Item 2 - The recommendation comes from Nidec, specifically for PC Fans.... the quote was from the Nidec site. When a fan manufacturer who has to warranty something says "don't do this", I'm not going to do it. It must be said however, that I think they have little to worry about, if a fan fails, user will generally write it off and most won't fail within the typical design life of the PC.
Item 4 - the 2 rule rule of thumb does not exist. MoBo manufacturers specifically state that the header rating is based upon the continuous voltage rating". We see the same thing with cooling where folks look at an instantaneous peak of a video card and begin insisting that the PSU will blow up or the cooling will fail if you don't size them for twice the TDP. No ... the TDP is the TDP, yes you may see peaks up to twice the TDP ... that's what caps are for. I have demonstrated this on the test bench by measuring the power output at the wall with a power meter. These short peaks simply do not register on the meter.
The 2 x peaks are ms in length and do not change the average required cooling load. We have 6 temp sensors in the test rig along with the power meter. We can easily measure the wattage at the wall and the corresponding increases in water temps in the loop... with and without the cards in play.
Let's do the math...
Peak power output of Asus 970 = 268 watts (add 10% when MSI AB slider power limiter slid to 110% = 295
http://www.techpowerup.com/reviews/ASUS/GeForce_GTX_780_Direct_Cu_II_OC/24.html
780 GFX w/ 26% OC = 295 x 2 = 590 watts
CPU @ 4.7 Ghz (1.385 VID) = 130 watts
Water pumps = 2 x 23 watts
MoBo = 40 watts
RAM = 10 watts
2 x SSDs = 8 watts
2 x SSHDs = 20 watts
16 fans = 2 watts
http://www.tomshardware.com/reviews/power-supply-psu-review,2916-8.html
BTW, if ya read the article... you see any mention of worrying about "inrush current" ?
Now lets compare that 846 watts with the power meter ... the meter tops out at 745 watts under Furmark.... allowing for the efficiency of the PSU, let's be real generous here and call it 90%, ... 745 a 90% = 680 watts. So wait ... why isn't it up at 850 watts and instead 170 watts less ?
If the GFX cards are drawing 590 watts ? ... that means everything on the PC is drawing only 90 watts .. that simply "isn't real".
On the heat measurement side, we see the same thing. We don't see Delta T rising in an amount that corresponds to 590 watts.
Not only I have I had success, repeatedly and w/o exception, I have never heard of a single failure. Inquiries to Tech Support from multiple MoBo manufacturers say the rating is for continuous voltage. Now if the people who build the thing say say that only the continuous rating matters and they say that they support up to 1 amp of continuous power, that they will have to replace the board if it fails with < 1.0 amp of continuous voltage .. where does this 2 fan thing come from ?
item 5 - Like the 2 fan rule of thumb, the warnings you speak of simply don't exist ... they are not "in the manual" as claimed.
Connecting the 4-pin to other 4-pin header (besides the CPU_Fan)
Other 4-pin connectors can be found on modern motherboards besides the “CPU_Fan” connector (e.g. “CPU_Fan2”, “CHA_Fan”, “OPT_Fan”), however not all motherboard manufacturers implement a true PWM signal modulation onto these connectors. These type of 4-pin connectors modulate the RPM by voltage, which has a smaller control range compared to modulation by true PWM signal.
The 12V SATA power cable can not be used to power the PWM hub if connecting to these types of 4-pin connectors, due to the interference with the RPM regulation by voltage (resulting in the fans running on full RPM). The PWM hub will draw its power from the 4-pin connector, which is limited to a total device consuming 30W in total.
So no the statements below are false...there are no such warnings ... it merely says that PWM offers a greater control range and not to connect the power cable if using DCV.
They emphasize that, to ensure control, the Hub's connection to a mobo port MUST be to a 4-pin header that IS using PWM Mode....
Simply not true as the manual states. PWM range is typically take as having a bottom end of 25% speed... it's generally somewhat higher with DCV... usually in the low to mid 30s so the control range is slightly smaller. However, with Asus FanXpert doing the fan profile, it sets the bottom with Phanteks DCV fans from 282 to 326 fpm .. that's 22 - 26%.
Why would they tell you how to use a the power cable when connected to DCV if you MUST use a PWM Header ?
they recommend that ALL users connect their Hub only to a CPU_FAN header.
Where ? The manual does not include this recommendation. It explicitly says that if you want to do this, don't connect the power cable and don't exceed 30 watts.
Point of Information ... they are not "fake" 4 pin headers. You stated repeatedly that PWM can be used on DCV headers ... Though fan manufacturers discourage this when used in critical applications, the resulting impact of the practice will not frequently manifest itself within the fan's typical warranty period. In addition, Z87 motherboards typify this 4 pin header but only DCV voltage control phenomenon. The origin of which becomes evident when you read the MoBo manuals. And, for example, the RoG line manuals describe these headers as PWM.
