Can anyone who's done outsourced hardware comment on how hard that part actually is? I mean, if I met with someone from Foxconn and said, "listen, I need to you to take this 3d model of a plastic case, slam in an Android SoC and some I/O ports and a power supply and have test units ready in two months and another 50,000 ready to go by March, I can give you $80 a piece up front" -- could I plausibly get something of decent quality? What does this process actually require on the retail end in terms of time and money and expertise?
Seems to me the controller might be the hard part -- more UX testing, less standard parts. But I'd love to hear from someone who's solved this kind of problem.
It's somewhat amusing to me to see several comments on this thread wave away the actual hardware as an "outsourced design" or "just call Asus". I have to assume it's a result of people not understanding the complexity of modern hardware.
First, I've been unable in my (admittedly, cursory) research to discover any prebuilt systems that even come close to their promised hardware specs. So until someone proves me wrong by posting a link to an OTS solution even close to what they're promising, lets assume that the system needs custom design.
This is a whole massive can of worms. I have some experience with hardware design (though not as much as a full-blown EE), and I'd ballpark hardware design alone as a half-year effort with multiple full-time EEs. And that's a best-case optimistic estimate. People who work in software don't seem to understand how hard this stuff is. It's not like you can simply take another schematic, slap the new parts you want on, delete the old parts, autoroute it and call it a day. With parts of this complexity there's a whole host of other design issues that come out of the woodwork. High-frequency switching means if one of your autorouted traces is longer than another, you're going to have timing issues. Crosstalk between layers, capacitance and inductance of the wires themselves can wreak havoc on your busses. You need to intelligently design the ground planes, separate your signals (analog and digital components), etc etc etc.
These are only the issues I'm aware of. I'm sure people experienced with creating boards this complex can fill me in on more. It's not like a hobby arduino project - hooking up traces to debug problems can increase the capacitance of a line enough to cause a new set of problems.
This isn't really directed just towards you, but it feels like a ton of people here are missing how difficult hardware design with components this complex is.
Price point: $384. With an LCD. This isn't even close to comparable to the product they're trying to make. So to close the gap, you say:
> Drop the camera, the touch screen, the battery and the enclosure, I bet you I could get a deal for these boards well under $50 for 10,000+ lots.
Engineering tightly integrated consumer electronics products like these isn't a menu. You can't just "unplug" the monitor "remove" the battery and then repackage it. Redesign of a processor board of this complexity isn't a picnic; it takes significant EE expertise.
And with an initial price point almost 4x that Ouya promises? Please. This isn't "comparable" at all. You want a comparable piece of hardware? See http://trimslice.com/web/order-direct. Twice the price point with significantly less horsepower. That's what people are probably going to get in the end with this project.
Firstly, the question was "there is no comparable system on the market with those specs," I was pointing out there was a system with those specs, which has already been cloned.
Secondly, the screen, battery, and about 50% of the time the camera, are NOT integral on the board. They plug into headers. I'm not sure who's misinformed you of this, but not everyone designs like Apple. And yes, I can "just remove the screen, battery, and likely the camera." I just don't plug them into the headers. How many EEs do you think are working in China for these manufacturers?
Thirdly, it's interesting that you have both insight into the margins of several different hardware manufacturers and also into the internal R&D discussions of Ouya to know that they are going to end up selling a $200 Tegra2 instead of a $99 Tegra3 device. Gosh, they should've just hired you, no?
Why so aggressive arguing that they can't do it? It's one thing to be suspicious (and I, am suspicious as to their ability to hold reasonable overall margins unless they intend to make over 50k of these), it's another to spread FUD that something is impossible.
You CAN'T just unplug the screen, battery, and camera from the board and then stuff it into a new form factor. The board will have to be redesigned to fit into a new form factor and accommodate new peripherals. This will take significant time and effort, regardless of whether the engineering is done in the US, China, or Zimbabwe.
> Thirdly, it's interesting that you have both insight into the margins of several different hardware manufacturers and also into the internal R&D discussions of Ouya to know that they are going to end up selling a $200 Tegra2 instead of a $99 Tegra3 device. Gosh, they should've just hired you, no?
This is just silly. I point out a system on the market with much closer specs to what Ouya is trying to accomplish (no screen, similar outputs, etc) and point out that's still twice their price point. I have seen no evidence from Ouya, you, or anyone else that they can do anything different. Until then, I'll remain skeptical that they'll deliver on their promises.
> Why so aggressive arguing that they can't do it?
It'd be awesome if they pulled it off; I'm just pointing out that they probably won't. And if I came off as aggressive, it's because of all the armchair engineers around here who think consumer hardware design is a walk in the park that can be completed on a whim in an instant.
> it's another to spread FUD that something is impossible.
Go find where I said "impossible". My entire post just points out how unlikely it is this project is going to deliver what it promises.
Comments
Can anyone who's done outsourced hardware comment on how hard that part actually is? I mean, if I met with someone from Foxconn and said, "listen, I need to you to take this 3d model of a plastic case, slam in an Android SoC and some I/O ports and a power supply and have test units ready in two months and another 50,000 ready to go by March, I can give you $80 a piece up front" -- could I plausibly get something of decent quality? What does this process actually require on the retail end in terms of time and money and expertise?
