# Hi-End low pulsating AC-DC power supply, principal circuit

User Rating: 0 / 5

Published: 08 March 2016
Hits: 757

To start working with any electronic device you need a power supply. For our vacuum tube project we need DC output for 250-300V; our electrical grid is 220V AC 50 Hz (European standard). A lineal laboratory supply is an expensive device. It might be uncomfortable to invest about 1500\$ at the beginning of your hobby project.

Let’s describe some basics principles and difficulties you may have with AC/DC supply.

Fig 1 – the elementary schema of AC-DC supply

D1,D2,D3,D4 – the diode bridge

C1 – the electrolytic capacitor

V1 – the sinusoidal source.

In our case, we must use the secondary winding of the transformer, to provide the galvanic isolation from theelectrical grid. Secondarywinding -is our sin source.

Before we start discussing some features of high voltage DC supply, let us remind some basic information for you.

The capacitor charging schema is presented on fig. 2.

Fig 2 – charging the capacitor

After the time period

(1)

the voltage С1 augmented to 63%, it’s relative to the voltage of the voltage source.

After 3 it increase up to 95%. After 5 the capacitor charging process is complete. The equation (1) is also valid for the unchanging process.

For our purpose, we need a capacitor that will not be uncharged lower than requested pulsation level during 1/50 second.

The resistor R1 ( fig 1) voltage will be aspiringto the amplitude of AC source in case of infinite resistance. The voltage is root mean square (RMS) value on low resistance (high-loaded output).

(2)

Disadvantages of the basic circuit presented on the fig 1 are:

- if we requested high-quality DC on high-load, we need a capacitor with the big capacitance. In the case, we get the long time period on the start of the short-circuit current (We have to consider that fact that, at the power on instant we get the shot circuit current on C1).

- if we use multiply R-C filters the efficiency of our system is decreasing, since we have to transform electrical power to heating (on resistors).

Above mentioned disadvantages are inappropriate for the low consumption of the energy and for the low voltage circuit (for example: 5 V), if we design 250 V DC power supply using the basic schema, we have to use devices with the overlapped current characteristics, it might be too expensive and might have inacceptable dimensions.  Also, a supply with high consumption on start might have some negative effects for electrical grid (warming up, activating a protection system).

We propose to charge our big electrolytic capacitor using the current-limiting resistor R2 fig 3.

Fig 3 – charging C1 capacitor with the artificially limited current

When we estimate 70-80% of the source voltage , we turn on SW1, and we shunting R2.

Calculations

If we expect infinite small resistance at the charge start, we get the peak value: 250 V, so R2=1 KOhm :

250/1000 = 0,25 {A}

To select the proper R2 (as all others resistors at the article) we also have to calculate the power value by the formula:

W=U*I                      (3)

Let’s use our values:

W=250*0,25 = 62,5     {Watt}

Charging process takes the following time (robust):

Tau=1000*0,001 = 1 {Second}

Since our peak current takes a short time (relative for temperature processes into materials of the resistor) and our resistor become cold before the next start of the supply, we propose to use 20W resistor (the resistor is smaller and cheaper).

Our Microcap model of the fig. 4 circuit shows that at the moment of turning on SW1 (80% C1 charge) when C1 voltage 200V our diode bridge gets the maximum permissiblecurrent. To decrease the amplitude of the current burst we use the inductor L1.

Fig 4 – Using the inductor as shunt

It’s recommended to use an inductor with the large cross section of the wire, the active resistance of the wind must be less than 10 Ohm (otherwise, our inductor allows pulsations and might warm up). We use 300mH , but, selected inductance is overall.

It is possible to implement VM1-SW1 system on electro-mechanical elements, but, we decide to use transistors.

Fig 5 – The AC/DC power supply based on the big electrolytic capacitor and the automatic shunting for the current limiter resistor

Let’s concentrate on the "trigger" part of the schema, presented on fig 6.

Fig 6 – the trigger block

This is a well known solution which works into the following principle: at the beginning of the process, R2 and R3 works as one common resistor and occur the voltage divisor with R4. Into the point #2 voltage becomes negative relative to emitter of VT3 (into a charging C1 process). The R4, R3 and R2 values are selected to open VT3 when the voltage C1 become 200V or more.

When VT3 is open the divisor R6 and R5 start working. Increasing R5 voltage opens VT2 and the R2 will be shunted. Divisor at point #2 becomes different and now if we decrease voltage on C1 the trigger block will save the state (the solution) – "open" on the output (point #1). The solution allows to hold shunting state into the definite state, ignoring pulsations from the diode bridge.

Since VT3 opens on the negative voltage we need PNP transistor. The transistor VT2 opens on positive voltage, so we use NPN type. In our case, we use high-voltage 2N6520 and 2N3439. For the energy efficiency it’s better to use R3- R4 100 KOhm and R4 600 Ohm.

Fig. 7 presents the modeling results.

Fig 7- R2,R4, R5 and C1 voltage at the first 6 seconds

Fig 8 – Current and volrage of C1 at the moment of shunting (at triggering)

Fig 9 – Modeling C1 current with the L1 inductor

We use the 300 mH inductor, this is an overall inductance. It is possible to calculate a smaller inductor, using the formula:

(4)

where

– voltage (V)

– delta time (s)

– current delta (A)

L – inductance(H).

For the 300 mH inductor, peak current pulse is lower 0,5A, which compiles to the nominal mode for the Russian KC405B.

P.S. I’ll be so gratify to get some feedback about the article. Thank you!