootstrap Driver, made from a capacitor and a diode. I recently learned about it on my other forum. So that will pull up the gate voltage and activate the 1st Tr. Very good that you tell me this stuff ! Im not that good with mosfets. I know something, but not much enough.

 

[_q12x_]

Here are some observed data I collected, using my other cct with BJT NPN in it.
All this data is not related to the problem in question with the mosfets right now.
But is some useful information for a better context.
This is from the other forum:
Well, at 5V, the 4wire steppers are consuming a constant 300mA. Well the entire cct. I read it directly from the PSU screen. They barely rotate but they do it and is good enough to tell if they work.
At 6V the amperage rises to 480mA. (The speed is a bit better but not very)
At 10V the amperage rises to 630mA but the speed (and most possibly the torque) is greatly improved and visible.
I didnt test them at 15V or 20V yet. But the google says that a Nema17 like the one I have, can be run at 12-24V.

And I know this specific search I did is not the best to do but it does get me some close result.
The problem I face is simple. I want to run them at 15V or 20V ! This means at higher amperage as well. Im not sure if 20V with 1.7A will be fine for my new motor or not. But the NPN's are definitely burning if keep them too long to work. The temperature rises too much for my comfort. Over 100dgrC. And the aluminium radiator on each transistor will not really do much difference. Maybe 2or3 dgrC lower. Really, its too high. So we need to attack it with something else. With other weapons of war. Soon the L298 IC will arrive. Probably less than 1 month - I hope. Then after that my P-ch Mosfets.
Until all of them arrive, I will try to make the same circuit here but with the mosfets that I have, only N-ch, IRFZ44N.
Here is my Nema 17 stepper model number: 17HS4401 data:

All this being said, we will concentrate on the mosfet problem. I only put this here as a reminder for me, and possibly helpful for you.

 

[marconi]

You explained this one very well ! I completely understand it ! Well done.
I only have the IRFZ44N and not the IRLZ44N. So a 10V for the full gate opening we will require.
I believe it is used a Bootstrap Driver, made from a capacitor and a diode. I recently learned about it on my other forum. So that will pull up the gate voltage and activate the 1st Tr. Very good that you tell me this stuff ! Im not that good with mosfets. I know something, but not much enough.

Just like me !

 

[_q12x_]

Ok. I did this and here are my conclusions:
Nema17 17HS4401 measured coil resistance = 2.4R per coil
I know the Resistance, I know the Voltage, I want the Amperage to know
So I =V/R = 5V/2.4R = 2.08A
and P=VI=5V*2.08A=10.4W


In datasheet, 17HS4401 has a maximum of 1.7A.
This means that V=IR=1.7A*2.4R=4.08V
and P=VI=5V*1.7A=6.9W
So nominally is 4.08V @ 1.7A with 6.9W

 

[marconi]

The maximum power produced by Ohmic heating in the mosfet is Pm = Isquared x Ro where Ro is the resistance source to drain when the mosfet if fully conducting. The datasheet has a figure for Ro. It will be a figure in milli-Ohms.

We then consider the duty cycle ie for what percentage of the time the mosfet is on compared to off. If the ratio was 1 to 3, ie off three times longer than on, then the the average power will be Pav = 1/(1 + 3) x Pm.

 

[_q12x_]

Nonono, all this Heavy math I did here is for (1 coil or motor) that I have brand new.
This is a very important math to be done! Is essential. I am not doing it every day. So thats my excuse not knowing it - not using it. I had to discover it myself from a movie I watched last night on youtube. And I didnt discover it, but I remember it, and I said aaaah, that thing... haha. And I quickly adapt it to my thing here.

 

[marconi]

Yes I understand your maths was for one coil. Mine was to give you a 'leg up' for the mosfet power to size the heat sink/radiator.

 

[_q12x_]

Yes I understand your math was for one coil.

But I am driving 1 coil at a time, right? I dont need to *2 for both coils. Right?

 

[marconi]

But I am driving 1 coil at a time, right? I dont need to *2 for both coils. Right?

Yes you are right.

