Automotive 12v to 20v converter Audio Amplifier

The impediment of car supply voltage 12v powers to change over the voltages to higher keeping in mind the end goal to power audio amplifiers. 

Truth be told the max audio power x speaker (with 4 ohm impedance) utilizing 12v is (Vsupply+ - Vsupply-)^2/(8*impedance) 12^2/32 = 4.5watts every channel, that is bizarre...

For driving effectively an intensifier the best is to utilize a symmetric supply with a high voltage differential. for instance +20 - -20 = 40v.

indeed

40^2/32 = 50 Watts every channel that is respectable.

This supply is proposed for two channels with 50w max each (obviously it relies on upon the amplifier utilized). Despite the fact that it can be effectively scaled up or the voltages changed to get distinctive qualities.

It is a fantastic push-pull outline , taking consideration to acquire best symmetry (to keep away from flux walking). Remember that this circuit will adsorb numerous amperes around 10A so fare thee well to fortify power tracks with solder and utilize substantial or heavy wires from the battery or the voltage will drop excessively at the input.

The transformer must be intended to diminish skin effect, it could be possible utilizing a few insulated magnet wire single wires soldered together yet leading independently. The regulation is carried out both by the transformer turn proportion and shifting the duty cycle. For my situation i utilized 5+5 , 10+10 turns acquiring a venture up proportion of 2 (12->24) and downregulating the voltage to 20 through duty cycle element alter performed by the Pulse Width Modulator controller TL494.

The step up ratio must be somewhat higher to overcome winding resistance, diode losses and so on  and input voltage drop because of wire resistance (battery to converter).

The transformer must be of right size so as to convey the power required, on the net there are numerous diagrams demonstrating the power in capacity of frequency and core size for a given topology. My transformer size is 33.5 mm length, 30.0 stature and 13mm width with a cross area zone of 1,25cm^2, useful for power around 150w at 50khz.

The windings , particularly the essential must be substantial gaged, yet as opposed to utilizing a solitary wire it is ideal to utilize

various wires in parallel each one protected from the other with the exception of at the finishes. This will decrease safety increment because of skin impact. The essential and optional windings are centertapped, this implies that you need to wind 5 turns, centertap and 5 windings once more. The same strives for the auxiliary, 10 turns, centertap and 10 turns once more.

The critical thing is that the transformer must not have air holes or the spillage inductance will toss spikes on the switches overheating them and giving a voltage higher than anticipated by turn proportion expectation, so if your voltage yield (at complete duty cycle) is higher than Vin*n2/N1 - Vdrop diode, your transformer has gap (obviously allow me saying you that you are blind in the event that you miss it), and this is went hand in hand with a drastical productivity decrease. Utilize non-gapped E cores or toroids (ferrite).

Yield diodes, capacitors and channel inductor

For rectification i wanted to utilize shottky diodes because they have low forward voltage drop, and are staggeringly quick.

I utilized the shoddy 1n5822, the best option for low voltage converters (3a for current ability).

The yield capacitors are 4700uf 25v, not enormous, subsequent to at high frequency, the voltage ripple is most because of inward top ESR luckily broadly useful lytics have enough low esr for a little ripple (a few several millivolts). Additionally at high duty cycle they are nourish very nearly with immaculate DC, giving little ripple. The channel inductor on the auxiliary centertap furter expands the ripple and helps the regulation in deviated transients.

