Console design and manufacturing comparison - effective cooling yada yada *spawn

[goonergaz]

You lost me. What's stopping Sony or anyone else designing a unit that can be slipped out and in, like any removal drive bay? I'm not talking about the existing PS4 design, but a theoretical console design where several components (HDD, optical, PSU) are supplied in discrete, interchangeable units, making replacement as easy as replacing batteries or a SIM card. It'd add a little bulk to the unit and cost per item, so may not be the best idea, but it's surely possible and not difficult for either the engineers to design or the home user to use.

well duh, profit/money/however else you want to put it

Seriously tho, surely such a solution would cost more to implement (time and material) and need to be passed onto the customer - Sony wanted to keep costs down. Not only that but by making consoles easier to fix there is less need to buy a replacement when it dies...I'm sure they make more on a console than a drive

 

[Shifty Geezer]

The increased cost would/could result in decreased warranty costs. I know of three PS4's who all suffered a drive failure. That required Sony paying for the return of the old unit, fixing it, and replacing with a new. Sending a replacement user-serviceable drive unit would be cheaper. Just as we have the above discussion on XB1 PSU's showing that identifying and replacing a faulty PSU instead of a whole console is cheap and fairly convenient for all.

I'm not going to say it'd be net win because I've no idea what the economies are really like. It'd be interesting to hear though, and maintain that the idea is sound even if the economies make it the less preferable choice.

As for profit, you can charge a higher margin on a lower cost component like a drive. A $20 drive can be sold for $40-50 and the consumer will be happy as they aren't having to shell out $300 on a whole new machine, which itself might be sold without profit margin. Plus the user wouldn't have to do some dumb PS4 transfer hoop-jumping as they'll be keeping the same device, same HDD, same content already installed. Hence the markup comes with added value, allowing it to be pushed a little higher.

 

[Lalaland]

Speaking as an employee of a PC OEM any kind of tool-less design costs extra and in the oddest ways, I was discussing this before with someone who knows these things and asking them why we don't make 'is easy to service' as a USP actross all of our market segments (i.e. if I've misunderstood what they said please do contradict me below). For example we have tool-less hd replacement in more expensive notebook models but not in cheaper ones and at least part of the reason we don't is that the hdd itself can be used as a chassis stiffening element if it is firmly attached to the chassis by screws.If attached by a latch, or some other tool less method, the strengthening has to be a part of actual chassis which makes that more expensive and more complex to manufacture. As another example every gap or service opening in the base of a notebook imposes additional strengthening costs elsewhere making the unit ore expenive overall (as well as the fixed manufacturing costs of having a multi-piece chassis now).

Generally speaking in any IT kit the more tool-less and easy to repair a thing is the more expensive it is due to both the higher manufacturing cost and because you have de facto declared 'I have more pressing needs than capital cost and will pay for them' and thus will accept paying a higher margin. This can be seen across all IT from the cheap NAS boxes that are a jumble of sharp edged metal and screws to the lovely smooth yet crushingly expensive 12 bay units with nice easy to use caddies. In many ways blade servers are the ultimate expression of this with their relatively higher purchasing cost being a direct consequence of their more advanced engineering allowing for lower end user servicing costs.

I doubt Sony or MS would see any benefit from tool-less chassis, they don't offer onsite engineer support so they have no need to maximise the number of calls they serve (and field folks are the most expensive folks) and their take in service is usually a 'power on -> if yes return to user, if no swap for service unit' rather than an in depth investigation that could be made materially cheaper by having easier to replace parts. By the time anyone is doing that kind of indepth service the unit has often been shipped back across to the ODM for them to refurbish for use as the next service unit, as ODMs are all based in low labour cost countries the cost savings from making them more time efficient are simply not there.

 

[Prophecy2k]

A little colder air is equivalent to more warmer air. PS4 regulates the air temperature.

There is no battery in there. That's not how surge suppression works.

A little colder air can be equivalent to more warmer air in a real system. It depends on the overall heat transfer coefficient, thus essentially how well designed the system is.

