Hello,
There is a little debate going around on using the L293D and SN754410 H-bridge for driving motors. Both devices are used on the Adafruit Motorshield, but there is some disagreement as to the direct replacement of the L293D with the SN754410 due to the inductive kickback of a motor.
The L293D calls out on the front page of the datasheet that the internal diodes can protect the device from inductive kickback. Figure 3 shows the L293D driving a motor with no external protection. The SN754410 only mentions "ESD diodes" and Figure 3 shows the SN754410 driving a motor with external protection.
On the surface this makes sense. However, the clamp ratings in the specification table for the L293D are 600mA with a forward voltage of 1.3V and the SN745510 are rated higher at 1A with a higher forward voltage of 2.5V maximum.
I would assume (if I am allowed) that if a given application used the L293D and any inductive transients where within the stated ratings then the SN754410 would have no problem either in that same application with no external protection (same layout, supply levels, motor, etc.)
The equivalent output structures are also illustrated to be the same in the respective datasheets.
Is this a matter of additional testing (L293D says High Reliability), process limitation, lack of datasheet updates (they are old parts)? Can you help shed some light on this matter?
-Ken
Hi Ken,
I may not be the definitive holder for the answer you seek, but here is some input to continue the debate:
1. Seems to me both devices have diode outputs which would take care of recirculating current when the device enters asynchronous fast decay, so as long as the currents are not too high, they should both be able to withstand the transient.
2. The reason why they may recommend external diodes could easily be for thermal dissipation constraints. When you do asynchronous fast decay, the current will flow through the diode. However, this implies a not so efficient path for the current, which in essence means more heat. If you want to remove some of the heat to outside of the package, using external diodes is a fairly common technique. This only applies for asynchronous decay modes, however. For synchronous mode we would use the FETs, so there is no way to remove the associated heat to outside of the package.
These devices are rather old and to fully determine the "back then" reasoning can be quite tough. Luckily, there is no need to suffer! Because today we have way much more efficient motor drivers and I invite you to look at the DRV8x devices which will allow you to control Brushed DC, bipolar steppers, BLDC, PMSM motors with a fair variety of voltages and current capabilities. The parts are recent enough that we should be able to answer any question, as well as settle any debate, you may have ;-)
Hope the info helps. Best regards,
Jose Quinones
Ken,
Please check out www.ti.com/motor for the latest products from TI for motor drive and control. We have an extensive portfolio to choose from.
Regards,
Ryan
Ryan Kehr
Motor Drive Application Manager
Thanks for the input. It is what I expected that both can withstand commutation currents from inductive loads providing that the specifications are adhered to. Obviously easier said than done and can be quite complicated: motor, layout, component ratings, etc.
The DRV8xx family may be a little too complicated for the application, but it is a good alternative for those who feel comfortable with them.
I understand some of the DRV8x devices may seem intimidating at first, but there is a great deal of them that make life considerably easier! I have done designs where I have 2 control lines and the device does the rest. The devices are more complex inside, but that is what makes it much more simpler outside.
Please, do not hesitate to contact us if you still feel there are ways to make it easier, as that is what we strive for!
Best regards,