-1

I was looking at a selection of sensors on Digi-Key and I notices that while there are definitely a lot of expensive ones, there's also plenty of really cheap ones. For example, there's a gyroscope chip available for 1.61 euros, a current sensor available for the same price and many more other chips, be it sensor or not, that are available with same prices or lower.

The shop also has to profit, so I assume the cost of making those chips is actually even smaller. Why are they so cheap? How do companies that design and manufacture them manage to turn a profit on the chip sales?

JRE
  • 67,678
  • 8
  • 104
  • 179
Mu3
  • 482
  • 1
  • 4
  • 15
  • When you make stuff in the millions, it gets cheap to make. If you sell them and other chips in the billions, even <1p profit per chip quickly adds up. – Tom Carpenter Sep 07 '17 at 09:19
  • 5
    It's not chips that are cheap. It's money that's expensive. – Schizomorph Sep 07 '17 at 09:21
  • What makes you assume that these ICs should be more expensive ? How much do you think they should cost ? IC production relies on mass production, high volumes, low cost per unit. To really understand this requires you to study all the aspects involved many of which you can only learn when working in or studying the Semiconductor industry. – Bimpelrekkie Sep 07 '17 at 09:21
  • @Bimpelrekkie the factories that produce those chips are really high-tech, so I thought the costs of maintenance is very high - shouldn't that be reflected on the price? – Mu3 Sep 07 '17 at 09:34
  • 3
    Price is also relative. For me, a chip that costs 5 dollars and saves me hours of time on a project is cheap. For a guy who earns 5 dollars a day and has to feed his family out of that, that chip is damned expensive. – JRE Sep 07 '17 at 09:39
  • *so I thought the costs of maintenance is very high* **WHY ?** Maybe it is so high-tech it maintains itself ? (not true). What about the **cost** of purchasing the equipment, a modern Wafer Stepper costs tens of millions $, a whole IC production factory can cost many millions. But these small chips are made using older equipment which has often been written of already. The chips are **very** small, perhaps 10.000 fit on one wafer. One wafer including processing might cost $ 2000. That covers factory cost as well. You might have to buy 100.000 wafers per year to get such a low price though. – Bimpelrekkie Sep 07 '17 at 09:42
  • @Bimpelrekkie yeah, that's a good point. I didn't really think about reusing old machines. Thanks for the answer! – Mu3 Sep 07 '17 at 09:46
  • If you have to store ICs, storing them costs money. If you're in the business of selling ICs, (a distributor) that cost pushes your prices UP but if you're in the business of making things (and these ICs are excess stock for any reason) that cost pushes their value to you DOWN - and it's worth selling them cheap to save that cost. See https://electronics.stackexchange.com/questions/194914/how-are-some-modules-priced-ridiculously-low/195087#195087 –  Sep 07 '17 at 10:38

3 Answers3

2

Once the equipment is setup it is very easy to produce ICs in huge quantities as you can construct billions of transistors, for example, on a single wafer of silicon. You are right that the equipment is expensive and high maintenance, I remember reading that the estimated start-up cost of a cost-effective semiconductor plant is in the several billions of dollars, but if you think back to the automobile industry before Ford only the very richest people could have cars. Now cars are available to even some of the poorest in developed countries. This is really the result of mass production, and obviously the raw materials are far more expensive for cars than ICs.

jramsay42
  • 1,182
  • 1
  • 6
  • 11
2

How do companies that design and manufacture them manage to turn a profit on the chip sales?

Spoiler: they don't make a profit on all chips!

But in the end it is simple economics. If it costs you $5000 to get the machines, $1000 to configure, and $100 to run one hundred parts. Then making 1 part is $61 per part. Yet, making a thousand is $7 per part.

However, in order to compete some chips are sold without profit, or even at a loss. Simply because the market needs the parts, but the competition is strong. Diodes are an example of these parts.
The higher feature parts (eg: micro-controllers) can be sold at a higher price to compensate.

For large nanometer IC's these prices go in the 6 digits only for the lithography mask set. You still have to pay machine time.

An additional factor is that old equipment is not scrapped. It's sold.
The initial investment to develop the machine and get the supply chain is a large investment. But when you decide to retire or move on, this supply chain isn't going away. It becomes accessible to the rest of the world.

Jeroen3
  • 21,976
  • 36
  • 73
  • Can you shed some light why IC bases, connectors (screw terminal) are costly than ICs itself, is the material (metal) responsible for this or low volume? – Moons Mar 11 '20 at 14:47
  • @Moons Connectors require a complicated mix of disciplines to manufacture (metal, plastic, manufacture, testing). If you cannot share those disciplines with other parts in your catalog the parts will be expensive in low quantity. – Jeroen3 Mar 11 '20 at 20:15
0

Often the cost of testing the IC is higher than the other costs.

And for some ICs, the mechanical structure, the leadframe, costs more than the silicon.

How? old production equipment, long ago written off, needing only occasional maintenance.

analogsystemsrf
  • 33,703
  • 2
  • 18
  • 46