 

[marconi]

https://www.farnell.com/datasheets/2303998.pdf

I had in mind last night the use of these opto-isolaters to control the mosfets.

 

[_q12x_]

The problem is the very high power over the driving Transistor. That's the key for everything to work.
I did the Heavy math there. So you can do the easy stuff and make it work. Hahahaha.
----
Someone from the other forum, but also form UK, because I asked, told me this:

Notice that they can never get more than +5V, because they are being driven by the 4017 at 5V, regardless of how high your motor supply is. So their emitters only go to ~4.3V, and all the rest of the voltage is dropped across the transistor, causing them to heat up.

And my answer was this:
This is the theory I dont know at all ! This is the first time Im reading it ! Formidable.
So... a BJT transistor is voltage dependent? I never knew. I was having the impression that is behaving like a mechanical switch, also in the sense of letting whatever current and accepting whatever voltage on its C-E to drive higher voltage components/loads. Also in any Transistor datasheet is specified a maximum voltage over C-E, I was sure I got it right. Very curious !
For example:

Or maybe you know this thing in a context I dont know.
Try to understand me here- I really dont know what your logic is and Im trying to get it.
It has sense for the case you described. But also is weird from the datasheet point of view. I cant say I get it.
----
I want some light over this information here. Like I said, I believe what it said, but Im confused.
Try to explain.
Thank you.

 

[marconi]

The problem is the very high power over the driving Transistor. That's the key for everything to work.
I did the Heavy math there. So you can do the easy stuff and make it work. Hahahaha.
----
Someone from the other forum, but also form UK, because I asked, told me this:

And my answer was this:
This is the theory I dont know at all ! This is the first time Im reading it ! Formidable.
So... a BJT transistor is voltage dependent? I never knew. I was having the impression that is behaving like a mechanical switch, also in the sense of letting whatever current and accepting whatever voltage on its C-E to drive higher voltage components/loads. Also in any Transistor datasheet is specified a maximum voltage over C-E, I was sure I got it right. Very curious !
For example:
View attachment 102981
Or maybe you know this thing in a context I dont know.
Try to understand me here- I really dont know what your logic is and Im trying to get it.
It has sense for the case you described. But also is weird from the datasheet point of view. I cant say I get it.
----
I want some light over this information here. Like I said, I believe what it said, but Im confused.
Try to explain.
Thank you.

You know that a diode has a peak inverse voltage rating. Well a transistor, npn or pnp also has diode like junctions to form the collector, base and emitter on semiconductor substrate. Being diode like these junctions have ratings for reverse/inverse voltage to prevent so called avalanche breakdown and damage. Two reverse voltage ratings are specified Vcbo which refers to the cb junction with emitter unconnected ( Ie = 0) and Vceo which refers to the ce junction with the base unconnected (Ib = 0). In the forward biased direction the usual voltages associated with the space charge region apply between transistor junctions for base emitter and base collector which are 0.7V for silicon and 0.2V for germanium.
Have a look at these links:

Transistor terminal voltages - https://www.physics-and-radio-electronics.com/electronic-devices-and-circuits/transistors/bipolarjunctiontransistor/transistorterminalvoltages.html

What is Avalanche Breakdown? | Electrical4U - https://www.electrical4u.com/avalanche-breakdown/

 

[_q12x_]

I know all that basics. Your explanation didnt help. And is ok, dont worry.
Lets concentrate over the bigger problem.
- How to make a more powerful transistorized driver for a 1.7A motor coil? The answer might be using mosfets. Now is a question of how.

 

[marconi]

So you want me to design a driver for the transistors I select from those you have to pass 1.7A in wave mode to the 2 phase bipolar stepper 17HS4401 using the switching waveform produced by your circuit (CD4017)?

 

[_q12x_]

yes
Design from scratch or copy a circuit that is already made and tested in time.
Hopefully one that is not overheating anything, not the motors, not the Transistors or anything else. To run cool.

 

[marconi]

What is the maximum current your regulated bench power supply can deliver?

 

[_q12x_]

What is the maximum current your regulated bench power supply can deliver?

