My Test Bench Project Log.
 

I just as well concede that I cannot give up this hobby so I just as well expand on it. What I want to do is put together a decent test bench for water block performance evaluation and comparison. Being I travel as apart of my job and have little free time while I am home the best way to go IMO is to try and setup an automated system so I don't have to babysit the bench and watch and record numbers manually.

The main goal is to set it up so all I have to do is mount the block and set a flow rate. Equipment I am planning on using.

I like the looks of the Adlink NuDam modules. http://www.industrialpc.com/nudam/nudamintro.htm

The modules I would get:
ND-6530 USB to RS-232/RS-422 /RS-485 Converter
This will be the module that makes a USB port the path from the modules to the computer.
---------------------

ND-6018 Channel Thermocouple Input Module
This is the module that will take the temperature measurements and send them to the PC for data logging and analysis.
---------------------

ND-6080 2 Channel Counter/Frequency
I THINK this will work to hook the flow meter to?????
---------------------

Other Parts of test bench:

Swissflow FS800 Flow meter
To take mutliple flow rate tests.
---------------------

Adjustable power supply (No specific model selected yet).
For adjusting the power for the heater cartridge in the die sim.
---------------------

Die simulator.
This will be based of Bill Adams design. I have a prototype slug made and will have to figure out how I want to insulate it. Final version pending insulation method.
--------------------

Needle valve.
For adjustment of flow rate.
--------------------

Analysis software (not sure what yet).
Something to take the data and convert it to something useful.



That is the main stuff. Cost will be anywhere from $1,000-$2,000 I am predicting.

Would this work or does anyone have a better solution?

Do you want a CrystalFontz 633 Like pHeastus has? I need to sell mine.

What is the accuracy and resolution of the thing? I think he had it setup for data logging?

I dont honestly know. pH should though. heres his review. http://www.procooling.com/reviews/ht...updated_2_.php

Jaydee does this mean you're repairing the mill too? I'd still like to get a good die sim made too if that's so :)

The main problem with the mill is the CNC part. There is so much backlash I can't compensate for it with the CNC software and it gets a little sloppy. I am fairly certain I can make a decent die slug if I just do it manually on the mill instead of with the CNC. It will take more time but I think it will come out better than that prototype I made before that was off a little.

Whenever I get back to the mill I will give it another go manually. Might be a while though. Headed to Georgia for a couple weeks and then it looks like right out to another job after that in Cali. Then the daughter will be moved back and I will have her on the weekends I am home. :D What I should do is drive up there today and pick the mill up and find a place to use it here....


What do you think of the CF633? Is it possible to get decent results from it? From what I looked up on their site it has an absolute accuracy of .5C, that is ok but most of my own blocks are less than .5C different :D. The main problem I see is the size of the thermo probes.

CrystalFonts 633
 

What do you want for that unit and what style is it?

It is a black drive bay kit with a red LCD with extra dow sensors (8 i think total). I used it for about a week before my videocard died, so its practically brand new.
Cost me 120 bucks plus customs and shipping.
$110 bucks shipped would be nice.

You got PM.

For temperatures, I'd still have to recomend that you get a reference thermometer, like I did, from Cole-Parmer (or anyone else you see fit). That way, you can calibrate your own instruments.

The one I got had a range of 19C to 27 C, with 0.1 increments. Accurate to one increment, NIST traceable.

The modules look fine. Calibration is going to be what makes your setup a good one.

You can add amp/volt measurement later, which is nice. I probably have a spare 2.5 volt reference IC, good for part of a calibration (when you get there).

As for a PSU, as you know, I got an HP/Agilent unit (pretty old) off eBay, relatively cheap. There might be some lag (voltage drop) as the unit powers up, but it should be rock steady after a 20 minute warm up (I figure). I have yet to test that one too.

How far off are you now from your new year's resolutions? :p

New years resolutions are shot to hell.:D

I am having secoond thoughts about the NuDam modules.
That is the specs for the thermocouple module. +-4% is not any much better than the CrystalFontz unit. I would prefer .05% - .1%. Especially if I am going to spend a chunk of change on it.

...then you will follow my path... come to the darkside...

:D

I'm in eBay 5 days a week, I'll spot a good unit for you. Just send me a good email addy, in case I come across something you'll want.

I have been searching e-bay religiously. I bid one of these: http://www.coleparmer.com/catalog/pr...=1&sel=0850216

Dosn't do data logging to PC but it does data log by itself. 4 days left on that bid.
The e-mail addy I use for this forum should be good. jaydee116 at customcooledpc.com if you find something useable.

It could come in handy.

