Performance Vs Noise - Experimentation

[UNDERBYTE]

I spent a day or two investigating building my own set up about 15 years ago, But ended up buying a 400 cfm set up with a large contract I got.

Most fan manufacturers bulid their own like the one in the link below, a box, some nozzels, couple of flow straightners, pressure taps, manometers, blast gate(to vary airflow) counter blower. Get ahold of the AMCA specs it will just take a day for you to figure out how to build it. Pretty simple actually, probably build it in a day or two for a couple of grand or less

I have seen homemade flow chambers washing machine sized to garage sized for measuring 40,000 cfm

Nozzels at the time were 300 to $400 USD apiece

I prefer the liquid manometers, you get more of a physical sense when the fan starts slip or when your flow test has flow anomaly over the digital ones

http://www.torringtonresearch.com/te...ent.html#tests

[UNDERBYTE]

I spent a day or two investigating building my own set up about 15 years ago, But ended up buying a 400 cfm set up with a large contract I got.

Most fan manufacturers bulid their own like the one in the link below, a box, some nozzels, couple of flow straightners, pressure taps, manometers, blast gate(to vary airflow) counter blower. Get ahold of the AMCA specs it will just take a day for you to figure out how to build it. Pretty simple actually, probably build it in a day or two for a couple of grand or less

I have seen homemade flow chambers washing machine sized to garage sized for measuring 40,000 cfm

Nozzels at the time were 300 to $400 USD apiece

I prefer the liquid manometers, you get more of a physical sense when the fan starts slip or when your flow test has flow anomaly over the digital ones

http://www.torringtonresearch.com/te...ent.html#tests

[BillA]

thanks
did/do you condition the air temp/moisture ? or just record it ?

comments on post#20 ?

[UNDERBYTE]

If you are at some elevation like Denver I would correct for that, also temp and barometric (radio shack meter or local weather internet source) you can do wet/dry bulb humidity as an exercise but generally I do not bother as impact is usally minimal. Although it it is good to start that way to get a feel for the process

[UNDERBYTE]

Re post # 20

On some airconditioning projects I worked on I found that condensers/radiators can act sometimes like that highspeed Delta fan with the downstream tuning vanes and increase/decrease flow through the condenser. PQ curve generally only puts you in the ballpark. OK for comparative devlopment to a point , then I confirmed/tweaked design by leasing Torrington Research's or EBM/pabst flow chambers for final evaluation on fans that my set up was not large enough for.

I never really get "to the bank" numbers until I test with a AMCA SET UP. Bootstapping airflow test set ups in my case usally ended being a hodgepodge. Too much wasted time trying to figure out whether the numbers were good or not.

Time is better spent putting together a reliable test set up rather than customizing each new test and then having to validate it. It takes me all of about 5 min to run a full PQ curve and less than a min. to take a system point like you would on a static radiator fan test. After all what is your engineering time worth? Mine is several hundred per hour

[BillA]

"Mine is several hundred per hour"
lol, Liar Gabe put my notions of such to rest; my 'effective pay rate' considering hours worked was <$10/hr
patsy BillA

[UNDERBYTE]

I quit taking consulting contracts several years ago although the money seemed good, the benchmark was always a strain to hit. After hours worked effective pay did not justify the labor.

there are better ways to make money

[Ls7corvete]

IDK about all this but I would like to see the delta EFB1212LE and the new coolermaster alum fan tested.

[bobo5195]

I am assuming we are talking about the ASME (acma?) spec here. I pointed out in another thread (I think it was at spcr) that building a test chamber for one of these is relatively easy and would give quite good results but someone said no for some reason. There are issues with low flows and measurement accuracy but ghetto/diy-able liquid in glass manometers are very good at measuring flows (as underbyte who probably knows a lot more about these things that i do) they do give nice consistent results. I’m not sure if a restriction might be better to create back pressure, just to save on cost and complexity.

Air properties might be a factor bill if your after your usual rigorousness. It’s a good statistic to have at least as it shows if a damp and cool basement is doing anything for example (thinking mike C at spcr here), although the results are only likely to have a small change. If your measuring to 2 sf they are likely to be just about noticeable (max 5% error).

10 dias have a small amount of friction which is known and the air flow distribution at the end is known as well, so its an error that can be factored out. More importantly the 10dias is a for a general flow a fan flow could well only need of the order of 1 dia.

Heat balance could work, indeed this is what thermal anemometers are, but you need a good test rig to confirm how the rad performs as well and looking at the variables involved it is too hard to get good numbers compared to other methods. It may well be non linear results as well which is annoying.

I disagree with testing fans at close distances to rads or obstructions as the variables involved explode ******ds. In reality you could never test all situations and variations between similar (same manufacturing line) fans would be a factor as well.

Im currently doing contracts for abit over $100 an hour, while still at uni but when you work out all the additional work its more like minimum wage and they will be feeling the force of my lawyer if they don’t pay me soon!

[BillA]

and I had hoped a technical field might buffer me from economics, fat f***ing chance that

re testing fan/rad assys, not so many combos really
base (for heater cores only today) is fan on fins
2ed is ~1/4" standoff (built into all 'computer' rads)
3ed is greater than 1/4" (no measurable dissipation increase however, why bother ?)

some w/o data may theoretically contest #3, but I've confirmed such now with 2 separate benches and numerous rads
the convential rot is 1/2 the fan dia, but also at that distance I can measure no increase over 1/4" (our little muffin fans ?)

my question was: if measuring the differential air pressure (in 1/4" plenum to ambient/rad backside), will that nominally equate to a point on the fan's PQ curve ?

