Shop OBEX P1 Docs P2 Docs Learn Events
Lm2940 — Parallax Forums

Lm2940

Bill ChennaultBill Chennault Posts: 1,198
edited 2007-09-22 19:11 in BASIC Stamp
All--

On the LM2940 datasheet, there are two filtering capacitors. (Forgive my lack of an ability to use the correct terms and I would appreciate correction.) One is a .47uF capacitor between Vin and ground and the other is a 22uF capacitor between Vout and ground.

The Parallax BOE Rev. C schematic shows a 1000uF capacitor between Vout and ground.

Folks on this board have recommended I use a 1000uF capacitor between Vout and ground.

My first experiments will DEFINITELY be congruent with the folks on this board and the Parallax schematic. But, my curiosity forces me to ask why the vastly different values on the data sheet? Are the extreme low-end values necessary to make the part function correctly in minimal conditions? (Also, should my capacitors be 100v devices?)

Another question, if I may?

Say I want TWO amps of continuous service. Can I use two LM2904's and two 1000uF capacitors to achieve this outcome?

Thanks!

--Bill
ps I bought six LM2904s from Parallax. They arrived today.·This may not have been their biggest sale. But, I got to talk to Wendy so it was a big deal on my end. (Thank you, Wendy!) Tomorrow, depending on what you say here, I am going to Radio Shack for capacitors.

▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
You are what you write.

Comments

  • Mike GreenMike Green Posts: 23,101
    edited 2007-09-20 23:31
    The capacitors mentioned on the datasheets are usually minimum values.· Although some devices will not work properly if these are increased indefinitely in size, most regulators (like the LM2940) will have better transient response if these are a lot larger than the datasheet would indicate.· Keep in mind that tantalum capacitors will provide excellent filtering at lower values than aluminum electrolytics.· In the case of servos, the higher value provides more current initially for the peak draw of the motor (than a lower tantalum value).

    In order to parallel regulators, the special quirks are unique to each regulator family.· Because there's some slight variation from regulator to regulator, you can't just parallel them.· Sometimes a small value series resistor in the output leads will help.· Sometimes you have to connect the regulator inputs (sensing terminal)·through a "balancing" circuit.· Check the datasheet and application notes.· I haven't worked that much with the LM2940 myself.· Perhaps someone else has experience in parallelling this particular regulator.
  • Bill ChennaultBill Chennault Posts: 1,198
    edited 2007-09-20 23:53
    Mike--

    Thanks a lot, as always.

    I have read about transient response and its relationship to capacitor size. I freely admit that I did not understand much.

    IF I want more current available, THEN I will figure out how to size the voltage regulator properly. (So, everyone should be aware of more ignorant questions!)

    --Bill

    ▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
    You are what you write.
  • RDL2004RDL2004 Posts: 2,554
    edited 2007-09-21 00:19
    Generally speaking, it's considered to be "not a good idea" to parallel 3 terminal regulators. It can be done, but I have no personal experience with the 2940s.

    If you need 2 amps and its not for one single device, then you could split the current between several regulators - for instance if you had 2 motors that each needed 1 amp, just feed each motor with its own regulator. You could also look into whats known as a "series pass transistor".

    Also, when working with 3 terminal regulators be aware of power dissipation ( V in minus V out times the current in amps = watts), these little regulators are often more limited by power dissipation than by current handling capability because they basically take the excess voltage and turn it into heat. Do the math and imagine them as tiny incandescent lamps [noparse]:)[/noparse] They don't like to get too hot.

    ▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
    - Rick
  • Mike GreenMike Green Posts: 23,101
    edited 2007-09-21 00:22
    From the Wikipedia:
    Capacitance (Farads) times Voltage (Volts) = Charge (Coulombs)
    and
    Current (Amperes) = Charge (Coulombs) / Time (Seconds)

    substitute to get

    Current (Amperes) = Capacitance (Farads) * Voltage (Volts) / Time (Seconds)

    solve for time:

    Time (Seconds) = Capacitance (Farads) * Voltage (Volts) / Current (Amperes)

    For 1000uF capacitor at 5V with an instantaneous current of 1A, you'll get current flow under ideal circumstances for 5ms. You'd need 20,000F to supply current for about 100ms. That's not hard to supply with a small SuperCap. [noparse][[/noparse]oooh, math is fun!]
  • allanlane5allanlane5 Posts: 3,815
    edited 2007-09-21 00:26
    Except SuperCaps have HUGE internal resistance, so they won't source (or sink) 1A for any amount of time.
  • RDL2004RDL2004 Posts: 2,554
    edited 2007-09-21 01:10
    FYI, from the data sheet:
    attachment.php?attachmentid=49446

    attachment.php?attachmentid=49447

    ▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
    - Rick
    898 x 253 - 16K
    419 x 416 - 12K
  • Bill ChennaultBill Chennault Posts: 1,198
    edited 2007-09-21 03:22
    RLD2004--

    I have absorbed that datasheet you posted to the limit of my limited ability. My remaining question is, Does the 1000uF capacitor need to be a 100v device or something else?

