On using SN754410 in place of the L293D's...

[pogcarrOld]

So, just thought I would toss in my experience of trying a set of SN754410 chips in a Motorshield:
I have a set of stepper motors, Frame 17 size, 6V and 1A per coil. Kinda much for the L293D's, especially if you forget in your code to release() them. So, since I had them available, I figured I would put in a set of SN754410 that are rate 1A continuous and 2A momentary. Well...
They were momentary all right. They lasted about three seconds under load before one blew up! So, while these may be rated for 1A, they sure can't handle it in the current board design. I am thinking if you were to solder on some external kickback diodes they might be fine. 4 diodes per H-bridge, but I have no idea what spec diode to use. Me? I decided to order some more L293D's and try them piggybacked.

[ImaginaryAxis]

This question comes up a lot here on the forums. Previously, I mentioned that the SN745510 requires external protection diodes because the internal diodes are only for ESD. I also know there has been feedback from Texas Instruments themselves regarding this matter stating as such.

However, I no longer buy into it. As long as you adhere to the voltage and current ratings of the internal diodes you should be ok. External protection diodes are always an option due to the voltage and current requirements of your motor - meaning if they exceed the internal diodes then you must also use external diodes rated at least to the voltage and current of your motor.

I say that with confidence because I know that the people you talk to on the phone or email with TI are not the application engineers who work with specific devices. They are general support engineers who often never have direct experience with the part you are asking about. I am not knocking the tech support, after all, TI has over 25000 parts.

[adafruit]

that is not true, we specifically spoke to the applications engineer for that specific part about those specific diodes.

[ImaginaryAxis]

With all due respect, whom did you speak to? Unless you have an FAE in your area - most likely you do since you are in New York, and spoke directly to them, it is highly unlikely you spoke directly to a factory applications engineer - they are far busier dealing with customers like Sony, Samsung, HTC, Visteon, etc. etc. If you call in or email a question, those questions are dealt by front-end engineers who have no direct experience with the devices in question.

If this is the person whom you refer, yeah, I know Eddie, and it also confirms that you did not speak directly to a factory apps engineer. I don't want to belabor the point or knock TI support but that is the why it was for several years. TI has since moved to a forum based support now where, surprisingly, some of the Factory Apps managers chime in (Don Dapkus is one - FYI).

Point is, there is no difference in the L293D and the SN745510 besides marketing fluff on the first page, "ESD protection diodes." They both have clamp ratings and they will survive inductive kickback as long as the ratings are considered, but again, external diodes are often used as an extra measure of protection. One reason is that layout plays a roll and the IC design cannot account for all variations of PCB routing affects - it is far easier for a system designer to pick and choose the diodes to match his or her specific application.

[pogcarrOld]

OK, so I can state that definitively, there IS a difference. As I mentioned in the starting msg, using those high torque frame 17 stepper motors on a regulated 24V supply, the L293D work fine (albeit get a touch on the warm side and REQUIRE the release() be used after each movement), and the SN745510 lasts less than 5 seconds. That my friend is not marketing fluff, but a repeatable, real-world experience.

[ImaginaryAxis]

Electrically yes, conceptually no. If the ratings are kept within the electrical specifications of the device - specifically clamp ratings then the internal diodes offer adequate protection.

[tastewar]

...Point is, there is no difference in the L293D and the SN745510 besides marketing fluff on the first page...

Electrically yes, conceptually no.

Wow.

[chuckm]

OK, so I can state that definitively, there IS a difference. As I mentioned in the starting msg, using those high torque frame 17 stepper motors on a regulated 24V supply, the L293D work fine (albeit get a touch on the warm side and REQUIRE the release() be used after each movement), and the SN745510 lasts less than 5 seconds. That my friend is not marketing fluff, but a repeatable, real-world experience.

I never dispute facts, but I might call into question the interpretation of the facts. Have you conisdered the possibility that your application is outside the specified capabilities of both devices, however the L293Ds may have a wider margin under which they are tested?

--Chuck

[adafruit]

With all due respect, whom did you speak to? Unless you have an FAE in your area - most likely you do since you are in New York, and spoke directly to them, it is highly unlikely you spoke directly to a factory applications engineer - they are far busier dealing with customers like Sony, Samsung, HTC, Visteon, etc. etc. If you call in or email a question, those questions are dealt by front-end engineers who have no direct experience with the devices in question.

If this is the person whom you refer, yeah, I know Eddie, and it also confirms that you did not speak directly to a factory apps engineer. I don't want to belabor the point or knock TI support but that is the why it was for several years. TI has since moved to a forum based support now where, surprisingly, some of the Factory Apps managers chime in (Don Dapkus is one - FYI).

Point is, there is no difference in the L293D and the SN745510 besides marketing fluff on the first page, "ESD protection diodes." They both have clamp ratings and they will survive inductive kickback as long as the ratings are considered, but again, external diodes are often used as an extra measure of protection. One reason is that layout plays a roll and the IC design cannot account for all variations of PCB routing affects - it is far easier for a system designer to pick and choose the diodes to match his or her specific application.

great! if you get an official response from TI about the chip, we will update the page with the email or forum post

[ImaginaryAxis]

...Point is, there is no difference in the L293D and the SN745510 besides marketing fluff on the first page...

Electrically yes, conceptually no.

Wow.

Nitpick much? I clarified what I meant.

[ImaginaryAxis]

great! if you get an official response from TI about the chip, we will update the page with the email or forum post

I call into the technical support line and the "engineer" cannot answer the question so he will go "research" it. A few days pass and I receive an email stating "for technical questions please use the forum" and honestly after receiving that type of blow off email, I haven't bothered since.

But if I were to design a motor driver then I would adhere to the clamp spec and feel comfortable with that. If someone were to follow up again and ask for technical details of the clamp in the output stage besides "Can I do this" I would certainly like to know.

I understand your situation in that you received an official response from TI regarding one of your products and you will stick to that. No problem. But I know the front end technical support has little to know direct experience with the, admittedly large, product portfolio. They will answer conservatively in line with legal reasons rather than technical reasons - which I also understand.

[tastewar]

...Point is, there is no difference in the L293D and the SN745510 besides marketing fluff on the first page...

Electrically yes, conceptually no.

Wow.

Nitpick much? I clarified what I meant.

Well, if it's picking a nit to call out the difference between "marketing fluff" and electrical differences, then I guess yes. Yes I do.

[pogcarrOld]

Actually, I'd like to get back to one part of my original question: Not being an electrical engineer or having any real idea of how to do the calculation...
What external diodes should I use? Someplace someone mentioned Schottkey diodes... is that right? What size or rating should I look for?

I As for the great debate, am quite sure I am trying to drive the device WAY beyond anything the factory intended.
My experience has simply shown that the L298Ds seem to be able to handle being abused in the hands of dangerous, sophomoric tinkerers better than the SN754410s.