Regardless, the general thrust of your point still stands

 

[MrFox]

Yes I agree, it's the ability to reach a specific parts temperature that is important. Sometimes it logically can't work. Best real world example is that they can't keep the HDD below it's rated 40C using any amount of 40C air, which is why the HDD is in the entry air path in the PS4. But in the case of power supplies, it's quite easy to keep a capacitor significantly below 85C using either some 40C air, or less 25C air.

 

[Prophecy2k]

Good example. I was trying to think up one but I eventually gave up. It was too late in the day.

On a tangental note, are HDD's actually only rated fro 40°C?

 

[MrFox]

Good example. I was trying to think up one but I eventually gave up. It was too late in the day.

On a tangental note, are HDD's actually only rated fro 40°C?

Ah sorry no, that's not the specs. I used 40C for years as a rule of thumb, but I looked at some recent HDDs and they say anywhere between 50C and 60C maximum. I guess 40C remains a good limit, google have some research that shows anything above this produces accelerated failure.
http://static.googleusercontent.com/media/research.google.com/en/us/archive/disk_failures.pdf

I'm having a lot of fun with this so I tried to be thorough with the capacitors life using this calculator:
http://www.illinoiscapacitor.com/tech-center/life-calculators.aspx

The PS4 main capacitor (450v rating and I assumed it's operating at 311v on a 220v AC, whatever):
http://media.digikey.com/pdf/Data Sheets/United Chemi-Con PDFs/KMQ-VS_Rev0703.pdf

I measured the outlet air temperature on my PS4 after hours of gameplay and I got 45C.
Even with 50C of air temperature I get 130,963 hours. That would be playing 8 hours per day for 45 years.

The output capacitor are the KZH series:
http://www.digikey.ca/product-highlights/ca/en/united-chemi-con-kzh-series/3483

I got 362,039 hours at 50C for those.

 

[brunogm]

In tropical countries there is a bunch of PS4 failing due to insufficient cooling. Mostly the fan/ temperature table is optimized to quiet operation. Then here in Brazil is really popular a PWM amplifier to double whatever Sony puts. Also why the fan cannot work for 1 min or 30s at shutdown to keep to remaining heat damaging BGAs.

 

[iroboto]

In tropical countries there is a bunch of PS4 failing due to insufficient cooling. Mostly the fan/ temperature table is optimized to quiet operation. Then here in Brazil is really popular a PWM amplifier to double whatever Sony puts. Also why the fan cannot work for 1 min or 30s at shutdown to keep to remaining heat damaging BGAs.

??? Never seen these reports online. Can you link something?

 

[brunogm]

??? Never seen these reports online. Can you link something?

Well, because of the gray sales market most will not report it. but many that bought in US paid import tax and have all legal papers aren't well treated or supported by Sony Brasil even if willing to pay for repairs.
http://www.reclameaqui.com.br/9479590/sony-brasil-ltda/cuidado-ao-comprar-ps4/
Then most accounts are from local repair shops. The most trusted one have many complaints about the design and personally repaired more than 150 consoles. And analysed another 100 dead consoles.

Also many problems with thermal paste in new consoles, in short what the design assumes isn't what happens in practice.

 

[brunogm]

Now for Xboxone let me link a thermal paste video:

at 20min.

 

[brunogm]

Sorry couldn't edit in time.

This PWM kit allows the PS4 to always stay 5c above ambient at least is what FLIR thermal camera shows at this moded console with 28c ambient.

Original would be running KZSF at 70c

 

[TheWretched]

In tropical countries there is a bunch of PS4 failing due to insufficient cooling. Mostly the fan/ temperature table is optimized to quiet operation. Then here in Brazil is really popular a PWM amplifier to double whatever Sony puts. Also why the fan cannot work for 1 min or 30s at shutdown to keep to remaining heat damaging BGAs.

I am not sure if I undestand correctly. But the fan does keep running if you enter standby. No ideal, but still a work around.

