causes of pressure/flow fluctuations

[BillA]

looking to identify the source of pressure and coincident flow fluctuations in a WCing system

the pressure and flow indicators cycle in unison so it is assumed not to be gauge related

under steady state conditions at high flow rates (~3gpm) there is a flow variation of ~±3%, less at lower flow rates
-> BUT under some (undefined ?) conditions the variation is ~±0.2% (obviously desirable)

the pump is an enclosed oil bath motor with a direct-drive centrifugal pump;
line voltage varies ±2V max, running the pump off a SOLA with a ±0.2V swing only slightly reduces the flow variation

had/have a 3qt vented res in the system, removing it eliminated the variation;
reinstalled it on a hard plumbed loop brought back the variation
(system capacity ~1 qt w/o res)
- sealing the vent seems to make no difference)

the variation seems to consist of 3 independent cycles;
1 minor of ~±0.2% with a period of perhaps 5 sec.
2 'major' of ~±2% with periods of 10 to 20 sec., not quite the same period and not in sync

doubt that it is the low modulus of the tubing;
much is copper and the hose is thick walled reinforced silicone (heater hose)

ideas anyone ?

be cool

[Tecumseh]

It makes no difference if the res vent is sealed or not? Can you
tell if the water level in the res is fluctuating?

Also changing the pump made no difference, right?

At the high flow rate is the pump operating within the
mfger's recommended limits?

Since most everything else has been eliminated, it must have
something to do with the plumbing of the res itself.

[The Overclocker]

it definatly sounds like the pump to me, although the voltage draw only changes by 4v (+-2v) which is a very large ammount it its self, what is the current draw, and does this fluctuate.

chances are the changes measured are down to dirty AC power - not only would this cause changes in just one component (say pump) it could also be magnified if the test equipment to measure the changes are powered by AC

[BillA]

dinking around for another half day w/o any progress

removed the copper res bypass loop and replaced with a straight section
- still the same variation

when running the pump off the SOLA (±0.2V variation) I see the same fluctuations
don't see how the motor's resistance is fluctuating - so the current is moot
(one might expect it to vary, but as we are talking about a 0.2psi range the variation will be pretty small)

I agree that the pump seems most likely, but bear in mind I have 2 pumps doing exactly the same thing
- which says it is:
not the pump
or
a pump characteristic

now I'm quite befuddled as to how I attained that previous stability

EDIT: addl info
no change if res vented or not, full or not
no apparent surging in res
pumps well within specs (given the always present 'resistance' of the 0.25" ID flowtube)

[gone_fishin]

Ah, a new section and a problem that is contemplatable by all, perhaps my first and last post here:)
It goes away when the res is removed, so it must be res related or at least highlighted by the res. Try a different size res, say double size, and see if the "waves" change in duration. The suction in the res could be affected by drawing off of the swirling water in the res?
Not an educated guess but it is logical. If you change the size of the res and the frequency changes, then you may be able to control it by inlet outlet positions in the res and res size.
Still does not explain how you attained a better equalibrium at that one time you mentioned.

If you research instabillities in incompressible flows, you turn up a lot of egghead (no slights intended) indepth studies. example (http://arxiv.org/PS_cache/cond-mat/pdf/0109/0109084.pdf)
This is where the engineers need to step in and start posting away:D

Edit: disregard the res ramblings, I misinterpreted your description:)

[BillA]

re-check description g_f (or I am unclear)

initially it went away when the res was removed
so I hard-plumbed a bypass
now I have it with or w/o the res
and even with or w/o the whole bypass assy

most perplexing

piping/hose with resonant and variable standing waves, nooo

be cool

[nihili]

Bill,

The only idea I'm coming up with is that there is some turbulence set up by one of the plumbing joints involving the resevoir and bypass which is itself sensitive to pressure variations. Perhaps if you could give a more detailed description or diagram of the system it might spark some ideas?

