NZXT Switch 810: 3 radiator, single loop questions

[Ladamyre]

I put this here instead of water cooling, if a moderator thinks it's better there, please move it.

I asked a builder who was doing a YouTube video review on this case, did he think about using the top front intake fan mount to mount a single 140mm radiator, keep the dual fan 280mm radiator on the bottom and the triple fan 420mm radiator on the top. He said it would look funny and people would laugh at it.

Well, I could care less what people think it looks like, function is what I care about. My question was about a planned single loop cooling plan I had in mind and he blew it off with that 'laughable' excuse.

So, I'm here asking the nerd squad virtual world superheroes their opinion.

Question/poll: On a single loop cooling circuit that involves one single, one dual and one triple radiator, and the components being cooled are, CPU, GPU(s), Northbridge, MOSFETs and RAM, what would be the best way to run the loop?

My vote was: Reservoir/pump(s) to triple rad, to CPU, to dual rad, to GPU(s), to single rad, to N/B, to MOSFETs, to RAM and back to the reservoir.

This was taking into account that the single and dual rads are bringing in fresh cold air, whereas the triple is exhausting already warmed air from those two on this particular case.

What do you guys think?

 

[Cjsparky]

From what I understand, the order of the system does not really matter that much, if at all. Flow, and fans are more important than serial loop order.

 

[rinval]

+1 for the loop itself which way it flows or whats where on the loop is not going to make a large difference maybe 1C-5C. the biggest thing is to make sure you have a good pump (like a swiftech 655). I would recommend on the flow is to keep you loops small for aesthetics as radiator placement is irrelevant. plus if you go cheap on tubing and it touches something hot you could run into alot more issues then radiator placement........

 

[Ladamyre]

As I have ruminated on these responses, all good observations, I think it would be best to remap the flow thus: reservoir to dual rad (coolest possible into the CPU), then to the single rad (blow off the heat added by the CPU), then to the GPUs, Northbridge, driver MOSFET's and RAM in that order, then on to the triple rad before returning to the reservoir.

What do you guys think?

In the end, if I do this, I'm gonna start with whatever consensus I get here, and test it out, then reconfigure the loop (hoses are cheap) and test that out. Repeat if necessary until I get the best overclock on my CPU. After all, this is all about the highest possible OC on CPU.

 

[rinval]

ok for the highest OC on your CPU put it on its own loop with refrigeration cooling.... but of coarse your talking alot of money...

your are putting 840mm of radiators in your system.... and average water cooled system only runs about 420ish mm of radiators for a full water cooled system... ie cpu gpu vrm etc. so once again it really doesnt matter where you put your cpu in the loop as everything in the loop all heats the same water... with that said running your cold air rads to your CPU will see a few degrees lower in the cpu... but will make the rest of the system hotter. so it is really a trade off.

edit. so the route i would go probably would be cold air rads to cpu then GPU to warm rad to the system... of coarse this will make your loops long and make your system look cluttered.

this is simply to get rid of as much heat from the CPU as possible.

so the loop will be COLD air-cpu (warm/hot)-Gpu(HOT)-Warm rad (warm)-System (warm/Hot)- back to cold.

 

[Cjsparky]

ok for the highest OC on your CPU put it on its own loop with refrigeration cooling.... but of coarse your talking alot of money...

your are putting 840mm of radiators in your system.... and average water cooled system only runs about 420ish mm of radiators for a full water cooled system... ie cpu gpu vrm etc. so once again it really doesnt matter where you put your cpu in the loop as everything in the loop all heats the same water... with that said running your cold air rads to your CPU will see a few degrees lower in the cpu... but will make the rest of the system hotter. so it is really a trade off.

edit. so the route i would go probably would be cold air rads to cpu then GPU to rad to the system... of coarse this will make your loops long and make your system look cluttered.

this is simply to get rid of as much heat from the CPU as possible.

Rinval,

Is there a point where more efficient radiator heat exchange occurs when there is a greater deltaT between loop water temps and ambient air temp, or does a system merely remove heat from the loop at the same rate regardless of deltaT?

