PDA

View Full Version : Please help me for an alternative IC..



ankur1986
02-21-2008, 01:04 AM
Hi,
I am almost thru with my project now which involves a Pick n Place robotic Arm . I have used 3 motors for the mechanism and the problem is with my shoulder motor.
My shoulder motor draws around 1.5- 1.8 A of current .
i have burned around 2-3 ULN2803's and have switch over to L293D .
To obtain more current i used two L293D in parallel . However even these IC's are getting heated up really fast and im afraid i wont be able to present my project.
So i was thinking if there was some similar IC which comes in a 16pin DIP which i could use. I have already made a PCB using the L293D so changing it will be a problem .
The shoulder motor is actually lifting the weight of another motor , so the constant current value should be more.
Please provide me with suggestions / alternatives.

What if I use MOSFET's in between the motor coils and the outputs of L293D's ? will the load still be on the IC's or the MOSFET's ?
also plz provide me with specs as to which MOSFET's to use.

Plz elaborate the connections for the Mosfets. My basic knowledge on it is that it consists of Gate - Drain -Source
So if i include 4 of them in betn my stepper coils and output of L293D will it work ?


o/p1 ------- G - D(gnd) -S(to coil 1) -- Coil 1

o/p2 ------- G - D(gnd) -S(to coil 2) -- Coil 2

o/p3 ------- G - D(gnd) -S(to coil 3) -- Coil 3

o/p4 ------- G - D(gnd) -S(to coil 4) -- Coil 4


Maybe u guyz can think of other cheaper alternatives . I am already over budget and can spare only about 150 bux.
Thnx .

Vivek
02-21-2008, 09:42 AM
No way and L293 can take so much current!!

Remember, a thumb rule, that the Real Current Rating of a driver IC is about 70% of what the datasheet claims.
(use this thumb rule if you dont want to get into complex thermal analysis.)
(anyway I wonder if you read the datasheet of L293)

Now your idea of trying MOSFEts is good.
The connections seem a bit wrong though.

Use any N-Ch mosfets -
IRF540N or similar

L293 o/p to Gate
Src to GND
Drain ----[ Coil ]-----VBAT/Supply+

I recommend driving gate with L293 so that -
you wont have to desolder them
and they would provide a better gate drive.

Good Luck!

cheers!

firoz
02-21-2008, 06:41 PM
are u over driving the stepper?
imean running the 5v stepper at 12v.this can be done unless u chop the current.use a current chopper or the big resistor with calculated value.
and yes mosfet is the best.
as vivek said any IRLZ or FZ will work just fine. and if i remember correctly each IRLZ44 costs some 10-12 Rs/- i guess.

what type of controller ur using for the robot.is it direct PC or some sort of micro.

ankur1986
02-21-2008, 08:48 PM
i donno if i am overdriving the stepper.. but i am operating a 12V stepper at 15V ... is that over driving ?

i used 4 IRFZ44N as instructed by vivek ... my problem is half solved now.. the L293D's are not getting that hot .. but my MOSFET's are heating up and the motor torque caves in ... Do i attach heat sinks to IRFZ's if i am drawing around 1.5-1.8 A of current ? Will attaching the heat sink solve my problem ?

Also i am interfacing the robot directly from my PC . I have installed an I/O card using which i am driving 2 stepper (both high Torque : base / shoulder) and on the onboard port i am driving the gripper... i hope attaching the heat sink solves my problem .

One more thing .. do u suggest i replace L293D's by L293 ? do i have to make any changes to my circuit ? i guess they are both pin -to-pin compatible ? The L293 datasheet says "On the L293, external high-speed output clamp diodes should be used for inductive transient suppression." Could u plz explain what this is ? thnx .
plz lemme know ..
Thnx for ur replies and support..

ice
02-21-2008, 09:48 PM
"On the L293, external high-speed output clamp diodes should be used for inductive transient suppression." Could u plz explain what this is ? thnx

Thats for suppressing transients or voltage spikes which may propagate back to the microcontroller or PC.
Read this
http://www.embedded.com/columns/technicalinsights/194300451

Part2 has solutions for suppressing transients, one of them being clamp diodes or zeners.
http://www.embedded.com/columns/technicalinsights/194400748

ice
02-21-2008, 09:58 PM
Just as a side note, according to this:

The shoulder motor is actually lifting the weight of another motor , so the constant current value should be more.
Please provide me with suggestions / alternatives.