My variable PSU is 30V @ 5A

 

[marconi]

Good morning. I may be a few days starting because I am overdue to deliver on a promise to my wife to install an outside light to cover the far end of the garden so that we can see our Jack Russell when she goes out at night. I am accompanying my wife on an outpatient appointment and this weekend we are away. Thus, next week I will begin.

I have in mind using optocouplers like the ones I mentioned earlier to make the transition between the 5V logic waveforms and the high side transistors. MOSFETs and perhaps some changes to your waveform generator. I have found the NEMA motor and will buy one.

 

[marconi]

I will use the IRFZ44N like you.

 

[_q12x_]

I made this movie last night and is completely unrelated to what we are doing here. Im braging with it and show it off. You can actually see on the table this project we are working right now with the H-bridge and the motors, but I push them away to make room for the new one. Is good to watch a technical video while you are away with your troubles. Especially when you are waiting for something and starts to be boring. See how much I care for you?

 

[marconi]

I am waiting with my wife at addenbrookes hospital Cambridge. The hospital is very large. Not had chance to watch your last video.

https://en.m.wikipedia.org/wiki/Addenbrooke's_Hospital

A question- what do you intend to use the nema17 stepper motor for? Knowing what for would provide me with some (extra) motivation to get started.

 

[_q12x_]

I am waiting with my wife at addenbrookes hospital Cambridge. The hospital is very large. Not had chance to watch your last video.

You will, when you will be bored out of your ears. Waiting in hospital, I know that story, I actually waited for days at some point in my life. I hope is not something serious, but a routine check, right? I wish you and your wife good luck. Good luck from Romania is a precious thing, you get it very rarely in your life. So use it well.
The motor eh? Haha. Well... not really-really a plan per say... but I do have something in mind. Its a very simple thing, but we will get there. Trust me. For now, I want a very serious and stable and reliable stepper motor driver/controller. That is capable (most importantly) to drive at full potential these motors ! Not like my dinky driver that is able to spin them but not really put them to serious work. You get the idea.
Another important thing is accessibility ! We, humans, in these modern days, we got used to already made boards that are performing very well, no doubt about that. But it should be correct to have the proper circuits to build those well performing boards they sell so expensive as well. Dont you think? I very much do. So from that aspect of things, I am very ----ed off and I want to correct it. I want a (very) good performant cct for driving these motors, available, online, on internet, and not hunt for help. That's important ! In my mind. I hope I transmitted to you the same feeling I have. And you should transmit it as well further.

 

[marconi]

I doubt my circuit will be the acme of what is possible. To do that requires considerable testing and real life trials to discover and fix problems which may at first be hidden. That is why in my view some of the best technology is often years old. But my breadboard may work reasonably well as a technology demonstrator which with your patience you can perfect.

 

[_q12x_]

Not acme as you say. Good enough is good for me. You know me.
My only concern is to work at a certain randament or in other words at a certain values I put it to work.
Simple.

 

[marconi]

Stepper and mosfets arrived. And I thought I would do some reading up since the last time I made any circuitry for stepper motors was 40 years ago at college:

https://www.ti.com/lit/ml/slua618a/slua618a.pdf?ts=1666798524312

If you dip into this reference you will see there is more to a good driver circuit than at first blush*. Maybe too mathematical for you but the text is reasonably understandable.

*
at first blush

phrase of blush

1. at the first glimpse or impression.
   "his next decision was at first blush disconcerting"

 

[_q12x_]

WOW- it must cost you a small fortune.
I have a success ! With the mosfets !
I managed to install a primary external PSU of 12V to drive all the gates and the voltage regulator and the logic circuit.
And a secondary PSU, my variable PSU that I can monitor the Amps and Volts on the motor itself.
Here is what I tried:

The 4wire NEMA17 - the new one, worked excellent at the lowest fv to the highest. I didnt reach 5V, but up to it, like 4.5, 4.8V. But the amperage was considerable high, at 1A, 1,2, 1.5. I keep it running most of the time somewhere at 300mA-to-600mA.
Some of the Tr were hot and some were Very-Hot.