 

[rinval]

Rinval,

Is there a point where more efficient radiator heat exchange occurs when there is a greater deltaT between loop water temps and ambient air temp, or does a system merely remove heat from the loop at the same rate regardless of deltaT?

really its all about the same give or take a couple degrees. the coldest point in any water system will be right behind the cold air rads. as i said though your only talking a degree or 2 but that could equal some extra OC. more or less it really doesnt matter... the best solution for having a cold point is a huge copper rad ie maybe 360x60mil cold air you would see that right behind the rad could be as much as 10C cooler ... but then again we are talking about an overkill rad in a heavily modded case. for his particulars throwing the cpu behind the cold air rads may give him a minuscule extra amount of OC which seems to be all he cares about.

 

[Ladamyre]

ok for the highest OC on your CPU put it on its own loop with refrigeration cooling.... but of coarse your talking alot of money...

I'm talking a lot of money here but if I were to go for a refrigerated cooling solution I'd go with a sealed case where I could add a vacuum pump to keep condensation out or do the radical "submerged in oil or Flourinert" AKA. 'fishtank' solution and put the refrigeration unit in another room. I want the system to be portable, that's why I'm not going there. Let me remind everybody I don't care what it looks like, or if there's cluttered coolant lines, maximum OC on the CPU is the object. GPU's can take more heat than CPU's, and by the time high-flow rate coolant gets to the RAM, none of those components are going to be needing high delta V since their temps will be incidental. The only reason the N/B, MOSFET's and RAM are in the loop is simply so their temps can be guaranteed to not be any concern.

Last on this point, if I do this, I will be running two pumps in tandem for high flow rate and redundancy. For these reasons, a two loop system is unnecessary, not to mention having to buy two extra pumps to keep redundancy, with no gain in efficiency for the extra money spent.

...your are putting 840mm of radiators in your system....

Yeah, ain't it cool!

...for his particulars throwing the cpu behind the cold air rads may give him a minuscule extra amount of OC which seems to be all he cares about.

Not to be hammering anybody or trying to preach but... Yeah! I'm trying to achieve the highest, stable CPU overclock possible, which has always been given the 90%, lions share of weight on system-wide OC values, with the remaining 10% divided between GPU and RAM frequencies. With 840mm of cooling capacity, and the proper configuration of the loop, I will be able to get the most out of the CPU without having to make trade-offs on the GPU's and RAM. All those objectives on a system I can pack up into the back of the van in 15 minutes.

 

[rinval]

Not to be hammering anybody or trying to preach but... Yeah! I'm trying to achieve the highest, stable CPU overclock possible, which has always been given the 90%, lions share of weight on system-wide OC values, with the remaining 10% divided between GPU and RAM frequencies. With 840mm of cooling capacity, and the proper configuration of the loop, I will be able to get the most out of the CPU without having to make trade-offs on the GPU's and RAM. All those objectives on a system I can pack up into the back of the van in 15 minutes.

with that in mind i would go one of 2 ways.

Cold air rads- CPU - Ram- NB - mosfets- warm air rad - GPU - pump/res -> this will keep everything relativly cool to allow good overclocking everywhere.

or

Cold air rads - CPU - Ram - NB - GPU - mosfets - warm air rad - pump/res -> this will make the CPU slightly cooler but your GPU will be hot and the mosfets will be very hot.

Keep in mind both configurations will only show about a max of about 5ish degrees difference on the loop b/c of how fast the water will be flowing.

which has always been given the 90%, lions share of weight on system-wide OC values

hahaha Lions are the laziest part of the pride rofl.....

 

[Ladamyre]

with that in mind i would go one of 2 ways.

Cold air rads- CPU - Ram- NB - mosfets- warm air rad - GPU - pump/res -> this will keep everything relativly cool to allow good overclocking everywhere.

or

Cold air rads - CPU - Ram - NB - GPU - mosfets - warm air rad - pump/res -> this will make the CPU slightly cooler but your GPU will be hot and the mosfets will be very hot.

Two possibilities worth a try, I might go with those first and then try out my idea and record the temps to see which is best for the CPU.

In the end, all the wattage these components dissipate is going to be blown off in the radiators, but the overarching consideration is "How low can I get the coolant temp just before it get's to the CPU?". Your idea of running through both of the cold radiators may just be it.

Let's remember, GPU's and driver MOSFET's today can take a lot more heat than CPU's, so as long as they are under 80 degrees celsius, they won't be compromised at all.

Now that I think about it, I should run coolant to the northbridge and RAM right after the CPU, and then on to the GPU's and MOSFET's.

