 |
| Author |
Message |
|
|
 Posted: Jan 18, 2012 - 02:17 AM |
|
Joined: Jan 30, 2010
Posts: 30
Location: Kitchener, Ontario
|
|
I have an h-bridge chip that I'm looking to implement on a robot controllers, and it has a few nice features, the L6205D. It's small, surface mount, has two channels and handles pretty decent amperage. Since my robot stalls at just under 2.5 amps/channel it seems perfect. Unfortunately it's one of these in1A in2A drive schemes where it seems to want the pwm on both input lines per channel, or in other words would overall need 4 pwm lines. It would work, but that seems sloppy.
I can implement a pwm/dir type pinout with some nand logic without too much difficulty, but board space is a serious concern. I would love to just pwm the enable pin and use a not gate for direction. Does this seem reasonable? I'm not sure from the datasheet if the reset timing would be too slow for a proper pwm, but it looks like it may be. I would also lose the option of breaking vs coasting, but I'm not sure that it really matters. From the truth table it looks like the disable goes into a tristate instead of a double ground motor braking state so that may even effect the pwm performance, but again no clue how much.
Does anyone have good experiences with this chip or is it pretty unknown? Should I use something different?
Thanks for any input, it's a huge help! |
|
|
| |
|
|
|
|
|
|
Posted: Jan 18, 2012 - 02:39 AM |
|
Joined: Sep 04, 2002
Posts: 21257
Location: Orlando Florida
|
|
| You can use the top fets for direction, enable one or the other, and the bottom one on the other side for pwm. |
_________________
Imagecraft compiler user
|
| |
|
|
|
|
|
|
Posted: Jan 18, 2012 - 02:46 AM |
|
Joined: Jan 30, 2010
Posts: 30
Location: Kitchener, Ontario
|
|
| Sorry I guess I should specify (though that is indeed useful if I end up just making my own h bridges) that it's two inputs per channel, so the two "A" inputs and two "B" inputs are already paired, with logic to prevent shoot through if both are enabled at once. but instead of having two inputs for dir and pwm it has inputs for forward and reverse. |
|
|
| |
|
|
|
|
|
|
Posted: Feb 15, 2012 - 04:45 AM |
|
Joined: Jan 30, 2010
Posts: 30
Location: Kitchener, Ontario
|
|
| For anyone else curious, it's definitely possible with a couple of quad nand chips, but to run two channels on an A/B input scheme needs 5 nands per or 3 chips total. Best bet, swap for the freescale MC33926 as it has PWM inpust, A/B inputs that can be PWM'd and an inverse pin. Almost a confusing number of control options. Only problem is needing two controllers instead of one, but still a net gain in board space and a nicer layout. |
|
|
| |
|
|
|
|
|
|
|
FAQ
Search
Register
Log in to check your private messages
Log in
Maximize
RSS
