Block Parallel vs Series (Let's find that point of equal)

[Martinm210]

So with the introduction of parallel tops, I thought it would be worthwhile to work through a bunch of examples of what would happen with different components.

Skinnee showed us that there is some benefit using the D5 when loops are relatively balanced in restriction, that's awesome.:up: I always ignored parallel myself, so it's cool that there is some untapped performance to be had with it.

However I don't think that's always the case, and I want to know:

When is parallel better using a D5 and when is series better?

I suspect there is a differential point in pressure drop difference between the two where they are equal. I want to find that point.

You could run actual tests to do this, but in interest of time and not having parts, I thought it would be worthwhile to simply use actual pump and pressure drop curves and calculate the theoretical results.

So with that, I'm going to use my flow rate estimator to solve these graphically. I'll start on the extreme unbalanced side of things and work my way toward a balance of pressure drop. With enough of these, it should paint a picture of the point of equality.

First up. An EK supreme and an MCR320, very unbalanced:
The colors are as follows:
Red is the CPU block plus tubing
Blue is the radiator plus tubing
Orange is the Series combined curve that the pump is working against(Added vertically)
Green is the Parallel combined curve that the pump is working against (Added horizontally)
http://img194.imageshack.us/img194/422/parallelmcrsupreme.png

Series is clearly better here. The CPU loop is around 7X more restrictive than the radiator loop.

Theoretical #1 Unbalanced 4.6PSI or 5.8/.8 = 7.2X
CPU Loop
Series = 1.22GPM
Parallel = .84GPM

Theoretical #2
Supreme in one loop
Danger Den Ione + Danger Den Ione + MCR320 in second
http://img54.imageshack.us/img54/7348/parallel2b.png
Theoretical #2 Unbalanced 3.2X
CPU block
Series = 1.18GPM
Parallel = .99GPM

Series still better here with a 3.2X difference in loops but getting closer...

Theoretical #3
Supreme in one loop
Danger Den Ione + Danger Den Ione + MCW30 + MCW30 + 480GTX in second
http://img110.imageshack.us/img110/869/parallel3.png
Theoretical #3 Unbalanced 1.7X
CPU block
Series = 1.01GPM
Parallel = 1.09GPM

Parallel is just barely better on this one, so I've crossed over.

It appears with the D5, the pressure drop difference between the two is around 1.8-2X. In other words, one loop can have nearly twice as much restriction and parallel will work roughly the same as series. If your two loops are more balanced than that...parallel a good option to consider and will likely produce more flow. If it's less balanced than that, or you simply don't have the pressure drop information to check...stick with series.

Anyhow, for rough basic planning, I would simply pick a flow rate like 1.5GPM and look up the pressure drop for your components in PSI. Add them up and try to look for balance when considering parallel, closer the better. You don't need to plot curves, just keep it simple and look up pressure drop at some fix flow rate...it'll get you close enough. Ignore the tubing since it's going to pretty much be there regardless of the setup.

At least this is how I think it works. Skinnee noted that he's doing some actual testing of this sort of stuff, so I would look for that to get the real test results.

Anyhow, that was interesting...I feel better now. I'll wait for skinnee to tell me if I'm close or out in left field on this. Admittedly, this is just my best guess..:D
Cheers!
Martin

[twwen2]

So if i'm understanding this correctly Martin, parallel is only worth considering if you're running balanced loops?

It's good that you're clearing this up.

[Martinm210]

So if i'm understanding this correctly Martin, parallel is only worth considering if you're running balanced loops?

It's good that you're clearing this up.

Yes, not perfectly balanced...but somewhat balanced.

I'm trying to firgure out just how unbalanced it can be before series is better.

It's also going to depend on the pump. I'm only looking at the D5 because it's flat curve is ideal for parallel. Not sure on the DDC but it would be different and likely not nearly as good with parallel.

Here is theoretical #3, I actually crossed the equal point line with this one.

The more restrictive loop is about 1.7X more restrictive than the lesser loop and parallel is just now barely producing better flow on the CPU block.

http://img110.imageshack.us/img110/869/parallel3.png

So I think it's probably around a 2X difference in pressure drop where they are about equal per this theoretical exercise. More balanced than that and parallel is better. Less balanced than that and series is better.