[UNDERBYTE]

$$$$my question was: if measuring the differential air pressure (in 1/4" plenum to ambient/rad backside), will that nominally equate to a point on the fan's PQ curve ?
Today 09:45 AM$$$$


Generally they only match about half the time ( within 5% or less) the other half of the time +/- 5-10% Rarely do I get a spot on match from a system/PQ curve

occaisionally I have had radiator fin configurations that straighten the flow in an optimal sense and in so doing actually increase flow as much as 10% and others as much as 10% less than the system point.

When you add a radiator to a fan, in a sense it becomes a diff. fan with pos-neg. or minimal effect.

Probably close enough -

[bobo5195]

underbyte - doesn't that go away if you push with the fan instead of pulling through the radiator . The radiator then acting as a lamina flow guide before air is pushed into the fan.

The procedure has a number of things going for it, providing a good rad can be found for testing and the shroud was well sealed with a some nice liquid manometers then this could be an excellent test rig. Within 5% is good enough for me and repeated tests / a few different rads could help things.

It has the major advantage that most of the stuff is readily available in the community and is cheap (no digital meters, a lot of stuff that ppl have hanging around). Most importantly it’s measuring real life, exactly how someone would run a rad with a shroud.

You should try at longer lengths bill just to confirm you could well just be hitting a resonance point caused by the waterfronts coming of the fan. ½ dia seems to close to me, but I could well be wrong.

[BillA]

ok, I'll do it
should be getting a 0-40"WC dP xmtr which I'll recal to 0-5"WC this week
get some small brass tubing, thinking it run into the plenum under a fan motor strut
experiment with the tube end to avoid a pitot tube effect I suppose

the atmospheric pressure sensor would not cal so I am waiting for its replacement
capacitive humidity sensor ok, eventually should have 'corrected' air data

[bobo5195]

careful do you know exact dp vs flow for a rad and you need to be very careful with the manometer entrance. Idealy sheilding it some how as all kinds of flow directions will be coming of the fan.

Need a duct on the end of the rad to help measure dp across it as well.

[BillA]

eh ?
for the liquid side, sure; air side about zip, never will 'know' the mass/volume flow
no, rad exit to air (as is presumably done); pressure in area of unconstrained exit flow is I think sufficient for a differential pressure reading

this dP can be had with some precision; how sensitive such is to the tip configuration and placement needs to be assessed, I agree
(same readings from several similar positions = 'ok' tip effect ?)

when I get some data I'll send it round

[UNDERBYTE]

BillA,

Tis the season- Bill if you would like I could run a system resistance curve and a couple of PQ curves on my Flow chamber to give you a baseline to work your set up.

Normally I get several huindreds of dollars an hour for this, but if you pay for shipping I will donate an hour or two to the cause.

I test to AMCA 210-85 and am accurate up to 400 cfm

[BillA]

aiiiii
super offer ub
let me get my hardware straight, and sort out a useful matrix
a bare rad, then with a couple of fans ?
possible to get the 2 fans at 3 voltages ?
a better method ?

(An aside to all; procooling, being on the extremest fringe of WCing, has an extreme audience as well - many of whom have knowledge, experience, and resources in exactly those areas of interest to us. We should strive to make the contribution of services an enjoyable experience.)

[Marci]

Someone gag me... afore I volunteer a set of PA series rads...

[BillA]

not really for data, is a means (hopefully) to cross-correlate to get a nominal CFM (or flow resistance vs. ?) on a bench w/o air flow measurement capability
a single rad and fan(s) are a PQ reference, dissipation is apart
as a reference it is the specific combination of those actual pieces
(I see ~5% variation in performance between the 'same' rads, several mfgrs' products considered - all with the same fan @12.00 ±0.01 VDC)

[UNDERBYTE]

Originally Posted by BillA

aiiiii
super offer ub
let me get my hardware straight, and sort out a useful matrix
a bare rad, then with a couple of fans ?
possible to get the 2 fans at 3 voltages ?
a better method ?

Bare rad system curve ok

Couple of fans ok

can run full 6-12 volt curves if you like-or single system points at different V

can run shroud variations(Distance?) on single point say 12V ?

[BillA]

let me first run my curves, I need some plenum pressure data to 'look at'
sounds great
say system points at 3 voltages (?)
skip the shroud, at ~1/4" the dissipation is optimized (by test)

thinking out loud . . . .
2 general rad cases; higher airflow resistance (2 core) and lower (single core)
- 2 core selected for performance, more powerful fan
- single core for lower noise, less powerful fan

is either rad configuration 'better' than the other ? (to establish an air flow reference)
seems that the air flow pressure range would be greater with the 2 core/more powerful fan combo
?

I should add

the actual rad/fan(s) combo could be preserved and passed around between testers as useful for cross referencing benches

not too many doing rad testing just now though

[gmat]

yep 3 voltages, 12V / 7V / 5V as they are the common data points used on SPCR and other sites to measure noise. So by crossing your flow data with the existing noise data we'll be able to assess a performance / noise ratio. Sounds promising.

[BillA]

can o'worms now, fan(s) ?
2 max
higher and lower, with 30% overlap of rated CFM ?

specific models helpful

[bobo5195]

Sorry bill I was thinking about something else. I’ll post something when I am feeling better (ie not had 1 hours sleep, don’t have a lecture in 2 hours and haven’t been drinking absinth and having various other naughty stuff  ).

Looks good I’ll see what the results are like. I’m sure Les will build you a model.

Fans are indeed tricky, my mind is not totally reliable this morning but CFM should be proportional to voltage cubed I think.

I agree with testing against SPCR data as they are ultimately the benchmark in this area and we should try and beat their data. The more data points the better though.

[gmat]

2 fans sounds good
the usual suspects will be fine - Panaflo L1, Papst 4412, Noiseblocker, etc.