Shop OBEX P1 Docs P2 Docs Learn Events
SN754410 question — Parallax Forums

SN754410 question

tobdectobdec Posts: 267
edited 2013-03-18 19:33 in General Discussion
SN754410.jpg
Ok so I have a robot project over in the projects area and have an SN754410 hooked up to drive my motors.....the Homework board im on can't supply enough amperage to drive the IC. So I am going to go to batteries plus today and pick up a rechargable battery. Heres my question: I know that I can basicly put whatever voltage of battery I want to the stamp bc it has a voltage reg built in easy enough.....and I know how to give the positive and correct voltage to my dual h-bridge IC....where do I out the grounds from the IC to? Do I connect them to the stamp's ground? or directly to the battery? Or does it even matter...I know it's a noob question but I simply don't know the layout of the HW board. Here is a schematic the 4 ground are the middle 2 pins on either side. If my thinking is correct these will go directly to the battery...correct me if imwrong I just don't want to fry anything.
365 x 283 - 17K

Comments

  • Duane DegnDuane Degn Posts: 10,588
    edited 2012-05-10 14:49
    You want the gounds connected to both the batteries "-" terminal and the Homework board's ground. Just connect it to the HW's board ground in a single location (at least that's what Phil keeps saying).

    I'd uses a heavy duty wire to connect with the motor and just a normal signal wire to connect to the HW boards's ground.

    I seen debate on whether or not the SN54410 needs external diodes or not. From what I've read, I think they should be used. I believe the sample circuit in the datasheet shows external diodes being used.
  • tobdectobdec Posts: 267
    edited 2012-05-10 15:10
    Are the external diodes for ESD protection? Or clamping? I won't be pulling that much currect through the IC it's a pretty small bot.
  • Duane DegnDuane Degn Posts: 10,588
    edited 2012-05-11 09:16
    The external diodes are flyback diodes.

    The coils of the motors resist changes in current flow. If you try to change the current too quickly the motor will create a voltage spike as it attempts to keep the current flowing.

    Since it is difficult to be sure you're not going to shut off the current too quickly, it's generally a good idea to use flyback diodes with inductive loads (motors and relays).
  • GordonMcCombGordonMcComb Posts: 3,366
    edited 2012-05-11 09:52
    The big difference between the L293D and SN754410 is that ST was willing to specify the diodes on the outputs would handle the kickback from the size of motors typically used for the chip, and TI was not. I don't recall ever reading (from a reliable source) that a 754410 was blown when using a properly sized motor. Datasheets for both products are riddled with errors anyway, and it's rather odd they'd specify the diodes, which are on low impedance outputs, are only for ESD. The chip was clearly intended to compete with the L293D.

    The 754410 is pretty inexpensive if you want to test it, and should it not work you can usually fall back to the L293D. But even then, there are so many $5-10 dual bridge modules now out that use more modern techniques, and are already well protected. A bit more than a bare L293D or SN754410, but better peak amps and often better PWM handling. Look, for example, at Pololu's TB6612FNG dual driver carrier. About $8, 3A peak, 1A continuous, itty-bitty board easy to integrate anywhere.

    -- Gordon
  • tobdectobdec Posts: 267
    edited 2012-05-11 10:39
    Well basicly it's running fine with no diodes...however with a simple high or low one motor is running faster than the other. Is this chip capable of handling PWM that my bs2 can produce to the point of being able to push my bot in a straight line? Im using one of Tamiya's twin motor differental trannys to push my bot....its about the size of a boe-bot slightly larger. I just cant get it to go in a straight line.....after switching the motor wires ive confirmed one motor is spinning faster. I don't rly need variable speeds for what im doing here just the ability to work out the speed difference.
  • Duane DegnDuane Degn Posts: 10,588
    edited 2012-05-11 11:33
    I'm officially back on the fence about using external flyback diodes with the SN754410.

    I based my earlier recommendation on this diagram.

    attachment.php?attachmentid=92437&d=1336760771

    I see now, it is for using the chip with a two phase stepper motor.