I am very happy that nothing burned out.
Well... in my first try I actually forget a link wire in the breadboard and that got extremely hot and the plastic got in thick smoke and I thought it was a Tr. But the Tr was fine. Haha. But that was the only incident and at relatively low power.
No motors , no Tr and no important components were damaged.
The Voltage regulator started to heat up quite high with a heat sink on it. BIGGER is BETTER was always my motto. Haha. But that can wait.
The only failure I have so far, with this current circuit that I just tested, is the second 4wire scrapped motor. It only trembled and never rotated. I switched its connection wires a lot and the motor could NOT start rotating. Only vibrating. I noticed the new NEMA17 was trembling until 4V but after 4V started rotating. So, I increased the Amps and Volts for the scrapped one, I reached 5.5V but at a very high Amp of ~2A, so I dialed down for safety. The Tr's got extremly hot at that point. I did all this experimentation as quickly as I could, in a couple of seconds, no more than 10s per high level of power. Usually I keep everything working at a very safe temperature and power of 300mA and somewhere at 2V. At this level Ican see trembling in the motors and also no heat on the Mosfets or the VoltageRegulator.
Like the BJT version, this mosfet version has the Voltage dependent to the Amps. But is a lot better because is not limited like the BJT was. I couldnt reach 1A with the BJT. If I remember right. At 600mA everything was starting to get extremely hot. So usually if I remember right I was keeping at 300mA or less. Back then the scrapped motor worked perfectly for the tiniest impulses. But now.....????
Now that I have 2 functional circuits, its a stage completed at this point.
I will try to build the BJT that give me the best results, as a stand alone board. As weak as it was, it did served very well for testing the motors. This more powerful mosfet circuit.... hmmm, I will come to it when I will finish the BJT one. If the scrapped motor worked with this mosfet circuit, I would have been very happy. So far im only 50% happy.

 

[_q12x_]

comments from the other forum:
I only look at the top Transistors. All the bottom ones are drive at 5V ! So 4 mosfets in the lower side. Hmmm. I didnt notice it. But it worked like that, haha.
Someone noticed 2 of the mosfets are still working at 5V, and I looked more closely and 4 of them , the bottom ones, are at 5V but I notice after all the thing worked as described. Hahaha
It worked fine....

 

[marconi]

Good morning. I have been studying your latest driver circuit, the one with some mosfets circled in red. I think you have made a wiring error between the cd4017 and the two H bridges. If you think of each of each H bridge as being made up of diagonal pairs of mosfets ie \ and /, then for wave switching the pairs are switched on as in my attachment to #7.

The sequence, looking at my #7 table is 1, 4, 2, 3, 1, 4, 2, 3, 1……..where these numbers refer to the pairs of Q transistors read left to right 1, 2, 3, 4.

Your pairs of diagonal transistor are being switched 1, 3, 2, 4, 1, 3, 2, 4, 1…….

If you study your diagrams at #1 and #36 in the way the 4017 is connected to the two H bridges you should see what I mean.

 

[_q12x_]

Thank you for your last comment.
My friend mister @marconi ... check this out, what I received today !

The far right IC that almost has no marking left on it is the one from you. I notice mine are bigger. I didnt realized it when I buy them. Eh...
Cool, yes ?

 

[marconi]

Thank you for your last comment.
My friend mister @marconi ... check this out, what I received today !
View attachment 103229View attachment 103231
The far right IC that almost has no marking left on it is the one from you. I notice mine are bigger. I didnt realized it when I buy them. Eh...
Cool, yes ?

Your chips At28c64 have 64k of memory whereas the At28c16 I sent you had 16k. Otherwise they have the same technology inside. There are two more pins for the address on yours A11 and A12.

 

[_q12x_]

There are two more pins for the address on yours A11 and A12 to address 64k rather than 16k since 2exp2 is 4.

Im not sure how your math works but here is my math:
2^16 = 65536 bit combinations (your chip)
2^64 = 1.8446744e+19 which is 2 quatrilions probably of bit combinations. And not 4 like you calculated. Haha.