You won't find an exact factor since it also depends a bit on where you're intersecting the curve, but it gives you an idea. The D5 has a fairly linear like curve, so it should be close.

You can be a fair amount unbalanced, but you definitely need to do some planning and have the pressure drop information if you intend to use parallel and expect to get benefits from that.

Parallel will also dump a touch more heat into the loop from the pump motor, but I think it may be small enough that you can ignore it.

[Vapor]

Good stuff Martin :up:

Parallel also has to take the 'hit' of having two Ys in serial (not very restrictive parts, but considering the overall pressure drop curve of a parallel config is meant to be low--it adds up).

Something like the T3 gets around that by having two ins/outs integrated into the pump--and also has its own design tweaks to make its PQ curve fatter above 2.75GPM.

EDIT: I'm not sure your math (at the theoretical level) is right....in a true parallel setup, a combined pressure drop curve is not determined by adding flow at any given pressure drop. It's a lot more complicated than that....let's see if I can dig up some of my old posts.

EDIT2:
Let's make it even simpler and translate an equation into a variable...R

R = resistance = ((Pressure drop) / (flow^2))

For serial: Rt = R1 + R2

For parallel: Rt = (R1 * R2) / (sqrt(R1) + sqrt(R2))^2

[SNiiPE_DoGG]

The way I see it, performance gain is not what should be touted from this parallel setup - just the ability to run a second loop off of a single pump (which is a very powerful aspect).

the only performance gain we do see is the higher flow in the lower restriction loop than if it were to be run in serial - even then I dont think you would see a degree of difference in temps.

[Martinm210]

Good stuff Martin :up:

Parallel also has to take the 'hit' of having two Ys in serial (not very restrictive parts, but considering the overall pressure drop curve of a parallel config is meant to be low--it adds up).

Something like the T3 gets around that by having two ins/outs integrated into the pump--and also has its own design tweaks to make its PQ curve fatter above 2.75GPM.

EDIT: I'm not sure your math (at the theoretical level) is right....in a true parallel setup, a combined pressure drop curve is not determined by adding flow at any given pressure drop. It's a lot more complicated than that....let's see if I can dig up some of my old posts.

EDIT2:

Yes, I was ignoring the fittings trying to get a ballpark.

OK, I just assumed it's added since they are added that way with pumps in parallel. At least everything I've read about pumps. I couldn't find much on grid type systems that would be more fitting of what we're looking at here.

Do you have any links you've found that shares that? Or did you figure it out from testing?

Keep in mind the T3 is going to have two different PQ curves (actually 3). There will be one with just one set of inlets/outlets and a different one for the two combined. You would need to retest with Y fittings to record that second one. So you can't really use the single inlet/outlet T3 PQ to formulate a generalized parallel/series equation since it operates under different PQ curves (It won't be that different...but somewhat different). There will be some sort of boost between one set of inlet/outlets and using two. That's probably that high flow gain thing you're noting.

Anhow, this parallel stuff just has my interest. Finally a good reason for folks to pay more attention to pressure drop. Always happy to see a little more science come back into this little hobby of ours...:)

[Ketzer7]

Great read, many thanks Martin

[Vapor]

Yes, I was ignoring the fittings trying to get a ballpark.

OK, I just assumed it's added since they are added that way with pumps in parallel. At least everything I've read about pumps. I couldn't find much on grid type systems that would be more fitting of what we're looking at here.

Do you have any links you've found that shares that? Or did you figure it out from testing?

Keep in mind the T3 is going to have two different PQ curves (actually 3). There will be one with just one set of inlets/outlets and a different one for the two combined. You would need to retest with Y fittings to record that second one. So you can't really use the single inlet/outlet T3 PQ to formulate a generalized parallel/series equation since it operates under different PQ curves (It won't be that different...but somewhat different). There will be some sort of boost between one set of inlet/outlets and using two. That's probably that high flow gain thing you're noting.