Preliminary configuration on that loop would then be: reservoir to dual rad - to CPU - to single rad - to northbridge - to RAM - to GPU's - to driver MOSFET's - to triple rad - and back to the reservoir.

I think we're getting close.

 

[rinval]

Two possibilities worth a try, I might go with those first and then try out my idea and record the temps to see which is best for the CPU.

In the end, all the wattage these components dissipate is going to be blown off in the radiators, but the overarching consideration is "How low can I get the coolant temp just before it get's to the CPU?". Your idea of running through both of the cold radiators may just be it.

Let's remember, GPU's and driver MOSFET's today can take a lot more heat than CPU's, so as long as they are under 80 degrees celsius, they won't be compromised at all.

Now that I think about it, I should run coolant to the northbridge and RAM right after the CPU, and then on to the GPU's and MOSFET's.

Preliminary configuration on that loop would then be: reservoir to dual rad - to CPU - to single rad - to northbridge - to RAM - to GPU's - to driver MOSFET's - to triple rad - and back to the reservoir.

I think we're getting close.

i think your on the right track. with your loop setup you will be going through the warm 420 and cold 280 rads to drop temp before the CPU.

as always remember unplug and discharge your mobo before changing configs... also make sure to wrap your drained hoses with a towel when you unhook them to pickup any left over water..

edit- and dont forget to check for leaks before connecting power to your mainboard.... ie you can jumper the green and black (3rd and 4th pins) wires in your 24 pin connector while your mainboard is unhooked to test for leaks..

 

[Ladamyre]

...dont forget to check for leaks before connecting power to your mainboard.... ie you can jumper the green and black (3rd and 4th pins) wires in your 24 pin connector while your mainboard is unhooked to test for leaks..

That would be the choir preaching to the minister.

I am a retired Institute for Automotive Service Excellence, Certified Master Technician and don't do 'leaks'. I will vacuum test the system (5 or 6 in/hg) before I put coolant into it. If air can't leak in, it's a certainty coolant can't leak out.

I plan to use hoses that are too long on most cases initially so I can test the possible configurations for the best scenario, disconnecting and rerouting hoses at the radiators. Once the tests determine that, it will simply be a matter of cutting to fit the hose ends that connect at the radiators and finishing those connections with one last vacuum test on a filled system on a running computer. No bubbles, no leaks. Plus with the vacuum pump attached to the reservoir, the system will bleed out the air inside the loop to the reservoir automatically.

 

[rinval]

That would be the choir preaching to the minister.

I am a retired Institute for Automotive Service Excellence, Certified Master Technician and don't do 'leaks'. I will vacuum test the system (5 or 6 in/hg) before I put coolant into it. If air can't leak in, it's a certainty coolant can't leak out.

I plan to use hoses that are too long on most cases initially so I can test the possible configurations for the best scenario, disconnecting and rerouting hoses at the radiators. Once the tests determine that, it will simply be a matter of cutting to fit the hose ends that connect at the radiators and finishing those connections with one last vacuum test on a filled system on a running computer. No bubbles, no leaks. Plus with the vacuum pump attached to the reservoir, the system will bleed out the air inside the loop to the reservoir automatically.

lol better to preach... then someone destroy $1000 in hardware... which unfortunately ive seen before . im glad though he had parents with more money then sense. but yes your method of test is the best. its just most people don't have a vacuum system to test with.

and yes i would go with long loop and throw some cheapo coolent or distilled water in there to test.

 

[Ladamyre]

Oh, not complaining. Good thing you do that warning for all concerned.

Distilled water to test. 50/50 automotive coolant/distilled water for final system configuration and seal it off. Should last 5 years before servicing.

 

[Motopsychojdn]

Not sure how I missed this thread hehe,
Too much anti-freeze there man, if you are using a/f for lubrication/anti algae/anti-corrosion purposes then I'd seriously drop the ratio to 5%, water is the active heat transfer medium here, if you reduce the amount of water in the loop, your temps increase, plus the pumps will have to work harder in order to cope with the increased viscosity of the a/f,
I'll post more from the main Pc, I just found this thread on the netbook and an essay is hard work on that hehe
Moto

 

[Ladamyre]

Thanks for the info Moto. Yeah, been reading too much about having to service PC watercooling for algae, corrosion, watch out for silver this, nickel that. I figured "Why so much trouble with this PC coolant? We solved all those issues on automobiles back in the 60's!"

