Golborne Vintage Radio

I intend to make a little PWM speed controller for controlling the speed of a 12V PCB drill.

I've already got one that I made 25 years ago, but I work indoors and in an outdoor workshop, so another would be handy to have.

Circuits for such things abound on internet and I've found one for which, fortuitously, I've got all the bits in my spares box. I've developed a component layout from the circuit, which is attached showing an 'X-Ray' of the PCB track layout from which I'll later make a PCB mask, but at this stage I'm after a bit of advice on heat sinking for the N-Channel FET used in the design. It's an IRL530N mosfet (TO-220 package), which I understand can handle current up to 27A. It was chosen by the designer of the circuit as it has fast switching and Low on-resistance. In this application, the PCB drill is unlikely to draw more than 2 Amps - 3 at most, and will be used for maybe up to half an hour, on load for drilling each hole in a PCB, which takes only a matter of a few seconds at a time, with a brief pause between each hole drilled.  

I'm wondering if the FET needs to be heat-sinked, and if so, would the small heat-sink in the attached pic be adequate if screwed down to the PCB with the FET laying flat, which would obviate the need for a larger heat-sink mounted onto say a metal back-plate on an enclosure - I'd rather avoid that if possible? I could of course always try it and see what happens but it makes sense to ask the question whilst at the development stage.

Some aspects of the component overlay in the attached diagram aren't to scale - the  78L05 voltage regulator for example is shown overly large. Initially I was going to use a7805 in a TO-220 package for that role, but it's only there to provide the 5V for the 555 IC, so  a 78L05 should be more than adequate for that.

Any thoughts on the heat-sink query anyone please?

A small heatsink should be fine. You may not need one at all. My inclination would be to lash it together on a bit of pinboard but I know you like things to look nice as well as work well.

It should be OK, but "+1" for prototyping it first

Any MOS-FET will be quick enough for this. The switching frequency doesn't need to be too high - anything above 10kHz will be fine, depending on your hearing! The switching time of the MOS-FET will be a tiny fraction of this - perhaps not even easily detectable. (If you go looking, use a resistive load - the inductance of the motor will dominate the observed wave shape).

Providing the MOS-FET is correctly driven into saturation when conducting, the R(DS-ON) parameter is the one that matters - this is the resistance between drain and source. This device is 0.1 ohms. With 2 amps flowing, that's 0.4 watts. In practice, it'll be less than that because you have to consider the duty-cycle of the waveform - if the MOS-FET is only conducting for 50% of the time, then the average current is halved, and the power dissipation is quartered.

One way to reduce the power loss in the MOS-FET is to pick one with a lower R(DS-ON). An oldie but goodie is the BUZ11, which is 0.04 ohms. For 2 amps continuous, the power would be 160mW. Last time I used one in anger, I used it to drive about 5 amps into some 12V halogen lamps. Without a heat sink, it ran warm, but was fine.

Try it and see, comfortable in the knowledge that the MOS-FET could be upgraded if needed. Even so, I'd definitely leave space for your small heat sink, just in case. Another tip - for these small-ish power demands, you can use the PCB as a heat sink - just leave a large area of copper leading away from the middle pin... It's obviously most effective with surface-mount devices, but it's "free" to try...

Looking forward to seeing the finished project - I only wish I could finish one of mine this decade

Small point. C2 and C3 should be right next to the 78l05... not for cosmetic reasons but to prevent RF oscillation....

Thanks for reading the thread Jeffrey & Mark, and for your prompt, helpful and reassuring comments. 

I'll report on progress in due course!

Hi.
I'd agree that little or no heatsinking will be needed. In my last job I repaired the big ride on floor scrubbers, they operated on 36v and the drive motor can draw 100 amps the speed of which is controlled by PWM. Even at this load the MOSFETS would run cool. So long as the on off switching is sharp and clean you'll be fine with the fet in free air.

A good place to harvest used FETs for this task is old cordless power tools. These would have been selected for the job.

(30-07-2015, 08:23 AM)Refugee Wrote: [ -> ]A good place to harvest used FETs for this task is old cordless power tools. These would have been selected for the job.

A very good point indeed.

Given that power tools are normally written off because of dead batteries, you'd also gain a working speed controller

Old PC motherboards might also yield some decent ones. The DC-DC converters on those are amazing feats of engineering (1.375V at 145 amps for a modern i7!)

Thanks for reading the thread guys, and for your helpful comments.

I've now tidied up the layout a little, and have finished the artwork for the PCB mask, which I've printed onto film, ready for exposing, developing and etching, when I've sprayed a piece of PCB with UV lacquer.

I've attached the updated component overlay and PCB layout for anyone who is interested.

I'll update the thread when I've made a bit more progress.