Swiftech MCP-35X & Reservoir

Posted: February 25, 2011 in Pumps
Tags: , , , , ,


Welcome to my Swiftech MCP-35X review, the latest in the extremely popular Laing DDC pumps. This latest incarnation just got smarter with PWM technology and has several factory high performance features included within. I want to give a special thanks to Gabe from Swiftech for sponsoring this pump.

If you have been watercooling for long, you’ll likely recognize this family of pumps that have many names and variations over the course of years in refinement. The Laing DDC series has been through various revisions from the DDC 1, DDC 2, and over the last view years the DDC 3.1 and DDC 3.2 being offered in two power flavors. The current common models would be the DDC 3.15(MCP 350) and DDC 3.25(MCP 355 or PMP400). They are exceptional and popular pumps because they are tuned for high pressure applications and carry a wallop of power in such as small package.

A common “Missing Refinement” of the “non 35X” pumps is that they typically come with a factory top that uses plastic 3/8″ barbs. The factory top typically comes with a sharp elbow at the pump inlet (which is poor in efficiency). Sooo..most water cooling enthusiasts have had a desire to use their own fittings which requires replacing the factory top with something after market. This provides users with standard G1/4 threading, improved looks, and higher performance. Unfortunately, removal of the factory top typically means you also forfeit the warranty.


It is very clear that switching out the stock pump top to an aftermarket top increases pumping power, but what many fail to recognize is it also increases current draw and pump heat. This in itself puts more stress on the pump motor and causes more heat buildup. Sort of like putting a turbo on your factory high performance car or truck, you exceed specifications and increase wear and tear. I also believe a large portion of pump failures are due to leaks by incorrect top installation. I myself have managed to flood out two different DDC pumps by incorrectly tightening the tops on (it happens). The way the pump motor sits on top of and inside the exterior shell, if the o-ring leaks, water will work it’s way down into the PCB area first. Leaking water will eventually fill up the bottom case and start spilling out of the holes in the bottom, but many times not until the PCB is completely under water. Anyhow, we’re only human and we will and do mess up. The manufacturers know this and you forgo your warranty privileges once the factory top is removed. I believe some of the newer packages will even start coming with a warranty void if removed type sticker, but my particular sample did not.


That’s where things start looking really good for this new 35X pump. It comes factory with a quality pump top that is already installed and ready to go. No need to remove the stock top and void any warranty, just plug it in, screw on your desired barbs and go. You get factory and warranty protected high performance right out of the box.

The top contains several new and/or improved features that are worthy of design/efficiency note. Of most importance is the spiral shaped volute. Not the best picture here, but the radius and space between the impeller and wall is continuously increasing. In addition the wall was cut with a radius bit which should help further reduce friction at this critical location. The top is also similar to the stock top in that there are centering feet to ensure proper inlet centering on the impeller. In addition the outlet port was specifically oriented in line or in plane with the impeller for a perfectly tangential alignment with the impeller (most of the current tops have a slight bend). This help keep the impeller running flat. And the best I’ve seen in matching the spiral volute to the exit port, the drilled hole lines up perfectly with the radius milled volute.

But the features and details of the volute are very clear that high performance was at the top of the priority list. I’ll include more on performance later..


This is the first I have seen of PWM pump control in the history of watercooling which offers you the ability to control pump speed either manually or automatically via MB core temperature. Prior to this pump, variable speed control was reserved to either voltage regulation, or buying the D5 vario (MCP655). Now you have that option here too. Furthermore, PWM offers a broader range of speed control AND more importantly and safe minimum and maximum speed. With voltage control, you have a fairly narrow and unforgiving window. Under about 7V OR over 13V and the pump will NOT start. If you hook any ordinary pump to a fan controller only to have accidentally bumped a control knob, that could leave you with a pumps sitting there and zero flow. It wont take long for the water in the CPU block to start boiling which eventually will lead to deforming and bursting of tubing nearby (I’ve almost done this myself).

