# Variable DC Linear Power Supplies

#### Michael Hlavaty

Joined Aug 16, 2015
13
I have a general question regarding DC variable linear power supplies. Why do most designs employ a multi tapped transformer? Why must the input transformer be changed when there is high output current? And how does it affect a pass transistor with no multi-tap transformer?

#### ronv

Joined Nov 12, 2008
3,770
I have a general question regarding DC variable linear power supplies. Why do most designs employ a multi tapped transformer? Why must the input transformer be changed when there is high output current? And how does it affect a pass transistor with no multi-tap transformer?
I think it is mostly about power disapation in the pass transistor at low voltage high current. Let's say you have a 5 to 25 volt supply at 3 amps. The input to the pass transistor is likely around 30 volts. So when set to 25 volts @ 3 amps the power in the transistor is about 15 watts, but if the voltage is set to 5 volts @ 3 amps it is 60 watts.

#12

#### #12

Joined Nov 30, 2010
18,210
Why do most designs employ a multi tapped transformer?
Fifty years of experience says I disagree with that idea. Most power supplies don't contain that much quality. ronv did a good job of describing the over-heating concerns, and that is the primary reason to use a multi-tapped transformer. I have even seen a power supply which used a Variac with a potentiometer on the backside of it to keep the bulk supply voltage just barely high enough to run the power transistors correctly.

Another scenario is about trying to be, "universal" with the primary windings of the transformer. The power grid in my neighborhood has changed over the years from using 110/220 VAC as the, "standard" through several steps to arrive at 120/240 VAC +/- 5% today. Then there is the option of trying to be ''universal'' by including European voltages and frequencies.

At my house, the voltage is almost at the upper limit. I have 125/250 VAC at the breaker box.

Clear as mud?

Last edited:

#### Michael Hlavaty

Joined Aug 16, 2015
13
Ronv, this is what I don't understand. Isn't higher power input to a pass transistor create a bigger power dissipation? How does lowering the input voltage to the pass transistor lower the power dissipation? Pwr dissipation = v-input × output current

AAC Fanatic!,
Thanks for your response. What circuitry methods have you seen in you experience on reducing power dissipation across pass transistors at high loads?

#### #12

Joined Nov 30, 2010
18,210
As a matter of formality, my moniker resembles, "Number Twelve".

Next, we need to define words. One does not supply, "power" to the pass transistor, one supplies voltage, hoping to have enough voltage for the pass transistor to do its job. The power dissipated (wasted) by the pass transistor is a matter of I x E (current through the transistor times voltage across the transistor). If you design a variable output LINEAR supply with only one possible bulk voltage, the voltage measured across the pass transistor (from collector to emitter) increases as you adjust the output to lower voltages than the maximum allowed. I think this is your point of contention.

In Ron's example, 30 volts available on the main capacitors is connected to the collector of a big transistor. If you adjust the output voltage to 25 volts, there will be 5 volts across the transistor. That 5 volts times the current flowing is the power dissipated by the transistor. Now, adjust the output to 5 volts and there are 25 volts across the transistor. Twenty-five volts times the current flowing is now the power dissipated by the transistor.

The multi-tapped power transformer is used to decrease the main voltage available to the transistor when you only need a low voltage as your output. You merely switch which transformer wire is connected to the rectifiers so the bulk voltage at the filter capacitors is less. Less excess voltage = less wasted power.

I already mentioned the Variac approach in which the Variac is used as a transformer with an infinite number of voltage taps. The potentiometer on the axis of the Variac is used to tell the power supply brain how much voltage you want for its output. The behavior of the two voltages is carefully calculated so the bulk voltage is just barely more than the output voltage. Thus the excess available voltage for the transistor is minimized. You want 5 volts for the output? The Variac is expected to provide 10 volts at the main capacitors. You want 10 volts for your output, the Variac is expected to supply 15 volts at the main capacitors.

That's about all I have. Many other people on this site might describe a method I overlooked or forgot.

#### Brownout

Joined Jan 10, 2012
2,390
How does lowering the input voltage to the pass transistor lower the power dissipation? Pwr dissipation = v-input × output current
Looks like you answered your own question. Wouldn't the product be lower if v-input were lower?

Actually, the equation you're looking for is (v-input - v-output) x output current. Still, lower v-input means the product is lower.

#### #12

Joined Nov 30, 2010
18,210
I just remembered another one. If there is a known, continuous, minimum load current, you can place a resistor from the filter capacitors to the output and use that to carry the known quantity. The transistor only has to provide and regulate the variable amount of current which is above and beyond the capability of the resistor.

#### AnalogKid

Joined Aug 1, 2013
8,534
Many low-cost bench supplies have a front panel range switch, something like 0-15 and 0-30. On some it is a voltage range only, and the max available current is the same for both ranges. This indicates that the switch is SPDT, switching between two secondary taps into a common regulator section. but on some supplies the lower voltage range has half the voltage but twice the max current. This indicates a DPDT switch connecting two independent secondary windings either in series or parallel. I've built more than a few of those.