Anhow, this parallel stuff just has my interest. Finally a good reason for folks to pay more attention to pressure drop. Always happy to see a little more science come back into this little hobby of ours...:)I was working on an estimator and figured it out from skinnee's numbers (not the sample loops, the dual loop PQ curve testing he did)....then went online in search of confirmation and the only thing I could find in terms of parallel pressure drop was this: http://www.eng-tips.com/viewthread.cfm?qid=156852&page=8 (which matched what I got, fortunately). Mine was really messy though since I didn't think to make "R"....much cleaner when I did that :rolleyes:

And yeah, there are two totally different PQ curves for single loop (both sets of ports have the same numbers it seems) vs. dual loop. Single loop is very Koolance-like and dual loop is very EK/DetroitAC-like (except at low flow, it's actually kind of weak there). It's basically the opposite of adding Ys :p: :shrug:

I'm not sure about testing dual-loop with Ys....I think dual flowmeters/manometers is the way to go. Or at least single manometer with dual flowmeters with controlled controlled secondary flowrates (like what skinnee did).

Anyway, yeah...it definitely is really interesting that with some planning, you can get much higher flowrates through key components (and sometimes higher flowrates in general) with the use of parallel loops.

[Sadasius]

Where would the flow rates be measured at in parallel loops? At the 'Y' where the loops converge or in the middle of each loop to see if the flow would be the same or near the same. In two separate loops it is easy to measure each loop and determine the performance of each loop. However with the parallel loops I am sure measuring the performance wll be very different and would be especially limiting on restrictive parts which would require considerable planning on the individuals part to ensure that the loops are balanced by calculating each and every part in various ways since fluid is very dynamic. Soon people will be selling 'Balanced kits' for people...:rofl: Interesting stuff to say the least but seems to be very tedious for any beginner to begin to understand.

OMFG what the hell is wrong with the website? Takes forever to load and has multiple database errors that keep popping up.

[Waterlogged]

So if i'm understanding this correctly Martin, parallel is only worth considering if you're running balanced loops?

It's good that you're clearing this up.

I thought it had been pretty clearly stated in some of the T3 topics that flow needed to be balanced for it to work properly. :shrug: You have to wade through a bit of crap to find it but it's there.

For some strange reason, parallel has always had this mysterious attraction from water coolers and it has also always had the caveat of needing the flow balanced to work right. In the past, it has been done with valves and inline flow meters but that gets expensive and adds a certain amount of unwanted restriction as well as bulk in some cases.

[Boogerlad]

Sticky!

[rambler358]

Very interesting thread to see. I was going to post a seperate thread before I saw this one. I should have more information regarding this topic and my setup in a couple days. I previously had a single loop for my components (see sig) with a D5 pump. The D5 netted me around a 1.5 LPM flow rate for the serial loop with a single 1080 rad. Changing out the D5 for dual DDC 3.25s with a Watercool top, still in the same serial loop increased flow rate to around 3.5 LPM.

I now have the XSPC res/dual DDC top unit with parallel loops configured as such:
Loop1 - Res/pump -> 1080 rad -> CPU -> flow meter -> back to res/pump.
Loop2 - Res/pump -> RX360 rad -> GPUs -> NB -> VReg -> flow meter -> back to res/pump.

I opted to use the 1080 rad strictly for the CPU because my 920 D0 gets very hot, and *hopefully* should get better cooling from this rad.

I have a flow meter for each loop, placed right before water goes back into the res. I also have a water temp sensor for each loop. The CPU sensor is located directly at the exit of the CPU block. The 2nd water temp sensor is located directly at the exit of the VReg block.

It will be interesting for me to see the different flow rates for each of my loops with this XSPC unit. It will also be interesting to see if there are any temperature variances between the two loops, as well as seeing if there is a CPU temp difference with my standard 4.0GHz overclock. I'm currently leak testing, and should have everything in place by the weekend. :)

EDIT: My sig doesn't show all my blocks which are the CPU, dual GPUs, Koolance Classified chipset (NB/SB) block and a Koolance Classified VReg block.

[skinnee]

Amp draw for all the D5 configurations I've tested... 1.94-1.95A looks to be the ceiling. :shrug:

I should really test the Koolance top for it's Buffet Fruit Salad numbers.



http://www.skinneelabs.com/Pumps/D5/D5-AmpDraw.jpg

[Martinm210]

I was working on an estimator and figured it out from skinnee's numbers (not the sample loops, the dual loop PQ curve testing he did)....then went online in search of confirmation and the only thing I could find in terms of parallel pressure drop was this: http://www.eng-tips.com/viewthread.cfm?qid=156852&page=8 (which matched what I got, fortunately). Mine was really messy though since I didn't think to make "R"....much cleaner when I did that :rolleyes:

And yeah, there are two totally different PQ curves for single loop (both sets of ports have the same numbers it seems) vs. dual loop. Single loop is very Koolance-like and dual loop is very EK/DetroitAC-like (except at low flow, it's actually kind of weak there). It's basically the opposite of adding Ys :p: :shrug:

I'm not sure about testing dual-loop with Ys....I think dual flowmeters/manometers is the way to go. Or at least single manometer with dual flowmeters with controlled controlled secondary flowrates (like what skinnee did).

