Page 1 of 1

How to make Class A 10 watts Power Amplifier?

Posted: Thu Jan 18, 2018 3:51 pm
by Dohangout
In making this work, I thought that I wanted to make the transistor 2N 3055.
It is a representative power transistor that is everywhere.
It is also a nice thing that it became possible to purchase MJ 2955 in Malts recently.

Since I feel that hFE is insufficient on the catalog spec, I will use 2SA1015 / 2SC1815 for Darlington.
power-sche.png (17.33 KiB) Viewed 1875 times
mdd2955-sche.png (19.72 KiB) Viewed 1875 times
Idling current
Set it so that the collector current does not become zero even at the maximum output or it can cut off a little at the maximum output.
The idling current Iq is roughly
Iq (A) = R11 / (R11 + R12) × Vf × (R2 / R11) / R4 (Vf is the forward voltage of the diode).
If you connect speakers of 6 to 8 Ω with ± 15 V power supply, 1A ~ 1.2 A will be the proper line. I am designing the above circuit diagram as well. If it is made too shallow without countermeasures, not only crossover distortion increases, but it may destroy it exceeding the emitter-base breakdown voltage (VEBO) of the transistor.
To ensure stability
Insert an isolator for when the capacitive load is hanging at the output.
A coil with 0.8 mm UEW wire tightly wound tightly around a cylinder 1 cm in diameter and hardened with an adhesive or a high frequency varnish can be formed.
For the time being, Zobel network is also included, but this may not be necessary.

A 22 pF capacitor is connected to the operational amplifier as phase compensation.
As a comment made actually, I feel that even a little smaller it does not matter.
This place is effective in general operational amplifiers, but it seems that it may be better to add it in parallel with the feedback resistor instead of the current location in the case of a high speed operational amplifier that applies special internal compensation.

Bill of Material (Amplifier Board)
number Item name article
C1 0.047 uF the film
C3 0.1 uF the film
C4 100 uF Recommended for general aluminum electrolytic acoustics
C5 0.1 uF Multilayer metallized film MMT, MTF, ECQV, etc.
C6 0.1 uF Multilayer metallized film MMT, MTF, ECQV, etc.
C7 100 uF Recommended for general aluminum electrolytic acoustics
C8 3.3 uF Recommended nonpolar electrolysis for sound or MMT, MTF etc.
C9 3.3 uF Same as above, although it does not need to be installed to operate
C10 470 uF Recommended for general aluminum electrolytic acoustics
C11 470 uF Recommended for general aluminum electrolytic acoustics
C12 22 pF Low dielectric constant ceramic
C13 22 pF Low dielectric constant ceramic
C14 0.1 uF the film
D1 1S1588 For small signal switching, 1N4148 etc.
D2 1S1588 For small signal switching, 1N4148 etc.
IC1 OPA 2134 PA JFET input 2 circuits OPA2604, AD823, LT1169 etc.
L1 1 u H Hand winding air core
Q1 2 SC1815-Y
Q2 2SA1015-Y
Q3 2N 3055 Replaceable with appropriate complementary Tr
Q4 MJ 2955 Replaceable with appropriate complementary Tr
R2 10 kΩ Recommended for audio 1/4 ~ 1 / 8W
R3 100 Ω Recommended for audio 1/4 ~ 1 / 8W
R4 0.22 Ω (3 W) High-precision product recommendation (5% or less)
R5 10 Ω (3 W) Acid gold, winding etc.
R6 10 Ω (3 W) Acid gold, winding etc.
R7 0.22 Ω (3 W) High-precision product recommendation (5% or less)
R8 100 Ω Recommended for audio 1/4 ~ 1 / 8W
R9 10 kΩ Recommended for audio 1/4 ~ 1 / 8W
R11 510 Ω Recommended for audio 1/4 ~ 1 / 8W
R12 27 kΩ Recommended for audio 1/4 ~ 1 / 8W
R13 510 Ω Recommended for audio 1/4 ~ 1 / 8W
R14 27 kΩ Recommended for audio 1/4 ~ 1 / 8W
R17 3.3 kΩ Recommended for audio 1/4 ~ 1 / 8W
R18 3.3 kΩ Recommended for audio 1/4 ~ 1 / 8W
R19 22 kΩ Recommended for audio 1/4 ~ 1 / 8W
R21 100 Ω Recommended for audio 1/4 ~ 1 / 8W

The operational amplifier can be used with JFET type 2 circuits. For the time being, we are assuming that if the varieties with ordinary characteristics are used for slightly better audio such as OPA 2134, they will move roughly. However, careful attention should be paid to negative feedback stability if it aims at broadband.
The resistance of the analog section seems to be about if you choose by your preference. However, if you insert R4, R7 with a large error, it will come out as output offset voltage.
The idling current is determined by R12 and R14. By decreasing by 3 kΩ step, the final bias can be adjusted.
Since the power transistor generates heat, consideration such as setting an electrolytic capacitor to 105 ° C in case of a case where heat is trapped may be necessary.