Well, no worries using PWM control, if connected to a molex and you set the PWM via MB controls, the pump will always start no matter what the PWM setting. Even disconnecting the PWM signal wire, the pump will simply default to 100% pumping power. PWM control means several things:

  • Wide RPM speed control (DDC Vario)
  • Safe Speed Control (Pump will always start)
  • More Low RPM range than Voltage Control
  • Ability to scale pump speed with CPU temperature (On Demand)
  • No extra controller necessary, typically possible via MB and MB related software (Free smart controls)

For PWM control, you will ideally have controls on your motherboard, but the level of control will depend on software and the motherboard used. For example, my older DFI board from a few years back can only set one temperature point on the CPU fan, and the motherboard ramp up down is fixed. However on my son’s newer Gigabyte MB and using EasyTune 6, I had full control over PWM from 10% up to 90% in 1% increments. This is what I used to run my tests as it turns out measuring duty cycle at higher frequencies is reserved for those who own a scope. My multimeter in Duty mode is simply incapable of measuring. With that said, I must put a large grain of salt on my PWM factor and suspect it may be off by as much as 10%. The performance PQ results are good, but the PWM Duty may be off a bit due to only having the MB software to go by.

Here are my PWM vs RPM scaling results. I increased PWM value from 10% to 60% in 10% increments and pulled RPM maximum values from those test to plot out the following chart.

And here you can see the PWM wires and PCB. At the surface level the PCB looks identical to that of the DDC3.25, but what hides within is a very different motor. Don’t assume you can simply solder in a green wire on a DDC3.25 and create a 35X motor, it doesn’t work.


It comes as no surprise to me when I opened the boxes that it was yet again a very complete accessory package, Swiftech is always very generous in the accessories. The package comes with:

  • 35X Pump Motor/35X Performance Top Assembled
  • Swiftech 1/2″ Highflow Barbs
  • Swiftech Clamps
  • Vibration (Noise Reduction)Neoprene Double Sided Sticky Pad
  • Corse threaded screws for solid mounting
  • 3 Page User Manual

The user manual is very well written and complete, it should be easy to follow for most users.

Overall an excellent and complete pump package with accessories above and beyond the factory package. It really is everything you need plus some..


To compliment the pump, there is an accessory option you can buy in the form of a combo reservoir unit that mounts directly to the pump top. In general I have found that combining a pump + reservoir is a good thing for pump performance as you generally eliminate extra connections plus tubing that creates additional restriction. The reservoir installation was a little different in that you have to install a little orientation pin that fixes the barb orientation relative to the base. Then you have to carefully place a sealing o-ring in place on the top. Next you set the reservoir on the o-ring and on the pin, and finally you get to use a long large flathead screwdriver to tighten the retention base fitting. I did this several times and each time went without a hitch although it was a little bit interesting.

After that is done you install a wire mesh and sponge over the inlet port on bottom. I presume the wire mesh is there to prevent any sponge or other material from accidentally getting into the impeller blades. The combo unit looks like this which is a nice compact unit only measuring roughly 5.5″ tall.

All fitted and ready for action!

Although I should note that I was not able to fit a 3/4″OD compression barb in the upper side inlet hole. The flat spot molded into the side was about 1mm too low and impacted the barb edge. I believe this was just a fluke in my particular sample, the holes were just drilled/tapped slightly too high. This could be fixed easy enough with a little dremel work, but I don’t have plans to use compression fittings so I didn’t bother.

The sponge systems allows you to utilize the side ports for one or two inlets. Depending on how you setup the system. I tested the sponge system with the lower inlet port.

But as a side experiment, I also tried creating my own top inlet/downtube setup which actually worked better both for bleeding and for performance. I simply screwed in another barb in the bottom of the top cover and cut off about a 3″ piece of tubing and tried to center the bottom part of the tube to line up with the pump inlet. Ideally you could also use some clear 1/2″ acrylic tube or something a bit more rigid/straight.

The system is very modular for inlet options.


In case you wanted the fine print details, here are the charts for pump only (no reservoir) for 10% through 60% in 10% increments as well as 100%. I’ll thumb these out below. This actually took a fair amount of time, but considering I’ve never tested anything with PWM control, I wanted to get some experience with it and did full 10% incremental tests.

And the summary Chart of those results:

Generally I think 30% is about the minimum you would want to go with even a low restriction loop (1GPM or better), and the majority of the speed change occurs between 10% and 50-60%. My 60% and 100% run came out about the same which is probably due to my PWM software and/or motherboard not being very precise. The PWM frequency is too high to allow multimeter measurement of duty cycle, so I could only go by what Easy Tune told me which appears to be approximate at best.