nihili

[BillA]

and that is just what I have done several times, 'go back' to the stable setup
-> but now I can't do so
(made some change I can't recall ??)

the setup is a tad complex, 3 hose sizes, over 80 fittings, over 15' in length
not too sure how to even describe its parts, some in the articles

I'm thinking to go 'back' to a simple loop with a pressure sensor only - and then start incrementally adding back in the bits of the system

jeez, real high tech here, eh ?

be cool

[The Overclocker]

Originally posted by BillA
and that is just what I have done several times, 'go back' to the stable setup
-> but now I can't do so
(made some change I can't recall ??)

the setup is a tad complex, 3 hose sizes, over 80 fittings, over 15' in length
not too sure how to even describe its parts, some in the articles

I'm thinking to go 'back' to a simple loop with a pressure sensor only - and then start incrementally adding back in the bits of the system

jeez, real high tech here, eh ?

be cool

sounds fun :rolleyes:

going back to the problem described:


the variation seems to consist of 3 independent cycles;
1 minor of ~±0.2% with a period of perhaps 5 sec.
2 'major' of ~±2% with periods of 10 to 20 sec., not quite the same period and not in sync


they seem to be two independant events, seperated into the major and minor fluctuations

it may be possable to ignor the first fluctucation as it is within the error margin of your equipment ( ±0.2%) HOWEVER as the flucuations seem to occur with a set pattern (every 5 seconds) this would point to something affecting it.

originally i thought it was related to changes in air pressure in the res -(air sucked into hosing, causing more water to be pulled though the system to even out the pressure - but as you have a vented res this should not be a problem)

although i realise that your watercooling system is very complicated - the same ideas should appy to a 'normal' watercooling system saying that with a res - especially a vernted one - there will be NO suction before the res - this furver eliminates the pump

however the problem now is very nhard to find - due to the lack of mechanical compenents in the system - is it possable testing quipment could be effecting it?

i am at a loss - i just hope yuo find the problem with a unfortunatly tedious strip down

[PCphreak]

Originally posted by BillA
the setup is a tad complex, 3 hose sizes, over 80 fittings, over 15' in length
not too sure how to even describe its parts, some in the articles


The flow rate of the system is determined by the head capabilities of the pump, as well as the tubing's total linear length & it's ID (inside diameter). The horizontal lenths affect this flow rate, as much as verticle lengths. They're both resistence in one magnitude or another. Smaller ID tubing also has more resistence than large ID tubing. This resistence is pressure.
AS AN EXAMPLE, your pump may be rated at a max of 100 gal/hr @ 1ft head @ 3.0 PSI.
Pretend your piping is only on the horzontal plane for the most part, and your total linear footage is 18 ft. Even though the piping is not on verticle inclines, the 18 ft of resistence may be enough to equal 5 ft of verticle head.
Lets say at 5 ft of head, your pump is at it's absolute edge & is just barely able to keep the water moving.
But when you lessen the linear footage the water has to travel (like removing the reservoir), you elimate enough resistence in the system & it will now give you zero fluxuation problems. Afterwards, you then plumbed in a bypass. This bypass probably added back in the equal amount of resistence as the reservior did when it was installed- therefore the fluxuations came back.
You may just need a bigger pump to fix this issue. At 18 linear foot, the pump may be operating at it fullest capacity.

-PC

[nihili]

Originally posted by PCphreak


With the rating on any pump, the flow rate is determined by the tubing size, length & head capabilities of the pump. The horizontal lenths affect this flow rate, as much as verticle lengths. They're both resistence in one magnitude or another. Smaller ID tubing also has more resistence than large ID tubing. This resistence is pressure.
AS AN EXAMPLE, your pump may be rated at a max of 100 gal/hr @ 1ft head @ 3.0 PSI. Your piping is all on the horzontal plane for the most part, and your total linear footage if 18 ft.
Even though the piping is not on verticle inclines, the 18 ft of resistence may be enough to equal 5 ft of verticle head.
Lets say at 5 ft of head your pump is at it's absolute edge & is just barely able to keep the water moving. But when you lessen the linear footage the water has to travel (removing reservoir), you elimate enough resistence & you have zero fluxuation problems. You plumbed in a bypass, probably adding back in the equal amount of resistence as the reservior did.
You may just need a bigger pump to fix this issue. At 18 linear foot, the pump may be operating at it fullest capacity.

-PC

PC,

Can you explain why you think this would result in the described periodic variations, particularly 2 major but independent cycles, rather than simply a decrease in flow? Even if you take into account that resistance varies in proportion to the square of the velocity of the fluid (if memory serves), why would you expect cyclic change rather than tendency to equilibrium?

nihili

[PCphreak]

Originally posted by nihili


PC,
why would you expect cyclic change rather than tendency to equilibrium?

nihili


Another comparable pump may continue to build pressure with the expense of flow rate since there is so much resistence. While, I believe this pump is showing variations because it's at the maximum operating conditions. Therefore, pressure suffers with flow rate but not necessarily in proportion. I've seen poor quality pumps exhibit this.

Example:

Good Pump:
Resistence increases > pressure increases > flow rate decreases

Bad Pump
Resistence increases > pressure decreases > flow rate decreases

-PC

[BillA]

PCphreak
you have failed to read, or perhaps understand, the problem description
a centrifugal pump will pump that which it can against the system's resistance
that is understood, but thanks

after a day of component substitution, the following seems to be indicated:

- the 'minor' cycle is a pump characteristic
- the 'major' cycles are suppressed or nonexistent only with a very short loop not containing any soft thin wall tubing
- the addition of even fairly 'stiff' hose, res, flow meter, or wb assy causes the cyclic fluctuations
(the hose is reinforced silicone heater hose, ~3/16" wall; no change with a vented or sealed res)

'obviously' the flow is surging; the low modulus of the tubing, and the amount of such in the system suggests it is the enabling component

tomorrow I will hard pipe the system and see if that stabilizes the flow

be cool

[PCphreak]

Originally posted by BillA
PCphreak
you have failed to read, or perhaps understand, the problem description
a centrifugal pump will pump that which it can against the system's resistance


I understand what your saying. I work for a Plumbing/Mechanical company. I install both boilers (centrifugal circ. pumps) & Well water systems (centrifugal pump). A centrifugal pump will continue to build pressure as greater resistance is met until an equilibrium is reached. This is a symptom I look for in a boiler heating system. This will tell me when an expansion tank is 'logged' for instance (creates resistence & head pressure).
Having tubing of a low modulus will have greater resistance per length than that of ridgid pipe. The reason is low mod tubing has a profile more like an oval and can pinch in bends; where as ridgid will be circular having a greater overall inner volume.
I'd like to think that the resistence in conjunction with the pump no coping with this, is causing system fluctuations. Ridgid piping, or increasing the ID of the low modulus pipe (within reason) should stabilize the 'surging'....
You could also try installing a balancing pipe equipped with a ball-valve, between the supply and return. Use the ball-valve for tuning. Then again, this all depends on how your currently plumbed. BTW, what brand are the pumps?

-PC

<edit>

Hell, nevermind...I completely overlooked that this was tested in an open state.

Originally posted by BillA

(...no change with a vented or sealed res)


</edit>

[Greedy Guido]

Bill,

It may be a suction problem. I have seen this on large pumps before where as the flow rate increases it reaches a point where there is not enough NPSHA (net positive suction head available) and the pump experiences transient cavitation for a very short period (sometimes it may not be detected by ear). As flow rate drops NPSHR (net positive suction head required) is reduced to a point where it is less the NPSHA and the pump runs on again. This happens mainly with very low discharge heads (high flow) and when the pump is pushed beyond design.

With a closed system NPSHA tends to by higher as you recover some from the discharge side. Running a reservoir opens the system (even with the vent closed) and NPSH is dictate by pump suction only.

Try running a larger suction and if possible the reservoir on the suction side with a very short hose to the pump. If the problem disappears then you have an NPSH problem.

If not can you draw up the system giving all the details and I will do some calculations for you. Please include pump details and curve if possible.

Hope this helps.

[BillA]

re my system
I have about 1m of 3/4" hose/copper tubing feeding the pump w/o the res,
with the res as originally configured ~1/2m of 3/4" from the res
(but a 1/2"NPTF outlet on the res with a 90° fitting to the 3/4" hose ??)
the pump has a 1/2"NPTF inlet in which there is a 1/2"NPTMx3/4" barb

the NPSHA comments make much sense (20 yrs since I used !)
I will replumb and close couple the res w/o the 90° ell

certainly a lot easier than replacing the hose with copper tubing
if this does not resolve the problem I'll work up a circuit description

[BillA]

no good news from the plumbing exercise

- installed a vented res 6" from the pump inlet connnected with 3/4" Cu tubing
(through two carefully chamfered 1/2" NPT straight connections, same as the pump inlet)
- and replaced ~8' of silicone heater hose with 1/2" Cu pipe

slightly higher flow rate (~3%)
very slightly lower pressure cycle amplitude (~3% vs. 4%)

so wtf is going on ?
bummed out in Texas !

be cool

[gone_fishin]

Could be a pump symptom. Have a look here (http://www.mcnallyinstitute.com/CDweb/c-html/c071.htm). Pressure pulsations are mentioned.

Dig around at that site, there is a lot of info on pump operation and cavitation problems.

[BillA]

as Guido mentioned, it seems to most probably be a NPSHA problem
but if so, and I cannot eliminate same with a close-coupled vented res, don't quite know what to run at next

I would conclude that it was an impeller characteristic,
EXCEPT that with a very short loop the pressure/flow is virtually rock solid

??
be cool

[nihili]

Originally posted by BillA
I would conclude that it was an impeller characteristic,
EXCEPT that with a very short loop the pressure/flow is virtually rock solid


Hmmmm, what if we turn that on it's head. What if it is an impeller characteristic, but one that is compensated for with very short loops. For example, if the loop were shorter than the wavelength, then perhaps the wave couldn't form.

Try testing with an open pipe, so that there's no connetion between the output and input of the pump, what happens then?

nihili

[BillA]

is that not what I have when supplying the pump from a vented res ?

be cool

[Tecumseh]

Now I wish you did have a DAQ card so I could look at the
FFT of the waveform.

What bothers me is that you did not observer this problem
before you made that high temp clear-out run. It did not
exist, correct? Since you have run another pump that was
in storage and the fluctuation still exists, doesn't that
point back to the modulus idea. Your plastic pipes got
soft and thus started to "breathe" maybe?

[BillA]

nah, this is/was reinforced heater hose, rated at 250°F for 150psi - stout stuff
and now I've Cu pipe and still the fluctuation (some hookup hose but less than 3 feet)

I'm even thinking of putting both pumps on line

in theroy I should also be able to reset the time constant of the Danfoss FloMag
but it's sooo damned complicated I can't figure it out, Monday I'll call the rep

but the better solution is to reduce the amplitude, not smooth the meter

not cool

[nihili]

Originally posted by BillA
is that not what I have when supplying the pump from a vented res ?

be cool

You're right, and you still got the fluctuation. So why couldn't it be an impeller caracteristic that is just compensated for when run with a short closed loop?

nihili

[BillA]

ok, I see what you're asking

because for over a year and a half, with dozens of different setups - it was as stable as could be

but after a 150°F run with Simple Green (to get rid of that GD WW gunk), the cyclic fluctuations started
- which now only disappear if the loop is very short (~4' OAL)

not cool

[Tecumseh]