Anyway, yeah...it definitely is really interesting that with some planning, you can get much higher flowrates through key components (and sometimes higher flowrates in general) with the use of parallel loops.

Thanks!

I read through that link, it is for HVAC, but I guess it should still apply (I'm still not comfortable with air). I did find about a 20 page chapter in an old fluid mechanics book I had buried in the shelf from 15 years ago that may help me make better sense of this. The way they do it in my fluid mechanics book is really tedious and complex, sort of a manual iterative process of constantly checking stuff like Reynolds number to then find the friction factors, looking up Moody's diagram, computing velocities, and finally flow rates. They go even further into triple and grid type water systems stuff, but it seems computer models are more typically used for that sort of stuff.

It's still not clicking for me....going to have to drink a beer and get back to this..:D

It's all your fault, now I'm going to have to break out the manometer and flow meter just so I can see it before my eyes. :rofl:

If you do get that parallel flow rate estimator all working, you may as well provide the option for multiple branching. That'll be the first thing someone asks "What if I want to put in three loops"...Ahhhh:eek::D

We also still need to answer the magic question of "WHEN" is parallel better than series. I'm convinced it's not always and some sort of balancing needs to happen, I just don't know yet, especially if I've got my pressure drop parallel combination stuff all screwed up...

[Martinm210]

Where would the flow rates be measured at in parallel loops? At the 'Y' where the loops converge or in the middle of each loop to see if the flow would be the same or near the same. In two separate loops it is easy to measure each loop and determine the performance of each loop. However with the parallel loops I am sure measuring the performance wll be very different and would be especially limiting on restrictive parts which would require considerable planning on the individuals part to ensure that the loops are balanced by calculating each and every part in various ways since fluid is very dynamic. Soon people will be selling 'Balanced kits' for people...:rofl: Interesting stuff to say the least but seems to be very tedious for any beginner to begin to understand.


In a parallel loop setup, you'll have three different flow rates. If you measured before the Y or after, you're in the portion of the loop that the pump sees. This is the pump flow rate. If you measure after the Y in each loop then you will get the individual loop flow rates.

That's a BIG part of the problem when discussing flow rates in a parallel setup. WHICH ONE? There are three different ones. We never had that problem with series so it didn't matter where you were measuring and it also didn't matter what you called it because they were all the same. My vote is to clearly identify the flow rates somehow when talking parallel. Pump flow rate, Parallel Loop 1 or Block 1 flow rate would be a good term for one of the loops and Parallel Loop 2 or Block 2 flow rate would be helpful to keep them all separated.

I agree on the complexity issue, hopefully Vapor will work up this new estimator to help out, but I still see a big gap in data available. There are soo many parts out there that have never been tested and don't come with pressure drop data. You're basically screwed if you have one of those parts. You can ASSume a like part, but you can get what's coming to you. There's alot of stuff that's very suprising when it comes to pressure drop. Some stuff that seem low in restriction like some mosfet blocks are actually really high. Some chipset blocks are very low and some fairly high. It's just pretty hard to guess sometimes.

That's the big issue I see with parallel, it's just not user friendly if it turns out we really need to balance pressure drop somewhat close when the data isn't out there for everything. The good news is, at least it's a good challenge. I have to admit, I'm a bit dissappointed and bored when it comes to a standard series loop. There's no challenge with it and it's pretty hard to screw up unless you have a leak. I have managed that a few times, but I'm ready for something more..lol!

[Sadasius]

Then onward....You take the lead because I am an idiot and will kill my system. :rofl: Okay maybe not but I learn better by watching others make mistakes first!:D

[Martinm210]

Then onward....You take the lead because I am an idiot and will kill my system. :rofl: Okay maybe not but I learn better by watching others make mistakes first!:D
I have killed my system by spilling into my PSU once. Would you like a write up of that one? What about when I poked holes into my MCR320 tubes with long bolts? What about when I nearly caught fire when my radiator test ran dry because my screwed up tightening a compression fitting?

