Electronics > Beginners

My stereo amplifier has some heat issues.

(1/3) > >>

WyverntekGameRepairs:
I have created and prefboarded a class-A amplifier that uses a Darlington pair consisting of a MJE3055T connected to a D1453 Horizontal Deflection Output transistor. I have two of these pairs, one for Left audio signal and one for Right audio signal, thus creating a stereo amplifier. I have chosen to use a 19K dual potentiometer to control the gain (or volume) via the bias on the 3055Ts. (It has 2 potentiometers in one package basically, and they remain isolated from each other. This allows the use of a single potentiometer dial without accidentally joining the signals from both sides).

The power supply I am using is 12V (it reads about 13-14 volts though, as these inaccurate linear power supplies tend to do) at around 1.5A. I'm dissipating a lot of the energy through a huge resistor capable of handling ~200 Watts of power. It has a value of 20 Ohms, and is placed at the positive voltage input of the amplifier circuit. I can understand this sucker getting quite hot despite dissipating around only 8.45 Watts (assuming the power supply is delivering 13 Volts). Now, here is where it gets tricky. I also have a 1KW resistor with a value of 2.2 Ohms. This beast is bigger than my forearm, and is most definitely overkill for the application I'm currently using it in.
Since we get the power dissipation of a resistor by calculating the current across it and then using that to calculate the power deing dissipated across it, let's do some calculations here.

Assuming our power supply is delivering 13 Volts steady, let's calculate the current across the resistor. I = V / R, so 13 / 2.2, which gives us 5.9091A. Now we use this to calculate the power dissipated by the resistor. P = I * V, so 5.9091 * 13, which gives us 75.8182 Watts. Holy mother of God. It draws so much current that the power supply drops down to 6V to compensate for the current draw (because the transformer is used to delivering about 1.5A, not 5.9-6A). That's pretty intense.

With this huge resistor in place, it has a few benefits and drawbacks. The benefits are that the lower supply voltage allows less distortion, as well as posing less of a risk to the signal capacitors that provide isolation between the signal inputs and the power circuitry. The drawbacks are that the resistor is sucking up a lot of current and wasting quite a bit of energy in the form of heat. If you leave the resistor on long enough, you can actually feel it getting about 7 or 8 degrees warmer than ambient. It also is bulky, and very hard to put in a discrete place (obviously, as this resistor is supposed to be big, mounted on a panel, and not-so-discrete in the first place). However, the biggest drawback that is very obvious here is quite naughty: Heat. Not in the resistor itself, but in the transistors. The transistors have less resistance preventing them from entering thermal runaway or getting dubiously hot. This causes the transistors to get up to around 10 degrees above ambient. I have had to mount them on some pretty good sized heatsinks to help dissipate that heat.

Using the smaller but higher resistance resistor (which dissipates ~8.5 Watts) allows higher voltage and thus more distortion, but cooler transistor operation (the transistors only get around 2 or 3 degrees above ambient when using that resistor). However, using the 1KW resistor with lower resistance allows less distortion due to more current draw reducing voltage for compensation... But also causes more heat to be dissipated by the transistors.

What would you do in this situation? I'm curious to see if there is a sort of "sweet spot" I can achieve by putting resistors in series or parallel to fine-tune the balance between heat dissipation and distortion.

NOTICE: This project is meant to be incorrect. Things are supposed to go wrong. I'm practicing troubleshooting methods, as well as my soldering, by building this small little project. This is for amusement purposes only, and experimentation. This is not meant to be a real project. Do not post comments saying I built it the wrong way, I already know it is not the right way to build an amplifier. As I said, this is intentionally designed to mess up.

madires:
A circuit diagram would be helpful.

mikerj:
Class A amplifiers are horribly inefficient, and using a resistor as a collector load is the worst thing you can do in terms of efficiency.  Switching to an active load would give some improvement, but if you want to stick to a linear amplifier and efficiency is a concern then consider a Class A/B design.

Gyro:
You need to check the SOA of a horizontal deflection transistor very carefully before attempting to use it in linear mode. Also, if you're overloading the transformer that badly then it is going to die, imminently!

See here: http://sound-au.com/tcaas/index-1.htm to see how to do it properly - you're clearly not at the moment.

WyverntekGameRepairs:

--- Quote from: Gyro on December 19, 2019, 06:57:17 pm ---You need to check the SOA of a horizontal deflection transistor very carefully before attempting to use it in linear mode. Also, if you're overloading the transformer that badly then it is going to die, imminently!

See here: http://sound-au.com/tcaas/index-1.htm to see how to do it properly - you're clearly not at the moment.

--- End quote ---
Not to worry, I'm replacing that adapter with a transformer that was meant for audio amplifier applications. It's a very interesting one, using a toroidal core. Anyway, it has high-current outputs that are capable of satisfying the power-hungry resistor.
Also, there's no such thing as "doing it right" when you are messing around experimentally and trying unconventional methods out. You can't learn until you screw something up!


--- Quote from: mikerj on December 19, 2019, 05:47:54 pm ---Class A amplifiers are horribly inefficient, and using a resistor as a collector load is the worst thing you can do in terms of efficiency.  Switching to an active load would give some improvement, but if you want to stick to a linear amplifier and efficiency is a concern then consider a Class A/B design.

--- End quote ---
I agree, class A are actually terrible when it comes to efficiency. But think about it, this is supposed to be inefficient. It's a very experimental thing. I'm not planning on releasing it to the public, if that is what you are saying. God no, I'd never release this monstrosity as a professional or even hobbyist-grade product. This is just for my amusement.
However, I will definitely try an AB amplifier. I'm curious to see the results.

Navigation

[0] Message Index

[#] Next page

There was an error while thanking
Thanking...
Go to full version
Powered by SMFPacks Advanced Attachments Uploader Mod