The Dual Stepper Driver Conundrum

One of my readers was kind enough to comment on the current affairs on stepper drivers and how many we can cram on a single device. Jim is one of the many stepper driver users who like the idea of using a device which has two stepper drivers, as opposed to two devices with a single stepper driver each.

This is a question I see a lot and in fact we may get greedier and ask, why not place three stepper drivers in one chip? Heck, why not place four stepper drivers in a single chip?

Whereas those two last questions may seem sarcastic, I want to point out they aren’t. Had I ask why don’t we place 50 stepper drivers in a single chip, then I would have been the ultimate sarcastic idiot, but the truth is a chip with 3 or 4 stepper drivers would actually come quite in handy!

I can immediately think of two applications which could benefit from a 3 or 4 stepper driver chip: CNC machines and 3D printers.

Today we tackle these applications with multiple modules. But wouldn’t it be awesome if we had a single chip capable of driving all the motors on our machine? Is it even possible? Well, it is quite possible as the laws of physics do not forbid it.

Then, why aren’t there any quad stepper drivers out there?

Whereas the laws of physics do not forbid it, they certainly don’t make it easy. But the real culprit here is not the laws of physics as much as it is the laws of economics. In fact there are two main reasons why I can assure you it will be decades before we see a commercially available chip with all this greatness, and it all boils down to the “greens”! Not talking about pees… OK, that joke may only make sense in North America (as in USA), so I am talking about money!

The first immediate problem is cost. For any of the semiconductor companies out there to come up and design a quad stepper driver chip, the investment would need to make sense.

Contrary to what many people may believe, making chips is not cheap at all! Sure, you log into digikey and you find the chip that you want and it hardly is $30,000. Heck, it is not even $500.00. Maybe the latest CPU is in the hundreds, but most chips out there are either in the cents or couple of dollars range. As a result, it may be tempting to think that making these chips costs very little.

The truth is, however, that making a chip costs MILLIONS of dollars. Not one, not two but way much more than that. The cost of making (and marketing) a stepper driver could easily be somewhere in between 1 million and 10 million dollars, depending on complexity and feature set. How on Earth can it be sold for $5-$15? Well, because ideally, millions of these will be sold and eventually not only the cost will be recuperated, but a profit will be made in the process.

So now it all becomes a question of how much will it cost me to make this chip and how much will I be able to make after selling a bunch.

Oversimplifying this tale, I can tell you that if you are going to spend 5 million dollars to make 5 millions dollars, I assure you NOBODY on this planet will go through the hassle. To make 10 millions? OK, it now starts to make a little bit more sense. To make 20 or 50 million? SIGN ME UP! That’s a sound investment!

Here is the deal. Say I design this quad stepper driver and I spend 5 million dollars in the process. I of course would need to cash in multiples of that, which means I would need to find many applications which could employ them. Since we are parting from the premise this chip would be GREAT for CNC machines and 3D printers, we would need plenty of those to get this investment cranking.

INTERRUPTION: Notice a 3D printer using 4 individual stepper driver chips help my economics as 1 million 3D printers represents a 4 million chips order. However, if I cram everything into a single chip, 1 million 3D printers represent a mere 1 million chips. Not good for any of our economic models.

And I go on…

I don’t want to bug you down with details on how nobody ever gets 100% market share and such, but let’s just continue making up numbers and say that for me to be able to sell this chip enough times to recoup my investment (and my sanity), I would need the whole planet to be requiring between 1 and 2 billion CNC machines and 3D printers.

Sounds like a lot! If it was smart phones, it would be believable. But CNC machines? It would require for pretty much every single household on this planet to have one and we all know that is not going to happen. I don’t think we all need a small CNC milling machine to make PCB’s and robotic parts.

And so no company in their sane mind would dare to design such a chip!

The only way in which I see this ever becoming a reality is when Skynet becomes online, takes over the world and we are all transformed into Terminators (if we are that lucky…). Now it would make perfect sense to design this chip because pretty much everybody will have a CNC machine at home. Heck, how do you pretend we are going to reproduce? I bet we will all want Baby Terminators!

Skynet would also not be interested in satisfying stock owners, as its sole goal would be total world domination! So spending 5 million dollars to make 0 dollars makes as much sense as it does to need all of those CNC machines cranking out more Terminators to aid in its total world domination goal.

So say Skynet needs this quad stepper motor driver. It doesn’t care about money because it will create a money-less world. It doesn’t care about resources because Terminators are the perfect moon and asteroid miners. They are impervious to radiation and can work without an oxygen rich atmosphere. Solar System exploration solved in a single shot! But skynet still has one dinky little problem…

Putting four stepper drivers in a single chip is 16 times harder than putting a single stepper driver into one! How could this be?

Well, like I said before, whereas the laws of physic doesn’t forbid this project, they still make it hard! And the problem at hand here is one pesky little equation I have named countless times:

P = I^2*R

Putting 4 stepper drivers in a single chip means you would need to pack 32 high current and high voltage FETs, which is the same as 8 Full H Bridges. WHAAAAAAAAAAAAAAAAAAAAAAAT!!!!

That puppy is going to get so hot, Skynet might as well use the chip to get some fusion started.

For human based enterprises, this problem alone is a deal breaker. Even, if every single human being does decide they want to own a CNC machine per household, how much would they be willing to pay? In order to make this chip feasible, the FETs would need to be huge in order to decrease the RDSON to a level in which driving all steppers at once would generate very little heat dissipation. It would not be (AND IT ISN’T) cost efficient.

A chip like this could be way much more than $5 * 4. Chances are a chip like this would be in the $50 to $100 range just due to silicon costs. Not to mention the device would be humongous…

Again, for Skynet this is not a problem because Skynet would have no issues with turning all of our beaches and deserts into one big lump of silicon from which massive dies could be constructed. These silicon wafers could look like flying saucers! (A joke possibly only semiconductor folks will get).

This is why when I see a movie where a computer is trying to take over the world and the humans fight it until they win, I am so disappointed. I tell you! If you want truly cool chips to become widely available, let the machines win!

And so the moral of the story is that if a T1000 or a Borg knocks on your door with an offer to add your consciousness into their collective, it is not only futile to resist, it is in our best interest to comply!

PS: Skynet, if you are reading this post, me and my family will HAPPILY join your collective!

Leave a Reply

Your email address will not be published. Required fields are marked *