    This next diagram shows the internal diodes.

    attachment.php?attachmentid=92438&d=1336760783

    I've read debate about whether these internal diodes offer enough protection. I'm not sure if the internal diodes provide adequate protection (what do flyback diode protect? The uC or motor driver chip?)

    I'm inclined to trust Gordon's opinion on this matter. I believe Robert also said the internal diodes should be enough.

    I'm not sure if the BS2 can provide a PWM signal these chips can use. I believe there is something a bit odd about the way the BS2 does its PWM that limits it usefulness in some applications.
    903 x 642 - 58K
    796 x 434 - 40K
  • GordonMcCombGordonMcComb Posts: 3,366
    edited 2012-05-11 12:35
    I'd think if the motor has so much back-EMF that it races back and knocks off the MCU it was too big for the chip anyway! Still, I always like to put in series resistors for BS2 projects. That way if the motor chip fries the much more expensive BS2 chip has a better chance of survival.

    But as I said, I think the overall best approach is just skip these late 20th century relics and go with something newer and more efficient. For just a couple more bucks you get a lot more functionality.

    For $25-30 you can get a serially-controlled motor driver, which will then do PWM, dynamic braking, etc. Parallax resells one or two of the Pololu drivers, though I'm not sure they carry any of the serial ones. These are nice because they are set and forget.

    I've used this one in several projects:

    http://www.pololu.com/catalog/product/1110

    -- Gordon
  • GordonMcCombGordonMcComb Posts: 3,366
    edited 2012-05-11 12:44
    tobdec wrote: »
    Well basicly it's running fine with no diodes...however with a simple high or low one motor is running faster than the other. Is this chip capable of handling PWM that my bs2 can produce to the point of being able to push my bot in a straight line? Im using one of Tamiya's twin motor differental trannys to push my bot....its about the size of a boe-bot slightly larger. I just cant get it to go in a straight line.....after switching the motor wires ive confirmed one motor is spinning faster. I don't rly need variable speeds for what im doing here just the ability to work out the speed difference.

    The stock Tamiya motors are of pretty low quality. They're also only for 3V operation, and at 5-6 V they pull several amps when stalled.

    I highly recommend upgrading the motors with these:

    http://www.pololu.com/catalog/product/1117

    These are meant to be used at 6V and even higher, and at 6V have a stall current of 800 mA, well within the range for the 754410, or the only "new-and-improved" bridges I've been describing.

    Even with these better motors it is not uncommon for one to go faster than the other. Fixing this depends on your application. For a line-following bot it's of little or no consequence.

    -- Gordon
  • tobdectobdec Posts: 267
    edited 2012-05-11 13:47
    Awesome recomendations Gordan thanks! And normaly I would think it wouldnt be a big deal...however this particular case its kinda extreme....like to the point I think a motor may be bad idk.

    edit: I actualy got it pretty damn straight now...I simply swaped the motors...voila! It ran over 4 ft in a straight line....huge improvement! So basicly I just gota get my black line sensors hooked up and the dvd-rom drive and its done. Gordon ima go ahead and order a few of those motors anyhow, im sure theyll come in handy..plus theyre cheap and a whole lot better with the voltage rateing than my stock ones.
  • Duane DegnDuane Degn Posts: 10,588
    edited 2012-05-11 14:44
    But as I said, I think the overall best approach is just skip these late 20th century relics and go with something newer and more efficient. For just a couple more bucks you get a lot more functionality.

    For $25-30 you can get a serially-controlled motor driver, which will then do PWM, dynamic braking, etc.

    I can see the advantage of using one of those controllers if your using a single core microcontroller, but with the Prop, it's not hard to send PWM signals to a controller.

    I switched from SN754410 chips to L298N chips in my Mecanum wheeled robot.


    attachment.php?attachmentid=90949&d=1332600679

    While the SN754410 chips used to get hot (even doubled up) the L298N chips don't even get warm running the Rover 5's motors. SF's price for the L298N chips is pretty good (better than Dig-Key's price) so the L298N chips don't cost a whole lot more than the SN754410s.