Seems to me the controller might be the hard part -- more UX testing, less standard parts. But I'd love to hear from someone who's solved this kind of problem.
It's somewhat amusing to me to see several comments on this thread wave away the actual hardware as an "outsourced design" or "just call Asus". I have to assume it's a result of people not understanding the complexity of modern hardware.
First, I've been unable in my (admittedly, cursory) research to discover any prebuilt systems that even come close to their promised hardware specs. So until someone proves me wrong by posting a link to an OTS solution even close to what they're promising, lets assume that the system needs custom design.
This is a whole massive can of worms. I have some experience with hardware design (though not as much as a full-blown EE), and I'd ballpark hardware design alone as a half-year effort with multiple full-time EEs. And that's a best-case optimistic estimate. People who work in software don't seem to understand how hard this stuff is. It's not like you can simply take another schematic, slap the new parts you want on, delete the old parts, autoroute it and call it a day. With parts of this complexity there's a whole host of other design issues that come out of the woodwork. High-frequency switching means if one of your autorouted traces is longer than another, you're going to have timing issues. Crosstalk between layers, capacitance and inductance of the wires themselves can wreak havoc on your busses. You need to intelligently design the ground planes, separate your signals (analog and digital components), etc etc etc.
These are only the issues I'm aware of. I'm sure people experienced with creating boards this complex can fill me in on more. It's not like a hobby arduino project - hooking up traces to debug problems can increase the capacitance of a line enough to cause a new set of problems.
This isn't really directed just towards you, but it feels like a ton of people here are missing how difficult hardware design with components this complex is.
Their specs read like I'm reading the ASUS TF300T tablet, and there are several clones on Alibaba right now. Here you go:
http://www.alibaba.com/product-tp/127889823/Asus_TF300T_B1_B...
http://www.alibaba.com/product-tp/127047344/10_1_TEGRA3_1_32...
http://www.alibaba.com/product-free/127258720/MID_Laptops.ht...
Drop the camera, the touch screen, the battery and the enclosure, I bet you I could get a deal for these boards well under $50 for 10,000+ lots.
Also: right now, they'd have no problems getting a bridge loan for a few million dollars to start the manufacturing rolling.
None of the links you list have prices available. Let's go with Amazon:
http://www.amazon.com/Transformer-TF300-T-B1-BL-10-1-Inch-Ta...
Price point: $384. With an LCD. This isn't even close to comparable to the product they're trying to make. So to close the gap, you say:
> Drop the camera, the touch screen, the battery and the enclosure, I bet you I could get a deal for these boards well under $50 for 10,000+ lots.
Engineering tightly integrated consumer electronics products like these isn't a menu. You can't just "unplug" the monitor "remove" the battery and then repackage it. Redesign of a processor board of this complexity isn't a picnic; it takes significant EE expertise.
And with an initial price point almost 4x that Ouya promises? Please. This isn't "comparable" at all. You want a comparable piece of hardware? See http://trimslice.com/web/order-direct. Twice the price point with significantly less horsepower. That's what people are probably going to get in the end with this project.
Firstly, the question was "there is no comparable system on the market with those specs," I was pointing out there was a system with those specs, which has already been cloned.
Secondly, the screen, battery, and about 50% of the time the camera, are NOT integral on the board. They plug into headers. I'm not sure who's misinformed you of this, but not everyone designs like Apple. And yes, I can "just remove the screen, battery, and likely the camera." I just don't plug them into the headers. How many EEs do you think are working in China for these manufacturers?
Thirdly, it's interesting that you have both insight into the margins of several different hardware manufacturers and also into the internal R&D discussions of Ouya to know that they are going to end up selling a $200 Tegra2 instead of a $99 Tegra3 device. Gosh, they should've just hired you, no?
Why so aggressive arguing that they can't do it? It's one thing to be suspicious (and I, am suspicious as to their ability to hold reasonable overall margins unless they intend to make over 50k of these), it's another to spread FUD that something is impossible.
You CAN'T just unplug the screen, battery, and camera from the board and then stuff it into a new form factor. The board will have to be redesigned to fit into a new form factor and accommodate new peripherals. This will take significant time and effort, regardless of whether the engineering is done in the US, China, or Zimbabwe.
> Thirdly, it's interesting that you have both insight into the margins of several different hardware manufacturers and also into the internal R&D discussions of Ouya to know that they are going to end up selling a $200 Tegra2 instead of a $99 Tegra3 device. Gosh, they should've just hired you, no?
This is just silly. I point out a system on the market with much closer specs to what Ouya is trying to accomplish (no screen, similar outputs, etc) and point out that's still twice their price point. I have seen no evidence from Ouya, you, or anyone else that they can do anything different. Until then, I'll remain skeptical that they'll deliver on their promises.
> Why so aggressive arguing that they can't do it?
It'd be awesome if they pulled it off; I'm just pointing out that they probably won't. And if I came off as aggressive, it's because of all the armchair engineers around here who think consumer hardware design is a walk in the park that can be completed on a whim in an instant.
> it's another to spread FUD that something is impossible.
Go find where I said "impossible". My entire post just points out how unlikely it is this project is going to deliver what it promises.