Shop OBEX P1 Docs P2 Docs Learn Events
PWM to Linear — Parallax Forums

PWM to Linear

ArchiverArchiver Posts: 46,084
edited 2004-01-14 06:11 in General Discussion
I have seen circuit designs for converting a PWM signal coming out of a stamp to
a linear voltage. The problem is I'm not sure how to go about adapting the
design to do what I want. Further more I'm not sure which parts would be best
suited for the job.

What I want to do is linearly control the input voltage to 12V DC fan that uses
around 4 Watts of power. Like I said I'm not totally clear on how to adapt the
conversion circuit design for this job. Also if you could recommend parts that
would be up to this task? As I'm not sure if a MOSFET, Transistor, etc. would
be best suited for this.

The reason I need to use linear instead of direct PWM is that the fans on tach
feed back on them and the tach sensors share common power and ground with the
fan's motor. When I use PWM and try to read the tach I end up reading the PWM
instead of the tach. It might be worth noting I'm using one of Al Williams's
excellent PWM coprocessors to do the PWM work for the stamp.
Any help is appreciated!

Thanks,
Chris Shuster

[noparse][[/noparse]Non-text portions of this message have been removed]

Comments

  • ArchiverArchiver Posts: 46,084
    edited 2004-01-10 21:53
    For some reason this e-mail never made it to the list the first time I sent
    it... weird... anyone have any idea why? Anyway here's my question...

    I have seen circuit designs for converting a PWM signal coming out of a stamp to
    a linear voltage. The problem is I'm not sure how to go about adapting the
    design to do what I want. Further more I'm not sure which parts would be best
    suited for the job.

    What I want to do is linearly control the input voltage to 12V DC fan that uses
    around 4 Watts of power. Like I said I'm not totally clear on how to adapt the
    conversion circuit design for this job. Also if you could recommend parts that
    would be up to this task? As I'm not sure if a MOSFET, Transistor, etc. would
    be best suited for this.

    The reason I need to use linear instead of direct PWM is that the fans on tach
    feed back on them and the tach sensors share common power and ground with the
    fan's motor. When I use PWM and try to read the tach I end up reading the PWM
    instead of the tach. It might be worth noting I'm using one of Al Williams's
    excellent PWM coprocessors to do the PWM work for the stamp.
    Any help is appreciated!

    Thanks,
    Chris Shuster

    [noparse][[/noparse]Non-text portions of this message have been removed]
  • ArchiverArchiver Posts: 46,084
    edited 2004-01-10 22:03
    You need to use an RC circuit to take the Stamp's PWM output and convert
    it to a voltage. Perhaps you could use that voltage to control a MOSFET
    ... but I'm not very experienced in those things. The original Stamp
    App Notes show how to buffer the voltage from the RC with an op-amp to
    "stiffen" in, but that will still be in the 0 - 5V range. If the op-amp
    could handle the power, you could use it to multiply as well, perhaps.

    -- Jon Williams
    -- Parallax


    Original Message
    From: Chris Shuster [noparse]/noparse]mailto:[url=http://forums.parallaxinc.com/group/basicstamps/post?postID=d_S7TvISuAKheFYXMqOl6PiWUaBf3shSRdE1C26mIal4Qnb6WNSc4icCCNzxlht8j47NoE0_9Z_NJvw]cpuman3@h...[/url
    Sent: Saturday, January 10, 2004 3:54 PM
    To: Stamp Mailing List
    Subject: [noparse][[/noparse]basicstamps] PWM to Linear


    For some reason this e-mail never made it to the list the first time I
    sent it... weird... anyone have any idea why? Anyway here's my
    question...

    I have seen circuit designs for converting a PWM signal coming out of a
    stamp to a linear voltage. The problem is I'm not sure how to go about
    adapting the design to do what I want. Further more I'm not sure which
    parts would be best suited for the job.