    Do you know what Mike and allanlane5 are talking about? Surely, they were not talking to me! [noparse]:)[/noparse] (Maybe Mike told me to use a 100v capacitor?)

    Supercap? I read about an Ultra Capacitor the other day, but it powers a car. I don't need quite that much.

    I have had a BS2 OEM and a BS2 Stamp module running for over a week on my dual, 4200 aH, 7.2v, parallel battery packs. They are in my "master-slave" configuration doing something extremely sophisticated; blinking an LED. But, it is fast and only takes a single pin and no SEROUT/SERIN statements. (I think this will be good for my application.)

    The battery packs mentioned above started out with a full charge. I measured them at 8.4vdc. (Gee. This newbie would have expected something like 7.2vdc, but I guess not!) After a week, they are down to 7.6xvdc . . . still above full rating.) I don't know if I have either the time or the inclination to run it much longer. Obviously, they are going to power four or five Stamps of whatever variety I choose for a few hours.

    --Bill; aka Hyper Capacitor

    ▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
    You are what you write.
  • Mike GreenMike Green Posts: 23,101
    edited 2007-09-21 03:49
    Bill,
    The 1000uF capacitor has to be rated for a voltage at least as high as what you're going to apply to it. Assuming you have a 5V regulator, you will need a 5V or 6V rating for the capacitor. You may find one at 6V, if not, 10V will do. You don't want the capacitor to have too high a rating (like 100V) since the higher the voltage rating, the physically larger the capacitor (for the same capacitance).

    The SuperCap capacitors (as allenlane5 mentioned) have a high internal resistance and won't really work for this application. I was just having fun with numbers.

    Battery packs are really a great way to power motors. They have a high discharge current. Some of them can be float charged ... like Gel Cells and (to some extent) NiMH cells without significant deterioration. Lithium cells last longest with only partial discharges and a limited number of charges, but need complex charging and protective circuitry.
  • Bill ChennaultBill Chennault Posts: 1,198
    edited 2007-09-21 13:00
    Mike--

    Thank you. Great explanation for me.

    I belong to many forums concerning many totally unrelated subjects. The folks on this one are the nicest and by far the most helpful.

    --Bill

    ▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
    You are what you write.
  • allanlane5allanlane5 Posts: 3,815
    edited 2007-09-21 18:17
    Hmm. 4800 amp-hour battery pack, eh? (Well, really, it's more likely a 4800 milli-amp-hour battery pack).

    At 6 mA average, they should last for... (4800 / 6 == 800 hours. 800 hours / 24 == 33 1/3 days.) thirty-three days. If you parallel two of those puppies, you should get 66 2/3 days. Very nice.
  • Bill ChennaultBill Chennault Posts: 1,198
    edited 2007-09-21 18:25
    allanlane5--

    That is the second time I've made that error. I caught it the first time in another thread.

    Thanks for catching this one!

    --Bill

    ▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
    You are what you write.
  • Bill ChennaultBill Chennault Posts: 1,198
    edited 2007-09-21 18:33
    Mike and allanlane5--

    All Radio Shack had were 1000uF, 35 volt, electrolytic capacitors. (Guess those are the most common movers for them.) Anyway, they appear to work fine except, as Mike mentioned earlier, they are physically too large. In the same breadboard powered by the dual 4200 maH battery pack (ha! Got it right that time!) that is currently running two BS2s in my master/slave test, I built a voltage regulation circuit that is currently running a BS2px24 . . . blinking an LED. I know these things are dirt-simple, but everyone must do them the first time and this is mine!

    Now, I need to build the RS232 connectors. I have .1uF, 50 volt ceramic capacitors. Will they work in place of the 100 volt capacitors specified in the various Parallax schematics for the RS232 connector?

    Thanks.

    --Bill

    ▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
    You are what you write.
  • allanlane5allanlane5 Posts: 3,815
    edited 2007-09-22 00:52
    Yes. Most 'modern' RS-232 circuits work around +- 9 volts. Even "classic" RS-232A was +- 12 volts. So a 50 volt capacitor should work fine.

    Note I once put a 500 volt capacitor on a 115 VAC line (to couple X10 signals across). Worked great, until there was a lightning strike, and the capacitor went off like a fire-cracker. Apparently, a surge protector in those applications helps a lot. On the other hand, the Cap wasn't that expensive...

    It's also nice to parallel a small capacitor (22 uF) with a larger one (100 uF or more). The smaller capacitor tends to catch 'small' quick transients, while the larger capacitor will work for longer transients.
  • Bill ChennaultBill Chennault Posts: 1,198
    edited 2007-09-22 19:11
    allanlane5--

    Thanks . . . and thanks for that hint about pairing caps!

    --Bill

    ▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
    You are what you write.
Sign In or Register to comment.