One thing to note about the summary chart, I also included approximate voltages that equaled the same RPM, but since the pump startup minimum is around 7V, I was not able to duplicate the lower PWM speeds.

PWM provides more low speed options than analog voltage control.


If you sit down and think about the likely purchase scenario’s for a DDC pump/top combo, it might look something like this:

Option A would be the Swiftech MCP35X with factory performance top.

Option B would be some version of the DDC3.25 such as the MCP355 or PMP-400 and you would add a performance aftermarket top of your choosing.

To compare these two options I selected the XSPC top as it was my top performing standalone top from my pump top shootout some time ago. I compared it directly on the same motor as the MCP35X and it performed equally as well. So the following comparison is more about pump motors than it is tops:

The 35X option has a decent performance advantage, moreso than most top differences. A low restriction loop may see about .2GPM or 10% improvement to flow rate over the DDC3.25+top alternative, and ultra high restriction loop will be approximately the same. An average system would see about 5% more flow rate.

So if the tops perform about the same, why does the 35X perform better you might ask?

The answer is RPM limiter programming, which is fairly obvious in the following chart:

The RPM limiter programming in the DDC3.25 is much more aggressive leaving the DDC35X to have a simple RPM advantage particularly with lower restriction scenarios. Note that this does come at some expense in power consumption though and possibly some increase in pump heat.

Regardless, the performance is simply hard to touch with that RPM advantage and as far as I know, the only way to get the 35X pump motor is to buy a 35X pump. And considering you get everything in one package with a backing 24 month warranty, that’s a good thing!


Along with any reservoir, you’ll want to know at least something about it’s ability to bleed air out of a system and how the performance of that reservoir compares to the stock top.

I tested the factory screen/sponge system with side inlet, pump alone, pump plus separated reservoir, and top downtube alternative.

Regarding bleeding performance, the sponge system worked ok for flow rates below 1.5GPM, but above 1.5GPM the inlet flow disturbance was enough that the surface water was fairly disturbed and I was getting air sucking back down the pump at the 2+GPM mark. This could be mitigated by simply turning the pump down a bit, but it should be noted.

With the sponge not working as well as I hoped, I set out to look at the alternatives. I found that the reservoir works perfectly using a top down inlet plus down tube configuration. This is extremely easy to build, all you need is an extra barb and about 3″ of tubing. Even better would be some sort of clear rigid tube, you just need something to protrude down through the surface of the water in the reservoir and point at the pump inlet. I tested with a piece of 1/2″ x 3/4″ tubing as I assumed that would be readily available by most users.

Here are the performance test results comparing the various reservoir alternatives tried:

While we are splitting hairs a bit here, I did measure a slight performance advantage to using the combined reservoir vs separated reservoir. This makes sense because you have one less fitting plus less tubing loss, etc. I also found the top down inlet tube idea to be superior to the sponge system in performance. But the key benefit to the top down tube is very still surface water clear out to 3GPM and more. It’s also more resistant to varied fill height levels.

I would recommend the TOP INLET INTERNAL TUBE ALTERNATIVE, it works the best in my tinkering experience both for bleeding and for performance


For those that haven’t used a DDC series pump motor, you should be aware they get hot on their base and the 35X is no different in that regard. I measured up to 55C in a 21C ambient which feels toasty to the touch. Many users of these pump go to an extra level of effort to provide some air flow over the base of the unit to mitigate that heat and while I’m not sure it’s necessary, I can’t imagine it hurts. I’ve run serveral DDC pumps, some with air flow, some without and never had any failure type problem. In this case, the pump motor and pump top were designed together and warrantied together, so I wouldn’t let heat bother you. Just don’t be surprised that they get hot to the touch.

Many people know this, but efficiency is also very important in water cooling. You want the least amount of heat transfered to the water as possible for maximum cooling. Many might think that the pump heat would not matter because it is too small. While it is true that the pump heat numbers are small, it’s also true that the gains by flow rate are small. Because both values are small, it doesn’t take long for heat dump to counteract the net flow rate gain by more pumping power. While we are splitting hairs, efficiency should not be ignored. For the same flow rates with two different pumps, the pump with the better efficiency will generally dump less heat into the loop and net fractions of a degree better temperatures.

