Question Replacing Intel CPU Heatsink Plastic Push Pins With Spring Screws? Downgrading from Asetek 714256-001

BlaineB

Distinguished
Oct 28, 2014
50
1
18,535
My HP Envy Phoenix system is going on 9 years in a few months. I have an Intel I7-4790 processor that was previously paired with a Asetek 714256-001 liquid-to-air cooler. Unfortunately the pump must be wearing out as every 30 seconds or so a very loud crackling is heard that intensifies in volume and speed until it suddenly stops making noise all together, but then not long after it repeats this process over and over again, all day long.

I "downgraded" and purchased an Intel E97379-003 heatsink/fan. This heatsink normally attaches to the motherboard with those plastic push pins that everyone hates.

Unfortunately my motherboard does not have sufficient holes for this to work, as there are "nutserts" in the motherboard that appear to be non-removable and molded into the plastic, that formerly secured the water pump for the Asetek 714256-001 liquid-to-air cooler.

I currently have the Intel heatsink/fan somewhat "jerry rigged" on my processor, using a combination of the screws from the Asetek 714256-001 cooler and some washers to ensure that it fits snuggly against the processor.

However I would like to clean this up a bit and make it more professional. Are there any particular spring screws that would work for my setup much more efficiently than what I have?

I have searched and found that some other individual removed his motherboard from his case and drilled out the holes so that he could insert some BOLTS through the motherboard from the backside. I would rather not do this as, in my book, DON'T FIX IT IF IT AIN'T BROKE! I'd rather avoid the possibility of damaging further components by disassembling the entire case down to bear bones and would really like to avoid using a drill to make holes in my motherboard!
 
Last edited:
I will say that you have answered your own question.

I.e., "DON'T FIX IT IF IT AIN'T BROKE".

And this:

"However I would like to clean this up a bit and make it more professional. "

I certainly understand that frame of mind.

On the otherhand if what is now in place and is working then leave well enough alone.

Do not let form get in the way of function.

Consider that the current assembly is indeed a result of professionalism. And nothing wrong with using washers.

Just my thoughts on the matter.
 
  • Like
Reactions: Nathkrul
I will say that you have answered your own question.

I.e., "DON'T FIX IT IF IT AIN'T BROKE".

And this:

"However I would like to clean this up a bit and make it more professional. "

I certainly understand that frame of mind.

On the otherhand if what is now in place and is working then leave well enough alone.

Do not let form get in the way of function.

Consider that the current assembly is indeed a result of professionalism. And nothing wrong with using washers.

Just my thoughts on the matter.
Here are my main concerns;

From what I've seen heat sinks are not smashed down onto the processor, they are all mounted with some sort of somewhat flexible mounting platform, either rubber bushings or a spring type of setup. I am concerned that my "jerry rigged" attachment may be putting too much stress on the motherboard and/or the processor itself. I am also concerned that the heatsink is not sitting squarely or centered on the processor. When I removed those plastic push pins I also had to remove their respective rubber bushings, and the bushings I am temporarily using from the Asetek water cooler do not fit properly on the Intel fan driven heatsink neither.

Secondly the thumb screws that I am using from the Asetek water cooler are too long, as the bushings used with the Asetek cooler are much longer (of course I they don't properly fit the Intel cooler regardless.)

Is there a proper cooler that I could use instead that would be compatible with my motherboard which uses a screw retention method for secure attachment rather than push pins that I should buy instead of what I had?

If there isn't, what thread size do you suppose those permanent nutserts in the motherboard are for securing the heat sink? I'd really love to replace those lengthy thumbscrews with something more appropriate.
 
When doing modifications of this sort, hard mounting is actually a good idea. You just have to be careful and count the screw turns and not overtighten. I often use precision screwdrivers for this since you can't get as much torque. If all you have is a normal screwdriver, stick to your finger tips.