I could go on forever....:rofl:

[skinnee]

So what are the tests we really need to run here...because the last time I made an attempt at running tests with two flow meters on each leg and logging data I was nearly drug into town square and stoned to death. I still am unsure with all the chaos whether the PQ tests were accurate...maybe differential pressure should've been taken? :shrug:

Another thing to remember when testing, don't get attached to a part or piece of equipment, because you will ruin it in one hell of an accidental and creative manner. :D

[Sadasius]

I have killed my system by spilling into my PSU once. Would you like a write up of that one? What about when I poked holes into my MCR320 tubes with long bolts? What about when I nearly caught fire when my radiator test ran dry because my screwed up tightening a compression fitting?

I could go on forever....:rofl:

Dude....:shocked: I spilled Coke in my first Ultra X3 1000 PSU. Thought I got it all out but when I used it, it went POW! I think I took off a couple years off of my life with that one. :up:

This whole new era into parallel loops though should be interesting. Might be a good idea as well like BOXGODS to start thinking products that can utilize this and better yet actually help balance the loops like a liquid dynamic gyroscope or something.


So what are the tests we really need to run here...because the last time I made an attempt at running tests with two flow meters on each leg and logging data I was nearly drug into town square and stoned to death. I still am unsure with all the chaos whether the PQ tests were accurate...maybe differential pressure should've been taken? :shrug:

Another thing to remember when testing, don't get attached to a part or piece of equipment, because you will ruin it in one hell of an accidental and creative manner. :D

Yeah that's not cool what happened to you and your review. I very much appreciate and like your reviews a great deal. I don't think people understand the work that is involved.

[Martinm210]

So what are the tests we really need to run here...because the last time I made an attempt at running tests with two flow meters on each leg and logging data I was nearly drug into town square and stoned to death. I still am unsure with all the chaos whether the PQ tests were accurate...maybe differential pressure should've been taken? :shrug:

Another thing to remember when testing, don't get attached to a part or piece of equipment, because you will ruin it in one hell of an accidental and creative manner. :D

I'm just trying to figure out how balanced the loops need to be for parallel to work in benefit over series. To find that point I think you need to test parallel vs series of different setups that have different levels of being unbalanced. Start on the extreme where you have one very restrictive loop and one very low restriction and work your way to the middle where the two loops are balanced.

Just like your other series vs. parallel flow only tests, but with a couple of unbalanced load scenarios to try and understand what if anything unbalancing does. If we had something like 3 or four tests that were spread out from highly unbalanced to very balanced, it should show if there is a crossing point (a point where series and parallel are just as good).

And the comparison would look specifically at the flow rate on the more restrictive loop (The high restriction block flow rate). You could also record the pump flow rate and other loop, but I think just recording the most restrictive block loop is good enough.

But whatever is cool:up:

[zeropluszero]

sorta off topic, but can someone point me to data on parrallel vs series gpu cooling?

[SNiiPE_DoGG]

here you go man :up:

http://www.xtremesystems.org/forums/showthread.php?t=223129

[zeropluszero]

thanks for that.

[skinnee]

I'm still digging a bit here...these are the same loops from the Typhoon III review, only this time nomenclature is more to folks liking. I still need to do parallel PQ curves for the two D5's and the DDC, but here is what I have thus far and I need some sleep. :)



http://skinneelabs.com/Pumps/SysLoopFlow-table.jpg

http://skinneelabs.com/Pumps/SysLoopFlow-Loop1.jpg

http://skinneelabs.com/Pumps/SysLoopFlow-Loop2.jpg

http://skinneelabs.com/Pumps/SysLoopFlow-Loop3.jpg

[alacheesu]

Do you have any links you've found that shares that? Or did you figure it out from testing?

Check out this book (http://books.google.com/books?id=Uo9gpXeUoKAC&pg=PA307&lpg=PA338&source=bl&ots=z5kmaj99xb&sig=gYlOE6hvmfZ78KoA1LKFxAX9DhY&hl=en&ei=N7h2Ss77JIje-QazhZDMBw&sa=X&oi=book_result&ct=result&resnum=4#v=onepage). Pages 307 and onwards go through calculations of resistance in parallel and series flow. Some pages are not shown, but piecing it together for the lazy:
dp: pressure drop
Q: flow rate
R: resistance

R = sqrt(dp)/Q

Series: R_total = sqrt(R_1^2 + R_2^2)
Parallel: 1/R_total = 1/R_1 + 1/R_2

This is mathematically equivalent to what Vapor posted, except his R was the square of their R.

