Excel Experiment Radiator Estimator

[Martinm210]

Since I had all three RS series radiators tested now, I decided to use that data and fiddle with a different estimator. Granted it's still just a rough estimate, (different fans will provide different performances) but this time I created power curves for the c/w vs fan rpm data points I had and made the estimator have the ability to select any fan RPM at 50RPM increments.

In addition, I made it possible to add up to four radiators in series to see thermal performance benefits.

Just playing around with format, what do you think?

I'm still going by the assumption that 5C or lower is Extreme Performance, 10C average, etc..but I'd be interested in what people think about that too.

Attached is the 97 excel format file zipped.

If it's recieved well, I might start making individual ones for each radiator, etc.

What would be helpful to add or change with this?

Update 7-27-08 Newer version:
Added a percentage user control shroud gain and did some chart formatting to make it a bit more clear.

Also made several other versions of radiators I had tested.

[Eddie3dfx]

Very Cool Martin!
Thanks.
I just tried it but something is wrong with the drop box by the heat load.. hmm maybe it's just my program acting finicky.
Geez, just my cpu alone is 280 watts of heat.

[Martinm210]

Very Cool Martin!
Thanks.
I just tried it but something is wrong with the drop box by the heat load.. hmm maybe it's just my program acting finicky.
Geez, just my cpu alone is 280 watts of heat.

Yeah, I just left that one a regular old entry(Click the cell and type in the number), I can make it a drop box though. Something like every 5 watts?

Martin...

Would it be possible to get one for the almighty PA120.3...???...

Yep, I just wanted to start with these to get the formatting down, then I can just save as and enter in the curves for different rads I've tested.

[WoZZeR999]

I tested it a few days ago with my MCR320, and it seemed to work correctly. I saw it on your site a few days ago.

[Eddie3dfx]

My mistake. I initially opened it up as read only. Works now.
Thanks, this will be fun to play with.

[Martinm210]

I tested it a few days ago with my MCR320, and it seemed to work correctly. I saw it on your site a few days ago.

Great!

This one is a bit different in that you can select any RPM and you can also add multiple radiators together, but same concept.

[majestik]

Very nicely done! Probably the best looking excel calculator i've used =)

[jkresh]

great idea, maybe someone could take the results from vapors fan tests and either add all the fans or modify the calculation so instead of rpm it could be cfm and pressure? That combined with your flow calculator (and maybe a little scripting) could come close to automatically generating a system (or at least some suggestions) for a given load.

[Bobly]

OOO nice Can't wait to see the all popular PAs though

[Xilikon]

Great work there, would help people pick the right radiator.

[disruptfam]

nice work martin!!! you have done alot for us watercoolers!!

looking forward to it being updated with the other big name rads

[Martinm210]

Thanks guys,

I'm curious about the shroud performance boost. I've experimented and found I can add a "Shroud YES/NO" pull down that'll automatically give it a performance boost based on some curve, I'm just not sure it's enough data to apply to all the radiator types (Too much error). Would it be worthwhile to have a should "boost menu" switch?

But if I make a separate file for every manufacturer, then I could have a separate shroud performance extra for each one.

[JaD]

I thought flow rate made a major difference in radiator performance, unlike waterblocks. Is that wrong or just too complicated to add? Or, probably, a pita to test :P

Forgot to say thanks for the great work

[Xilikon]

I thought flow rate made a major difference in radiator performance, unlike waterblocks. Is that wrong or just too complicated to add? Or, probably, a pita to test :P

Forgot to say thanks for the great work

The flow vs thermal dissipation curve is mainly flat above 1 GPM so it's safe to remove this from the equation and use 1.5 GPM as a reference point.

[Barb]

I think there are several more inputs that need to be considered to even achieve a gross estimate. This type of analysis is touchy even when all the conditions are considered and well measured. I understand the need to keep it simple but this is just a fancy way of saying more Radiator = better performance, Higher fan speed = better performance.

I noticed the heat load doesn't effect the system Cº/W, It's been a while but I thought that number would vary with the heat load. Since the higher the load means higher temps and a higher heat transfer rate between the system and ambient.

Also, Is it valid to simply add the C/W for multiple radiators? I would think you would get diminishing returns pretty quickly.

[Xilikon]

Also, Is it valid to simply add the C/W for multiple radiators? I would think you would get diminishing returns pretty quickly.

Martinm210 did a estimator a while back with this method and the results isn't really valid with lots of discrepancies. Now, he used actual curves from each radiator size to make a more valid estimation. Like you said, adding the c/w numbers doesn't always work since it's not really linear.