    What I want to do is linearly control the input voltage to 12V DC fan
    that uses around 4 Watts of power. Like I said I'm not totally clear on
    how to adapt the conversion circuit design for this job. Also if you
    could recommend parts that would be up to this task? As I'm not sure if
    a MOSFET, Transistor, etc. would be best suited for this.

    The reason I need to use linear instead of direct PWM is that the fans
    on tach feed back on them and the tach sensors share common power and
    ground with the fan's motor. When I use PWM and try to read the tach I
    end up reading the PWM instead of the tach. It might be worth noting
    I'm using one of Al Williams's excellent PWM coprocessors to do the PWM
    work for the stamp. Any help is appreciated!

    Thanks,
    Chris Shuster

    [noparse][[/noparse]Non-text portions of this message have been removed]


    To UNSUBSCRIBE, just send mail to:
    basicstamps-unsubscribe@yahoogroups.com
    from the same email address that you subscribed. Text in the Subject
    and Body of the message will be ignored.


    Yahoo! Groups Links

    To visit your group on the web, go to:
    http://groups.yahoo.com/group/basicstamps/

    To unsubscribe from this group, send an email to:
    basicstamps-unsubscribe@yahoogroups.com

    Your use of Yahoo! Groups is subject to:
    http://docs.yahoo.com/info/terms/




    This message has been scanned by WebShield. Please report SPAM to
    abuse@p....
  • ArchiverArchiver Posts: 46,084
    edited 2004-01-10 22:43
    I was about to ask why you didn't just PWM the fans and then I read the
    bottom :-)

    An RC network will create the voltage. The PAK-V won't stop making PWM, so
    the response does not droop and the C value is not critical. BUT, you still
    only have 0-5V and no where near 4W.

    So... you need an op amp. There are a few problems with that, though. Let's
    say you decide to run your PWM from 0 to 4V and X3 to get 0 to 12V. Great
    idea! But the op amp will need more than 12V unless it is a special "rail
    to rail" op amp. The more current you draw from an op amp, the harder time
    it has approaching the supply rail. Many op amps can go to the "low rail" so
    you can use ground as V-, but fewer can hit the +rail. So you either need a
    higher supply (15V? 24?) or you need a real rail to rail op amp. I don't
    have a specific suggestion, but something off of this page would probably be
    appropriate: http://www.linear.com/pdf/rr_inputoutput.pdf

    Now, you still don't have your 4W. Well, that's another problem. I would
    lean towards an emitter follower here. Essentially, you'd take a transistor,
    put a higher voltage on the collector, the op amp output on the base, and
    connect the fan as the emitter load. But as you expected, there are some
    issues here too. First, there is some Vce drop, so the collector still has
    to be above 12V (therefore, maybe you don't need a rail to rail op amp).
    Second, the Vbe drop will be about .7V. If you aren't very critical, you can
    just jump the gain of the op amp up so that you get 0-12.7V and live with
    the .7V drop (this means you will generate .7 to 12.7 to get 0 to 12).

    An even better idea is to put the transistor in the op amp's feedback loop.
    Look at the schematic on page 9 of
    http://cache.national.com/ds/LM/LM124.pdf.

    That amp has a gain of 10 but you could adjust it by changing the resistors.
    Obviously, the transistor has to be big enough and maybe even heat sinked.
    Putting the transistor in the loop takes care of the offset problems
    assuming the op amp has enough "overhead" to comply. In other words, the op
    amp won't be happy until the output is at 12V (assuming a 4V input and X3
    gain). So the op amp will have to put 12.7V on the base of Q1. Since the op
    amp can't reach the top of the rail, if you had, say, a 12V supply going to,
    it will just be unhappy and won't do what you wanted. Even a 13V supply
    would probably be too little. Why? Because the more current you draw from
    the emitter, the more the collector current and the base current will be
    proportional (but smaller) than the collector current (by beta). So if you
    are drawing 1A and the beta is 100, then you are still drawing 10mA and with
    a 13V supply you might achieve a 10V output.

