TL494 PWM Shunt Regulator Project

[Opera House]

I was at a garage sale last weekend and picked up a box with three computer power supplies for $3.   Thought this might be a good opportunity to build a demonstrator PWM shunt using a TL494 switching regulator chip.  This chip has been in production for more than 20 years and can be found in many power supplies.   A cutoff grinding wheel was used to cut out a section of the circuit board containing the TL494, this makes a convenient breadboard.  Excess components are removed and an Exacto knife used to cut excess traces.   These are all single sided boards which makes the traces easy to figure out.  There are very few components to add, mostly jumpers between pins.   As shown, the circuit is designed for a 12V solar or wind generator.  In standby mode the current draw is 8mA.   Maximum voltage for the chip is 42 volts, but the FET gate limit is 20V and it would be nice to protect the gate with a zener. As shown it does not have one and with internal voltage drops this will tolerate an input up to 23V.  For 24V operation use a 7815 three terminal regulator to supply the chip.  Internal load resistor is limited to 3A and an external 5A load can be connected.  The pot adjusts from 13.1V to 15.5V.   Modification can be made to turn this into a motor speed control, light dimmer, battery desulfinator, voltage booster, or power point regulator.

[Opera House]

The TL494 has two separate op amp inputs and the outputs are ORed together.   The second one is wired so it has no effect.   Tthe internal 5V reference is connected to  the + input, pin1.  Pin 2 looks at the input voltage through a voltage divider and compares that with the reference on pin1.  A 0.1uf cap provides a little noise filtering, the value is not critical and can be up to 1uF.  The output of the op amp, pin 3,  is filtered with a 1uF cap. Just like an op amp, a  resistor can be added between pins 2 & 3 to reduce the gain if you wanted to start shunting at one voltage and slowly start increasing the duty cycle over the next volt increase. Other applications for a feedback resistor would be a motor speed control.  

R and C on pins 5 & 6 set the frequency.   Normally these supplies operate at 30-40 KHz.  C was changed to 0.1uf to reduce the switching frequency to about 500 Hz.  This eliminates a lot of noise problems  associated with wiring and driving the FET.

[Opera House]

The internal load resistors are four 5 ohm 10 watt resistors in a series/parallel configuration for an effective resistance of 5 ohms at 40 watts.  This represents a 3A load at 15V.  The resistors are mounted just in front of the vents in the rear of the box.   This fan only has mounting holes on one side of the frame.  It would have been nice to turn it around and have the air blow into the box and mount the FET on this insulated plastic frame.    The male IEC power connector is used to bring power into the box and provides a convenient disconnect.  Don't use this in situations where someone might connect 120V to it.  The female IEC power connector is connected across the load resistors so an external load like a lamp can be connected to increase the current.  The FET is a BUZ71 which is rated at 50V @ 14A,  0.1 ohm.  I would limit external current to an additional 5A with a single FET.   The FET is isolated from the IC driver by a 120 ohm resistor.  With this, the FET can short out without destroying the IC.  Additional FET gates can be driven in parallel using a separate resistor for each gate.  Keeping the loads separate on each FET is preferred.  Switching frequency is very low so there is no concern about turn on time heating of the FET.  Output transistors of the regulator chip are ORed together and are capable of sinking/sourcing over a quarter amp.  With a current limiting resistor these can drive transistors instead of FETs.

 The fan is driven by the 12-15V voltage pulse across the load resistor   A 500uF capacitor is used to supply fan voltage and charged through the 1N4002 diode with a 10 ohm resistor. The resistor limits inrush current and drops the voltage down a couple of volts to match the fan voltage.   Computer fans are 12V brushless DC motors that draw about 100mA.  The motor starts automatically when the shunt load current reaches about a quarter amp.  This same circuit design for the fan motor could also be used to turn on an external relay for load diversion.

[Opera House]

A picture got lost.....

[oztules]

Nice job.

It never ceases to amaze me how much useful technology you can get from a computer psu. For a dollar or so, you get a cooling system, a tl494 with the drivers attatched if you want them, driver transformer with optional pos feedback winding for self start, lots of bits and pieces, inc high voltage caps and mains filter caps and line filters and varistors, transistors,..... and you get the project box as well.... now thats value!

..........oztules

[willib]

Yes nicely done !!

nice equipment too.

what model scope is that?

ps your bench is too clean !

[stephent]

Nice project Opera House.

Simple and easy to make.

It does have several possibilities for modification and other uses as you mentioned, again probably easily implemented.

I ain't going to throw away another old computer PSU.

Those things are a treasure trove of usable parts n pieces...even the housing, fans..

Thanks for taking the time to type this--draw the schematic, take the pictures, etc.

But you might want to take another look at those power resistors in series/parallel combination and end wattage, but it could be just me and "math", I didn't pay much attention way back when I should have.

[dinges]

Shame on you,

for making other people aware of the nice things that are inside a PC PSU, or the things you can do with it.

If everybody starts hacking their PC power supply, soon there will be none left that gets thrown away anymore. Surely you don't want that to happen?

"Modification can be made to turn this into a motor speed control, light dimmer, battery desulfinator, voltage booster, or power point regulator."

Stop it! You're doing it again.

[Opera House]

Math.... 5 = 5 = 10   10/2 = 5  15V/5 = 3A  4A x 15V = 45W  x 95% duty cycle max = 43W.  40W worth of resistors, some points for a fan, amd a good manufacturers safety margin...close enough.

The scope is an older analog HITACHI V-1100 100MHz scope.  While not digital it does have on screen frequency counter and cursors you can move around to get voltage and period.  When I got rid of my slide rule, I lost the ability to hold a decimal place in my head.  That period calculation is nice, counters are so arbitrary in what they read.   That was the only clean spot on the bench, everything got swept to the floor.

The best deal out there are these battery UPS for the computer.  I'm making a power point regulator this week out of one.  Running out of time before I go to the camp  for to weeks.

[kitno455]

right- i knew eventually i would be able to justify all these dead psu's i have kept. around 2 dozen of them at this point, gives me a few to experiment with i've got to go tell the wife i was right for a change

allan

[(unknown)]

Does anyone knows why 500uF capacitor and 10 ohm resistor that are powering up the TL494 cip are heating during PWM? I am actually using a IRFZ44 mosfet that I checked and should't be a difference to BUZ71.

[(unknown)]

A blocking diode is necessary to be inserted in the above picture, between 500u capacitor and dump load. Otherwise the dump load will suck the current back from the capacitor and so it will heat up together with 10ohm resistor that is connected after.