Wow I think this person may have voided their warranty

[shap]

...except D510 is definitely by far NOT the best low-power embedded system CPU that's out there.

My other OS guy and I were discussing just this in the car today. Started with me replaying for him Dan's correction of my ARM v ATOM performance statement. We wanted to understand whether the Akai type of product (where you plug in your iPad) is likely to emerge as a threat. Eric knows a lot more about the low-level hardware these days than I do. We don't think it happens soon, for two reasons:

• -Devices like iPad are extremely power constrained and therefore have limited memory bandwidth. Widening that bus would involve a significant power hit. It also involves a larger number of edge pins on the A5 package, which would result in a higher reject rate. Taken in context of the overall iPad target space, we don't expect the memory bandwidth issue to be a design focus for them. Of course, we're speculating.
  - ARM is fumbling very badly to get its multicore story working right. Fixing it requires changes to their hardware-level memory consistency model. Those changes entail cache design changes whose power budget implications that will be very hard to swallow. They're going to have to do it, but it may not happen until their 64-bit architecture emerges. They have a lot going on right now.

KRONOS is in a somewhat unusual position as an embedded system. Compared to conventional embedded devices, it's a very low-quantity design. It's thermal and power constraints are much richer then most embedded or hand-held devices. Finally, it needs the horsepower. Because of the memory model issues and the desire for concurrency, ARM isn't really an option.

I'm a little surprised that Korg didn't go fanless, but I'm sure they had a reason. Given the fan, yes, they might have gone with a low-power core-i3, but that would put them up into the desktop chipset pricing scheme. The ATOM design still has a lot of low-hanging performance headroom, so all in all, it's a good choice. And since Korg had most of this stuff running on the OASYS, and the ATOM 510 performance isn't that far off in terms of performance, choosing the lower performance processor (compared to core i3) wasn't an unreasonable risk.

Looking back now, in July 2011, it's easy to say the D525 might have been a better choice, but remember that the D525 was announced in August 2010, and they originally planned to ship in March 2011. Allowing for manufacturing lead time at KORG and availability ramp-up delay from Intel, that would have been pretty risky.

[madbeatzyo111]

I don't know if the D510 can go fanless at 13 watts TDP. Maybe with a massive heatsink and a lot of ventilation. Let's not forget about liquid cooling too

I suppose the Oasys owners didn't complain hard enough about their fan noise at the time, so Korg thought it was ok to put it in the Kronos too. It's the dangers of fanboyism; the honest criticism that could help make a better product in the future tend to get muted.

[NuSkoolTone]

I don't know if the D510 can go fanless at 13 watts TDP. Maybe with a massive heatsink and a lot of ventilation. Let's not forget about liquid cooling too

I suppose the Oasys owners didn't complain hard enough about their fan noise at the time, so Korg thought it was ok to put it in the Kronos too. It's the dangers of fanboyism; the honest criticism that could help make a better product in the future tend to get muted.

Liquid cooling? God no. Simply not worth the hassle and maintenance these days IMO. You still need a pump, and that makes noise too. As long as a low db fan is used, I don't see the issue.

[cello]

I suppose the Oasys owners didn't complain hard enough about their fan noise at the time, so Korg thought it was ok to put it in the Kronos too. It's the dangers of fanboyism; the honest criticism that could help make a better product in the future tend to get muted.

Or alternatively, OASYS users didn't complain about fan noise because there is none to complain about... My O fan can heardly be heard at all; I have to go the end grate to actually hear if it's running or not.

[synthjoe]

I don't know if the D510 can go fanless at 13 watts TDP. Maybe with a massive heatsink and a lot of ventilation. Let's not forget about liquid cooling too

I suppose the Oasys owners didn't complain hard enough about their fan noise at the time, so Korg thought it was ok to put it in the Kronos too. It's the dangers of fanboyism; the honest criticism that could help make a better product in the future tend to get muted.

Liquid cooling? God no. Simply not worth the hassle and maintenance these days IMO. You still need a pump, and that makes noise too. As long as a low db fan is used, I don't see the issue.

Heatpipes to the metal casing would have been a nice and modern, proven solution.

[madbeatzyo111]

Or alternatively, OASYS users didn't complain about fan noise because there is none to complain about... My O fan can heardly be heard at all; I have to go the end grate to actually hear if it's running or not.

It's interesting. The fan noise db level of the Kronos and Oasys were both measured recently and found to be similar (albeit weighted differently in frequency), yet there are much more complaints about the K fan noise. Searching through the O forums I found only 2 threads about it. So have people become pickier about fan noise in the last few years? Or were people just more forgiving of it in light of the sheer awesomeness of the O? Conversely are people less forgiving of the K? Something to ponder...

I don't have a Kronos myself (yet?) but I believe it when some people say they have issues with the fan noise. I don't think they are being overly picky; rather I think the fans being used have a wide variation in noise output. In other words, the specs are not very tight so some people get quiet fans and some get loud ones (29db +/- 10db for example). Maybe Korg just didnt specify a fan with tight enough specs. Then again tight specs always mean more cost...