PWM circuit for SMA wires help

[Hal_Emmerich]

Hi, I was hoping someone could help me figure out how to make a 30V-40V Pulse Width Modulation circuit (without a microcontroller). Or even a circuit that has a maximum output of around 10V would work, and I could just use several of them.

I am trying to power a bunch of SMA wires, and I need a circuit that has an adjustable duty cycle and power output. I have a circuit that has a very low power output, but I don't know how to increase the output without burning any components.

If anyone has any links or instructions that could help me, I would greatly appreciate it! Thank you! 

[TrickyNekro]

That's kinda easy.... Search for 555 circuits...
Then use it to drive a transistor... if you need bid loads, then use a transistor with a mosfet... or simply a mosfet....
depends on what you want...

But first step... the 555 circuit.... It's so so classic.... ;-)

[Hal_Emmerich]

Thank you! I will search for it right now!

[Soeren]

Hi,

I have a circuit that has a very low power output, but I don't know how to increase the output without burning any components.

If you have the PWM circuit you need, just use a transistor or two.
Tell us what you've got (and output voltage and current) and you can get the rest here.

Else, as TrickyNekro mentioned, a 555 could be used (again with a current booster transistor), or like half a 4093. There's several ways to shave that goat.

[Hal_Emmerich]

Hi,

I have a circuit that has a very low power output, but I don't know how to increase the output without burning any components.

If you have the PWM circuit you need, just use a transistor or two.
Tell us what you've got (and output voltage and current) and you can get the rest here.

Else, as TrickyNekro mentioned, a 555 could be used (again with a current booster transistor), or like half a 4093. There's several ways to shave that goat.

Thanks for the post! The circuit I am using is kind of complex (it uses 4 nand gates), but I will try to get a diagram of it up soon. For the 555 timer, I was thinking about using the circuit at the bottom of this webpage: http://www.dprg.org/tutorials/2005-11a/index.html


Can you tell what the maximum output of this circuit might be? My circuit analysis skills are pretty poor.

Thanks for offering help! I really appreciate it!

[Soeren]

Hi,

The circuit I am using is kind of complex (it uses 4 nand gates)

4 NAND gates, complex... You're being ironic, right?

, but I will try to get a diagram of it up soon.

If it's just a PWM circuit then never mind, is it just a single gate output?

For the 555 timer, I was thinking about using the circuit at the bottom of this webpage: http://www.dprg.org/tutorials/2005-11a/index.html

Can you tell what the maximum output of this circuit might be?

Don't use that circuit, it's somewhat flawed.

• It's spec'd to work from 3V to 18V - at 3V the MOSFET will not open correctly. At 18V the 555 is on the verge of dying.
• Extremely little gate current for the MOSFET, just what goes through the 10k resistor (i.e. very lossy switching).
• No pre-resistor on the potentiometer, which means it could go to (theoretically) 0 Ohm.

Amazing that they succeeded getting 3 grave errors into such a simple circuit.
The MOSFET can stand ample current if just driven correctly.

Here is a couple of circuits that will do better.
I just realized that I didn't get any values on R2 and R3 in both circuits. R2 should be 10..22 Ohm and R3 can be around 100..220 Ohm. In the 4093 circuit, it is probably not needed at all (I just copy-pasted the circuit and changed it to use the 4093). The 555 have the highest drive capability of the two (200mA vs. around 30mA), but I don't think you'll notice the difference, so go with what you have or can get, or choose from what it must interface to.

You can use the IRFZ46N in place of the IRF540, they share the same pinout (so does a lot of MOSFETS).

If you want another circuit to turn it on/off, both circuits can easily be changed to accommodate this, as long as the driving circuit is known (V and A)

How much maximum output... Output what?
The max voltage they can handle is: With IRFZ46N, less than 55V - with IRF540, less than 100V.
The max current they can handle is:  With IRFZ46N, less than 53A - with IRF540, less than 33A.
The R_ds_on (when driven fully with 10V on the gate): 16.5mOhm (IRFZ46N) and  44mOhm (IRF540).

But this is turning things upside down. You tell us what you need and then we find the solution.