    Which brings up another point. 4W @12V is around 333mA. Remember that beta
    mentioned in the transistor's data is at DC with nothing changing. You will
    be turning the thing on and off at a pretty good clip, so I would expect
    beta to be effectively reduced. Let's say the transistor has a minimum beta
    of 100. I'd guess as we put PWM through it that could drop to 50 or 60. Now
    with a beta of 50 a 333mA Ic will give an Ib of 6.6mA, so that's probably
    within the output range of the op amp. Of course, if beta reduced more...
    And some power transistors have pretty low betas to begin with.

    If this were a problem (and I suspect it isn't), you could use a darlington
    in the loop. Now you lose 1.4V, but since you are in the loop you don't care
    (as long as that +rail is high enough). The betas of the pair of transistors
    inside the darlington multiply. So even if both had a B of 10, you'd get an
    effective Beta of 100!

    My practical suggestion: an RC, an LM324, a 24V supply, and a power
    transistor. Your mileage may vary.

    Oh well, I've gone on too long as usual. Hope that helps.

    Regards,

    Al Williams
    AWC
    *NEW: Universal PCB for Stamp, SX, or practically any processor
    http://www.al-williams.com/gpmpu40.htm





    Original Message
    From: Chris Shuster [noparse]/noparse]mailto:[url=http://forums.parallaxinc.com/group/basicstamps/post?postID=WClKg9CUEyow-vTxXstpYlMOaOiMxzbRmFv7WX8UEnegkPVhW1eLJOOoGjNNb8t5u5IQkuRoU5I_]cpuman3@h...[/url
    Sent: Saturday, January 10, 2004 3:54 PM
    To: Stamp Mailing List
    Subject: [noparse][[/noparse]basicstamps] PWM to Linear


    For some reason this e-mail never made it to the list the first time I sent
    it... weird... anyone have any idea why? Anyway here's my question...

    I have seen circuit designs for converting a PWM signal coming out of a
    stamp to a linear voltage. The problem is I'm not sure how to go about
    adapting the design to do what I want. Further more I'm not sure which
    parts would be best suited for the job.


    What I want to do is linearly control the input voltage to 12V DC fan that
    uses around 4 Watts of power. Like I said I'm not totally clear on how to
    adapt the conversion circuit design for this job. Also if you could
    recommend parts that would be up to this task? As I'm not sure if a MOSFET,
    Transistor, etc. would be best suited for this.

    The reason I need to use linear instead of direct PWM is that the fans on
    tach feed back on them and the tach sensors share common power and ground
    with the fan's motor. When I use PWM and try to read the tach I end up
    reading the PWM instead of the tach. It might be worth noting I'm using one
    of Al Williams's excellent PWM coprocessors to do the PWM work for the
    stamp. Any help is appreciated!

    Thanks,
    Chris Shuster

    [noparse][[/noparse]Non-text portions of this message have been removed]


    To UNSUBSCRIBE, just send mail to:
    basicstamps-unsubscribe@yahoogroups.com
    from the same email address that you subscribed. Text in the Subject and
    Body of the message will be ignored.


    Yahoo! Groups Links

    To visit your group on the web, go to:
    http://groups.yahoo.com/group/basicstamps/

    To unsubscribe from this group, send an email to:
    basicstamps-unsubscribe@yahoogroups.com

    Your use of Yahoo! Groups is subject to: http://docs.yahoo.com/info/terms/
  • ArchiverArchiver Posts: 46,084
    edited 2004-01-11 04:38
    -snip
    An RC network will create the voltage. The PAK-V won't stop making PWM,
    so
    the response does not droop and the C value is not critical. BUT, you
    still
    -snip
    say you decide to run your PWM from 0 to 4V and X3 to get 0 to 12V.
    Great
    idea! But the op amp will need more than 12V unless it is a special
    "rail
    to rail" op amp. The more current you draw from an op amp, the harder
    time
    -snip
    connect the fan as the emitter load. But as you expected, there are some
    issues here too. First, there is some Vce drop, so the collector still
    has
    to be above 12V (therefore, maybe you don't need a rail to rail op amp).
    -snip
    Putting the transistor in the loop takes care of the offset problems
    assuming the op amp has enough "overhead" to comply. In other words, the
    op
    amp won't be happy until the output is at 12V (assuming a 4V input and
    X3
    -snip
    beta to be effectively reduced. Let's say the transistor has a minimum
    beta
    of 100. I'd guess as we put PWM through it that could drop to 50 or 60.
    Now
    -snip

    Man, when I see a post like that do I feel stupid! You guys are
    unbelievable. What seems 'obvious' to you couldn't be further from the
    truth. If I read this list, every day, for the next 10 years, will I
    begin to understand posts like this?