@skinnee: Thanks. Very interesting.

[skinnee]

Now that I have a bit of sleep and a whole lot of coffee in me...

The System flow Rate comparisons I ran will be added to my Pump Comparison pitching both D5's and DDC's in one set of scatter plots, we'll finally get to see PQ's together. Its going to be an eye chart, but it will show the full sprectrum of PQ. I've already been IM'd numerous times by a certain someone who wants RD-30 testing included and really wants this whole Y/parallel loop discussion to be over. But there is still more testing and information that needs to be done before this parallel discussion can be boxed up and set on the shelf...I hate it when I open a can of worms like this. :D

Back to the system flow rate comparisons...

As I stated in the T3 review, the Single Loop (TITO used in the charts) performance on the T3 isn't earth shattering and does not match the DDC w/ XSPC V3 top or PQ boasting D5 tops. Parallel in general shows how seperating restriction from a serial loop to two legs can boost flow rate for each individual leg, but there is quite a bit of testing that needs to be done to either support or rule out the level of balancing. From what I am seeing on a purely flow rate perspective it doesn't appear that the balancing or planning is as crucial as we once thought...but we'll find out for sure soon enough.

[Martinm210]

Check out this book (http://books.google.com/books?id=Uo9gpXeUoKAC&pg=PA307&lpg=PA338&source=bl&ots=z5kmaj99xb&sig=gYlOE6hvmfZ78KoA1LKFxAX9DhY&hl=en&ei=N7h2Ss77JIje-QazhZDMBw&sa=X&oi=book_result&ct=result&resnum=4#v=onepage). Pages 307 and onwards go through calculations of resistance in parallel and series flow. Some pages are not shown, but piecing it together for the lazy:
dp: pressure drop
Q: flow rate
R: resistance

R = sqrt(dp)/Q

Series: R_total = sqrt(R_1^2 + R_2^2)
Parallel: 1/R_total = 1/R_1 + 1/R_2

This is mathematically equivalent to what Vapor posted, except his R was the square of their R.

@skinnee: Thanks. Very interesting.

Awesome, thanks!

I worked on making some sense of my fluid dynamics book, but it didn't seem to click for me. Sounds like that's the trick to calculating the result and what's needed in an estimator tool..:up:

My guestimations would be wrong by simply adding then.


Now that I have a bit of sleep and a whole lot of coffee in me...

The System flow Rate comparisons I ran will be added to my Pump Comparison pitching both D5's and DDC's in one set of scatter plots, we'll finally get to see PQ's together. Its going to be an eye chart, but it will show the full sprectrum of PQ. I've already been IM'd numerous times by a certain someone who wants RD-30 testing included and really wants this whole Y/parallel loop discussion to be over. But there is still more testing and information that needs to be done before this parallel discussion can be boxed up and set on the shelf...I hate it when I open a can of worms like this. :D

Back to the system flow rate comparisons...

As I stated in the T3 review, the Single Loop (TITO used in the charts) performance on the T3 isn't earth shattering and does not match the DDC w/ XSPC V3 top or PQ boasting D5 tops. Parallel in general shows how seperating restriction from a serial loop to two legs can boost flow rate for each individual leg, but there is quite a bit of testing that needs to be done to either support or rule out the level of balancing. From what I am seeing on a purely flow rate perspective it doesn't appear that the balancing or planning is as crucial as we once thought...but we'll find out for sure soon enough.

Thanks!

Awesome work again as usual. I see that in Loop 1 that you'd be better off in series. Which parts were part of loop A and which parts were in B? Maybe you already said that.

Either way as you've mentioned, it's not like one way or the other even if losses do occur like loop one are going to be anything really problematic....probably not even measurable from a temperature standpoint.

Anyhow, that's great to see that the once taboo/ignore/misunderstood parallel does have options. It also appears the T3 must have a fairly stronger PQ curve (Especially on the right side) under parallel which makes it less likely to see benefits in series over parallel.

Good stuff here, thanks for all the awesome work!:up: