Discussion in 'Electronics & Electrics' started by Omarko, May 3, 2014.
Plenty of 1.5A+ diodes around (that dont seem to need heatsinks (since they are cylidrical).
I'm curious to see how it goes, let us know!
You're right but; The 1N5400 will handle 3A with no heat sink however at 1.5A it'll be over 100 degC. By comparison a 5W resistor at the same current has a temperature rise of under 70 degC. So yes, 2x 1N5400 diodes in series or a 5W 1.2 ohm resistor will achieve the same outcome however the 5W resistor will have a lower case temperature because of it's larger size. More area means it can dissipate the same power with less temperature rise. Less temperature means less risk of melting something that touches it e.g. wires.
All of that said the best solution would be the buck converter. The OP is assuming less pump speed will give less noise. That's perhaps only generally true as in part the noise from spinning things such as pumps and fans can be due to harmonic vibration. Get the device to spin at just the right speed and vibration and hence noise can be much lower than a slightly higher or lower speed. Having a pot to give fine control of the speed makes the job of lowering the noise much easier.
If it was me, I would use a LT1074 or LT1076 http://cds.linear.com/docs/en/datasheet/1074fds.pdf
Nothing wrong with that family, but there is quite a lot of work designing a PCB around them and the datasheet doesn't offer a reference design to get you started.
As the OP doesn't yet know the required output voltage, an adjustable regulator would be mush simpler.
thanks guys, I got the regulator today, looks really cool!
the only thing I am worried about atm is short circuit (eg against metal part of the case in the PC). Can I cover the regulatort in electrical tape or does it dissipate heat?
The LM2596 may get quite hot, so tape wouldn't be my first choice.
Self adhesive stand-offs are the easiest, but nylon and brass have their advantages.
thats a great tip, thanks for that!
ok, so I hooked it up, works beautifully !
I'm running the pump at 9V, cant hear it ! it dropped the pump RPM from 4500 to 3200.
next is :
1. measuring the temps (CPU loop)
2. isolation of the regulator from metal casing
It does get bit warm to touch, but not too hot. I might put a small heatsink on.
In general with electronics if you can hold your finger on it, it's not too hot.
The hottest most people can put up with is around 60°C, which sounds reasonable as an upper limit.
For the finger tips of adult males probably a bit more than that. If I'm in any doubt I lick the end of your my finger and tap it lightly and if it sizzles it's over 100. Unless it's a wire would resistor in free air it might be time to think about a heat sink.
ok no worries then, its not that hot at all.
I ordered another one for the 2nd pump.
So far its been working beautifully!!! I don't understand why computer shops like PCCG dont sell them !
Because of the possibility of fucking them up.
For example, if you wound the voltage down to far, the pump wont spin up, take a large amount of current, and burn out.
Most regulators have thermal protection but it often kicks in at over 100 degrees, so there is the possiblity of burns.
However, you will find that some fan speed controllers might do the same job, and thus, they do sell them really.
If I wasn't supposed to be doing uni work - I'm mighty tempted to design a switching regulator (with fuse/protection), and then stick up the design and the component list for OCAU'ers to reproduce..
The WC pumps I have seen are electronically commutated, surely their controller will shut down before it frys itself?