Originally posted by BillA
nah, this is/was reinforced heater hose, rated at 250°F for 150psi - stout stuff
and now I've Cu pipe and still the fluctuation (some hookup hose but less than 3 feet)

I'm even thinking of putting both pumps on line

in theroy I should also be able to reset the time constant of the Danfoss FloMag
but it's sooo damned complicated I can't figure it out, Monday I'll call the rep

but the better solution is to reduce the amplitude, not smooth the meter

not cool

You are right. I have seen people ready to give up on stuff like
this "tune the meter" and it's not pretty. It's actually too easy
to do, but will come back to bite you every time. Do not go
down that path. The Dark Side clouds everything.

Maybe it's time to look on Ebay or LabX for a different kind of
pump. :bang head

[Tecumseh]

Bill, I just had an awful thought. What if the high temp run
damaged your instruments? They were in the loop for the
high temp run, right? Maybe we can't find the cause of the
fluctuation because it does not exist. :eek: :eek: :eek:

[BillA]

not likely, 3 instruments involved
Foxboro 832DP
Foxboro ICS ss transducer
Danfoss 1100 ceramic/platinum flowtube

they are all rated far over 150°F, and their readings are consistant and the 'same' as before
except for the added fluctuation which shows on all three

not cool

[nihili]

I take it that you've run 150 degree Simple Green through pumps before with no ill effects?

nihili

[BillA]

no, so I assumed I'd ruined the pump as it is rated only at 120 (+/- ?)
so I bought ANOTHER GD pump -> same damned thing !!

pump and impeller are glass filled pp, should be ok for short exposure

not cool

[nihili]

Ok, this is a bit of a stretch, I'll admit. Have you tried gunking it up with WW again to see if the fluctuation smooths out? Perhaps the WW was smothing things out?

nihili

[BillA]

lol

you got me with that one nihili
I guess I could, but I sure DON"T want to

in thinking about it, you may have hit it
the stuff builds up on the walls with decreasing 'density' as it gets thicker
-> a perfect viscoelastic dampner

sh*t

ok, but why then is there no cycling in a clean system whose length is short ?
(pump/impeller ??)
you're on a roll nihili, don't stop now

be cool

[nihili]

Originally posted by BillA
lol

you got me with that one nihili
I guess I could, but I sure DON"T want to

in thinking about it, you may have hit it
the stuff builds up on the walls with decreasing 'density' as it gets thicker
-> a perfect viscoelastic dampner

sh*t

ok, but why then is there no cycling in a clean system whose length is short ?
(pump/impeller ??)
you're on a roll nihili, don't stop now

be cool

I already covered that one, when the length of the system is less than the wavelength of the fluctuation (or something like that) then there isn't enough room to get the waves going.

I'll bet if you calculate the wavelength, you can predict the action of shorter systems and maybe even calclulate the secondary waves on the larger systems.

Do you find that the characteristics of the minor wave and one of the major waves change with changes ins overall system length?

nihili

[BillA]

no way to do so with the requisite precision
need that DAQ system
Guido offered to calculate it for me but I'm going to have to generate head loss curves for the heat exchanger and flowtube
and while I'd 'like to know', my immediate goal is simply to eliminate (or at least significantly reduce the amplitude)
- kind of pointless for Guido as there is nothing I can actually do with the info

I guess just for ducks I'll experiment with the 2 pumps and a bypass valve tomorrow
perhaps I can 'de-tune' the system depending of where I place the bypass

ideas ?

be cool

[Tecumseh]

Bill, have you always used WW in the coolant? If there was a
time when you did not use it and you also did not have the
fluctuations then another theory bites the dust.