I utilized d2pak 70v 80a 0.004 ohms ultra fets (Fairchild semiconductor), extremely costly and elusive. On a fundamental level any fet will work, yet the bring down the on-safety, the bring down the on-state conduction losses, the bring down the hotness created on the fets, the higher productivity and littler the heatsinks required. With this fets i am ready to run the fets with little heatsinks and without fan at full appraised force (100w) with a productivity of 82% and distinguishable warming and with little warming at 120w (a few degrees) (the center begins to soak and the effectiveness is a bit lower, it's around 75%)

Attempt to utilize the most minimal resistance mosfet you can put your filthy hand :-) on or the productivity will be lower than appraised and you will need even a little fan. The fet driver i utilized is the Tps2811p, from TI, evaluated for 2a crest and 200ns. Is essential that the gate drive is streamlined for insignificant inductance or the exchanging losses will be higher and you hazard clamor coupling from different sources. Actually i feel that bent pair wires (gate, ground/source) are the best to keep the inductance little. Place the gate drive resistor close to the Mosfet, not close to the IC.

I utilized the trusty Tl494 PWM controller with freq set at around 40-60 Khz customizable with a potentiometer. I likewise actualized the delicate begin (to diminish powerup drifters). The modify potentiometer (feedback path) must be set to acquire the desired voltage. The yield signals is planned with two pull-up resistors on the collector of the Pulse Width Modulator chip yield transistor pulling them to ground each one cycle then again. This signal is sent to the double reversing MOSFET driver (Tps2811p) for getting the right waveform.

As i said before the power tracks must be substantial gaged or you have to scarify regulation ( it depends on stepup ratio of transformer & input voltage) and efficiency as well. Bear in mind to place a 10a (or 15a) fuse on the input because the auto batteries can supply high currents and flows if there should be an occurrence of shorts and this will spare you confront from a mosfet blast in the event of failture or short, keep in mind to place a fuse likewise on the battery side to expand the security (unplanned shorts->fire, battery blast, fire fighters, police and legal advisors around). Input filtering is essential, use no less than 20000uf 16v in capacitors, a channel inductor would be helpful as well (heavy gauged) yet i chose to abandon it..

This supply surrendered me to 85% productivity (at times even 90% at a few burdens) with an info of 12v on the grounds that i watched all these traps to keep it utilitarian and effective. An o-scope would be valuable, to watch the ripple and gateway signals (looking for overshoots), however in the event that you take after these rules you will dodge these issues.

The cross regulation is great however remember that just the positive output is completely directed, and the negative just tails it. Place a little load between the negative rail and ground (a 3mm light emitting diode with a 4.7k ohm resistor) to maintain a strategic distance from the negative rail getting lower then -20v. On the off chance that the heap is topsy-turvy you can have two cases:

1.More load on positive rail  :  no issues, the negative rail can be lesser than -20v, however it is not a main problem for an audio amplifier.

2.More load on negative rail: voltage drop on negative rail (to ground) particularly if the heap is just on the negative rail.

Luckily audio amplifiers are very symmetrical as a heap, and the output channel inductor/capacitors serves to keep up the regulation great amid awry transients (Basses)

Remember that THIS PROJECT IS NOT FOR BEGINNERS , IT IS EXTREMELY DANGEROUS IN CASE OF PROBLEMS.

For first testing use a small 12v force supply and utilization resistors as load observing switches hotness and current utilization (and yield) and attempt to focus efficiency, on the off chance that it is higher then 70-75% you are situated, it is sufficient. Conform the frequency for best trade off between switching and power losses, hysteresis and skin effect.

 R1,r2 = 10                         2

 R3 R4 R6 R7 = 1k               4

R5 = 22k                             1

R8 = 4.7k                            1

R9 = 100k                           1

                                      

C1 C2 = 10000uf            2

C3 C6 = 47u                  2

C4 = 10u                       1

C5 C7 C14 = 100n          3

C8 C9 = 4700u               2

C12 = 1n                       1

C13 = 2.2u                    1

 

U1 = Tl494                         1

U2 = Tps2811p                   1

 Q1 Q2 = Fdb045an            2

 D1-D4 = 1n5822                4

 D5 = 1n4148                     1

FU1 = 10a                          1

L1 = 10u                            1

L2 = FERRITE BEAD            1

Rv1 = 2.2k                         1

Rv2 = 24k                          1

T1 = TRAN-3p3s                 1