I'm curently tracking:
http://cgi.ebay.com/ws/eBayISAPI.dll...tem=3880320262
analog output

http://cgi.ebay.com/ws/eBayISAPI.dll...tem=3880327141
analog output

http://cgi.ebay.com/ws/eBayISAPI.dll...tem=3880915616
Relay output only

http://cgi.ebay.com/ws/eBayISAPI.dll...tem=3880919247
Relay output only

http://cgi.ebay.com/ws/eBayISAPI.dll...tem=3880930985
Relay output only

http://cgi.ebay.com/ws/eBayISAPI.dll...tem=3880705370
Relay output only

http://cgi.ebay.com/ws/eBayISAPI.dll...tem=3880988724
Relay output only (8 inputs)

http://cgi.ebay.com/ws/eBayISAPI.dll...tem=7500415699
Relay output only

I'm not planning on purchasing any of them, it's just for someone else's benefit (in this case, you! :D)

No good pickings here, unless one of the first two has some other kind of output.

I just got the info on the Flow Meter I just won on e-bay.
http://www.dwyer-inst.com/htdocs/FLO...-VFCIISpec.CFM

it is the one in the middle with end connections. Should help flow a little without having 90's on both ends.
I got the VFC-141 model. 2% accuracy through the whole scale (whatever that means). It is a big meter at 13.5" tall. 1" NPT conections...

Well I am off to Atlanta Georgia on Wendsday. Will be there for a couple weeks. Meter should be here when I get back and I will know if I win that other thermometer or not.

From the WBTA thread is sounds like the equipment in my first post should be ok? If so I need to find some logging software to use for it...

2%, but with 0.5 gpm graduations. It'll do, it's a start.

Good luck on the other unit.

Today's eBay pickings are much better:

http://cgi.ebay.com/ws/eBayISAPI.dll...tem=7500805342
RTD only, mA output.

http://cgi.ebay.com/ws/eBayISAPI.dll...tem=7500628037
RS-485

http://cgi.ebay.com/ws/eBayISAPI.dll...tem=7500628050
RS-485

http://cgi.ebay.com/ws/eBayISAPI.dll...tem=7500562667
TC or RTD, mA output.
(check the model# though).

I am actually going to be building 2 test benches. 1 on a computer and 1 with the die sim. I will use the lesser parts on the computer bench.


Looks like I am going to pick up that CrystalFonts 633 from 9mm. With that and the flow meter I picked up I can start some decent testing. With some luck they should both be here when I get back from Georgia.

If I win this http://cgi.ebay.com/ws/eBayISAPI.dll...e=STRK:MEBI:IT auction I will have another toy to play with. :D

Those auction look very interesting. I like the sound of the controler. What kind of pressure deal can hook up to it? Also I assume there is some kind of flow meter than can adjust flow rate automatically?

As for the SwissFlow flow meter i am disapointed to not find any place to buy one easily. I might opt for this one http://store.yahoo.com/pulseinstrume...ewimoinbo.html. I belive they have a version with just an anolog output instead of that LCD display. I asked them for specs and they actually REPLIED!
-------------------------------------------
Pressure 300psig/20Bar
Full Scale Accuracy +/- 1%
I hope that helps.
www.pulseinstrument.com

I haven't used thermocouples.

I guess I'm tolerably familiar with the Crystalfontz 633 and the Dallas Semiconductor / Maxim IC ds18b20 thermometers. ( BB2K is definitely correct in suggesting that you calibrate them. http://dbserv.maxim-ic.com/appnotes....te_number/1117 # App Note 208: Curve Fitting the Error of a Bandgap-Based Digital Temperature Sensor - DS18B20 The math in that is not something that I'm particularly versed in. My gerneral sense is that the ds18b20's varry... As I write: three taking temp into radiator are within .15 deg. C of each other, three on the rad out within a .12 deg. C range -- and one of 8 air temps always seems to read a whopping 3 C high no matter how I arrange them. The last time I orderd from maxim they were $2.28 apiece, so culling out a few deviants from a bundle is quite affordable. ( I have ~10 on the shelf ) Last time I tried to focus on calibrating a set, I was overwhelemed by how much more effect the mounting of the sensors had than the error in the sensors. In my case: how well was each sensor measuring the water versus measuring the that which held it in the water. I'm actually feeling thick-headed on the topic at the moment and am wondering what's on sci-fi ...