Try running your setup without connecting anything at the output shaft, rather disconnect your shoulder motor.

ankur1986
02-21-2008, 11:21 PM
Thanks for ur tips "ice" . i will go thru those articles and figure out what can be done .
Also plz provide me with a solution for heated IRF's ?
I am supplying arnd 15V o/p from the L293D o/p pins to Gate(1st pin) of IRFZ44n , i have connected the Drain(2nd pin) to the motor coil and the Source(3rd Pin) to ground. Is this connection correct ? Also will attaching heatsink's to the 4 MOSFET's solve my issue ?
i guess they are getting heated up pretty quickly and just giving up ..

i dont think i will be able to change the design now . I just need to present it for a few mins . Only thing is i need to test it b4 presenting but the duration of the tests wont be long ...
so needed a viable solution for d same.. thanks ...


P.S: Turning to robotics sure makes one insane, the heartbreak , the levels of Frustration , the burning grey matter and then the satisfaction of it all comming together and working is just something that words cant describe :D

ice
02-21-2008, 11:30 PM
Well it seems that your stepper is drawing to much current. It happens when theres too much load on it. Since you put in your mOSFETS, the mosfets are handling the current and heating up instead of the L293D.

So it looks like a mechanical issue, from all that you said above.

The heatsinks will help to an extent then throw up. So go with them for the moment if you cant change your design.

Good luck.

docel
02-22-2008, 12:25 AM
What is the great point in "presenting it for a few minutes"....??? Arent you "fooling" someone, maybe the examiner?????
The whole question is of DESIGN and engineering........
This has no place for "show" for a few minutes!!
If you make a project, the first thing you must do is to make sure the design is in accordance with the Manufacturers data sheet. and you should have done so , before wiring the components.

Also , doing a project without knowing the fundamentals is a cardinal sin!!

Why don't you just use relays in place of "incapable" ICs and even more exotic and over-capable mosfets, because someone just told you to??

If you cant read the datasheet and understand the current limit of the 2003 or the L293D, How do you expect an IRFZ44 to work in place of a IRF540n???
Vivek said IRF540 and you used something else.....
This means you have NO idea what you are doing..... :evil: It is like eating beef instead of the suggested mushrooms!

I'm sorry, but you cant expect anyone to help you when you dont know the basics nor follow advise to the point. You have GOT to Engineer things, not copy them, that too wrongly.

In my opinion, the entire project is wrong. An arm cannot be implemented without following the basic principles of middle school physics.

ankur1986
02-22-2008, 12:42 AM
@docel . I wont argue with you here coz ur a professional and i am a noob here.
I know vivek said to use IRF540N ... but by that time i had already made my 8th visit in the past 2 weeks to lamington road(mumbai) which itself is like a pain in d ...

Also he suggested to use any N Ch Mosfet , IRF540n or similar . i guess what i used is similar , again i dont have any idea on this but i read other posts here which said they used IRFZ44n when they needed to drive higher current motors.

Also engineering everything ? Well if that was the case then i would have come to this community a lot earlier :D .. .its my bad for being introduced to this field quite late .. U cant be the master of everything .. so when it comes to areas of electronics and mechanical designs i am outta ideas and options . Ask me to program anything or maybe anything remotely related to IT then i may be able to showcase what i can do :) .
And u were right about the "fooling" part . I just wanna make this work in from of him on D-day and give him an illusion that it works . I am being very honest here. I am being forced to eat the beef inspite of being a vegetarian :P

ice
02-22-2008, 12:55 AM
Ankur,

If you could explain your setup to us, we can be of help.

1>If possible post your schematic.
2>What steppers are you using
3>Are the steppers geared down.
4>Whats your power supply(rating in Volts and Amps)

If you really are short on time on your project, then I understand.
But if its something you want to understand whats going wrong, we are here to help.

ankur1986
02-22-2008, 01:14 AM
Thanks ice. Well i am short on time coz i have to wrap up and start preparing for theory subjects. Got 8 of them .
Also i have harnessed enough intrest in the past 2-3 months to keep me going after my project presentation .So may be once its all done i will be "RE-Engineering " the entire setup and pay special attention on the design part of it . As of now i just want this setup to perform 2 basic motions .
yaw motion of the base (left - right)
pitch motion of the shoulder/elbow (up -down ) not totally only a litte bit.
and a open close gripper action.
The base and gripper stepper are working fine . Only problem is with shoulder coz i have weld the gripper motor directly to its shaft.

The shoulder motor draws arnd 1.5A of current and is operating at 15Volts.
I dont understand what u mean by geared down steppers.

My power supply is a 15V , 2A transformer .

here is a pic of the setup :
http://asi.byethost9.com/snaps/DSC00357.JPG
http://asi.byethost9.com/snaps/DSC00355.JPG

I am using PC parallel port to drive my 3 motors.
For the shoulder motors i have connected 2 L293d in parallel . and then connected the 4 parallel o/p's to 4 IRFZ44n.

P.S : Plz copy and paste the image links in ur browser if clicking directly on them opens up a different tab .

ice
02-22-2008, 01:30 AM
Only problem is with shoulder coz i have weld the gripper motor directly to its shaft.