Didn't think about viscosity.

DOH! Now you got me seriously considering Flourinert. No additives needed and a much better thermal transfer rate.

Just added 300 bucks to the build though.

 

[Motopsychojdn]

The only nickel issue is from EK blocks, they had a plating issue, allegedly sorted now but they insist you use certain coolants or lose your warranty (guess whose coolants the endorse?)
I'm not too up on flourinert but I can check it out and compareto plain distilled
Moto

 

[Ladamyre]

Flourinert is what they used on NASA's Cray-2 back in 1985. * Here's a link. *

Some company tried to use it not long ago on a sealed system case, but I never saw it at market anywhere.

The stuff evaporates easily so it has to be used in a sealed environment, but basically it's a freon, a chloroflourocarbon, and a very stable, non-reactive molecule.

You can find a link to a selection guide on the 3M website about the stuff * here *, I have no idea what most of that stuff means. I am afraid to guess which type would be best because I'm not sure what a pressure build up in these waterblocks would do. Would it deform the blocks? I suppose it would, given enough pressure. So then a pressure relief valve would have to be engineered in, and the cost of the fluid would necessitate some way to recover gases, not to mention the ethical concern to recover the gases for environmental reasons. It's just too many problems if you ask me.

And since you did ask me, in a way, I'll tell you now that I was, as we say in the former colonies, "Doing some jack-assing." Kidding, ribbing, not serious.

 

[Motopsychojdn]

Lol, is it bad that I remember it from Abyss?
There wouldn't ever be enough pressure in a W/c loop to cause issues with it, but I share your environmental concerns over the stuff,
you wouldn't know you had a leak until you were breathing it in hehe and with three cats, I prefer straight distilled and silver coils, aided by UV cathodes, it covers me for algae and stuff
if all your fittings are the same brand you will avoid any mixed metal corrosion problems as well,
but back to the initial query, Loop order is largely considered irrelevant these days, the temps in a loop rarely vary +-5'C over the whole loop,
a 'normalstyle' routing is res>pump>Cpu>Gpu or rad>Rad or Gpu then back to res,
but with more and more people putting more and more hardware in, its getting to the point where more radspace is needed and folks are putting rads between blocks just to grab an extra drop before it hits the next block,
most consider mobo and ram blocks as pointless, ram certainly shouldn't need cooling at 1.5v, and Nb/Sb blocks are high restriction so often go on a separate loop from Cpu/Gpu blocks,
Ignore the response form Youtube guy, there is no 'right' way of approaching a loop, its finding what works in your set up and I personally have a 120mm fan at the top front blowing onto one of my reservoirs, partly for case airflow, partly to help cool the water in the res, so if you can do it, do it and forget nay-sayers
list the hardware and we can draw up a decent loop for a set budget, but with the mobo and ram blocks out of the way, you have more to play with on rads/fans, which will have a better effect on your Delta-T
Moto

 

[Ladamyre]

...I prefer straight distilled and silver coils, aided by UV cathodes, it covers me for algae and stuff....if all your fittings are the same brand you will avoid any mixed metal corrosion problems as well...

There's an idea I hadn't thought of, UV radiation. Given enough of that the silver coils may be unnecessary.

Silver is a good anti-microbial, but corrosion, more accurately oxidation, is always going to occur when dissimilar metals are connected by water. Whichever metal has the excess electrons will transfer those through the water, combining oxygen atoms from the water molecules with the other metal. Distance helps, but it's then only a matter of time before the process becomes a problem, hence the requirement to service the system. I want to avoid that problem if I can.

...ram certainly shouldn't need cooling at 1.5v, and Nb/Sb blocks are high restriction so often go on a separate loop from Cpu/Gpu blocks...

You'll see soon why I want to make sure those components can be written off from heat concerns so I can do anything I want with processor clock multiplier, bus speed, DDR3 speed, etc., be able to do whatever is possible with those components voltage wise so I can get the most out of those as well. I plan on splitting off to cool two parallel lines to the RAM, then one of those to the N/B and the other to the driver MOSFET's, then back to a single line. I may put a better passive heatsink on the S/B, but have no plan to W/C it.

list the hardware and we can draw up a decent loop for a set budget, but with the mobo and ram blocks out of the way, you have more to play with on rads/fans, which will have a better effect on your Delta-T
Moto

I thought you'd never ask.

