Pump Performance

The MCP655-PWM from Swiftech is a Laing D5 at heart and does, as such, already have a history of reliable and good performance behind it. But as with every product we receive, it underwent a full test suite, including that for the pump itself. To test a D5 pump, I use a reference D5 top from XSPC (XSPC D5 Bay Res) and the MCP655 provided with it. This is to help decouple the pump's performance from the pump top's performance, which will be separately tested below. Given this is a PWM pump, the first thing to test is its RPM response, and to do so, I used the MCP655 in a simple loop of reservoir-in to reservoir-out with pump power from an EVGA SuperNova 1300G2 PSU and PWM control from an Aquacomputer Aquaero 6 XT, which also provided RPM monitoring.


The Laing D5 PWM from as recently as last year was not really optimized for Intel PWM spec 1.3 despite the standard being over a decade old now. It lacks a pull-up resistor, which results in a non-linear RPM response near the lower end of the PWM duty cycle range and in the pump stopping completely instead of spinning at a fixed minimum RPM as with most PWM devices. A side effect of this is also a narrower range of RPM control and that the D5 PWM will spin at a speed corresponding to ~60% PWM duty cycle (~3100 RPM) if there is no PWM signal provided, rather than the usual 100%, which hurts when bleeding a loop as the pump is not pushing as much coolant as possible through and air out. Owing to these factors, the D5 PWM has not been as popular as it could have been, with the VARIO version ending up more popular overall. This is also why some manufacturers, including EK and Aquacomputer, created their own version of the D5 PWM with a different PWM motor for better RPM response curves.

When I contacted Swiftech about it, they mentioned that they had worked with Laing directly to address this issue and that newer batches had the updated PWM motor. After some checking, they confirmed that I had received an older batch since they still had some of those remaining and send a newer batch pump to check as well. The results above speak for themselves - the updated motor behaves much better, has a more linear response and more control in the lower end as well as a full-speed operation when no PWM signal is provided. Kudos to Swiftech for working it out with Laing.


An easy way to identify if you have the older batch or the newer one is to see where the pump specs label is located. It is on the sides on older batches and on the bottom on newer batch(es).


Here, I have mapped out the RPM response curves for some high-performance PWM pumps, including three from Swiftech. The MCP655 is the subject of this review, the MCP35X is Swiftech's take on the Laing DDC with a Swiftech-designed top on the Laing DDC PWM pump, and the MCP50X is an all-new Swiftech design and not a re-brand. I have also included the EK D5 PWM G2 (Gen 2), which is the aforementioned EK modification of the Laing D5 PWM pump. While the older Laing PWM pump would have had the one of the least linear plots, the newer batch does much better and is basically on par with the EK-modified D5 PWM. The DDC pump remains the worst of the lot now. Overall, the MCP655-PWM is a much improved Laing D5 PWM - provided you have one out of the newer batch.

In order to see how the pump performs, I used it in the same loop as before, but with a Dwyer 490-1 wet-wet manometer to measure pump head in PSI and a King rotameter to measure average flow rate in gallons per minute (GPM). This P-Q curve for a pump is extremely useful information as it dictates the average flow rate in a loop when any flow restriction is added from such components as blocks and radiators, under others. The P-Q curves below are thus for the pumps in a closed loop with minimal flow restriction present to produce an ideal and yet realistic data set.


I have included here some other pumps that have been developed within the few last years, and have at the same time removed the DDC style pumps from consideration to keep to a more D5/flow-optimized pump comparison. This helps keep the plots as clean as possible and gives direct comparisons. A lot of the more budget-oriented pumps introduced lately aim to match the Laing D5 Vario on Position # 3, which corresponds to the MCP655 PWM at 70% PWM duty cycle. As such, I have included that setting as well, in addition to 100% PWM. The Alphacool VPP755 is a brand-new "D5-replacement" pump sold by Alphacool, with a different design but based in a similar housing as the Laing D5, and it ends up outperforming the MCP655 (w/XSPC top) ever so slightly. The rest of the lot are hovering around the lower end with only the XSPC X2O 420 in the XSPC Ion and the Topsflo TDC doing better.