 

[marconi]

Im not sure how your math works but here is my math:
2^16 = 65536 bit combinations (your chip)
2^64 = 1.8446744e+19 which is 2 quatrilions probably of bit combinations. And not 4 like you calculated. Haha.

16k memory locations is organised as 2048 x 8 bit memory locations shortened to 2k x 8 bits = 16k. 2exp11 = 2048. Address word A0 to A10.

64k is 4 x 2048 x 8 bit = 8192 x 8 bits = 8k x 8 bits. 2exp13 = 8192. Address word A0 to A12.

 

[_q12x_]

Aha... I believe this number 2048 and 8192 are referring to the Addresses/or/MemoryRegisters of the IC.
I think I get your point. Im not used to think in these terms, I still lack some experience but im not far away. Thanks for your kind explanation !

 

[marconi]

https://ww1.microchip.com/downloads/en/DeviceDoc/doc0270.pdf

 

[_q12x_]

https://ww1.microchip.com/downloads/en/DeviceDoc/doc0270.pdf

- I open it before I buy it. Im always reading the datasheet, I got used to it and is a routine now for me.
Good thought but im doing it all the time. I didnt read it corner to corner, I only fly over it, I smell it. Haha. Probably I will, in the future.
- For now I showed you that your influence grow with me. I liked the IC especially after that HARD circuit I made, also with your help, I realized it is a very good chip to have. And it was quite cheap anyway so I made the step and took them. How I will use them... we'll see in time.

 

[marconi]

Actually there is another forum discussion until reaching this cct here. It was an entire evolution of it.
This cct is working ok but in certain parameters. It is limited to power and heat over the Tr's.
Is working fine in the 5V range for 2 of my stepper motors that I tested with it. But if Im bringing up the voltage, to 10V, I get 140dgrC on one Tr and about 60dgrC on a few others. At 5V, everything stays at a constant 50dgrC which I call it normal and safe.
One of the test motor is a Nema17 17HS4401. The other one is a scrapped stepper. Both work fine at 5V and even at 10V, no one was heating. Over 5V, especially at 10V, only the driver's Tr's were cooked to (almost)death. I stopped the power after smelling hot metal in the air.
I am also receiving a couple of L298N IC's. They are coming in the mail, not arrived yet. But I dont expect much difference from them, compared to what I have here. It would be nice to be better ! I hope so.
- My question:
- How to make this cct more powerful? To drive more power through the coils of the stepper motor? (and not blow up it's Tr's!)
I already have 2 answers to this question in my mind, but I dont want to influence your answer. I expect to be pleasantly surprised. Haha.
View attachment 102851Thank you !

Pondering the circuit at your #1, I am doubtful whether when the CD4017 output is logic 0 that the inputs to the base of the upper transistor and the input to the NOT gate are pulled down to 0V because those 4 diodes, one on top of the other, are reversed biased so no sinking current can flow through them. The diode path can only source current which raises the potential at the base and input to the NOT gate - not lower it. I would insert four 10kR resistors as I have indicated in the attachment to provide a current path to the 0V rail when the input to the diode is logic 0.

 

[_q12x_]

I would insert four 10kR resistors as I have indicated in the attachment to provide a current path to the 0V rail when the input to the diode is logic 0.

-I actually did that in my experiments. Here are my results: In the simulator, I got a very small performance boost in the motor spinning. Almost negligible. In reality, nothing really looked different. With or without those resistors, the motor(s) were driven the same. I put 100k indeed there, in all my experiments. But I put 10k and 1k and the result was the same so I decided to go quite up in the resistor value since the performance and the result was equal in any circumstance.
-Your explanation is very pertinent and I like it. I may try it again and see if indeed something improves.
Well done with figuring how to draw over an image ! I know it is hard to do it for you. But it pays off, doesnt it? Heh. I cna teach you some advanced image editing stuff, since you know me, Im actually a digital artist and not someone who aspire to be one. Anytime, my doors are open to you.

 

[_q12x_]

Project is finished and working !

 

[_q12x_]

Project is finished and working !

 

[marconi]

Well done!

You could turn it into a clock.