When I contacted Asus TS after determining that no PWM control existed on these headers, they advised that the Z97 boards would have dual control options .. the header would work with either PWM or DCV with either auto detect or BIOS setting. That would require a 4 pin header.
This was expected to be incorporated as a feature on the Z87 boards, but they never quite got the bugs out in time for release so it was scrapped ... unfortunately, either the manuals were already printed or no one remembered to go back and edit them when the feature was ultimately omitted.
they recommend that ALL users connect their Hub only to a CPU_FAN header. ... That recommendation causes some consternation to users who want to use the Hub solely for case ventilation fans and don't understand the limits and reasons for what the manual says.
Again, the manual does not say that.... anywhere.
Originally, when the Phanteks cases were released, there was no intention to release the hub as a separate accessory. We, and other users participating the the Enthoo thread on overclock.net, were building Enthoo based builds at a time when there were no or insufficient PWM headers available for water cooled systems. In my box, the CPU and CPU_OPT were taken by the twin water pumps. I had no headers available for **any** fans. TS advised to just connect them to the DCV headers.
After Phanteks was convinced to release the hub as an accessory, they went thru several months of testing as the plastic cover, alternate power source and universal mounting methods were evaluated. Phanteks TS was very active in the thread as this and "recommendations for version 2" of the cases / accessories were discussed.
We kept pestering them to sell us two "Phanteks case style" units as "replacements" w/o success. Finally, Phanteks provided us with two extra hubs "for testing" free of charge. We worked under the direction of Phanteks TS thru the entire testing procedure and thus are very familiar with their "recommendations". The ones you stated, do not appear in the manual, nor were they ever stated to us; the manual and everything we received is in direct conflict with your statements.
Item 6 - You are free to consider anything your choose, but a statement that a device does not exist should be qualified by saying "To my mind .." because the industry says otherwise.
My use of those terms is that a splitter simply connects two or more loads to a common source
Ya mean like a PC network hub that lets you connect multiple computers to a common source ?
"Definition Hub: 3. (Computer Science) computing a device for connecting computers in a network
"Definition Hub: 6. any one of the holes in an electrical panel, into which connections may be plugged."
Thermaltake, Swiftech, NZXT, Deepcool, Silverstone, Swiftech all have fan hubs on the market
http://www.deepcool.com/product/dcoolingaccessory/accessory/2013-12/48_663.shtml
http://www.performance-pcs.com/internal-power-distribution-pcb-10-way-3-pin-fan-hub-power-splitter-25054.html
http://www.performance-pcs.com/silverstone-cpf04-1-to-8-pwm-fan-hub-black.html
http://www.aliexpress.com/item/1-8-CE-ZIF-50pin-IDE-to-SATA-22pin-Adaptor-Converter-2in1/431004672.html?spm=2114.40010308.4.8.UsETPs
http://www.ebay.com/itm/1-4pin-to-4-Port-host-case-cpu-PC-Fan-Hub-Splitter-cable-for-Multi-Way-PWM-Fan-/311502444768?hash=item4886fe18e0:g:MZwAAOSwmrlUurTS
The generally accepted distinction between a splitter and hub is that a splitter is a cable with one 'in' connector and multiple outs, A hub is a "central point"... the classical definition of the word hub, where all the connections are in one place ... a hub has no wires. Like a bicycle wheel, the "hub" is where the spokes come and meet.
The Phanteks hub is simply more than just a hub as the classic definition of hub is just "central location where things come together". In addition to serving as a hub the device acts as signal converter and source of auxilliary power. No where are those additional functions described / defined as being necessary to be a hub. A network hub doesn't convert the signal .. it doesn't add auxilliary power to make the connected devises function.
A Hub, on the other hand, obtains its power for all loads from a different source that can provide all the power needed by the load, and separately provides some method of applying a common control to the loads. Most 4-pin fan Hubs uses simple sharing of the PWM signal, but the Phanteks PWM Hub does it differently, as we've been discussing.
So by your definition, the Phanteks device...
1. When it works as PWM with 11 fans and auxilliary power, it's a hub
2. When it works as PWM with 2 (or more) fans and no auxilliary power, it's not a hub
3. When it works as DCV* with 2 (or more) fans and auxilliary power, it's not a hub
* as specifically describe in the Hub Manual.
Why does Deepcool and all the others think their hub is a hub, it doesn't have power, it doesn't convert signals. The dictionary and industry standard definition differs from yours. The words splitter and PCB may also be used and they do accurately describe the devices... but so does hub.
I note that ModMyToys specifically reminds users to ensure their total load connected does not exceed the mobo maker's limit for the header, thus limiting their liability. Like most others, they do not discuss any detail about whether "load" means peak surge load or continuous running load.