In actual equipment, the input coupling capacitors (C8, C9) are not attached and they are short-circuited. It is OK if the input device can be trusted, but since it is a circuit that amplifies the DC offset of the input side by 20 times and outputs it, we recommend that you wear it properly and seriously.

Power supply

(5/12 rectifier diode for power supply for op amp, withstand voltage is not enough with 2 GWJ 42 of 40 V. It has been changed to 11 EQS 06 of 60 V withstand voltage)

24 V 2 A rectification smoothing with a midpoint of a transformer as a midpoint, ± 15 V Give power independently and give it independently.
I wanted to apply a slightly higher voltage to the operational amplifier, so we prepared a ± 17 V power supply stabilized in a separate system.
In the real machine, I use two of Sugano's SP-242 and one HTW-1802 of Toyofu.
I think that it is good to increase the voltage of each with a large heat sink and a high voltage operational amplifier such as OPA2604 and aim at high output.

Heat dissipation

Because it is class A operation, it needs a large heat sink.
The collector loss of the transistor is maximized when there is no sound, which is the product of the idling current and the supply voltage.
Shall I make 18 W at 1.2 A × 15 V per transistor?
For properly designed heat sinks, the thermal resistance can be estimated from the envelope volume (vertical × horizontal × height).

One that is exhibited as "heat sink (heat sink) Tr · SCR attached 200 mm × 115 mm × 28 mm", one for each channel. Two TO - 3 transistors and two TO - 66 transistors are attached, but I will remove them. 2SD 319 can be used as equivalent to 2N 3055, but I feel a little sorry, but there is no PNP transistor to be complemented, so it can not be helped.

Since it is 200 mm × 115 mm × 28 mm, it reads as approximately 1.5 ° C / W from the graph.
18 (W) × 2 (pieces) × 1.5 (° C / W) = 54 (° C)
It can be seen that it rises 54 ° C from room temperature.
If room temperature is 40 ° C, is the heat sink temperature 94 ° C?

Calculate the junction temperature of the transistor.
Temperature resistance of junction to case is 0.7 ℃ / W from data sheet.
Shall be
1.7 ° C / W as 1 ° C / W between case ~ (insulating plate and silicon grease) ~ heat sink ?
18 (W) × 1.7 (° C / W) = 31 ° C for each one
already Since the heat sink has already reached 94 ° C, 125 ° C will be the junction temperature by adding that amount.
In the meantime, the upper limit of the junction temperature is known to be 200 ℃ from the 2N3055 datasheet, so it is clear.
Let's check if it is within the allowable range of the power reduction curve on the data sheet.
When the temperature is high, it can not withstand the large electric power, so please discount it, it's a graph, but there is no problem.
The ASO curve is also inside and it is affordable.

A little test

First check the DC offset voltage.
Left is 11 mV, right is 3 mV. There should not be any problem.
Afterwards check the idling current. Measure both ends of 0.22 Ω and find it.
It changed to 27 kΩ → 24.9 kΩ, 510 Ω → 499 Ω, but it became 1.13 A.

When a sine wave was input by connecting a 6.7 Ω dummy load, visible clipping appeared at about 9.2 Vrms.
Is it about output 12W? feel well.

Try entering a 10 kHz square wave.
mdd-sqw1.jpg (9.73 KiB) Viewed 1875 times
mdd-sqw2.jpg (10.04 KiB) Viewed 1875 times
mdd-sqw3.jpg (10.45 KiB) Viewed 1875 times
(Top: 6.7 Ω, Medium: 6.7 Ω + 0.15 uF, Bottom: 0.15 uF only)
It is quiet. I do not need to worry.

Try to ring
No pop sound comes out when the power is turned on / off. It is a good thing.
It is a straightforward sound that you can listen calmly. It seems unlikely that the band is clinging or clouding somewhere, and since it does not feel any strange protrusion, it is connected naturally from top to bottom.
The bass is relaxed and has a dense density. It is different from the hard bass line that I have been pointing with my hobby so far, but I thought this is ant for this.
It is not an analytic direction. For the sound image localization that I think is personally important, do you call that the field is ringing and not a feeling of ringing at a point? Sound separation is okay.
Listening to a CD that is said to be good is a lot of fun, but if it is not, it will get something. It seems to be moderate and may be unpalatiously difficult.
Since there is not enough extrusion when trying to make violent sounds, there is a song that is not good. Would it be better to make the power supply a bit more robust.

If you change the op amp from OPA 2134 to OPA 2604, it turns moist and gentle, and if you like this kind of sound, I strongly recommend you to recommend it. It sounds good even in low volume, so it is suitable for regular use without giving too much volume.

I was dissatisfied with the speaker in use for a moment, so I'd like to make a final evaluation a bit earlier.
mdd-top.jpg (17.35 KiB) Viewed 1875 times
mdd-power-photo.jpg (21.52 KiB) Viewed 1875 times