Thats the problem. You should never connect a heavy load directly to your shaft. It's taking so much load that its drawing too much current.
Gearing is something like this:
http://www.machinegrid.com/content/view/33/46/
Go down to the Drive trains section.

When the motor is unable to move, theres a short circuit in the motor, which leads to too much current draw.

Theres a workaround, if your willing to go to Lamington again.
Get a 30rpm gear motor and connect them to the L293D.

You should be good.

EDIT:
I saw your pics. Theres no way that thing will work. There has to be gearing between the motor shaft and what you are trying to move.
If you really are short on time, I suggest you get rid of the arm motor immediately and make something of the 2 motors you have. MAKE sure that none of them are lifting heavy loads.

On the side:
IF you remove the arm of the shoulder stepper, Your setup will work and the L293D will not heat. Try it.

ankur1986
02-22-2008, 01:41 AM
Thnx gerry.. i will try that .. u have any ideas if i buy a 30 rpm geared motor then how can i attach my gripper motor to it ?

ice
02-22-2008, 01:45 AM
I would suggest you get rid of that motor too. Because of your design, the DC geared motor wont have enough torque to lift the stepper. It could but it will be heavy. If you could just use a light toy motor or solenoid( look it up ) for your gripper, you should be good.

docel
02-22-2008, 02:54 AM
Ankur!
Try this:

1. Extend the shoulder piece to the same length as that to the gripper motor.
2. Fix the 30rpm motor on the base motor( near the base motor shaft.)
3. Tie a fishing thread (no 100) or even thick thread to the tip of the extended shoulder piece. Fix this to the 30RPM motor shaft, so that it can wind on it.
4. Now, the 30RPM motor will wind the thread and the thread will pull the gripper motor using the shoulder motor as the fulcrum.
4. Reverse the motor to bring it down. This is easy due to gravity and motor weight.
5. The middle stepper may be burnt out due to the welding heat. Disconnect and use it as a pivot bearing.

http://img104.imageshack.us/img104/3332/dsc00357modbq8.th.jpg (http://img104.imageshack.us/my.php?image=dsc00357modbq8.jpg)

Vivek
02-22-2008, 10:11 AM
Most likely this is not a Mechanical issue.
(of-course the mechanical design is trashy. sorry Ankur, to be frank, i dont have time to sugarcoat words today.)

But this seems like the Mosfet's are not 100% ON or OFF.
(they are in ohmic region.).
I could be wrong, and it could be just due to Rdson heating of the mosfets due to high current of motor, as others have pointed.

But these mosfets are capable to burn your coil, if driven well.

So, If you have an oscilloscope, look at Drain Voltage.
When a MOS is On, its Vdrain should be about 0~0.5V max
If its 1V or somewhere above..., then its not being turned ON 100%, and thats the cause of its over-heating.

Similarly, during off state, you should see 15V on drain, and nothing lower. If not then its not turning off completely.

Usually its the partial turn-on, (due to insufficient gate drive), which causes such sudden and drastic heating.

If I dont have oscilloscope/multimeter, i guess it by touching the tab of the mosfet. If it makes a red mark or boil on skin immediately, means it is indeed a half-on condition. You can try this at your risk.

I will try to find a schematic of mofet gate drive which i made few months ago.

But a tip to a good programmer-
When your gripper is supposed to be down (hanging as in the pic.) keep the shoulder motor driver off (all mosfets).
That way the motor coils dont apply holding torque and your mosfets remain cool more often. (so turn of driver at that angle for less power consumption.)

Meanwhile, try connecting the MCU output pin directly to the Gate and disconnect L293 o/p from gate.
This direct drive form MCU works for some people but never worked for me.

if out of time, add big honking heatsinks....

cheers!

Vivek
02-22-2008, 10:21 AM
and whatever be the cause of this problem, you should a-tleast follow Docel's good suggestion to modify it. if you have time, do this.
If the prof has some knowledge of physics and practical mechanics, he will reduce your marks just for mech.

firoz
02-22-2008, 02:14 PM
see its simple.

first of all design and make ur mechanical 's first.then u will know which shaft or ball screw need how much torque or force etc.then choose ur steppers or servos according to it.
u build a heavy arm and expect a small stepper to lift is not good.if u force it using high voltage and hight current,it will work but u will face problem like what ur facing now.
for ur solution.first detach that motor from the setup and run it under no load condign.if still the electronics getting hot(little warm is ok to over come the rigidity) then the problem is in electronics.
if not then as i said,torque required is not enough.

and mechanically u can reduce the weight using aluminum and nylon material.

and always remember one thing,copying the circuit and making it work is ok.untill u learn something from that experience.this experience helps u to make a better design next time and help others.which is the most IMP point if u ask me.

college project i can understand(NO project--no marks--no MNC JOB :wink: ).but when u start building stuff for ur self,take ur own time so that outcome will be very satisfactory.

and lots of blablabal...... :lol:

if only the mostfets getting hot,then get a aluminum extrusion and bolt the mosfet heds to that extrusion.worst case scenario place a 12v fan.

is this circuit step and direction type?if yes then there r some plan with just simple PIC and mosfets.(of course its not for u.as u almost finished ur project)

docel
02-22-2008, 03:13 PM
1. Most likely the Stepper motor Magnet & / or the armature wires are fried inside. Welding the shaft generates enough heat to ensure that.

2. The Mosfet is being overdriven. You should not direct-drive the mosfets through the L293D without current limit and bias/Gate clamp.
3. L298 would have been better.

ankur1986
02-24-2008, 12:51 AM
hey guyz ! thnx a ton for ur support ! well i can say that my problem of the shoulder motor and the up n down movement is solved. I tried that arrangement as suggested by docel and it worked !
I still have to get the dimensions for that new piece right but that thing actually works !
I used another stepper which i had as spare ... i have mounted it temporatily as of now .. will do the final fixatures on monday..
still need to finalize the connections and test everything once... I attached the stepper to the 2 parallel L293D's i made .. and its working fine ..
Also the fishing rope was not winding on the stepper shaft and hence i used a nylon thread which is not thick but strong enough to lift the load up and down ..

here are some pics ..i have kept that blue dhakkan so that the thread is visible..
http://asi.byethost9.com/snaps/DSC00375.JPG
http://asi.byethost9.com/snaps/DSC00376.JPG
Copy and paste these links.. .wont work if u directly click on them .

Lemme know of any other changes which i may be able to do by tomm .
And once again thanks a lot everyone !!!!

docel
02-24-2008, 01:21 AM
Good Work, Ankur!! Nice to hear that it works :lol:
I actually rigged it up here and took some pics. For some reason , I have failed to post the link :(
http://img206.imageshack.us/img206/7079/moddsc00357dt9.th.jpg (http://img206.imageshack.us/my.php?image=moddsc00357dt9.jpg)
Like I said, your stepper motor will be dead, due to the welding heat. The Rotor is a ferrite magnet and loses its strength when heated excessively. This will reflect in very high current draw.

If you move the new motor fully back, to the tip of the horizontal rod, it will work even better. (It will then be able to lift the discarded stepper too!! :lol: )
Also, the horizontal base stepper will be relieved of strain due to the one sided weight of the other 2 motors + the arm metal.
The Lever is a part of a circle......
Dont you have ANY DC motor?????
...can guide you to a good design before the night is over......

ankur1986
02-24-2008, 01:36 AM
well i dont have a DC motor as of now .
however i can tweak it according to the changes u told. i may face problems with my gripper .. so will have to re-invoke this thread again ...

u know i was just wondering that if i had concentrated on the mechanical aspects first and not gone ahead with that bulky stepper motor and the various costs that came along with it .. then i would have saved around 900 buxs !!!!!!!!!!!!!!!!!!!! :o :(

Anyways i guess u learn from ur mistakes... same is the case with me :D

docel
02-24-2008, 02:10 AM
Hi Ankur!!
Well, learning IS an incentive for a mistake.... :wink:
I liked your answer to my 'biting' remarks. You were quite contained and level headed with your reply, even at the time of adversity, the D-day, etc.,
And........ quite frank too....

So, good luck and best wishes!

ice
02-24-2008, 12:24 PM
Looks good. Glad to know it worked out.

ankur1986
02-28-2008, 12:10 PM
Sorry for revoking this thread again :oops:
The arrangement which docel suggested works pretty fine . The motor lifts up perfectly , but when reversing it makes the gripper go down a liittle and then the thread just keeps unwinding on the shoulder motor , the extented bar to which the other end of the thread is attached doesnt go down.
Could u suggest what would be wrong ? and any solutions for it ?
may be my dimensions for the extended bar is wrong . It is extended approximately the same length from the pivot as is the length from the pivot to the gripper arm.

ankur1986
03-15-2008, 12:21 AM
Hello Folks !
well i dont know if starting a new thread for this would be appropriate. but i would like to convey my heartiest thanks to all the members at robotics india who have helped and guided me through my project.
I feel that words are hard to describe the kind of support rendered to me here.
I hope that under the guidance of the members here i will be able to take more advanced and interesting robotics projects in d future.
Once again i am Indebted to everyone here. Thanks a ton for ur support.

Also here is a link of the final thing.
http://www.sendspace.com/file/d6p9s8
Its a 7MB video file.

You could browse through the various stages of my project through here:
http://asi.byethost9.com/snaps/

Thanks once again.
Hope to return to this forum around May when my exams are done [:)]