[Hal_Emmerich]

To Soeren above:

Thank you so much! I had one question on that circuit you gave me for the 555, it's probably obvious but I'm worried I'll make a mistake: is it true that one end of the SMA wire at a 30-40V source, and the other attached to the MOSFET, and a 6Kohm resistor and LED are there to help control the current flow? And also, would I need to change the 5-12V or 30-40V sources if I was using an IRF510 instead (that's the only MOSFET I have at the moment)?


Thanks for all the help! It's true the circuit I'm currently using isn't actually complex, I just meant for me it would be too complex to describe without a picture! My word explanantions tend to confuse people.   I got it from a book and I haven't signed up for any free webspace services yet to upload a picture here.

And thank you again for those PWM circuits! I didn't even notice those errors in that circuit I linked. I'm sorry for not being more specific with the values I need. The voltage I need for my wires and the PWM timing I need keeps changing, and I've never actually built a PWM circuit without a ton of help. I never expected to get such great guidance here!

[Soeren]

Hi,

[...] is it true that one end of the SMA wire at a 30-40V source, and the other attached to the MOSFET,

Well yes, that's what you asked for (it does sound a bit high, but I assume you have a reason).
If you need less voltage over the wire, just substitute the lower voltage.

and a 6Kohm resistor and LED are there to help control the current flow?

No, it was just meant as a help to see when there's power on the wire (and a very rough estimate of the duty cycle) and it can be removed (or not installed at all) when it's running as planned.

And also, would I need to change the 5-12V or 30-40V sources if I was using an IRF510 instead (that's the only MOSFET I have at the moment)?
The IRF510 has a much lower current handling, only 5.6A, its max. voltage is 100V, so no problem there. R_ds_on is a bit high, 0.54 Ohm, but most important, you need at least 7..8V to drive it properly.

For testing with less than 5.6A it's OK (assuming you use the needed voltage, i.e. 7V absolute minimum, 12V if possible).
I don't know your application, but 5.6A at 30V (40V) is a healthy 168W (224W) and that's quite a bit of heat for an SMA app.

How did you arrive at the figure of 30..40V?

(Basing this on numbers from Dynalloy)
If you use a 0.203 mm diameter wire with the following data:  29.1 Ohm/m, 570.0 g pull force, ~660 mA for a contraction time of 1s ("safe" value).
With 30V (40V), you need to use 1.5m (2m) to keep the current from going higher.
The IRF510 is able to handle 8 such wires in parallel
Just an example, your wire is probably different.

The voltage I need for my wires and the PWM timing I need keeps changing,

Now I feel it in my bones that this is a top secret   But if you could reveal just some metrics, like how thick a wire, how long, how hard does it need to pull (min/max) and such, it might be easier to find the best solution.

Do you have the wire?
If you do, do you have a datasheet on it?

If the specs keeps changing, wouldn't a microcontroller be a better option, as it could automate the regulation?

[Hal_Emmerich]

To Soeren above:

Thank you!

I assumed I would need 30V to 40V because I am using 0.015 inch diameter wire from Dynalloy to bend a robotic arm at the elbow, and then later I need to have it swivel side to side at the shoulder. I was worried because I have to build my arm out of HDPE plastic (I can't use wood) and I thought it might be kind of heavy. I was planning on using maybe six wires maybe around 14-20 inches to pull the lower arm from wires connected at the upper arm (longer is probably better because I think that naturally means more contraction).

I got some info from Dynalloy that the 0.015" LT wire I am using has the following properties: current required to actuate it with a 1 second pulse is 2.25 Amps; resistance of 0.21 ohms/inch. I don't know if I should use different wire.

I really don't have an exact picture of what I need to do to finish this arm. Every time I try to experiment, I start thinking I'm using wires that are too short, but I'm also nervous about handling high voltages from one source.

I was thinking I could just use several sources that are 20 V or less, where each source is a PWM current through the wire so I can keep the arm bent by sending periodic pulses to the wires until I turn it off and the arm lowers when the wires cool (I hope that makes sense).

I really wanted to use less than ten 9 V batteries to power the whole arm, but I don't know what I will ultimately need. I really appreciate all your help, especially since I'm building this thing with little help and no experience. Thanks again!