With that said, you should know that the Swiftech MCP-35X pump is very efficient electrically compared to the D5 series pumps. I like both pumps very much, but it should be noted that you’ll have a little less heat dumped into the loop using a pump like the MCP-35X.


I will include this pump in my pump noise thread for further noise analysis, but considering it’s a DDC pump with variable speed, it sounds a lot like a DDC pump except it runs a bit higher in speed at max setting and has variable adjustment to get really low speeds as well. This is obviously advantageous to have via PWM control because you could scale RPM and Noise via thermal demand if tuned properly. I will do more of this later in my noise thread.

In the mean time, I’ve included this video that explores noise somewhat informally using a custom PWM controller. This gives you a sense of the noise scaling that is possible via PWM.


I really like the pump/reservoir combo for all out DDC performance in a smart variable and warranty covered package. My only complaint would be the sponge system in the reservoir being difficult to bleed at really high flow rates, but that can be mitigated by either dialing down the PWM setting or using the top inlet with an internal tube.

I look forward to experimenting with this and more “Smart” water cooling products that can scale noise on demand. Swiftech has taken one of the most popular water cooling pumps, added smart PWM speed control features, reprogrammed the RPM scaling for even higher performance, and included a very well designed pump top. All of this comes in a factory warranty package, very nice!


  1. Walter Bak says:

    i got the same pump and reservoir.
    works just gread, but i have to say tx for the top inlet/downtube setup.
    im gona try thad.

  2. LipschitzWrath says:

    Sorry to resurrect an old thread, but I have a question for Martin. If one desired to, say, put every fan (case and radiator) in his case AND this pump under PWM control off the mobo header, would it be better to pull the RPM signal off the pump or a rad fan (I’m guessing a case fan is a poor choice). This pump obviously goes up to ~4500 rpm, but the fans I want to use max out at like 1200 rpm. Theoretically, is the rad fan or the pump a better source for the RPM signal?

    And yes, I know you can’t power 4+ fans off the mobo header. I plan to use one of those fancy Akasa cables that sends PWM/RPM to the mobo yet pulls power from a molex.

    Thanks in advance, I know if anyone knows, it would be you Martin!

  3. PepeLapiu says:

    Hi Martin,
    you basically shrugged off the idea of cooling the pump as it is designed to operate with some degree of heat. But I would like to add that Swiftech sells a heatsink to place at the base of the pump with provisions to mount an 80mm fan to it.

    I think it’s a good idea because whatever heat a fan and heatsink can remove, that is heat not absorbed by the coolant inside the pump.

    • Martinm210 says:

      Good point. At the time I think I did this review, they didn’t have their heat-sink in place. They do run hot if 100% all the time and especially when the system doesn’t have much restriction. But it never hurts to do some extra cooling and their heatsink is a simple bolt on way to do that.

  4. Sprucemoose says:

    Hi, I feel the need to just add a cautious reminder about the black reservoir sponge that shipped with the reservoir. After a year or so running my loop with anionized water + PT Nuke, this sponge had become so brittle that it collapsed into tiny small pieces when touched. It seems that the sponge had given off a small amount of these particles without being disturbed too, and some of these particles have probably been lodged in the waterblocks. I would advice anyone to remove the sponge and only go with the metal grating as soon as the loop is bled properly.

  5. wrogad says:

    good work and the top inlet i already have from an old EK reservoar that i took the acrylic tube from.

  6. Tim Chartier says:

    Hey Martin, great stuff as always. I have a question concerning the top inlet/downtube mod.. Did you use the top inlet for the return (with a fill port integrated as well)? Or, did you use the side inlet as the return. I just want to confirm the best port to use for return and where your fill port was when going with the mod. Thanks

  7. Olo says:

    “Under about 7V OR over 13V and the pump will NOT start.”

    My pump (ek dcp 2.2) starts at 4.6V it is merely audible from a few centimeters
    kind regards

  8. Olo says:

    “If you hook any ordinary pump to a fan controller only to have accidentally bumped a control knob, that could leave you with a pumps sitting there and zero flow. It wont take long for the water in the CPU block to start boiling which eventually will lead to deforming and bursting of tubing nearby (I’ve almost done this myself).”

    True ! using voltage regulator you should be very careful – especially if your loop is designed with no up to CPU radiator – then warm water will stay at the top of the loop (your CPU block 🙂