PSU - Enermax Platimax EPM1200EWT
M/B - ASUS Crosshair V Formula
Mem - G.SKILL Ripjaws X Series 32GB (4 x 8GB)
SSD - Samsung 830 64Gb (boot drive)
H/D - Seagate SV35.5 Series ST2000VX000 2TB
GPU - Nvidia GeForce 690 GTX (two for quad SLI)

Waiting for Vishera in September for the CPU.

Add some decent peripherals and Windows 7 Ultimate and I'm looking at $4000.

I'd spend another grand for the W/C.

Hard drive speed isn't a concern since I'll have 32G of memory for a RAM disk, any game I want to play can load into it before I play it, so MTBF is the main concern there.

Feel free to suggest better parts like HD, SSD, PSU, just don't try to give me all the reasons why Intel is a better CPU. I'm a hard core AMD 'fanboy' if that's what you want to call me.

I will never do business with those fascist nazi's at Intel.

Maybe you see now why I want to overclock the RAM, the AMD CPU's respond well to higher speeds there.

If I can get a stable 5 - 5.1 GHZ out of the CPU I will be happy but the more the merrier.

I don't see AMD catching up to the 690 GTX in either performance or power consumption this year. Those two I will parallel cool using the Koolance setup.

The M/B has eight 4-pin fan headers and the UEFI bios will be able to control those quite well. My initial idea is to push-pull the single and dual radiators (the two filtered intakes) and let the triple rad fans pull the air out with the remaining exhaust fan at the back varying it's speed based on the S/B temp. I need the quietest fans I can get on those four so any recommendation there will be appreciated. With this air management plan I think I can count on a positive pressure inside the case and not have dust sucked into the optical drive.

Oh, and this M/B also has 3 temp sensor headers and I might use one to control those 3 fans on the triple rad to control the reservoir temp.

I like the Black Ice GT Extreme radiators as they are 60mm thick and have high fin counts.

Let's see, anything else?

Oh, I think the girl on Newegg showing off the case this thread is all about, the NZXT Switch 810 is hot!

 

[Motopsychojdn]

The Bix rad have higher fpi so they will want highspeed fans to perform at their best, which means louder,
Xspc Rx series are 60mm rads with 8Fpi, I have one Rx240 on top of my case and the four Gelid blue wing 120mm fans on that run at 600rpm,
I barely hear my Pc
but certainly check out other low fpi rads as well, I'm just stating from my experiences
theres also a fan roundup at the top of Cpu section that compares a few popular fans for noise/performance
Have you considered an external radiator box btw?
a couple of us here have had really good results by taking some rads outside the main case
I won't suggest an Intel build, I built my rig on BD's promisse and ended up stepping to the backup plan, a 975BE which I've managed 4.5GHz out of, golden chip hehe, I'm hoping Vish is good as well as you
I think if you are insistent on the mobo/ram cooling that a dual loop res would be in order, that way the ram etc isn't restricting the cpu and Gpu's flowrate
two pumps, each feeding their own loop but sharing a reservoir makes a larger volume of water in the loop which means more time before your thermal mass heats up, which lets your fans work slower to balance it,
let us know your views on the external box because that opens up a world of quiet, powerful cooling for you
Moto

 

[Ladamyre]

The Bix rad have higher fpi so they will want highspeed fans to perform at their best, which means louder...

That objection has always made me grin. Yes, to get the same amount of airflow through a high fpi radiator, the same fan will have to run faster than through a low fpi radiator. But we're not talking about moving just air, we're talking about moving heated air. Boyle's Law says if you heat a gas, it expands and with higher fpi the air is hotter than with low fpi and therefore more expanded. When it's more expanded it's also thinner. I have no tests to prove it, but knowing what I know about these principles of physics, I'd bet $1000 (what I'm willing to spend on this system's cooling) that the same fan will run just a bit slower through a high fpi radiator, than through a low fpi radiator to remove the same amount of heat (BTU's) from the coolant. Besides, even if my reasoning is wrong, high fpi will remove more heat from the coolant than low, and since the main objective is the coldest possible coolant as it enters the CPU W/B, high fpi radiators are non-negotiable.