Edit: Just wanted to add that I'm terrible at microcontroller programming, and I have no time to learn, so I can't use one.

[Soeren]

Hi,

I assumed I would need 30V to 40V because I am using 0.015 inch diameter wire from Dynalloy

OK.
According to Dynalloys data, this is 0.21 Ohm/inch, so 14" (20") will be 2.94 Ohm (4.2 Ohm) total. For 2.25A through it, you'll need 6.6V (9.5V).
No reason to go to 30..40V, 12..15V is more than enough, even if you wanna speed up the contraction some.

I was planning on using maybe six wires maybe around 14-20 inches to pull the lower arm from wires connected at the upper arm (longer is probably better because I think that naturally means more contraction).

You'll have 4.42 lbs with a pull of 0.56" (0.8") each wire so 26.52 lbs for 6 wires in parallel (don't bundle them tight, or the recovery time will go up).
That equates to 237.6 ozf-in (339.5 ozf-in), which you can "gear" up or down, but more movement means less power and v.v.

I don't know if I should use different wire.

Don't know if the above got you any closer to a decision, but do you have to use muscle wire? Did you consider "air muscles" (not talking about bodybuilders here ).

I really don't have an exact picture of what I need to do to finish this arm. Every time I try to experiment, I start thinking I'm using wires that are too short, but I'm also nervous about handling high voltages from one source.

I was thinking I could just use several sources that are 20 V or less, where each source is a PWM current through the wire so I can keep the arm bent by sending periodic pulses to the wires until I turn it off and the arm lowers when the wires cool (I hope that makes sense).

One source is enough... As long as it got the omph needed.

I really wanted to use less than ten 9 V batteries to power the whole arm,

6 wires of 2.25A = 13.5A
9V batteries are for currents of up to 10mA or so.
13.5A / 10mA = 1,350 of 9V batteries 
I'd suggest using an SLA battery (a small 12V/7A would enable around 20..30 minutes of constant agitation).

but I don't know what I will ultimately need. I really appreciate all your help, especially since I'm building this thing with little help and no experience. Thanks again!

You are welcome!
Is this a school project where you have to do it a certain way, with certain materials and methods?
There are lighter materials than HDPE and, as I mentioned, I'd really consider air muscles (or DC motors with worm drive).
The trouble with SMA in this app. is that you need plenty of power just to hold a position, seemingly idle, but consuming your 9V batteries faster than you can change them.

With air muscles (which is cheap to DIY, you need a small compressor (perhaps from a car horn) and a valve that can either feed from the compressor, be closed or let air escape (manual or electrical, depending on what you want it to do).
To contract the muscle, you pump air into it and to relax it, you let the air escape.
One added benefit with air muscles is, that they have a very natural movement, as they work very similar to the "meat muscles" in us - look at eg. one of your biceps - when you contract it it gets thicker and shorted - just like an air muscle.

To make an air muscle you need:
 Inner tube from a bicycle or a moped
 solution (for gluing the inner tube
 Wire tubing weave of a suitable diameter (depends on the inner tube diameter).
 a few tie-wraps

  plus a small compressor (can feed several muscles) and the valve mentioned.

[Hal_Emmerich]

To Soeren above:

Thank you for all those calculations! It's helping me immensely!

Yes, it is a school project! I must use SMA actuators to move the arm, but everything else is up to me. The direction I'm going in now is simply because I thought it was easiest, and I've seen similar PWM-powered SMA wire applications in places like the Muscle Wire Project Book (where I got my current PWM circuit) and some random things on the internet. I'm definitely open to new design choices from anyone! I chose the HDPE plastic because I saw it suggested on these forums, but I'm willing to use any other material that won't catch fire, is light, and is not too expensive (I have only a few hundred more in my budget).

Thank you for that battery info! It really didn't occur to me that just a bunch of 9 V batteries wouldn't work. I will definitely check out SLA batteries instead!

Those air muscles sound really cool, I've never even heard of them. I'm becoming more interested in building my own robots for fun someday, and those air muscles seem like they would be great to build with. Thanks again for all this info!