A flowmeter with a 4-20 milliamp pulsed output in a range compatible with fan rpm outputs I've found to be convenient for automatic data logging. http://www.omega.com/ppt/pptsc.asp?r...000&Nav=gref02 is what I'm using. Mine is a tad askew -- I ran watter wetter through it and I think something in that kept causing a low grade nylon bushing to swell and sieze. I haven't had one in my hands or talked to anyone who has used one but I've been window shopping for a doppler or vortex-shedding or magnetic induction flowmeter. The zero moving parts appeals to me. I can imagine ( quite hypothetically for me as a hobbyist here ) swapping in the 5th waterblock into a test bench for the day, wondering if the flow seems low as reported by a turbine or paddlewheel meter, finding some sliver of hose trimmings or whisp of teflon tape stuck to the turbine, and asking myself "now how much data should I throw out?"
things on my window shopping list:

http://www.flocat.com/products/index...eries&sid=161&
http://www.clarksol.com/html/prodspecsDM_01D.htm
http://www.thorntoninc.com/products_pfa.htm

Thanks ToasterIQ2000. I been up and down Omega's site and did not come across that meter. The accuracy is 3% which is probably the worst part. Only $86 is good though. Those other meters you linked are nice.

Calibration is certainly a part of the setup and maintenance program. :)

Sorry, I meant to tip you there; email the people at the Swissflow website:
http://www.swissflow.com/

(Thank you WebLinks!)

Yeah, mine's mechanical too. I'll be adding a filter to make sure I don't gum up that flowmeter!

I'll be using a simple copper mesh screen for my PC, and I'll probably do the same with my testbench.

Linky!
http://www.twpinc.com/twp/jsp/produc...pe=3&page=data

These are 2" by 2", but you can fold/bend them into a cone, trim the edge, JBWeld a flange to it, and slip it inside of a 1/2" NPT fitting.

Follow the manufacturer specs to size the filter mesh.

(Thank you Weblinks!)

I already did e-mail them. No response for days now.

I'm not well rounded enough or experienced enough on this to say something decisive or final on this, but I since I mentioned that Omega flowmeter I would like to say that I've blown two fan headers using it. I'm not sure ecctly why, but I have not blown anything since I wired it up to power the flowmeter directly from the PSU via a spare floppy connector, and sending only one signal wire to the pulse / 'RPM ' / flow monitoring fan header.

It is possible it is soley do to some mistake in my wiring harness: something grounded out...

My other 'best guess' is that it only blows resistors near the fan header when the flowmeter is powered but not revolving for a long time. I've only blown the fan headers on my Crystalfontz 633, which does fan speed / voltage adjustments and such. ( I run experiments off something like the 633 so I do not trash motherboard components as often. ) The omega requires a pull-up resistor. Do all computer RPM sensing fan headers have pull-up resistance built in. ( I have done no research and a simple answer here would be neat, btw.) Does the 633 use pulse width management (PWM), I can't recall either at the moment. Anyway, I have this idea that when the flowmeter is stopped it may create some feedback with a fan header ... and pop: one more fan header on the 633 stuck at < 5v forever...

I haven't blown anything since I rewired it as above: unmanaged power from the PSU and signal only to RPM monitor.

I'd never assume that a resistor is not required.

The metering device may very well send pulses that are not "clipped" or otherwise "raw", requiring either some kind of diode or a resistor.

What you describe here is a blown input, due to a high output signal from the meter, burning out your fan header.

Unfortunately, I have no info on the input limits of fan headers, but it can't be too hard to find.

Back from Georgia finally. 2 packages were here awaiting for my arrival. :cool:
I was under the impression the plumbing config for the flow meter was a connection on the top and bottom. I was pleasantly surprised to find the multiple plumbing options. This will make plumping it much easier.
The Cole Parmer Thermometer looks to be in great shape. It has 3 YSI RTD probes with some kind of seals on them marked Jan 2005. Calibration looks to be in order thought. 2 probes read within .03C of each other while the 3rd is 1.5C off the others. I am still waiting on the CrystalFontz 633 although I am not sure how I will end up using that. If I can calibrate this Cole Parmer unit then is has all the temp readings I need. WB IN, WB OUT and die temp. That with the GPM measurement I should be able to get some better results.

http://www.customcooledpc.com/testbench/001.jpg
http://www.customcooledpc.com/testbench/002.jpg
http://www.customcooledpc.com/testbench/003.jpg
http://www.customcooledpc.com/testbench/004.jpg
http://www.customcooledpc.com/testbench/005.jpg

I got the CrystalFontz 633 today in the mail. Thanks 9mm. Will try to figure out how to hook it up tomorrow. Also working on plumbing the Flow meter and sensors into the loop.

Looking at CrystalFontz's website trying to figure out how the hell to hook this USB cable to my mobo I discovered they are located here in Spokane! :D

Anyway I have no idea how to get this USB hooked up.... The cables that came with it have a small ass 2mm connector that goes to nothing I can find on the 633 unit...

EDIT: I found it hiding under the USB PCB. :D

I spent the day playing with the CF633 and plumbing in my test bench. I got 8 Dow sensors with the CF633 and found 4 probes read real close and 2 are almost identical. I am using the two closest one's in the bench. 1 under the CPU and onw water in. The flow meter works good. Attached is a graph of the two sensors I picked.