[BillA]

over the course of several years its hard to speak in absolutes,
and this pursuit of precision is so evolutionary in nature that comparability is a real issue
- not to mention that seldom has a specific setup lasted over 6 weeks

but generally speaking I've always used WW,
probably in excess as with the makeup water I occasionally added a bit of WW also
- the gunk buildup was always noticeable, and of late quite objectionable as it was beginning to slough off and making control impossible

so I cannot say how a WW free system responded last year,
for ALL of my instruments are also different (and much better)

anyone care to predict 2 pump effects ?

be cool

[Greedy Guido]

Bill,

Are you saying that you have two pumps of the same type and both have the variation.

If so I might have an answer.

Give me a day to think about it a bit more.

[BillA]

yup, 2 pumps; one 2 yrs old, one new
both singing the same song

the pump specs are here (http://www.electricmotorwarehouse.com/PDF/1P323.pdf)

the brazed plate heat exchanger is an Alfa Laval CB 12-36 from Sweden using o-ring connectors and the metric equvalent of 1/2" copper pipe (a bit smaller ?)

I'll work up a circuit description for you Guido
Thanks

be cool

EDIT: and the flowtube is a Danfoss FlowMAG 1100, size: DN 6
(I'll look for a pressure drop curve from Danfoss)

[Tecumseh]

Originally posted by nihili
Ok, this is a bit of a stretch, I'll admit. Have you tried gunking it up with WW again to see if the fluctuation smooths out? Perhaps the WW was smothing things out?

nihili

nihili, I think you might get the Sherlock Holmes award for this.
The famous quote:"It is an old maxim of mine that when you have excluded
the impossible, whatever remains, however improbable, must be the truth."

Bill, put the WW back in so we will know for sure.
We gotta know! :D

[BillA]

WW back in ???
you gotta be crazy

in digging up the pump specs for Guido, I noticed that Little Giant offers a variant with a 3/4" inlet, for submerged apps only

so now I'm modding one of these 1/2" inlet pumps to be direct coupled to the res with 1 1/4" Cu pipe

still thinking its a NPSHA problem
the symptoms are all there

be cool

[BillA]

I heard ASARCO is going to open a new pit in Chile,
and NIBCO a new copper fitting forging plant in Houston
my consumption of fittings is amazing, and going to continue I suppose

I am now quite convinced it is an intake suction problem
I modded the pump intake to a gentle contour eliminating the 1/2" NTP threads, and bonded on a 1 1/2" tee with the branch facing up on which I put some clear hose

a differential manometer when compared with the res 6" away,
and as expected there is a progressive drop of up ~ 4" at the max flow rate
AND this level cycles ~1/8" as the pressure/flow goes up and down

despite all my re-plumbing, I was unable to eliminate the 1/2" connection at the base of the res feeding the pump;
tomorrow I'll install a split return line to both the res and the 1 1/2" stand pipe at the pump inlet

AND I recalled another bit of the past with this pump:
until 8 weeks ago when I installed the heat exchanger, the pump had always run submerged in the bath (which was why I used WW) and without any intake piping

I do believe its all a suction problem which I did not 'see' because the system was so 'gunked up' with WW

be cool

[Tecumseh]

Just for the record could you characterize this "gunk?" Like is
is a red tar or what exactly?

Will you ever use WW again?

[BillA]

never

apparently it is the synthetic oil coming out of solution

a beige semi-polymerized 'film' that builds up on the surfaces until it is thick enough, or otherwise disturbed, to slough off

it is not an algae, I had some amount of it in a baggie for over a month and it did not smell or change

it can be broken up into smaller pieces, but apparently will not go back into solution

and this is the 3 time in 2 yrs I've observed this
- but I have a big open system that takes a fair bit of makeup water

[Cathar]

So are we seeing a case here against the long term use of Water Wetter?