[Martinm210]

I thought flow rate made a major difference in radiator performance, unlike waterblocks. Is that wrong or just too complicated to add? Or, probably, a pita to test :P

Forgot to say thanks for the great work

I started the whole flow rate estimator and to be honest the more I got into it the more I realised flow rate doesn't have that much influence. Generally if you have over 1GPM, gains and differences are too small for most users to see. Most radiators are similar to blocks on flow rate effects. 1.5 GPM is pretty easy to attain +- so I picked that for all my radiator testing. It's also very hard to measure, because I found that as flow rates are increased in my test system the added heat by the pump for the higher flow rates essentially cancels out any gain. That tells me it's too small of a difference to matter and in the end, the pump heat is offsetting alot of that anyhow.

I think there are several more inputs that need to be considered to even achieve a gross estimate. This type of analysis is touchy even when all the conditions are considered and well measured. I understand the need to keep it simple but this is just a fancy way of saying more Radiator = better performance, Higher fan speed = better performance.

I noticed the heat load doesn't effect the system Cº/W, It's been a while but I thought that number would vary with the heat load. Since the higher the load means higher temps and a higher heat transfer rate between the system and ambient.

Also, Is it valid to simply add the C/W for multiple radiators? I would think you would get diminishing returns pretty quickly.

Right now there is nothing at all to help people figure out how to size a radiator, and this is all based off of my own testing, much of which is based on more than one heat load.

Not sure on the heat load vs. c/w. I've ran several now at 300 and 600watts and it's generally linear "enough" to be within 5-10%, some are spot on while others are a little off, but nothing really consistent.

And yes there are diminishing returns, try it and see. I find it really educational and at least on the few real world tests I've ran it's doing pretty well.

This is freeware and open source type stuff, so if you see a particular area where it went wrong and you have some results to compare it to and show me, I'd really like to see it..

I realise it's not perfect, there are always other variables, that's also why I think it's a good reason to keep manufacturers separate. I don't want people to use it to compare radiators, I want to help them understand how heat load and fan selection is critical to each radiator's performance and give them a rough start. It's really hard for anyone to understand how multiple radiators in a system work, this helps give you some idea.

It's an educational tool that more than anything should help people understand how important heat and fan performance is..

I say don't make a generalized "shroud boost" button...

Just make seperate files for the rad's most used, like the PA120.3, for example.....

Great work as always.... Your the UBER god of Excel Spreadsheets.....

Martin's Excel Skills.............FTW

OK, maybe what I can do is just make a pull down for some percentage adjustments, then I can leave it up to the user to make their own judgements about performance. The only data I really have is the two tests where I gained about 7% in pull with shroud at 1350RPM and 15% with pull in shroud at 2000RPM. But that was all using these fans and the XSPC raditor. I think I'll just make a 0-15% in 1% increments adjuster, then as we get more data around shrouds and different radiators, you can adjust accordingly.

Great ideas...

[Martinm210]

Uploaded a newer version with some cleanup and added in that shroud gain by percentage idea.

[TouGe]

I've only been in this section for a few weeks and I've already noticed the wealth of info you provide here.
Another great contribution!

[Martinm210]

Thanks!

Alright, I've got the PA, TFC, HWLabs, and the Bonnie Heatercore versions done too.

Enjoy!
Martin

[SiGfever]

Outstanding Martin. What ever your wife pays you it is not enough.

[leo_bsb]

Martin, you don't cease to amaze us all.
Great work as always.

[disruptfam]

cheers martin thanks alot

[twwen2]

Wow thanks once again Martin, this'll be usefull

[voigts]

Martin, your work ALONE out of all of the watercooling community really is making the fuzziness and guesswork of figuring out how to actually calculate what watercooling components to use and performance to expect into a palatable subject. I find myself referring to your material on your website often. I am unaware of anyone else in years who has tackled so many areas of watercooling and come out with easy to understand testing information so that people can make educated decisions on what to buy, and so that people can give educated advise to others on what kind of setups to configure based on their wants/preferences/needs.

I for one hope you continue to find the interest and time to keep up the good work, but I hope that you don't burn yourself out in the process and just give up as I can only imagine the amount of time it takes to test and come up with all of these great tools. I just donated $10, which isn't much, but I wish everyone else here who benefits from your work would do at least the same.

I do have one puzzling question though. According to the flowrate calculator, I should be well around the 1.5gpm range at least. Given maybe a 200w heatload (oc-ed at 3.2ghz, DDC 9w original with washer and XSPC top), I should be seeing a Delta of around 5c or less according to the radiator calculator. I am using yate loon d12sl-12 fans at about 7v, so I'm not sure exactly what the airflow is. However, I am seeing about an 11c delta between idle and running Prime 95. I'm not sure where the difference is coming from.