    Humbly,

    -John
  • ArchiverArchiver Posts: 46,084
    edited 2004-01-13 13:31
    I've noticed I'm not getting all the posts in my e-mail (Yahoo problem?).
    Did you read my earlier reply at
    http://groups.yahoo.com/group/basicstamps/message/39364?


    Al Williams
    AWC
    * NEW: Universal PCB mounts Basic Stamp, SX, more...
    http://www.al-williams.com/gpmpu40.htm


    Original Message
    From: Chris Shuster [noparse]/noparse]mailto:[url=http://forums.parallaxinc.com/group/basicstamps/post?postID=sVEqyz44xDHEUfxh9UrvegVQDRwmmSlB5PY5DuEEcTh1eSZqQSUqWrbJNDYd7CC0-2YyM_npKDEBplc]cpuman3@h...[/url
    Sent: Tuesday, January 06, 2004 6:43 PM
    To: Stamp Mailing List
    Subject: [noparse][[/noparse]basicstamps] PWM to Linear


    I have seen circuit designs for converting a PWM signal coming out of a
    stamp to a linear voltage. The problem is I'm not sure how to go about
    adapting the design to do what I want. Further more I'm not sure which
    parts would be best suited for the job.


    What I want to do is linearly control the input voltage to 12V DC fan that
    uses around 4 Watts of power. Like I said I'm not totally clear on how to
    adapt the conversion circuit design for this job. Also if you could
    recommend parts that would be up to this task? As I'm not sure if a MOSFET,
    Transistor, etc. would be best suited for this.

    The reason I need to use linear instead of direct PWM is that the fans on
    tach feed back on them and the tach sensors share common power and ground
    with the fan's motor. When I use PWM and try to read the tach I end up
    reading the PWM instead of the tach. It might be worth noting I'm using one
    of Al Williams's excellent PWM coprocessors to do the PWM work for the
    stamp. Any help is appreciated!

    Thanks,
    Chris Shuster

    [noparse][[/noparse]Non-text portions of this message have been removed]


    To UNSUBSCRIBE, just send mail to:
    basicstamps-unsubscribe@yahoogroups.com
    from the same email address that you subscribed. Text in the Subject and
    Body of the message will be ignored.


    Yahoo! Groups Links

    To visit your group on the web, go to:
    http://groups.yahoo.com/group/basicstamps/

    To unsubscribe from this group, send an email to:
    basicstamps-unsubscribe@yahoogroups.com

    Your use of Yahoo! Groups is subject to: http://docs.yahoo.com/info/terms/
  • ArchiverArchiver Posts: 46,084
    edited 2004-01-13 13:41
    Hi John,

    If you want a great education on electronics, read the Art of Electronics by
    Horowitz and Hill. Big book and you do need a little math. But if you really
    read it you'll wind up knowing more about electronic design than you need to
    know :-)

    Read my review: http://www.wd5gnr.com/books.htm

    I will say that I have spent 2/3 of my life working in electronics (I'll
    have been a ham for 26 years in July -- a notable date since I passed my
    first ham exams on 26 July which happens to be my birthday). It is easier to
    learn now then it ever has been before. Why? First, of course, the Internet.
    Forums like this are great. Web pages are around for everything (good
    tutorials, for example at
    http://www.meridianelectronics.ca/circ/circuits.htm and
    http://tutor.al-williams.com). Google for "electronics tutorial" and you'll
    find the Navy's NEETS course online along with details of anything you want.
    One of the advantages of doing this professionally used to be being able to
    get your distributor to give you data sheets. Now they are all online in PDF
    format!

    Another reason is the widespread availability of SPICE and other circuit
    simulation tools. I did an article on Switcher CAD (a Spice program) for
    Nuts and Volts. Want to know if my circuit will work? Instead of building a
    copy, just simulate it in Switcher Cad! Even Eagle can be made to use Spice
    (I use it with WinSpice 3). Unless you are very handy with a soldering iron,
    learn Spice. You can do in an afternoon the experiments it took me 3 weeks
    to do as a kid. SwitcherCAD is free and http://www.linear.com/software/. Of
    course, it won't simulate a Stamp, but it will simulate the stuff around the
    Stamp nicely. I've often wanted to do a book on electronic basics using
    Spice. You know, talk about a voltage divider and simulate it. Talk about an
    opamp circuit, simulate it. You get the idea.

    Hope that helps.

    Al Williams
    AWC
    * New Kits:
    http://www.al-williams.com/kits.htm



    >If I read this list, every day, for the next 10 years, will I begin to
    >understand posts like this?

    >Humbly,

    >-John
  • ArchiverArchiver Posts: 46,084
    edited 2004-01-13 15:59
    AS you can see from Al's excelent write-up, there is no simple
    answer. that makes me ask more about the tach. Any transistors you
    use will have filtered DC power, so the tach should be able to be run
    off the non-pulsing power in the circuit.

    If the tach is just monitoring voltage, then it is actually easier to
    offer a smooth low current voltage for the tach and run the motor on
    pwm.

    As I see it, you can create a clean high amp power for the motor or
    you can create clean low amp power for the sensor.

    Dave
  • ArchiverArchiver Posts: 46,084
    edited 2004-01-14 05:58
    Al,
    Yah I read your post earlier. Very strange that the original e-mail shows
    up after the e-mail I sent the second time.

    What you have suggested would work but there is one problem. This fan
    control system is for a PC, so I'm pretty well restricted to 12V.

    BTW John: Don't worry I'm an engineering student and some of the stuff Al
    talked about gave me a headache. [noparse]:)[/noparse]

    Thanks,
    Chris Shuster

    Original Message
    From: "Al Williams" <alw@a...>
    To: <basicstamps@yahoogroups.com>
    Sent: Tuesday, January 13, 2004 6:31 AM
    Subject: RE: [noparse][[/noparse]basicstamps] PWM to Linear


