Pitch seems ok:

1046:
http://img94.exs.cx/img94/8858/1046.gif

D4
http://img81.exs.cx/img81/7694/d13.gif

Headset mic (I think it was?) + 4" desktop speakers may not (indeed) be optimal combo for reproducing 50Hz hum :D

Is it active voltage control by stepping the power transistors or just passive filtering with something like an LC filter?

Nice lolito, 50Hz harmonics are clear (~3000tr/min for France).

[mode fr :D]
C'est quoi ton soft? Goldwave gérait pas ca... J'ai bien envie d'updater le dossier avec ca.
[/mode]

Thx

Sorry I don't know ;)

Roscal, I used Adobe Audition (ex CoolEdit Pro)
no problemo si tu veux faire un copier/coller (même si jai fait ça un peu Ã* l'arrache, comme on dit :D )

C'est bon j'ai trouvé un freeware qui fait les FFT , ca rend très bien :) -> http://audacity.sourceforge.net/windows.php
Regarde la 1048 le fondamental est a 100Hz par rapport au 50 de la 1046, pourtant elle tourne a la même vitesse d'après leur moteur (3000 tr/min), bizarre que la fréq soit exactement le double, je vais réfléchir Ã* ca.

Bizarre...
1048:

http://img20.exs.cx/img20/2951/2705.gif

As Roscal says, no 50HZ fundamental for this one?!

Also note the cutoff at 7k or so (same for the other pumps) this is due to the mic and/or codec, and will have an effect on the perceived noise - of the D4 in particular

DDC:

http://img48.exs.cx/img48/2576/ddc.gif

small spike at 58Hz, moves up to 62Hz when the pump speeds up. 3480->3720rpm?

Yep, DDC spins from 3465 tr/min at full flow to ~3900tr/min at 0"head. Codec cut high frequencies I think because I must compress sound records for bandwitch.
The 1048 is perhap's a bad recording but strange because I didn't change anything, just a pump replacement. There's a little spike at 50Hz but very small in comparison of 100Hz.

Sounds like the D4 could become nearly as quiet as the DDC, given the right electronics (i.e. done as a product update/improvement at Laing).
Is it possible that this more sophisticated electronics might not get along so well with over/under volting? Might be why they went with square wave on the D4. Probably just a cost of goods issue, though...

Simple coil switching of the motor coils is definitely the simplest and cheapest solution. It reduces losses in the drive circuitry and it also gives you more power out of the motor compared to a sinusoidal drive. The cost is reduced motor efficiency, increased motor heating, and increased noise. It's certainly possible to design a sinusoidal drive that will undervolt well, but cost, efficiency, and drive circuit dissipation tradeoffs may weigh against it. A sinusoidal drive shouldn't add more than $1-$2 to cost of manufacture.

After thinking about things, I am now more curious about the actual physical cause of the 50Hz switching noise of the D4. It cannot be electronic in nature, since electrons don't make noise. :D

So, is it the electromagnets vibrating as they are turned on and off? Is it vibration of the shaft as it is subjected to changing torque when the electromagnets turn on/off? Is the amplitude of the noise caused by the required travel freedom of the D4 spherical rotor? Would a traditionally axle driven rotor be less noisy for a given drive electronics type because it might be less "sloppy" ?

When you switch the coils on it generates a sudden torque pulse on the motor. The rotor is given a sudden push, and the pump frame experiences the same torque pulse in the opposite direction. Sort of like hitting it with a hammer at 50HZ to make it go around . Something similar happens when the coils suddenly turn off.

You can modify the abruptness of the torque pulses by changing where in the rotation you turn the coils on and off. If you turn the coils on when the magnetic field is at it's strongest on the coils you will get the worst case, if you turn the coils on before the rotor rotates into the strongest part of the field the initial torque pulse is less.

You can also reduce the abruptness of the torque pulse by turning the coils on and off slowly which is essentially what a sinusoidal drive does.

You can also get noise from the metal laminations that are used to build the core of the electromagnets. As the magnetic field is turned on and off they are alternately pulled together and apart. This is not usually a dominate noise source.

Thank you ferdb. That was what I expected to be the answer, although you describe it much more clearly!

I guess the operative question is "Why is the Laing pump so noisy in comparison to other pumps?" Is it doing a trade-off of noise versus efficiency that other pumps don't? Or is it made with cheaper tolerances so that the vibration from torque on/off is less damped (i.e. accumulating the energy from torque accel over time to deliver a rapid impulse when the slop factor is reached)?

The reason for asking is to see what the probability is for the pump mfg to produce a decent (still high head and low temperature) pump that is quieter if a bit more money is thrown into production costs or if we are experiencing a natural consequence of the required performance. You get what you pay for, but you can't pay to remove the laws of physics.

I'm sure most of the noise comes from the brushless dc motor and little from the pump. So if you make a quiet motor, you'll have a quiet pump. It's certainly doable, the question is is the market for such a pump big enough to bother with and will that market pay the price premium necessary for such a pump.

Maybe you are missing the key(?) : The D4 coils are fed with simple square wave AC, all the others are sine wave driven
(apologies if you already knew this!)

I knew that most pumps are given 120VAC or 240VAC power, but I was making sure that this was the only reason that the D4 is noisy. I was wondering if it was possible to keep the high efficiency square wave impulse and keep it quiet via mechanical dampening. Not insulation, but sturdy construction or other mechanical designs.

Thank you, though.

All I want to know is where I can buy one. :)

EDIT: Ack. Current pump failed. Cooked a Radeon 9800Pro 256. I just ordered the MCP650 directly from Swiftech. :shrug:

Sorry if this is mostly off topic, but does anyone know how MCP350 and MCP650 compare to CSP-750 purely on the grounds of acoustics?

MCP350 (DDC) is apparently "inaudible"
CSP-750 should also be inaudible, but I have heard of people claiming it gurgles.
MCP650 (D4) is far from silent, but it's likely a new version will fix that.

I don't hear my CSP-750s gurgling. In a closed case sitting on a piece of foam I cannot hear them at all. If somebody is hearing gurgling I would guess they still have bubbles in their system.

It's not a trivial task to dampen out the motor switching impulses mechanically. When the motor coils switch the torque goes from zero to close to the maximum the motor generates and then off again. The stronger the motor, the stronger the impulses. The coils are switched on and off rapidly to avoid switching losses in the transistors that drive the coils. You can change the drive circuits to slow the turn on and turn off times of the coils (I've done this with brushless DC fans) and it reduces the noise considerably but the drive transistor temperatures go through the roof because of the switching losses. You either have to put up with that heat and provide adequate transistor cooling or run the motor at greatly reduced voltages and rpms, or use a more complex motor drive circuit. Another design option to reduce the torque pulse noise is to increase the number of poles in the motor, but that increases costs. Changing the coil switch timings to occur at a point where the magnetic flux on the coils from the permanent magnets is still weak, or spacing the magnets and coils farther apart to reduce the flux gradient will also help, but again these options reduce efficiency and increase heat in the drive circuit.

It's hard to get manufacturers to increase costs and decrease efficiency for something as nebulous as quietness.

My CSP-750 (which I bought from Ed) is definitely not inaudible. I run it a little over 7V to get rid of a whining sound not unlike a hard drive's. No gurgling noises, but my flow must be little more than 2LPM at 7V. The lower pump speed costs me about 2C on the reported CPU temp. And you need a quiet system/environment for a CSP to be a noise problem.

Do you have your CSP750 sitting on foam?

He's got it completely suspended, same way I have my two MkIIs.

heh :) http://swiftnets.com/products/mcp350.asp