help with 48v dc fan control

[Axher]

Hi, I'm currently in a design phase of a variable speed 48vdc control for a fan. I want to connect the fan on the ac line so i need a ac to dc converter. I found out that the best solution is probably to use a switch mode power supply for the firts stage and a pwm buck converter to control the speed of the motor on the second stage. I need about 200 watts from the power supply at 48vdc to drive 3 seperate fans. My question is for the firts stage the power supply. Is there some integrated ic that I could use to simplified the design process? Or is there a simple way of doing what I want? I'm new to SMPS and dont know much about designing a high frequency transformer for it. Is there such a think as a commercial high frequency 48vdc secondary transformer at 200watt? thanks

Michael

 

[itsmoked]

Hi Axher;

You should seriously consider buying the 48V supply. There is NO WAY you will ever get the money back on designing your own. They are actually quite complex and generate a ton of EMI unless that is correctly handled also.
Switchers are designed buy huge companies that only do "that". Most are now off shore designed and built. India is the switching supply Mecca presently. If you buy one you will get warranties, solid specs that include MTBF, EMI, temperature deratings, ripple, soft-start, fold-back, UL, CE and a zillion other important certs.

Focus on the rest of your job and ride the mass-production trolley for your front end supply.

For your fan control just use some hefty Mosfets and PWM them! That would be the typical straight forward method.

Oh and for fans you could probably just use an extremely cheap unregulated linear supply.

Keith Cress
Flamin Systems, Inc.-

http://www.flaminsystems.com

 

[nbucska]

if you happen to have 3 phase power, it would be cheapest
to use 3 phase rectifier -- it would have an output
changing between 311V and 267V -- without any expensive filter.

You could just PWM this with about 1/6 duty cycle.



Plesae read FAQ240-1032
WEB: <

http://geocities.com/nbucska/>

 

[Axher]

Hi, itsmoked

I'm not sure how to do this without a SMPS with PFC.
I try to use a simple power conversion with a rectifier and a capacitor and transformer but the peak current that I was getting at the ac line was to high because of my buck converter at the output. Is there a cheap way to eliminated this problem without expensive filter?

 

[bklauba]

How about a simple SCR or PWM drive, ala Minarek or KB Electronics? 120 s.p in, 5 - 90 DC out, if you want no ripple, go PWM, and spend 'tween 50 - 150 bucks and be done with it. Current limiting, adjustable acc and decel, the whole thing is built, on a heat sink; plug and play.

BK

 

[itsmoked]

200W at 120Vac is only 2A...
200W/48 = 4A on the DC side.

I'm surprised this is presenting enough "peak" to be a problem generally.

You can greatly reduce the rectifier peaking with some filtering, notably an LC filter down stream of the rectifiers as apposed to the more common RC.

Maybe we better back track.. I'm assuming this is a product you are designing for some typical level of production. Axher is this what you are doing or is this a one-off?

Keith Cress
Flamin Systems, Inc.- <

http://www.flaminsystems.com>

 

[Axher]

Ok your right I'm gona start over. It's for a product that will be controled from a microcontroller with pwm output. The microcontroller will get data from different sensor like temperature, humidity, etc. I need to control 3 dc fan from the microcontroler. each fan is about 60 watt. So I was thinking about a simple dc chopper because I only need one quadrant. I tried the chopper on a lab power supply that I have and it work fine. But I have some trouble to make my own power supply to power the chopper at 48vdc. I was also thinking about doing a controle rectifer for each fan but did'nt get to that because I tough that the riple in the current would be to high i'm right ? thanks for your help

 

[Axher]

Ok I'm sorry about this your correct I did some more test and I only get about 2 amp on the line but the power factor is not that good would this be a problem to certified it in north america ? I would also like to know if there's a better solution to this thanks

 

[itsmoked]

Hi Axher.

I don't believe this would be a problem in N.America.

I look at your task as being very similar to supplies for steppers or servos. Regulated supplies are often avoided because they are controlled by some form of regulator(obviously). Well that regulator looks at the output and then IN TIME makes a change to the control element to correct a shifting output. This time introduces a delay. Delays mean you can get into oscillation if the load is cyclical and periodic. Your load WILL be cyclical and period. Your PWM could feed the fan at the exact correct frequency to cause the regulator to beat with the power being called for. This is even more pronounced with switching supplies as they have even more delays inherent in their feedback control loops.

Because of this, stepper and servo controllers often specifically require unregulated power supplies. Besides it is a case of who cares really what the voltage is to the fans as along as it is "about" 48V. This is because you are feedback controlling the fans so your controller should command whatever PWM duty cycle is required to meet the desired fan speed. If the voltage is low your controller just cracks open the throttle a little more. This allows the prudent designer to build an inexpensive and tough unregulated supply and then build only one control system that controls the actual device of interest(the fan).

Take a look at this unregulated supply. This is VERY typically used in your type of application.

http://www.elpac.com/uploads/documents/datasheets/BFS200.htm

Here's a picture of one that is 10Amps output @ 48V for a little more than $100.

If you were mass producing a product you could just build your own and possibly save some money. Of course 100 piece prices would be far less.

If you think this solution can work weight and size wise you could buy one and get your control going.

Keith Cress
Flamin Systems, Inc.- <

http://www.flaminsystems.com>

 

[Axher]

Ok thanks that's what I will do and I will probably design my own because I already have a 48v transformer and everything else.