    > I've noticed I'm not getting all the posts in my e-mail (Yahoo problem?).
    > Did you read my earlier reply at
    > http://groups.yahoo.com/group/basicstamps/message/39364?
    >
    >
    > Al Williams
    > AWC
    > * NEW: Universal PCB mounts Basic Stamp, SX, more...
    > http://www.al-williams.com/gpmpu40.htm
    >
    >
    >
    Original Message
    > From: Chris Shuster [noparse]/noparse]mailto:[url=http://forums.parallaxinc.com/group/basicstamps/post?postID=EDob6wu3D1J_PG9AKEQyPgNRQ60ZoCPe0YOO5r66NpOQvC-FfXXqfO-yCi4oRhxOgwjx9xJInA_eHw]cpuman3@h...[/url
    > Sent: Tuesday, January 06, 2004 6:43 PM
    > To: Stamp Mailing List
    > Subject: [noparse][[/noparse]basicstamps] PWM to Linear
    >
    >
    > I have seen circuit designs for converting a PWM signal coming out of a
    > stamp to a linear voltage. The problem is I'm not sure how to go about
    > adapting the design to do what I want. Further more I'm not sure which
    > parts would be best suited for the job.
    >
    >
    > What I want to do is linearly control the input voltage to 12V DC fan that
    > uses around 4 Watts of power. Like I said I'm not totally clear on how to
    > adapt the conversion circuit design for this job. Also if you could
    > recommend parts that would be up to this task? As I'm not sure if a
    MOSFET,
    > Transistor, etc. would be best suited for this.
    >
    > The reason I need to use linear instead of direct PWM is that the fans on
    > tach feed back on them and the tach sensors share common power and ground
    > with the fan's motor. When I use PWM and try to read the tach I end up
    > reading the PWM instead of the tach. It might be worth noting I'm using
    one
    > of Al Williams's excellent PWM coprocessors to do the PWM work for the
    > stamp. Any help is appreciated!
    >
    > Thanks,
    > Chris Shuster
    >
    > [noparse][[/noparse]Non-text portions of this message have been removed]
    >
    >
    > To UNSUBSCRIBE, just send mail to:
    > basicstamps-unsubscribe@yahoogroups.com
    > from the same email address that you subscribed. Text in the Subject and
    > Body of the message will be ignored.
    >
    >
    > Yahoo! Groups Links
    >
    > To visit your group on the web, go to:
    > http://groups.yahoo.com/group/basicstamps/
    >
    > To unsubscribe from this group, send an email to:
    > basicstamps-unsubscribe@yahoogroups.com
    >
    > Your use of Yahoo! Groups is subject to:
    http://docs.yahoo.com/info/terms/
    >
    >
    >
    >
    >
    > To UNSUBSCRIBE, just send mail to:
    > basicstamps-unsubscribe@yahoogroups.com
    > from the same email address that you subscribed. Text in the Subject and
    Body of the message will be ignored.
    >
    >
    > Yahoo! Groups Links
    >
    > To visit your group on the web, go to:
    > http://groups.yahoo.com/group/basicstamps/
    >
    > To unsubscribe from this group, send an email to:
    > basicstamps-unsubscribe@yahoogroups.com
    >
    > Your use of Yahoo! Groups is subject to:
    > http://docs.yahoo.com/info/terms/
    >
    >
    >
  • ArchiverArchiver Posts: 46,084
    edited 2004-01-14 06:11
    Dave,

    And it's not that simple. I'm dealing with standard PC case fans. The tach
    sensor and the motor share the same power supply on a circuit board that I'd
    have to physically break the fan to get to, which makes it fairly impossible
    to modify the fan. They sell fans that have little pots attached to them
    for speed control w/o affecting tach I'm sure I could chop the pot off and
    use the connections for PWM to make it work but that would require buying
    all new specialized fans. I'd prefer to be able to just use standard 3 wire
    fans.

    As an added note I did not mention earlier the fans mainly need to be
    operated in this fashion... Off or 7 to 12 volts.

    Chris Shuster

    Original Message
    From: "Dave Mucha" <davemucha@j...>
    To: <basicstamps@yahoogroups.com>
    Sent: Tuesday, January 13, 2004 8:59 AM
    Subject: [noparse][[/noparse]basicstamps] Re: PWM to Linear


    >
    >
    > AS you can see from Al's excelent write-up, there is no simple
    > answer. that makes me ask more about the tach. Any transistors you
    > use will have filtered DC power, so the tach should be able to be run
    > off the non-pulsing power in the circuit.
    >
    > If the tach is just monitoring voltage, then it is actually easier to
    > offer a smooth low current voltage for the tach and run the motor on
    > pwm.
    >
    > As I see it, you can create a clean high amp power for the motor or
    > you can create clean low amp power for the sensor.
    >
    > Dave
    >
    >
    >
    >
    >
    >
    > To UNSUBSCRIBE, just send mail to:
    > basicstamps-unsubscribe@yahoogroups.com
    > from the same email address that you subscribed. Text in the Subject and
    Body of the message will be ignored.
    >
    >
    > Yahoo! Groups Links
    >
    > To visit your group on the web, go to:
    > http://groups.yahoo.com/group/basicstamps/
    >
    > To unsubscribe from this group, send an email to:
    > basicstamps-unsubscribe@yahoogroups.com
    >
    > Your use of Yahoo! Groups is subject to:
    > http://docs.yahoo.com/info/terms/
    >
    >
    >
Sign In or Register to comment.