[Hal_Emmerich]

Hi, this is the type of (12V,7A) SLA battery I could find that would suit my project based on Soeren's advice: http://www.batterymart.com/p-12v-7ah-sealed-lead-acid-battery.html#pr-header-SLA__12V7__F1

I'm not sure how I would attach this battery directly to my six SMA wires. Do I need alligator clips to attach to the batter? Or can I just wrap some stripped 22 gauge wire around the terminals to make a direct connection to a conductive surface on the other end of the wire? Would it be all right to attach the wire straight from the battery into a breadboard, or would it fry the breadboard?

Thank you very much for any advice!

[Soeren]

Hi,

Hi, this is the type of (12V,7A) SLA battery I could find that would suit my project based on Soeren's advice: http://www.batterymart.com/p-12v-7ah-sealed-lead-acid-battery.html#pr-header-SLA__12V7__F1

That's just for 20..30 minutes of runtime, so perhaps you need a larger one?

I'm not sure how I would attach this battery directly to my six SMA wires. Do I need alligator clips to attach to the batter? Or can I just wrap some stripped 22 gauge wire around the terminals to make a direct connection to a conductive surface on the other end of the wire?

You need spade terminals, as used in cars. Don't use clips or worse, you don't wanna risk a short.

Would it be all right to attach the wire straight from the battery into a breadboard, or would it fry the breadboard?

You should have a fuse as physically close to the terminal as possible in one of the lines (positive is the common line to fuse, but it doesn't matter).
Then add a switch to make life easier.
Make sure to isolate all exposed wires in the hookup.

[Hal_Emmerich]

Thank you again, Soeren!

I probably should use a larger battery, but I think I'll stick with this one for now so I don't get myself more confused!

Here are the steps I will try to take based on what you said, I hope I have them mostly right:

So spade terminals similar to these I attach to the terminals of the SLA battery, then I can put 22 gauge wires into the back of the spade terminal, and you suggest that I make a wire from the positive end of the battery connect to a fuse similar to this (making sure they can handle the right current), and then I continue a wire from the fuse into a switch like this (again making sure it can handle the current through my particaluar circuit) and then that continues to one end of the SMA wires? Do you have any recommendations on fuses and switches I can use? I'm trying to make everything fit on a small breadboard and then preferably fit inside the arm or attach to the outside.

And what did you mean by isolating exposed wires? I'm sorry, I don't know any of these simple terms.

Would it be possible for you to maybe modify those schematics you gave me for the 555 timer PWM circuit to help me understand how to wire up everything? If you can't, of course I will understand. I will try to ask some people at school if I can, but you seem more experienced than most people I talk to at my school! Thank you again for all the help you've given me.

EDIT: Added the fact that the SLA positive terminal goes to the SMA wires, not the breadboard.

[chelmi]

Based on the assumption that you will use 13.5A

then I can put 22 gauge wires into the back of the spade terminal

22 is way to small. 12 is probably better (Soeren can you confirm?)

I'm trying to make everything fit on a small breadboard and then preferably fit inside the arm or attach to the outside.

If you are talking about solderless breadboard, putting 13.5A through it is probably a bad idea.

[Hal_Emmerich]

Based on the assumption that you will use 13.5A

then I can put 22 gauge wires into the back of the spade terminal

22 is way to small. 12 is probably better (Soeren can you confirm?)

I'm trying to make everything fit on a small breadboard and then preferably fit inside the arm or attach to the outside.

If you are talking about solderless breadboard, putting 13.5A through it is probably a bad idea.

Thank you, Chelmi! From Soeren's help, it is true that the large voltage/current from the battery would be directly applied to the SMA wires, so I guess I can still keep my breadboard with the PWM, and the current wouldn't have to go directly into the breadboard. Sorry, I forgot that detail in my earlier post! But from Soeren's circuit diagrams for earlier, I'm not sure, would the negative terminal of the SLA battery have to be attached to the common ground of the whole PWM circuit?

I will try to get some designs ready and start ordering parts soon. If anyone has any more tips or info on parts or anything else, please post them here! Thank you for all the help!

