Subject:
|
Re: Power saving on L293D???
|
Newsgroups:
|
lugnet.robotics
|
Date:
|
Tue, 6 Jul 2004 19:50:25 GMT
|
Viewed:
|
1315 times
|
| |
![Post a public reply to this message](/news/icon-reply.gif) | |
In lugnet.robotics, Bert Weber wrote:
> Hi,
>
> I'm building my own controller to use with Mindstorms. I want to support 6
> motors, with 3 L293D IC's. But I found that this IC needs about 25 - 30 mA on
> the logic supply, when the chip is disabled. That means even if I dont use any
> motors the current is 75 - 100 mA!!! This is more than my µC with 128k SRAM!
> Is there a way to reduce the current? Maybe switching logic supply on/off of the
> L293D?
>
> Any ideas?
>
> Bert
Looking in an old Maplin catalogue from 1998, the entry for the L293D shows a
total quiescent logic supply current of 16mA per chip, so 48mA total.
Presumably you're using a 9V logic supply to match the RCX outputs?
The output supply current will be 2mA per chip when disabled, 16mA when enabled,
so there's a saving there when the motors are not in use. Since the data quotes
1.4V source dropout voltage and 1.2V sink dropout voltage at 1A output, you need
to add these to 9V to get the necessary output supply voltage - 11.6V.
It looks like the input and inhibit voltage levels are 0 to 1.5V for low and 2.3
to 7V for high - less than the 9V RCX output despite the logic supply voltage
being 4.5 to 36V, so you'll need to drop 2V with resistors or diodes.
You could switch the logic supply off as long as the output supply is already
off. However, it will need to be on before the driver chip is driven again.
For a university project I made my own H-bridges from transistors to be driven
from a micro board at 5 volts. This gets round the problem :-)
To design a more powerful motor driver to be driven from the RCX outputs I would
personally do the following:
1. Make a differential amplifier from an op-amp, with a gain of 0.5 or less, to
turn a signal of -9 to +9V into a single 0 to 9V signal. The output of this
will be 0V for full reverse, 4.5V for stop and +9V for full forward.
2. Make a window comparator where comparator A looks for the output to be below
+6V and comparator B looks for the output to be above +3V. This gives 0,1 for
+9V, 1,1 for 4.5V and 1,0 for 0V from A,B.
3. Make each comparator drive a pair of power transistors or MOSFETS (one side
of an H-bridge each) where a 1 from the comparator causes the lower one to be on
and a 0 causes the upper one to be on. Therefore 0,1 gives forward drive, 1,1
gives fast stop with both bottom transistors on and 1,0 gives full reverse.
N.B. I have not considered fail-safe design such that the motors turn off by
default!
You can get 4 op-amps in a 14-pin DIL chip and choose transistors to suit - you
don't need a high voltage rating but a current rating of at least 1-2A would be
useful as it might remove the need for a heatsink. The TIP31C is in a TO220
3-lead package suitable for use with veroboard and is rated at 100V, 3A, 40W. I
have used similar transistors successfully to drive Lego motors.
This scheme is OK for gearmotors, but train motors don't like PWM, so I would
integrate the RCX output to turn PWM into smooth voltage for that application.
Mark
|
|
Message is in Reply To:
![](/news/x.gif) | | Power saving on L293D???
|
| Hi, I'm building my own controller to use with Mindstorms. I want to support 6 motors, with 3 L293D IC's. But I found that this IC needs about 25 - 30 mA on the logic supply, when the chip is disabled. That means even if I dont use any motors the (...) (21 years ago, 6-Jul-04, to lugnet.robotics)
|
2 Messages in This Thread: ![Power saving on L293D??? -Bert Weber (6-Jul-04 to lugnet.robotics)](/news/x.gif) ![](/news/46.gif) ![You are here](/news/here.gif)
- Entire Thread on One Page:
- Nested:
All | Brief | Compact | Dots
Linear:
All | Brief | Compact
|
|
|
|