    The L298N can be persuaded to fit on a breadboard.

    This isn't much help with the problem of powering a motor from a BS2, but I think these chips are worth keeping in mind for other projects.
  • tobdectobdec Posts: 267
    edited 2012-05-11 15:16
    I actualy seriously considered those chips in my build.....however the SN754410's were on the shelf at out local supply shop so I grabed um. BTW Duane nice project I love it! Amkes me want to get videos of my project up to view! The biggest cool factor of my bot is the dvd-rom drive that gives our customers the business cards...I don't think neone else has done anything similar yet...correct me if im wrong.
  • GordonMcCombGordonMcComb Posts: 3,366
    edited 2012-05-11 17:33
    L298s are okay, and I use them a lot, because they're cheap and many folks (like SparkFun) still use them in modules for brushed motors. But the chip is really intended for stepper motors, which are predominately run at higher voltages. That's what the L298 really likes. Over-volting is often needed to overcome the higher-than-expected drop through the chip -- it can be as much as 2.5 volts at 1A. This means more batteries, which means more weight, which means more current consumption, which means...

    I have a real issue about that SparkFun L298 board -- one of their poorer designs, and I see they still have 1,000+ of them! There aren't many 2A bridges out there in the price-range, though, so we put up with it.

    -- Gordon
  • tobdectobdec Posts: 267
    edited 2012-05-11 18:05
    Funny story Gordon I originally bought my SN754410's to drive steppers in a desktop CNC mill. Thanks guys! I got my bot up and running the only issue now is my voltage regs don't have enough amperage throughput to serve my h-bridges...no worries though easy fix ordered some 5amp ones :cool:

    and Gordon this link is for you http://www.ebay.com/itm/Heathkit-Zenith-Hero-Jr-Robot-Vintage-Home-Robot-no-charger-included-has-batteri-/110875648227?pt=LH_DefaultDomain_0&hash=item19d0b420e3 I know ur into vintage bots...havnt seen any pic of this particular model thought you might be interested!
  • Duane DegnDuane Degn Posts: 10,588
    edited 2012-05-11 20:27
    I have a real issue about that SparkFun L298 board -- one of their poorer designs, and I see they still have 1,000+ of them!

    I can find several different L298 boards at SparkFun. Since you mentioned they have over 1,000 of them, I'm guessing you mean the little breakout boards I used above.

    I'd really like to know what you don't like about them.

    I don't like they didn't add any easy way to add diodes and the power traces seem on the thin side but I'm not sure what's not to like since the board just makes it easier to connect wires to the chip.

    I've recently taken up PCB design in order to make some boards for some magnetic encoder chips I like. Since I have a second Rover 5 chassis, I've been seriously considering making a PCB with two L298s, a Propeller and places to hook up motors and encoders of the Rover 5.

    The Rover 5 version SparkFun sells has quadrature encoders on all four motors. I haven't seen any examples (other than my robot) of all four encoders being used with the chassis. I'm hoping to make a board that would make it easy for Arduino users to take advantage of the Rover 5's encoders.

    If you think the L298 isn't a good match for the motors of the Rover 5, I'd like to know if you have any suggestions for a better motor control chip (hopefully inexpensive).

    As I mentioned earlier the SN754410 chips I had previously used with this robot got very hot after a few minutes of use. I haven't ever felt these L298 get even warm when I've used them with my robot. I took the lack of heat as a good sign.
  • tobdectobdec Posts: 267
    edited 2012-05-11 22:56
    I havnt had much of an issue with my SN75410's getting hot yet..well see after the new vreg's come in.
  • GordonMcCombGordonMcComb Posts: 3,366
    edited 2012-05-11 23:08
    Duane Degn wrote: »
    I'd really like to know what you don't like about them.

    I don't like they didn't add any easy way to add diodes and the power traces seem on the thin side but I'm not sure what's not to like since the board just makes it easier to connect wires to the chip.

    Exactly. The board is just not well thought out. For the same money and footprint they could have added pads for the diodes. Some pads for bypass caps would have been nice, too. Really nothing more than the basic parts for the reference design that's in the datasheet. And yes, they really need to go on the board. The closer to the chip the better.