[Soeren]

Hi,

So spade terminals similar to these

No, if anything they're pitchforks
The battery have the equivalent of this, which resembles a spade, hence the name (this store must be slightly confused or misinformed).
The part you need looks like this.
Or even this with full isolation.
And here's a transvestite.

Mind you, they come in different sizes, both in width and wire gauge (red is for the thinnest wire, then blue as a medium and yellow which is for around 6mm^2 wire - colors mostly used in the automotive industry).

According to this, you need 3/16" terminal connectors, which is here

On the PCB you use these screw terminals.

I attach to the terminals of the SLA battery, then I can put 22 gauge wires into the back of the spade terminal, and you suggest that I make a wire from the positive end of the battery connect to a fuse similar to this (making sure they can handle the right current),

Yes, AND a fuse holder like this or my personal favorite, use blade fuses like found in cars in a holder like this, which can then be used from the battery to the switch (it has a single loop of wire that you cut, so plan correctly where you need to cut it before the *oops* ).

and then I continue a wire from the fuse into a switch like this (again making sure it can handle the current through my particaluar circuit)

Yes.

and then that continues to one end of the SMA wires?

No, to a screw terminal - the 6 terminals/holes in the schematic is meant for screw terminals. SMA+ and the positive side should be joined with a heavy wire if you use stripboard.
This will make it much neater and easier to follow the wiring than wires running crisscross.

Do you have any recommendations on fuses and switches I can use? I'm trying to make everything fit on a small breadboard and then preferably fit inside the arm or attach to the outside.

Hopefully not a solderless breadboard?
If so (and don't say I didn't warn you, when it comes undone at demo time ) at least make a small soldered board able to handle the currents for the MOSFET and wiring terminals - the rest is not critical in that quarter, but the larger currents should be off the solderless board.

And what did you mean by isolating exposed wires? I'm sorry, I don't know any of these simple terms.

Where you join the wires with the switch and so on, you will have uninsulated (bare metal) that could potentially touch other stuff, shorting the battery in the process - a bit of electricians tape, gaffer tape or similar could save the day. I usually put a bit of sturdier plastic or carton over the bits first and then secure this with tape. That way pointy bits wont work through the tape.

Would it be possible for you to maybe modify those schematics you gave me for the 555 timer PWM circuit to help me understand how to wire up everything? If you can't, of course I will understand. I will try to ask some people at school if I can, but you seem more experienced than most people I talk to at my school! Thank you again for all the help you've given me.

Here's a 12V only schematic - (I changed the LED resistor for 5mA@12V as well)

If you want to make it on a PCB, I'll change that too, but if you don't wanna use it anyway...

You could use something like this for the larger currents and just connect +12V, Gate of the MOSFET and 0V to this.


And you would connect it like this:

(Better resolution)
Just make sure that connections are tight and not any longer than needed between terminals and breadboard.
The capacitor C1 should be physically mounted where the wires (+ and -) from the terminal row enters the breadboard.
Twist the gate wire around the ground wire a few turns if possible, it will shield it some.
And finally, I have shown the pins of the MOSFET in full length going through the terminal block, but you will have to form the leads to make it fit.
I hope you know how to bend leads?
(Hold with pliers near the body of the MOSFET and only bend the loose ends).

EDIT: Added the fact that the SLA positive terminal goes to the SMA wires, not the breadboard.

Let's settle for an in-between here

[Soeren]

Hi,

22 is way to small. 12 is probably better (Soeren can you confirm?)

I'm more accustomed to dealing with square-mm than gauge, so I had to look it up here.
I assume AWG here.

Gauge/max. Ampere
12/41
14 /32
16/22
18/16
20/11
22/7

Anything from 19 gauge and heavier should do according to the table, but the larger the wire, the lesser the voltage drop.
If 22 gauge is all that's available, it could just be used in pairs, i.e. 2 wires connected in both ends and perhaps twisted lightly to hold them together (a few strips of duct tape would do just as fine).

[Hal_Emmerich]

To Soeren:

I really can't thank you enough! I can't express how helpful all your information is. I've learned more from your posts than from any lab on building circuits I've ever taken. I will try my best to buy the correct parts and begin building what you've shown me. Knowing me, I'm probably going to have more questions, but I definitely know what direction to go now.

Thank you very much!