HELP: Swiftech GTZ - snap, crackle and pop

[JackOfAll]

Guys, either I'm doing something stupid or am stupid, but I could do with some help. I've taken this loop and GTZ block apart several times. When the pump is running the GTZ generates so much noise. What am I doing wrong? It sounds like bacon frying or the snap, crackle and pop you'd get from a bowl of rice crispies. Even with the case closed it can be heard some distance away. It's louder than all of the fans combined and very annoying.

The loop is an 18W DDC with XSPC res -> TFC 360 rad -> Swiftech GTZ -> res. 1/2" ID tubing and 45 deg TFC compression fittings on the pump and block, straight TFC compression fittings on the rad.
http://www.vacuumtube.org.uk/images/gtz_install.jpg

I've recorded the noise. Snap, crackle, and pop mp3 (http://www.vacuumtube.org.uk/gtz.mp3). The system is bled. There is no air in it. Any ideas?

[CmB]

Maybe it's a faulty pump? How long have you had it? Can't think of much else.

[JackOfAll]

Maybe it's a faulty pump? How long have you had it? Can't think of much else.

No, I don't think it is the pump. I was using this setup with a D-Tek Fusion V2 and Core2 Quad until I rebuilt with the P6T/i7 and Swiftech GTZ block.

Can someone confirm that the GTZ block is silent in operation?

[exemay]

It's air going thru the pump before bleeding out, happens everytime until the air bleeds out. Prefectly normal. It's going to take a little bit longer with the res below the block as you have it set up now, but will bleed out. Try cracking open the fill port a tiny bit to allow free air to escape.

andyc

+1 i had similar noises when i was bleeding my system, it took about 5 min's (and my 18W DCC pump/top/res was below my block and radiator) allow air to enter the res and just keep runnning the system till it bleeds entierly!

and yes, the block is silent

[JackOfAll]

It's air going thru the pump before bleeding out, happens everytime until the air bleeds out. Prefectly normal. It's going to take a little bit longer with the res below the block as you have it set up now, but will bleed out. Try cracking open the fill port a tiny bit to allow free air to escape, and fill up the res as much as possible. That helps keep the air from recircultaing back thru the pump.

Andy, thanks for the reply, but I'm not convinced that it is air in the loop. I've had the loop running outside of the case. I only mounted it back again this afternoon. I'm 100% convinced there is no air in the loop.

[Sadasius]

Yup....Air in the loop! :yepp:

[JackOfAll]

With a block like the GTZ, it's much more pronounced until it bleeds out.

OK, perhaps that's it. I've not used a Swiftech block before, only D-Tek and XSPC. In those cases, the noise made by air in the loop is different to what I'm hearing now. Not at all like an electrical static. With this Swiftech block, the noise that I'm used to caused by air in the loop has subsided after 12 hours, leaving this 'static' noise. When it doesn't get any quieter after 72 hours running, my gut feeling is that it is not going to go away. How long would you suggest it might take?

[yokomo]

+1 for air in the loop . youve got black tubing which looks ace but means you have no idea if theres air in it or not !
if you can slow down the pump do that , my loops wouldnt as after days of running bleed at full speed !
turn the pumps down and its all gone in an hour ish

[Hondacity]

the gtz takes the longest time to bleed in my experience...it will take awhile

[Grinch]

+1 air in loop :)

[JackOfAll]

Hard to say with a horizontal rad mount and the res being so low in the loop, but it will bleed out sooner or later. Try filling the res up as far as you can while the loop is running, and let the fillport stay somewhat lose to allow air out.

Done. I'll leave it for a few more days and see if it helps.

What I was trying to say before - but didn't exactly spell it out.... I always build the loop outside of the case. The initial fill was done with the block and rad below the pump/res and the rad vertical with barbs at the top to remove as much air as possible. It ran for at least 3 hours before it went into the case. In my previous builds in this case (with the res lower than rad/block), any residual air in the loop has been bled out only after a few hours of running.

Anyway, a tip of the hat to the water cooling experts here at xtreme. An expert I'm certainly not, with only 4 water builds to my name. Thanks all for your advice.

[Whilhelm]

Try remounting the block 90 degrees clockwise so that the swiftech logo is upright. Their instructions are very specific with regards to block orientation, maybe that has something to do with it.
I installed a GTZ today and it makes no weird sounds, my block is oriented as the instructions suggest.

[Sparky]

Unmount the GTZ and spin different directions while the pump it running, tapping it to try and release any caught air.

[JackOfAll]

Try remounting the block 90 degrees clockwise so that the swiftech logo is upright. Their instructions are very specific with regards to block orientation, maybe that has something to do with it. I installed a GTZ today and it makes no weird sounds, my block is oriented as the instructions suggest.

The instructions I received with the 1366 mount plate kit show that the logo should be orientated so that it is parallel to the locking lever. That's how I've fitted it.

[JackOfAll]

Unmount the GTZ and spin different directions while the pump it running, tapping it to try and release any caught air.

I've done that twice now. It made no difference. The block was still emitting the 'snap, crackle and pop' sounds in vertical and horizontal planes, while unmounted with the pump running. I'm still not convinced it is trapped air, even though otherwise it seems impossible to explain.

[Sparky]

FWIW my loop doesn't get quiet until about 24 hours.

[shazza]

Nice sound recording ... it sounds just like mine does with the GTZ and XSPC res/MCP355 when I first hook it up. Sometimes takes several hours to calm down.

Just to be clear ... you build the entire loop outside the case, and then install it without taking any of the tubing off any of the fittings? If so, I can see why you're pretty sure you have all the air out. Have you tried rocking the case, just in case there is air in the rad?

The only way I can think you'd be getting any noise from the GTZ itself is if there is something loose in there rattling around - but I think this is unlikely, so gotta go with the air bubble team on this one. Will be curious to see if it calms down over time.

[JackOfAll]

last bit of advise,
Relax a little and forget about it for a while. :up:

Advice is appreciated ....... BUT,
Relax!!! Chuckle. This has been driving me crazy for 5 days! <grin>


Enjoy your new rig and congrats on a great build :clap:

Thank you. Although, I wouldn't call it a great build, more like OK build. Temps are maxing out at 55 deg C, 100% load with HT enabled, overclocked to 3.6G (18x200) with 1.25V vcore. I can get it stable at 4G but anything past 3.6G requires a voltage jump, so I'm happy with the 3.6G as a 24/7.

I've ordered a 1366 bracket for my Fusion V2, so we'll see if the GTZ quietens down in the meantime, otherwise I'll go back to the Fusion.

Thanks for your input.

[Sly Fox]

Just for the record and future builds of anyone's, I found the EK 150 bleeds insanely fast.

Used it two days ago when I rebuilt my loop and I was shocked to see all the bubbles gone within 10 minutes. Much faster than my experience with the Swiftech MCRES.
I've never used the XSPC res top though, so I can't say if it actually bleeds better.

Then again maybe part of that is that my blocks aren't as restrictive and let bubbles get to the res easier, who knows. (Well, I'm sure SOMEONE will)

[gabe]

To all:

based on the recording that I listened to carefully, this doesn't exactly sound like air to me, sorry y'all, but I tend to agree with Jackofall on that one.

I obviously can't be exactly sure, but to my own ear, this recording sounds like a very sharp metallic noise, as if there was debris stuck in there somewhere and the pump seems like the most likely candidate IMO. Once I saw a staple stuck to a 350 impeller (magnet). It makes a hell of racket. similar to this.

I recommend opening up the pump if it continues for more than 24 hours.

I doubt that this comes from the W/B because the clicking noise has an amplitude suggesting that the "alleged" debris has room to move around, which you really don't have in the block, besides, it also sounds so metalic that a debris wouldn't make such racket hitting copper (soft metal) or Delrin (plastic dampening effect) again suggesting the pump....

As to rotating the block, please gentlemen, stop suggesting that. We recommend one direction for very good reasons. Whoever says he didn't see a difference just didn't take measurements precise enough.



Best to all.

[JackOfAll]

based on the recording that I listened to carefully, this doesn't exactly sound like air to me, sorry y'all, but I tend to agree with Jackofall on that one.

I obviously can't be exactly sure, but to my own ear, this recording sounds like a very sharp metallic noise, as if there was debris stuck in there somewhere and the pump seems like the most likely candidate IMO. Once I saw a staple stuck to a 350 impeller (magnet). It makes a hell of racket. similar to this.

I recommend opening up the pump if it continues for more than 24 hours.

I doubt that this comes from the W/B because the clicking noise has an amplitude suggesting that the "alleged" debris has room to move around, which you really don't have in the block, besides, it also sounds so metalic that a debris wouldn't make such racket hitting copper (soft metal) or Delrin (plastic dampening effect) again suggesting the pump....


gabe, thanks for jumping in. I'm more than 100% convinced that it is not air, but I can't explain it, so air bubble remains an option. I took the complete loop (including res/pump) apart on Saturday. Flushed the rad. Checked all tubing clear by holding it up to the light. My first thought was a piece of debris. If you can picture me with a 1 metre length of 3/8" tubing stuck in one ear and using the other end to locate the noise - that's how I'm pretty damn sure the noise is originating from the GTZ block and not the pump. The block is clear of debris. It's very, very strange. I have had one thought looking at a spare 45 degree TFC compression fitting, that I'm using on the block. The internal construction of this one is less than perfect. I can't really explain it so I'll try and take a picture later. Whilst the bend on the exterior is smooth, it's not on the inside - there is a ridge. Could this be causing cavitation (is that even the right word) and somehow that's being amplified when it hits the inlet plate. I'm clutching at straws .......

[Sparky]

Sorry gabe I meant rotate it around unmounted to try and release air from the block, not rotate its mounting :)

Jackofall, you could try changing the fitting to something else just to rule it out if nothing else :shrug:

[sabe]

I would be terrified if my setup made a sound like that, it sounds like something in a grinder... how the hell can air sound like that? Disassemble everything and check for stuck particles.

[gabe]

I would be terrified if my setup made a sound like that, it sounds like something in a grinder... how the hell can air sound like that? Disassemble everything and check for stuck particles.

air can sound like that because of resonnance. Martin could be right. but it just doesn't last that long.

never heard air sounding like this, unless he had the mike literally applied against the pump or the water-block.

where did you have the mike setup Jackofall?

[JackOfAll]

where did you have the mike setup Jackofall?

One inch away from the centre of the top of the GTZ. I was hand-holding it so it wasn't picking up vibrations.

[sanhacker]

Almost sounds to me like a wire being hit by a fan. Is there a fan on the NB ?
If you unplug your pump do you still hear the noise ?

[NameTaken45]

Almost sounds to me like a wire being hit by a fan. Is there a fan on the NB ?
If you unplug your pump do you still hear the noise ?


he said it was still making the noise with the loop entirely outside the case.
i just got my GTZ and ive heard the air rushing through the block and it doesnt sound anything like that. very strange, hope you figure this out.

[JackOfAll]

If you unplug your pump do you still hear the noise ?

If I cut power to the pump, the noise stops.

[JackOfAll]

i just got my GTZ and ive heard the air rushing through the block and it doesnt sound anything like that. very strange, hope you figure this out.

Exactly, initial fill - running with air in the loop does not sound like the noise I'm left with. That is more of a rushing sound (as it passes through the block) than this, which I can only describe as bacon frying, rice cripsies - snap, crackle and pop, or rain falling onto glass.

[shazza]

Gabe could be right, but I do get the snap crackle pop when I fill my loop - but that's filling it after everything is installed. Good thing is, if it's the pump, that's an easy if costly fix.

[NameTaken45]

thats the only thing i can think of that it could be, when my last MCP655 pump died it made that noise. although it only did that for about 4 hours before it died.
but thats a different pump and i think he ruled that out?

[sanhacker]

If I cut power to the pump, the noise stops.

That's to bad. If it were me, I'd pull the input off the GTZ and plug that tube directly into the reservoir. If the noise stops pull the GTZ apart and see if something is lodged in there. If the noise continues, it's your pump. My GTZ has never sounded like that. Matter of fact, I've never noticed it at all. Even right after filling the loop.

Best of luck.

[Martinm210]

Yeah, I agree there is some sort of debris causing this. It could be a metal shaving of some sort hanging on the impeller magnet or some stray piece of plastic spinning around in there. You need to pull the pump and top off and apart including the impeller magnet off the ceramic bearing. Take a close look at the ceramic bearing and spin the impeller by hand to ensure you don't have some sort of damage to the impeller bearing surface. It's most likely a noise that would generate at the pump, but I suppose there could be something similar in a block if the debris is an odd shape. Air will typically have that particular pump RPM increase as it encounters the air, I didn't hear that here.

Also make sure that pump top is correctly centered right over the inlet. Most pump tops can slide around a little and you could potentially install it far enough over that you'll get impeller impact during high flow rates, I've seen that before on several tops.

I tested two GTZ blocks and didn't have any sort of noise problem even with that exact same pump/top setup. I've heard some people complain about noise on the fuzions, but I didn't notice that either.

Anyhow, I'd look for debris, that sounds a little different than air. Air itself however can take up to a few days to completely bleed out in some setups.

[coolmiester]

Im going to stick my neck out and say its air.

Tiny bubbles bouncing around the block which in time should filter through to the res but for some reason the system doesn't want to fully bleed.

I've noticed on some XSPC res/top that the inlet plexi pipe which should be directly over the inlet impellor hole on the pump is not always straight and misses the hole slightly which could be constantly drawing in tiny air bubbles which is why the system won't bleed properly and probably never will.

Check the plexi inlet pipe is central to the pump and if its off just slightly maybe try loosening the res/top from the pump and adjusting so it is as central as you can get it.

EDIT for picture

Got a dozen or so of these res/tops sitting on the shelf and the first i pulled off is a prime example.

See how the plexi isn't directly over the inlet which would lead me to believe is could be drawing in air very slightly.

http://www.coolercases.co.uk/atcs_840/xspcres.jpg

[voigts]

I always try to keep a res at or as near as possible to the highest point as it is very hard for these pumps to get all of the air out otherwise.

If there is no debris in the pump impeller or block, then your only alternative is air.

[3Z3VH]

I dunno, on the Swiftech site the exploded view of that block shows a small plastic piece inside the block around the inlet. If that piece is snapped, or simply just too small to fit snugly, it would rattle around just like that.

I'd say pop open the block and check it out.

http://www.swiftnets.com/assets/images/products/apogee%20GTZ/gtz-exploded.jpg

See ? third piece from the bottom. It looks like something to direct the water out to the edges so it doesn't simply pass through the middle of the block and back out, but rather passes the water through the whole block. That looks like the culprit.

[Waterlogged]

I'm gonna say air as well. The XSPC res top is actually closer to being a T-line setup as far as bleeding and loop pressure is concerned, and it's all due to that section of acrylic pipe coming off the inlet. This is why is has the lowest pressure drop of any res, because the water doesn't really empty into a camber where air can compress...rather, most of the flow is directed back into the pump, leaving very little chance for the smaller bubbles to break free and float to the surface.

[JackOfAll]

OK. Today I stripped everything - the block and the pump. There is no debris in the loop. The noise is coming from the block and only the block. It doesn't matter whether the block is attached to the motherboard or free, or its orientation. When I stripped the block - it all looks OK to me. The inlet plate correctly fitted in the top housing, etc. etc. I can't find any fault with the block and the point I raised yesterday about the ridge on the inside of the TFC 45 deg compression fittings attached to the block - the fittings on the block have a tiny ridge - nowhere near the size of the spare I looked at.

What I've managed to do is isolate the noise a bit better in a sound recording. This time, the loop has only been assembled and running for an hour, their is some air in it, and you can hear the noise (background noise) that sounds pretty constant, like a ball bearing race or waterfall - small air bubbles suspended in water rushing through the block. What is not normal, is the two snaps that I have isolated in this small recording. I do not believe this is air. I'll eat my hat if it's air. The frequency and that fact that they stand out in amplitude as they do......

Here is a picture of the waveform .....
http://www.vacuumtube.org.uk/images/waveform.jpeg

2 snaps sound recording (http://www.vacuumtube.org.uk/proj1.mp3)

Guys, do you really still think this sound, (which is definitely originating from the GTZ block), is being caused by air bubbles in the loop?

[3Z3VH]

have you tried removing that inlet plate and running it ? Regardless if it fits properly.

[JackOfAll]

have you tried removing that inlet plate and running it ? Regardless if it fits properly.

Not yet. I'll give that a go tomorrow and also try using the barbs supplied with the block rather than the TFC's.

Having said that, I'm expecting the i7 bracket for the D-Tek Fusion V2 to arrive and chances are if that does arrive tomorrow, I'll call it a day and fit that instead.

[Sadasius]

Ninja's!!! You have Ninja's trapped in your block. They are very elusive and hide when you take your block apart. :up: They like water systems...Here is one on a sprinkler head.

http://images.quickblogcast.com/115838-108114/IMG_4578x.jpg

[hellcamino]

have you tried removing that inlet plate and running it ? Regardless if it fits properly.

My thoughts as well, the noise is too slow and too sharp to be a pump or rotational noise. I am guessing that the plate is flapping around inside the block. A bearing noise would be constant as would any other noise resulting from pump mechanical problems or trapped debris. Air will sound similar, not as sharp but generally faster and loudest at the pump itself with an accomanying rise in pump rpm when it hits the air bubbles. Try a different block or remove the plate and try again.

[JackOfAll]

Ninja's!!! You have Ninja's trapped in your block. They are very elusive and hide when you take your block apart. :up:

PT-Nuke doesn't kill Ninja's? Then that's going back for a refund! ;)

[hotdun]

I dunno, on the Swiftech site the exploded view of that block shows a small plastic piece inside the block around the inlet. If that piece is snapped, or simply just too small to fit snugly, it would rattle around just like that....

Very good idea...

[3Z3VH]

Did we ever hear what became of this ? I am still super curious.

[JackOfAll]

Did we ever hear what became of this ? I am still super curious.

I meant to update before now but never got around to it.

FACT: It's the GTZ block generating the snap, crackle and pop sounds.

I ran the block without the inlet plate and it did not make the sounds.

However, since then I received the D-Tek 1366 mount kit and mounted that instead. Silence and full load temps on average just a little (1-2 deg) better than with the GTZ (original mounts with inlet plate).

I've disassembled the GTZ and cant see why it would make those sounds or what could be faulty with mine. Go figure.

Anyway, I have no beef with Swiftech or their products. I can't say that my GTZ block doesn't work. It does. Just that it is noisy.

Swiftech has been good to me in the past - the time I snapped off one of the ports from an integrated pump/block. They sent me a new mid-plate from the US to the UK, and paid the postage. My initial enquiry to Michelle was about the availability of the part and I expected to pay for it. Gabe to Michelle -> "Please send replacement part to customer." Job done!

I intend to re-build a Core Quad system next week. I'll mount the GTZ on it and see. Not that I expect it to magically be silenced, but that case is sound-proofed, so we'll see. Otherwise, I'll fire an email to Gabe and ask if I can ship the GTZ back to them for a replacement.

[3Z3VH]

Glad to hear they are supporting you well !

It did seem that part was the only thing that COULD have caused the noise... mystery solved ! :D

[Tom128]

nevermind lol

[iandh]

My guess is cavitation caused by a combination of the 18w DDC and resonance of a slightly loose fitting insert. I have a hunch that if you undervolted the pump a bit you could knock the insert out of its resonant frequency and eliminate the noise.

Either way looks like you are set.

[Brother Esau]

Dude that aint AIR you hear thats a Fact!:eek:

SHUT DOWN OPERATION ....NOW!!!!! Like Gabe said examine everything very closely, This sucks that you have to break down your system and go through this but better get the jump on it right now before something gets damaged.

[soulesschild]

No, I don't think it is the pump. I was using this setup with a D-Tek Fusion V2 and Core2 Quad until I rebuilt with the P6T/i7 and Swiftech GTZ block.

Can someone confirm that the GTZ block is silent in operation?


Is that an ATCS 840? :D Your setup looks exactly the same to mine. I also have an UN Design fan bracket on the lower fan holding my mcp355 res top :D

To your original question, I have the same setup and my GTZ is silent. I had alot of noise in the beginning that only got silent probably around the 48hr mark.

[tomb18]

never mind...

[JackOfAll]

Is that an ATCS 840? :D Your setup looks exactly the same to mine. I also have an UN Design fan bracket on the lower fan holding my mcp355 res top :D


Yes, the case is an ATCS 840. The UN brackets are mana from heaven. (Well, at least if like me you have 10 thumbs and a drill phobia! ;))


To your original question, I have the same setup and my GTZ is silent. I had alot of noise in the beginning that only got silent probably around the 48hr mark.

That's pretty much done and dusted now. I proved to myself that the noise was from my GTZ block and it wouldn't have gone away. Snapping sounds aside, the GTZ does seem to take a lot longer to settle down compared to the Fusion. Anyway, I'm happy with where I am now and not really inclined to fiddle with this system any further. I had toyed with the idea of adding a MIPS block to the heatpipe, but installed the Asus supplied mini-fan instead and that has brought the chipset temps down. A couple more NF-P12's mounted on the back of the hard disk bays has dropped the temps of the 4 WE RE drives to mid 30's, and also means that the pump is being cooled by the fan to the side as well as from below. This system runs 24/7 and is my work PC - mainly to write and compile code. I'm going to buy another ATCS 840 for the Q9550 / Gigabyte EP45-Extreme rebuild. That I can fiddle with. It doesn't matter if it is working or in pieces.

http://www.vacuumtube.org.uk/images/c7p6t_d-tek.jpg