Depending on available height, coolers like the Hyper 212 Evo, Deepcool Gammaxx 400, and many others with a cross shaped mounting mechanism can be fit directly to the existing mounting holes. Just a matter of getting the right screw and some washers. (Typically M3)
 
When doing modifications of this sort, hard mounting is actually a good idea. You just have to be careful and count the screw turns and not overtighten. I often use precision screwdrivers for this since you can't get as much torque. If all you have is a normal screwdriver, stick to your finger tips.

Depending on available height, coolers like the Hyper 212 Evo, Deepcool Gammaxx 400, and many others with a cross shaped mounting mechanism can be fit directly to the existing mounting holes. Just a matter of getting the right screw and some washers. (Typically M3)

One of these days when I am not in a rush and I have some time (during the day) I will disassemble everything so I can go to the hardware store and purchase the appropriate screws. Do you think that Lowes or Home Depot will have screws that small such as the M3 (3mm) screws you suggest?

I have a variety of spare leftover motherboard standoffs and screws but these screws are very short and I do not believe they will be long enough to perform well in this particular application.

Is there any issue if a heatsink is not perfectly centered over the processor?

I really hate having to "modify" it like this. I understand I mentioned that I am a "don't fix it if it isn't broke" but at the same time I am anally OCD enough to only feel comfortable with a proper and reputable repair.
 
Yes, m3 machine screws are pretty easy to track down. For length you need to place your cooler on CPU and see the distance between what you are interfacing with on the cooler and the threaded holes on the motherboard/chassis.

With the coolers I mentioned you would be screwing down from the top, so too long is preferable to too short. You would add washers on the screw head or cut the screws to length. (You can also just thread some nuts onto the screw shaft to take up slack)

If you have to come up from the bottom, you need to verify that you can actually do that, and then you would need nuts to secure cooler to the shafts sticking up. This is how many liquid nitrogen/prototype/custom coolers are mounted.
 
Yes, m3 machine screws are pretty easy to track down. For length you need to place your cooler on CPU and see the distance between what you are interfacing with on the cooler and the threaded holes on the motherboard/chassis.

With the coolers I mentioned you would be screwing down from the top, so too long is preferable to too short. You would add washers on the screw head or cut the screws to length. (You can also just thread some nuts onto the screw shaft to take up slack)

If you have to come up from the bottom, you need to verify that you can actually do that, and then you would need nuts to secure cooler to the shafts sticking up. This is how many liquid nitrogen/prototype/custom coolers are mounted.
I would definitely be coming in from the top-down as the threaded nutserts are in the motherboard. I would strongly prefer not to disassemble everything and remove the motherboard from the case. The back panel of this HP enclosure appears to be riveted on and non-removable, to complicate matters of course.
 
Can you assist me with another issue -

I am receiving a BIOS error on startup ever since removing the Asetek liquid cooler and replacing it with the fan.

There are 3 fan connections on my motherboard.

CPU FAN - 4 Pin - Previously connected to the 80mm case fan that was mounted to the heat exchanger/radiator from the Asetek cooler. I am still using this same fan but have it mounted to the back of the case without the radiator.

PUMP FAN - 4 Pin - Previously connected to the Asetek water pump.

I am currently receiving BIOS error 512 - Rear chassis fan not connected.

It seems like regardless of which of these 2 fans (either the 80mm case fan OR the Intel air cooler heatsink fan) the PUMP FAN connection is not detecting the fan. I assume that this 4 pin PUMP FAN port is what is being referred to in the BIOS as the "rear chassis fan."

I currently have my Intel heatsink fan plugged in to the CPU FAN connection, and in the bios it is registering a proper RPM.

There is also a 3rd fan connection on my motherboard, but this is labeled SYS FAN and is a 3 pin arrangement. This was never connected to anything previously.

In my bios there is NO way to ignore any of these fans, so as of right now, my only solution is to press F1 to continue every time I want to boot!

I did notice that the Asetek water cooler has pins 1 and 4 jumpered on the pump plug, which are the outer most connections on the plug. There are only 3 wires which lead from this plug to the Asetek pump.

Are there any solutions for what I can do to eliminate having to press F1 to continue every single time I want to boot?

Both of my fans which are connected are 4 wire fans as well, plugged in to the 4 pin fan connectors on the motherboard.

From my brief research it appears that the Asetek fan is jumpering "PWM" and "Ground" for their pump.

Why is my motherboard not detecting an RPM from any of the 2 fans that I have tried connecting to the PUMP FAN connection, yet the CPU FAN port works properly? and displays an RPM?
 
Fans and pumps often have different coil arrangements so that the RPM pulse differs between them.

If they configured it to look for a certain range from the pump, you may be getting the warning because it is spinning too fast or too slow as read. I forget which way it goes. Why newer motherboards started coming with specific pump headers, both for current and the correct RPM sense.

Correction would be some sort of small circuit to change the pulse signal.

I would probably have to sit down with it and an oscilloscope to get you more details.

A quick search for a sensor adapter might have someone selling the solution already. I checked moddiy.com which usually carries such things and didn't find anything.
 
Fans and pumps often have different coil arrangements so that the RPM pulse differs between them.

If they configured it to look for a certain range from the pump, you may be getting the warning because it is spinning too fast or too slow as read. I forget which way it goes. Why newer motherboards started coming with specific pump headers, both for current and the correct RPM sense.

Correction would be some sort of small circuit to change the pulse signal.

I would probably have to sit down with it and an oscilloscope to get you more details.

A quick search for a sensor adapter might have someone selling the solution already. I checked moddiy.com which usually carries such things and didn't find anything.

What is the "benefit" of shorting PWM to ground as with the Asetek pump arrangement from the factory?
 
Not sure. I suppose it depends on the logic they were going for and how they designed everything. Perhaps they assume the pump is running and are only checking if the fan goes dead. It might be expecting a normally low output and they have something in the pump to drive high when the RPM is no longer there. The reasoning being that only the appropriate OEM pump/fan combo would register correctly.
 
Not sure. I suppose it depends on the logic they were going for and how they designed everything. Perhaps they assume the pump is running and are only checking if the fan goes dead. It might be expecting a normally low output and they have something in the pump to drive high when the RPM is no longer there. The reasoning being that only the appropriate OEM pump/fan combo would register correctly.
Could I do something similar with my fan, shorting PWM and ground as well? I would think this would damage something typically.
 
I would first try disconnecting the RPM line and than trying out a ground, just to see if anything changes.
As I think more about it (correct me if I am wrong) I do not think that will help as having the fan totally disconnected results in the same error regardless.
 
I was under the impression that the pump signal was the one shorted.
That is correct - Pins 1 and 4 were shorted together on the wire harness from the Asetek cooler, that I am no longer using. The pump was connected to the PUMP_FAN output on this motherboard.
 
I purchased a brand new 92mm PWM case fan to try as well. What I discover is that the PUMP_FAN header on the motherboard seems to supply full voltage to whatever fan you connect to it regardless if that's a case fan or the CPU heatsink fan, which makes sense if it's powering an AIO cooler pump, it runs the pump at full speed.

However what I am still confused about and what no one has been able to provide me an answer with is why shorting pins 1 and 4 together on the Asetek AIO cooler pump wire harness seems to "satisfy" the motherboard.

I still have the Asetek AIO cooler pump in my possession, and plugging it into the PUMP_FAN header once again satisfies the motherboard and it immediately ceases to give me the fan error on boot. It is only fans that produce this error when connected this PUMP_FAN header.
 
I will say that you have answered your own question.

I.e., "DON'T FIX IT IF IT AIN'T BROKE".

And this:

"However I would like to clean this up a bit and make it more professional. "

I certainly understand that frame of mind.

On the otherhand if what is now in place and is working then leave well enough alone.

Do not let form get in the way of function.

Consider that the current assembly is indeed a result of professionalism. And nothing wrong with using washers.

Just my thoughts on the matter.
I wish I could post pictures here without having to have the pictures hosted elsewhere. But I did find a Dell OEM heatsink that uses a screw with spring fastening system that matches the bolt pattern on my Motherboard. So there is no need to use the Intel heat sink with the plastic push pins!
Although I am still having the issue with "Bios Fan Error" ever since replacing the Asetek AIO cooler with a fan driven air-to-air heatsink.
 
Post pictures here via imgur (www.imgur.com)

Many posters and mods use imgur on a regular basis.

As long as the pictures do not reveal any personally identifying information (either directly or in the background) you should be safe.

Someone may well spot some system problem via the photographs.

Pictures of a motherboard, fans, connections, etc. otherwise being useless unless a crime has been committed and the authorities have seized the computer and made some incriminating physical comparisons that could prove, for example, that the computer is yours.

Hopefully not something that you need to be concerned about.
 
Post pictures here via imgur (www.imgur.com)

Many posters and mods use imgur on a regular basis.

As long as the pictures do not reveal any personally identifying information (either directly or in the background) you should be safe.

Someone may well spot some system problem via the photographs.

Pictures of a motherboard, fans, connections, etc. otherwise being useless unless a crime has been committed and the authorities have seized the computer and made some incriminating physical comparisons that could prove, for example, that the computer is yours.

Hopefully not something that you need to be concerned about.
The only crime I've made is perhaps utilizing less than top tier thermal compound. 🤡
 
Post pictures here via imgur (www.imgur.com)

Many posters and mods use imgur on a regular basis.

As long as the pictures do not reveal any personally identifying information (either directly or in the background) you should be safe.

Someone may well spot some system problem via the photographs.

Pictures of a motherboard, fans, connections, etc. otherwise being useless unless a crime has been committed and the authorities have seized the computer and made some incriminating physical comparisons that could prove, for example, that the computer is yours.

Hopefully not something that you need to be concerned about.
s-l1600.jpg


The only thing I noticed is that the pre-applied thermal compound seemed to be a bit hardened, so I replaced it.
 
As near as I can see the wire colors/pinouts are as would be expected.

FYI:

https://landing.coolermaster.com/faq/3-pin-and-4-pin-fan-wire-diagrams/

You can easily google for other similar diagrams and explanations.

However, still doublecheck as applicable for the host build - fans, plugs/connectors, ports. Ensure that all match as they should.

Look at both the CPU-FAN and the PUMP-FAN pinouts.

= = = =

That said, it appears to me that the four fan wires are really wrapped/twisted quite tightly. Maybe too tight...

And very kinked where the wires enter into the fan housing.

Consider that there may be some intermittent connectivity problems as a result.....

Just a thought.
 
As near as I can see the wire colors/pinouts are as would be expected.

FYI:

https://landing.coolermaster.com/faq/3-pin-and-4-pin-fan-wire-diagrams/

You can easily google for other similar diagrams and explanations.

However, still doublecheck as applicable for the host build - fans, plugs/connectors, ports. Ensure that all match as they should.

Look at both the CPU-FAN and the PUMP-FAN pinouts.

= = = =

That said, it appears to me that the four fan wires are really wrapped/twisted quite tightly. Maybe too tight...

And very kinked where the wires enter into the fan housing.

Consider that there may be some intermittent connectivity problems as a result.....

Just a thought.
Whichever fan I connect to the 4 pin CPU_FAN header works flawlessly and displays an appropriate RPM within the BIOS, and no errors are given.

However whichever fan I connect to the 4 pin PUMP_FAN displays 0 RPM within the bios and gives the 512 error "no case fan detected" at boot. Also I noticed that when any fan is connected to the PUMP_FAN header it runs at full RPM versus a variable RPM depending on heat and load. The CPU_FAN header controls the fan speed appropriately.
 
That suggests then some issue with PUMP_FAN pins 3 and/or 4.

Namely:

Pin 3 = Tach/Signal/Sense
Pin 4 = Control/PWM

Because both fans work via the CPU_FAN header the problem is likely within the PUMP_FAN connection.

Any access to an additional known working fan or two to test as well.

Objective being to confirm, in some manner, that the PUMP_FAN connection is the problem.

Or that some other factor is involved.