I have a plan for fan noise anyways and the Crosshair V's eight fan headers are gonna come in handy for that. Three of the four fans that come with the case will all be inside, two doing the pull side of the dual rad, one pulling on the single rad and the last one doing the exhaust at the back, it will be running slow most, if not all of the time. I'll get 3 fans that are the quietest I can find to do the push on those two radiators, and 3 more to push out the top of the triple radiator. The radiators should muffle a lot of the noise from the five fans on the 'inside' side of them, and those fans will do most of the work.

Have you considered an external radiator box btw?

I shudder even at the thought of those single radiator mount kits that attach to the back of the case. No, like I said, I want a PC I can pack up and put in the back of the van in 15 minutes.

I think if you are insistent on the mobo/ram cooling that a dual loop res would be in order, that way the ram etc isn't restricting the cpu and Gpu's flowrate...two pumps, each feeding their own loop but sharing a reservoir makes a larger volume of water in the loop which means more time before your thermal mass heats up, which lets your fans work slower to balance it...

Moto

It's been a while since I looked at the specs on it but I think I'm going with the XPC RASA for the CPU because it has a low restriction and will accommodate a high flow rate. My plan, in case you missed it, is to use high volume pumps in tandem. This will give me a high flow rate and redundancy in case one should fail. Two loops would mean four pumps, and since there is 840mm of 60mm thick, high fpi radiators, it won't, IMO, bring any noticeable difference in CPU cooling. Maybe you missed the plan to split the loop for parallel cooling on the RAM, N/B, MOSFET's and GPU's?

I'll run it down again, reservoir through the tandem pumps, to the dual rad, to the CPU, to the single rad, then split into two lines, each to it's own two RAM modules, then one of those to the N/B and the other to the driver MOSFET's, then the two back into one line, to the GPU intakes (the Koolance setup then splits into parallel and collects into a common exhaust), then to the triple rad and finally dumps into the reservoir.

Your suggestions are good and I appreciate the thought, but, and don't take this wrong, I have already thought of all of them...

...except the earlier ones about automotive coolant and using UV radiation. What do you think about a large custom made reservoir with built in UV lighting inside to irradiate the coolant and at the same time, bling out the inside of the case?

Maybe a little mirrored disco ball spinning inside.

 

[Motopsychojdn]

When i say external radbox, I mean something like mine, I have quick disconnects on the lines and can move that or the main pc independently,
http://www.tomshardware.co.uk/forum/page-274180_11_150.html
Page four has most pics to give you an idea of how they are,
I had forgotten the redundancy option you mentioned and I run two pumps for the same reason, so I think yes, parallelling the mobo blocks would be a plan,
I'm thinking to put the mobo blocks between the two pumps so as to negate any pressure/flow loss incurred, maybe:
Pump 1>Ram>Nb>Rad>pump 2>Cpu/Gpu's and rad then res>pump 1,
definitely no to the Glitterball lol, well its your rig and if you want to look at it hehe
large res with Uv's get the thumbs up yes
Have to call it for today, I need sleep before my nightshift, I'll check in though
Moto

 

[Ladamyre]

I don't know where people get the idea that the flow of coolant is going to somehow slow down further away from the pump. It's like thinking that if I am towing two cars going 40 mph, with a 50 foot rope on the first car and another 50 foot rope from it to the second car, somehow the second car will be going slower than the first. The coolant flow will be slowed by a longer loop, yes, but spreading out where the pumps are will make no difference in the flow at any point in the loop. Like in the analogy, if I'm trying to tow the two cars with a 4 cylinder Mini Cooper, I might not get to 40 mph, but since I'm towing with an 8 cylinder Mercury Marauder I'll make 40 easy, and both cars are gonna go just as fast.

Oh! You mean they make a thing where you can detach the lines from the computer? Do they leak? Oh, I get it. Like those air lines I used for so many years back at the car dealership.

Why didn't I think of that!

I'm being facetious. If I were to do an outside-the-box radiator, I wouldn't be using this case. I would use a much smaller case, maybe even one of those old timey (going back to my first 386 Tandy) desktop flat box, like this one here.

 

[Motopsychojdn]

Aaargh! beige, make it stop hehe
yes, I meant the connectors you are thinking of, some are not no-spill, some are,
I used koolances no-spill on my box and they don't leak even upon disconnection
Moto