[BillA]

yup
and given the large surface area and narrow channels in your wb, you may wish to encourage users to go with a copper/brass rad and skip the WW

what is needed is a comparison of the cooling 'hit' with WW vs. say 15% antifreeze

some day when I'm really bored . . . .

be cool

[BillA]

today I split the 3/4" return line to dump into both the res and the standpipe at the pump inlet
- it had the desired effect of eliminating the negative pressure drop at the inlet and there is no discernable fluctuation of the vented res/standpipe level(s)

BUT, I still do have some pressure/flow fluctuation although its down to about 2% (in terms of flow, really all that I am concerned with)

got to believe I'm now dealing with a pump/impeller characteristic

thanks nihili for identifying the role of WW in all this

be cool

[Cathar]

Interesting, in the technical details for RedLine Water Wetter, one of the things that it states that it does is form a thin film between the pipes and the water to protect the pipes from corrosion.

This would explain the residue that every complains about after using it for a while.

Perhaps Bill, Water Wetter is indeed doing just what it was designed to do, which for us CPU water-coolers is exactly what we don't want...

[BillA]

re WW

yes, but a thin film to me is not something that sloughs off in ropy tendrils
I know that I am done with WW

be cool

[Tecumseh]

Originally posted by BillA

thanks nihili for identifying the role of WW in all this

be cool

So do we all agree that nihili gets the first ever forum
Sherlock Holmes Award for solving a difficult mystery
in water cooling? You got my vote, nihili. :D

And now we may be seeing that Water Wetter may not be
our friend at all. This is a hugh change.

[BillA]

yes, a big S.H.A. for nihili

but we have only circumstantial evidence, no numbers, eh ?

be cool

[nihili]

I think this is one of those cases where not having a clue to begin with makes it easier to find one. ;)

[Cathar]

Originally posted by Tecumseh

This is a hugh change.

Almost as big as going from Liz Hurley to Divine Brown. :D

[Tecumseh]

Originally posted by Cathar


Almost as big as going from Liz Hurley to Divine Brown. :D

I did NOT need that picture in my head. :D
But that is a technical issue. :)

[Greedy Guido]

Billa,

I have talked with a very senior engineer where I work and he said that my idea may be what is happening.

This is what I think.

Unfortunately for you, I believe that this is a characteristic of your pump (3% flow variation).

This is mostly caused due to motor slip (variation in motor speed) or mag drive variations. With a res you have to accelerate the water so you will notice it more then if the system is closed. A closed system acts like a dampener and the changers are smaller (due to water momentum). My senior engineer could not recall every hearing or wittnessing this himself but he said my idea could be possible.

Have a think about it. It could be total bull****.

[BillA]

well here I specialize in testing theorys

with the present setup virtually all of the return flow is going into the standpipe
(the return line is on the T branch)

so if I close the system the momentum aspect should not be a factor
- I have bypassed the res several times but that was with the 1/2" NPT x 3/4 barb connector in the pump inlet

Guido, a question:

what will be the effect of enlarging the 'inlet' hole itself in the pump casing ?
(where it faces the eye of the impeller)
it is, for reference just about the same dia as a 1/2" fitting ID

be cool

[Greedy Guido]

Bill,

If it is indeed a NPSH problem, enlarging any of the suction will help.

[unreal]

thx for the help this thread helped me to not use WW in my system just distilled water and anti-freeze

[FizzledFiend]

must know what happend here. Did enlarging the suction side decress the fluxuation?

[CrystalMethod]

Just a thought, but could the comination of a few different chemicals be causing this? Possibly somethng no one has looked at as a possibility yet? The solder perhaps? Rosin from the solder? As of yet, we're all just speculating, as to the cause of this. The common demoninator being WW. But, from experiance in racing cars, I know that I have never encountered this problem with any wetting agent, including Water Wetter. There's a reaction with something we're using that's causing this. Something common to wetting agents that doesn't react well to higher pressure, and relitively lower temps. (Don't forget, we're using a product that's designed for automotive purposes, and not for cooling computers).

[eobard]

I came to the "chat" area looking for the "Merry Christmas" thread, and instead I find this which had lain dormant for over a year before being revived. Well, *yoink*