Group Purchase? PWM Voltage Regulation for Water Pump

D

dinofx35

Weaksauce
Joined
Jan 16, 2006
Messages
91
board.gif


I have a working prototype of the above circuit that controls the speed of my Laing DDC water pump. Using PWM from the motherboard, output voltage can be controlled from 7.5 to 12 volts (Speedfan or your BIOS controls the signal). This means your water pump (or non-PWM fan) runs quieter when your computer isn't doing anything.

You can hear the noise benefits in my system here:
mp3 audio

The kit (assembly required) would be around $15, perhaps less depending on the interest. Any takers?
 
I can't tell for sure- but does the motherboard supply the power directly to run the pump/fan, or is it supplied from the +12 line from the power supply?
 
is that a PWM to voltage converter or a high power buffer for the PWM?

anyways it might be more popular if you use non SMD components some people get scared off by having to solder such tiny stuff. little 1/4 watt resistors are not much bigger anyways.

I guess it depends what the IC is and if there is a non SMD version or equivelant.
 
I'd be even more interested in a fully assembled board. There's quite a few company out there that will make prototype boards for less than $100 and them smaller numbers of the finished product. I'd be in for that! :cool:
 
Troney1169 said:
Good idea I would take one but i dont run WC yet.
Click to expand...
I'm in the same boat. It's a good idea that unfortunately I can't use.
 
This is not a high power buffer for PWM (buck converter). That type of circuit is hard to design and get a linear response from the input signal. I was also concerned about high-powered switching generating RFI. So, I guess that makes it a PWM to voltage converter. Really, it's a voltage to voltage converter. The PWM is filtered to a steady "analog" input.

My thoughts are this: If you are going to use the larger "through hole" components instead of surface mount, you may as well build the circuit on a prototype board, which is what I've done already. The soldering was so much easier than I anticipated (previous experience limited to RC cars), I decided surface mount components wouldn't be much harder, and the end result is a much smaller, professional looking board. For my prototype, I was using a large tip (stock tip on a Weller), and the solder does not leak off of the pads. If anything looks too close, you can cut it.

All of the soldering should be simple except for the 8-pin chip. Which only needs 5 pins soldered. I'm sure someone would be willing to do this for would-be group-buyers.

The circuit is grounded from the motherboard, and receives +12V from the pump's power connection:
schematic.gif
 
willie92 said:
I'd be even more interested in a fully assembled board. There's quite a few company out there that will make prototype boards for less than $100 and them smaller numbers of the finished product. I'd be in for that! :cool:
Click to expand...

How much more are you willing to pay for assembly? Do you have any links to sites providing this service?
 
Good idea dude, I like the looks of this alot. It should help with lots of pumps that don't have a built in rheostat (or variable resistor, I don't know which one is used in the D5's).

If I ever build a system with a couple smaller pumps, I'd so be in for this.
 
OK, finally some real interest. I was worred after posting on another forum of so-called SFF modders.

Here's the deal, the minimum order just for the PCB is $160. The current design is 1.0x1.45 inches (see updated image at top of thread). I then need to order the HEXFET, heatsink, capacitor, and MOLEX KK parts for each board. I already have the SM parts. Probably 3 weeks turnaround.

I have spent lots of time simulating and prototyping this circuit, and designing the PCB. I'm just looking for buyers to bring down the cost of fabrication, and of course to share the idea. Otherwise, I'll just live with the working prototype.

Now let's see some committment. Even if you haven't started your $200+ water cooling setup yet, go ahead and support the cause. PM me or reply to this thread. I'm looking for pre-sales. If I can reach $100 in presales, I'll throw in the other $100 (in addition to what I've already spent). The price will be MAX( $100/#orders, $15), plus $4 S&H. We could use PayPal to set this up.

Buyers can have input to the final design. For example, how will this be mounted in a system? Currently it has 4 holes, which you'd have to mount spacers on and then back it with lexan, and stick it somewhere with double-sided tape. Another option would be to have a TAB sticking out to one side which could be used to hang the circuit off of the bottom of a drive bay or something.

In the interest of "full disclosure", this circuit generates about 3.7 watts of heat when your system is idle/pump running at low voltage. But, your 18 watt pump would only be using 8 watts @7.5 volts, so the combined power consumption is much less. Most of your pump's 18 watts would be converted to heat anyway. When your pump is at full speed, the circuit consumes only milliwatts. The heatsink gets hot, and is electrically connected to variable voltage output. You can use a fast-burn fuse or a thermally conductive isolating strip if you are worried that the heatsink will contact something in the case. I am not using either currently.
 
Arcygenical said:
It should help with lots of pumps that don't have a built in rheostat (or variable resistor, I don't know which one is used in the D5's).
Click to expand...

But, even that pump does not varry its speeds based on the needs of your system.
 
Put me down for 1 - I can see this running both fans on my WC system...
dinofx35 said:
Buyers can have input to the final design. For example, how will this be mounted in a system? Currently it has 4 holes, which you'd have to mount spacers on and then back it with lexan, and stick it somewhere with double-sided tape. Another option would be to have a TAB sticking out to one side which could be used to hang the circuit off of the bottom of a drive bay or something.
Click to expand...
One suggestion - Multiple on-board output plugs. I've got 2 120s that I'd like to run together. However, I'm sure this could be accomplished using a Y-connector, but would look much neater....
 
dinofx35 said:
How much more are you willing to pay for assembly? Do you have any links to sites providing this service?
Click to expand...

I'd probably pay $20 - $25 for it to be assembled.

Here's a place I was going to try when I was toying around with the idea of making a latching relay board:

Olimex
 
lamarth said:
Put me down for 1 - I can see this running both fans on my WC system...

One suggestion - Multiple on-board output plugs. I've got 2 120s that I'd like to run together. However, I'm sure this could be accomplished using a Y-connector, but would look much neater....
Click to expand...

Please note the usage in the above schematic. The 3-pin header is not a standard fan connection. It is 12-volts in, variable voltage out, and a tachometer. This is to avoid having a 3rd plug on the circuit for a high-powered ground. The "input" ground can not be used for the pump's ground because the motherboard header is not rated for 1.5 amps. The circuit is pretty much divided in two parts so that the real current is provided only by your 12 volt rail.

You could use a Y-connector, but you'd have to turn it into an X basically.

But, for just running 2 fans (0.2 amps per), I could power these directly from the motherboard, which is generally good for 1 amp. So you could draw the power from the motherboard with a single plug, then I could put 2 or 3 standard 3-pin connectors on it for fans.

I would basically order half as many circuit boards, each with two separate circuits on it and cut them apart on arrival. But I don't think the demand is 50/50.

The most flexible approach would be add another 2-pin plug for power, and then wire your pump the same as a fan. Schematic to follow...
 
The schematic diagram has been updated (press reload if you aren't seeing a change).

The new usage allows you to power the circuit from your motherboard header, or from a standard MOLEX power connection. In either case, the circuit now has 2 outputs, both of which conform to the standard 3-pin arrangement (ground - positive - tachometer). The tachometer for the second output is not reported back to the motherboard.
 
dinofx35 said:
But, for just running 2 fans (0.2 amps per), I could power these directly from the motherboard, which is generally good for 1 amp. So you could draw the power from the motherboard with a single plug, then I could put 2 or 3 standard 3-pin connectors on it for fans.
Click to expand...
That's my greatest worry- I understood that a typical MB can't source enough current to run large fans. I would rather have it control another board which did the current sourcing.

Both my 120s do use a 3-pin connector (Panaflos), btw.
 
You must log in or register to reply here.
Back
Top