    Apart from the occasional limited run board I see here and there, the SF breakout is about all we have to select from, and you're right it's better than nothing. But I just think they could have put a little more time on it.

    >>I took the lack of heat as a good sign.

    Or the high internal resistance of the L298. I'd be curious what voltage is actually reaching the motors. You might try a stall test to see how much current is drawn. The heat dissipated from the chip is really just a function of current. With one you get the other. You can always bolt on a piece of aluminum bar if you need to enlarge the heat sink area. That's a nice feature of the Multiwatt package.

    To be fair there are things I don't much care for in the Pololu designs. They take more time on their designs, and they produce very detailed documentation. But they're on a miniaturization kick that's so extreme many of their boards have no meaning for mounting. If you want something solid you have to resort to cable ties.

    -- Gordon
  • tobdectobdec Posts: 267
    edited 2012-05-12 00:15
    Well its all thrown down on a board...now to wait for the vregs..Gordon any suggestions on how to keep everything nice an d cool? I was thinking of putting my h-bridges in a project box with small sinks and a small fan to blow through.
  • GordonMcCombGordonMcComb Posts: 3,366
    edited 2012-05-12 07:57
    I'm not sure forced air cooling is needed if you upgrade to those 6V motors. With stall at 800mA even the 754410 won't get that hot. Stall is worse case and during runtime your bot should be pulling far less. Upgrading to better motors is more cost effective than a fan. Try to keep down the weight whenever possible. Otherwise your robot could spent more time recharging then running.

    -- Gordon
  • tobdectobdec Posts: 267
    edited 2012-05-12 08:09
    Words of the wise!
  • Duane DegnDuane Degn Posts: 10,588
    edited 2012-05-12 09:35
    tobdec wrote: »
    Words of the wise!

    Amen to that. (Or should it be "Words from the wise!"?)

    Thanks for the help Gordon.
  • tobdectobdec Posts: 267
    edited 2012-05-12 12:52
    lol Gordon if your ever in Ohio drop by i'll buy you a beer. U can have one too Duane!
  • WBA ConsultingWBA Consulting Posts: 2,934
    edited 2013-03-18 00:18
    If you are using the SN754410, it is best to make your own setup on a piece of prototyping board. That one you link to appears to be laid out nice, but it is impossible to swallow the $24.95 price tag for a breakout board for a chip that costs 98 cents at Arrow. Also, as Gordon mentioned in this thread earlier, there are much better options available at much lower prices if you are going to buy a finished motor controller. For my Rocka robot version that will have dc motors, I am steering away from the L298 and SN754410 options to go with one of three from Pololu: the DRV8835 unit for $5.95, DRV8833 unit for $6.95, or the Pololu TB6612FNG unit for $8.45. I would recommend looking into those further.
  • Duane DegnDuane Degn Posts: 10,588
    edited 2013-03-18 03:27
    it is impossible to swallow the $24.95 price tag for a breakout board for a chip that costs 98 cents at Arrow

    Particularly when L298N chips with diodes are so cheap.
  • Duane DegnDuane Degn Posts: 10,588
    edited 2013-03-18 19:33
    rocky79 wrote: »
    I agree at $24.95, it's a bit overpriced but they also offer the bare board version for few dollars. Not bad if you're trying to build your own driver.

    http://pcbgadgets.com/index.php?route=product/product&product_id=65

    This L298N board which is populated costs 15 cents less than the bare board you're selling.

    Parallax does have a classified forum which I think is a more appropriate place to promote your website (I'm not sure if it's completely appropriate even there). Posting here comes across as spam, particularly when it's a reply to a post made ten months ago.

    Edit: Actually, I don't think it would be a problem to promote your website here if it's done in a manner similar to the way Andrew does. No one minds Andrew including a link to some of the things he sells in his signature when it's part of a genuine (and generally very helpful) post. IMO, the way you're promoting your site comes across as dishonest.
Sign In or Register to comment.