Atmel ATtiny84 based platform to learn AVR programming.
Already have an account? Log in.
To make the experience fit your profile, pick a username and tell us what interests you.
Schematic:
tinyDriver has the following specifications:
Applications:
2WD Robot with ultrasonic sensor:
A laser display that syncs with audio input:
After a bunch of tests and deliberation, decided to skip breaking out the motor driver pins. It's a "driver" after all. You still get a total of 6 pins broken out after disabling the RGB LED.
Based on some feedback (From Lee and others) , I've decided to add 5 jumpers to disconnect the motor driver IC when needed. Unfortunately this means another round of prototyping. Sigh.
I think I might add a MOSFET based Reverse Polarity Protection scheme...
Trying to decide between motor drivers. TB6612FNG or L293D? DRV8836 also looks interesting...
Thank you for the feedback. Will look into this.
You cannot simply disconnect the DRV8835 power without disconnecting all of its input pins as there are internal diodes to its VCC pin. There will be power leeching from I/O.
Datasheet: Section 6,
Digital input pin voltage: –0.5 to VCC + 0.5 V
OK, thanks for pointing that out. Maybe I'll remove that jumper, then.
Section 9 says:
"There is a weak pulldown resistor (approximately 100 kΩ) to ground on the input pins.
VCC and VM may be applied and removed in any order. When VCC is removed, the device enters a low power
state and draws very little current from VM. To minimize current draw, keep the input pins at 0 V during sleep
mode."
So at a 5V output on a pin, the 100K draws only 0.5 uA. I have been testing loading on all 5 pins with VCC jumper disconnected and so far it looks OK.
I am talking about internal ESD protection, not power down. If the part could handle signals without power applied, the maximum parameter would be written differently. They would given an actual voltage e.g. 5.5V. The +0.5V means a diode.
One thing you could do is to make sure that all those I/O pins are always driven low when you pull the jumper. If they are high, then the ESD diode would conduct.
http://www.ti.com/lit/an/slaa689/slaa689.pdf
Were any I/O pin high? If yes, Have you measured the I/O pins to the part when you do that? Measure the current from you micro to the I/O pin? Measure the voltage at the disconnected power pin?
Also read the fine print about not following the max parameters about the part might get damage?
You are right, voltage from output pins is coming into the VCC pin of the IC. I am thinking of adding AND gates on each of the 5 pins, with respect to VCC after the jumper. That should work?
Just using the power with AND gate can be tricky as the signal level isn't going to be well defined.
Or you an use 74LVC/AHC gates that are powered from the DRV8835 side of the jumper to buffer those signals. These 5V tolerant series are do not have the diodes to the VCC so they don't leak power thus they are designed to be use in situations like this.
I think I'll just go with jumpers to disconnect all lines. Thanks for your feedback.
Mahesh Venkitachalam
This is an interesting project. One suggestion: in some cases one may want to measure the current fed through the driver (I'm reluctant to feed current to any motor without doing that). The poor man's option is to put a very small resistance in series with the ground return from the H-bridges and guess the load current knowing the switching state. I know this drops the efficiency of the bridge further and in this case may mess the voltage of the ground reference of the DRV8835 but if you find this feature of interest you may consider leaving a gap with a 0 ohm resistance and a few traces that may be used in case someone wants to use this.