#12, when did the U.S. power grid change from 110 VAC to 240 VAC? (post #3) I assume from the 110 V reference that you are describing wall outlet voltage rather than the actual distribution grid. Did I miss a memo?

ak

#### #12

Joined Nov 30, 2010
18,210
#12, when did the U.S. power grid change from 110 VAC to 240 VAC? (post #3) I assume from the 110 V reference that you are describing wall outlet voltage rather than the actual distribution grid. Did I miss a memo?
The delivered voltage at the breakers was 110 VAC in 1970, in this county. Now it's 125 VAC. Had to redesign my clothes dryer with Calrod so it wouldn't melt the nichrome heating coil. Had to add buck transformers to my outdoor lights to get the \$35 bulbs to last more than 3 years.

I had many experiences (during my job) measuring the power voltage when investigating window air conditioners that wouldn't start.

Best part of this? Being able to communicate with TS in terms he seems familiar with, and nobody corrects us.

Last edited:

#### Michael Hlavaty

Joined Aug 16, 2015
13
DUHHH voltage across pass transistor is VCE or Vout-Vin I feel retarded. So lowering the voltage lowers the total heat dissipation. Thanks for all the replies.

#### crutschow

Joined Mar 14, 2008
25,251
I just remembered another one. If there is a known, continuous, minimum load current, you can place a resistor from the filter capacitors to the output and use that to carry the known quantity. The transistor only has to provide and regulate the variable amount of current which is above and beyond the capability of the resistor.
Here's another way to use a resistor.
If the raw DC voltage from the rectifier is significantly more then the maximum output voltage than you can also use a power resistor in series with the pass element to reduce its dissipation.
You select the value of the resistor to give the minimum required voltage drop across the pass element for the minimum raw DC voltage and the maximum output voltage and current desired.

#12

#### Michael Hlavaty 1

Joined Aug 17, 2015
7
Thank you for all the replies, so if I want to build my own linear power supply from 0-30V, 0-5A.

I have a few reference designs:

1) The primary winding has a central tap with a relay that moves from low voltage(high current) to high voltage(low current) settings. It is labeled as S2A range but I don't see how the relay is switched from the schematic. Other things I don't understand is how U4A op-amp works and the reasoning of the BJT's connected to the MOS pass transistors.

#### jssamp

Joined Jun 19, 2014
3
Best part of this? Being able to communicate with TS in terms he seems familiar with, and nobody corrects us.
Lucky. In my experience with calling TS (mostly PC OEMs, cable & ISPs) I usually get the impression that they are unskilled minimum wage employees who, while they can read the steps on some flowchart, display a surprising lack of knowledge given their occupation.

#### #12

Joined Nov 30, 2010
18,210
In my experience with calling TS (mostly PC OEMs, cable & ISPs)
Wrong definition. This Website calls the Thread Starter the TS.

It used to be, "Original Poster" but then the site changed.
Meh.

#### KeepItSimpleStupid

Joined Mar 4, 2014
4,153
FWIW, I have two supplies at home that incorporate the tap changing.

One is a very nice universal linear battery eliminator. As you select the voltage, you select the tap. The wall warts caused way too much hum for an AM radio power supply.

Then I have a "Selectable 0-30 V" at 10 A power supply which was from a company that HP gobbled up.
"selectable" means you had a min control a max control and a front panel screwdriver adjust control AND you had to manually change the taps depending on the range.

Where I worked, we had two 100 VAC Japanese Electron Microscopes One had a Variac was set to 100 V and the knob was removed. The other had a transformer. I maintained the old one. It had lots of 741 OP amps in it.

#### #12

Joined Nov 30, 2010
18,210
#12, when did the U.S. power grid change from 110 VAC to 240 VAC? (post #3) I assume from the 110 V reference that you are describing wall outlet voltage rather than the actual distribution grid. Did I miss a memo?
"In the United States[8] and Canada,[9] national standards specify that the nominal voltage at the source should be 120 V and allow a range of 114 V to 126 V (RMS) (−5% to +5%). Historically 110 V, 115 V and 117 V have been used at different times and places in North America. Mains power is sometimes spoken of as 110 V; however, 120 V is the nominal voltage."

https://en.wikipedia.org/wiki/Mains_electricity

#### GS3

Joined Sep 21, 2007
408
In Europe it used to be 220 Vac in the continent and 240 in the UK. Some time ago they decided to unify it to 230 Vac but everybody still refers to 220 and 240, whatever they were used to.

I have a Chinese-bought bench power supply, 0-30 V, 5 A, which automatically switches some relays to select transformer windings depending on the output voltage. Very occasionally, at the edge of a boundary, the relay will chatter a bit but mostly it works fine. Nice adjustable power supply which I got for very low price.