 

[brunogm]

http://vr-zone.com/articles/sony-engineer-unveils-functional-beauty-ps4s-cooling-system/69852.html

Especially the image about airflow separation is a lie and most heat is found near the AC plug:
(these consoles are on few minutes in the main menu only!, only first 5min show thermal)

 

[MrFox]

I call bullshit on the FliR methodology.

I also call bullshit on the explanation that it doesn't have any reports on this because everyone have a gray market PS4.

I also call bullshit on the idea that the fan supposedly turns off and causes the PS4 to overheat, we can see experimentally that the fan continues to run. It behaves as it should for an active control based on air and parts temperature.

FliR shows the heat of the power mosfet heatsink, which is... (roll drum)... near the AC plug. And these are rated 175C. The air just outside the laminar flow across this heat sink, and the case area above it, will be the hottest one. By design.

This isn't even worth arguing it sounds like total bullshit trying to sell a snake oil PWM kit.

 

[brunogm]

I call bullshit on the FliR methodology.
I also call bullshit on the explanation that it doesn't have any reports on this because everyone have a gray market PS4.

I said in tropical countries. And yes for a correct FLIR one has to calibrate for emissivity and so on(emissivity controled stikers help). Then real temp delta is only a estimation error but people experience this since PS3 with repeated issues.

My PS3 case was modded to open the intake airflow, but is so hot after playing for some hours that i have to use 7X "max fan eject trick" at shutdown, until the case is cold enough to touch.

 

[Lalaland]

Hmm I'm sceptical of the idea that it wasn't designed for tropical countries, simply because the Middle East, SE Asia, India and chunks of East Asia are hardly Northern Europe like climates but are Tier 1. I'm pretty sure the TIM is applied with more enthusiasm than care but that's true almost across the board in all consumer electronics. Folks did the math and too much is less of a problem than too little so it gets over applied as a standard. If heat sink failure was a generalised issue I'd expect that to have been reported far more widely. It's not the first time I've seen people report 'issues' that they just so happen to have the cure for a modest fee.

 

[MrFox]

I said in tropical countries.

This misconception you're playing on (that the tropical zone is hotter than the latitudinal desert belts) is because high humidity prevents perspiration making it extremely uncomfortable. The tropical zones have a maximum temperature similar to the summer in many temperate zones. The sub-tropical zones are much worse and that includes the southern US states, and I think Australia.

And yes for a correct FLIR one has to calibrate for emissivity and so on(emissivity controled stikers help)

Emissivity have nothing to do with it, it's all the default 0.95 here (for fuck's sake it's only measuring black plastic). The problem is measuring plastic without any understanding of what the limits are, or what parts are underneath.

The correct methodology is to understand what the control algorithm does to the fan, where the sensors are, what the parts temperatures are, what their operating limits are, and whether or not those limits are reached. FliR doesn't tell us shit.

Then real temp delta is only a estimation error but people experience this since PS3 with repeated issues.

My PS3 case was modded to open the intake airflow, but is so hot after playing for some hours that i have to use 7X "max fan eject trick" at shutdown, until the case is cold enough to touch.

You are changing the subject to the PS3 which have nothing to do with the discussion.
The PS3 had a very different fan control with no airflow temperature feedback, and the 230 watts it generates makes it more difficult to cool under high ambient temperature. The fan had much less available margin. Even more important is that there was a leadless solder issue that was exacerbated by heat.

 

[brunogm]

Ok, i was just referencing to 46 celsius (115F) normal daily temperatures, blame urban heat islands or desert climates.
As presented by http://www.famitsu.com/news/201401/17046618.html, they review PS2, PS3, PS4 cooling solutions. So not changing the topic because as the presentation show the PS3 G and N were key to the current system.

Output temps of dumb plastic dont tell the whole history but junction-to-case thermal resistance in mosfet varies in 1.5 to 4C/W. My guess is at least 40C increase to the internal temps.

Going forward i would like to ask if someone knows the cost analysis in these decisions, like a 40 cent SiC heatsink could be used or is better a vaporchamber?

 

[Deleted member 11852]

This is like the weather channel!