They don't say anything about in rush current ... which **if it was a "concern" or the deciding factor in a warranty dispute, they certainly would be required to do so.
1. The header has a published rating
2. The fan has a published rating
3. The published ratings are available to the consumer.
4. The consumers responsibility, as established by the warning, requires him / her only to examine the published ratings. They are not expected to take courses in electrical engineering, read textbooks or perform calculations.
If inrush current was an issue, ModMyToys and the MoBo manufacturer would be required to make that information available to the consumer. Otherwise, if we accept your premise, when I blew up my MoBo header with the 6 fans connected to it, they would be required to replace it.
Again, the proof is in the pudding ...
1. We have three boxes, all under extreme loads, with more than 2 fans using continuous rate voltage as the limiting criteria.... no problems. I had 8 fans going for a month... if inrush current was a 'thing" to be concerned about, how did the 7.84 amp inrush current not blow the board ... or the 94 watt wattage not blow the hub w/ the 30 watt max rating exceeded by a factor of 3 ?
I like to de-rate the header to provide a cushion as 1) out boxes are put thru some rather extreme conditions such as torture test runs with no fans running, repeated restarts and we have our fans set to shut off when temperatures are low. When Asus FanXperrt does the fan profiling it repeatedly shuts the fans on and off to determine the minimum operating rpms. So how did my motherboard header manage to survive not just one but a continuous series of fan restarts with a 4.5 amp inrush current ?
2. All of our builds for users that we have done, none have had a problem
3. All of the users in the Enthoo thread over on overclock.net ... no one has blown a header.
4. No reports of blown headers on Phanteks web site
5. No reports of blown headers in discussions with Phanteks tech support.
6. How is it that we see fan splitters on the market with 3, 4 and 5 connections ? Wouldn't there be a wave of "I blew my fan header" posts on forums ?
7. Why do highly respected enthusiasts vendors make custom sleeved cable splitters with 3 ,4 ,5 connections ? Is it possible that if inrush current was an issue, that these people who "do this for a living" are not aware of it ?
8. With all these vendors selling these splitters and hubs for so many years ... why can't we find innunmerable posts on blown fan headers ?
One final point of max loading. The Phanteks PWM Hub manual specifically says that it is limited to 30 W total load, and also that it can be used with UP TO 11 fans connected to its six ports, using simple 2-output 3-pin Y-splitters. Even at a continuous load rating of 0.15 amps per 12-volt fan, 11 fans comes to 19.8 W. They appear to be saying you should NOT connect up to 30 W of continuous fan motor load to the Hub.
No, that's not what they are saying. They are saying, don't connect even 1 of these:
http://www.frozencpu.com/products/8147/fan-500/Delta_Mega_Fast_120mm_x_38mm_Fan_-_252_CFM_-_Bare_Lead_PFB1212UHE-F00.html?tl=g36c15s562
Or more than 3 of these:
http://www.frozencpu.com/products/7759/fan-471/Delta_120_x_25mm_Extreme_High-Speed_Fan_-_15033_CFM_FFB1212EH_Bare_Wire.html?tl=g36c15s60
or more than 8 of these
http://www.frozencpu.com/products/24857/fan-1365/Noctua_Industrial_NF-F12_120_x_25mm_PWM_Fan_-_3000_RPM_NFF12IPPC30P.html?tl=c365s936b32
I know what's coming .. that these fans are not normally used. But they are. Delta fans have been very popular over the years for radiator and even air cooler usage. A typical case 1200 rpm case fan may use 0.15 amps but high SP, high rpm radiator fans are higher.... much higher. Phanteks, as a manufacturer is specifically focused on accommodating the water cooling community so accommodation of radiator fans would be a necessity. And Noctua is certainly no stranger to the PC world.
Test sites evaluate coolers by attaching delta fans. Again, these fans are commonly used for radiator and air coolers to improve cooling performance. I wouldn't wanna use more than 3 or maybe 4 if ya stretch it a bit (7.56 watts) of those.
http://archive.benchmarkreviews.com/index.php?option=com_content&task=view&id=674&Itemid=62&limit=1&limitstart=4
Again, consideration of inrush current is a logical pursuit that would only come to mind by someone with considerable knowledge on the subject. However, marketing materials are not prepared on the basis of a highly knowledgeable consumer.The standard of care is to make it understandable to the average consumer.
The "MoBo Header Rating" is to be compared with the fans continuous voltage rating. The people who design the motherboards, fans and hubs, know all about inrush current. The ratings are based upon them being satisfied that the inrush current of fans, because of the extremely short duration, will not be an issue.
If all of this wasn't true, and inrush current was an issue, it would be quite easy to find a web site article on the subject. It would be quite easy to find forum posts with users ranting that that the complied with the ratings and they blew up their headers. But neither exists.
Facebook
Twitter
Instagram