Adding output transistors

Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
I am looking into making an amp centered around the tpa3001 chip. On page 14 of this .pdf it has a typical application. Now, if i wanted to get more power, how would i go about adding transistors to the output?

I will be using +-18V rails so i was thinking of using a current gain configuration. Are there are precautions that i should know about? Can i assume the output of the chip is like any other input?

I have looked at some applications for the lm391 (i think that's the number) and was hoping it was going to be that simple for any chip.

Thanks for your help.
 
AX tech editor
Joined 2002
Paid Member
Adding output transistors doesn't give you more power. You need more supply voltage, so that you can drive move output signal across the load, which means more power. As a result of the increase, you MAY need more output power transistors, or you can use heftier ones.

Edit: Just looked at the datasheet. This is a class D amp, which is a different beast than a regular analog amp. Not for the faint at heart! Its max supply is 21 Volts, so if you want to stay on the safe side, there is not much beyond 18V left, increase in power with say 20V probably inaudible. I think you're out of luck here.

Jan Didden
 
Yes it is (I*V) but only if you have control over the impedance of your speaker. For example car speakers are typically 4 ohms which allows for lower voltage power supply rails for a given power. If you are designing your speaker as well, you can parallel drivers and such to get a lower impedance.

So let's say your speaker is nominally 8 ohms, then power available to it before clipping of your 18 volt rail amp is going to be

P = I * V
V = I * R

so

P = V^2 / R

P = 18^2 / 8

P ~ 40 watts peak or roughly 32 watts RMS

That's still plenty of power for my (and I suspect many others) needs. In fact a common practice for people using Naim amplifiers to drive the orignal Quad ESL is to lower the rail voltages in the amp to about 20 volts to prevent the amp from being able to damage the speakers.

Sheldon
 
Ahh, i see my math blunder.

Now, the chip i have puts out 20W @ 10%THD.. now,

When i look at the charts, i can stay below 1%THD with a gain of 36dB into 8ohms. My output power is 15W.

Since the chip is near 90% efficient, i'm not really close to my power rails them am I?

now, for arguments sake, lets say that with 18V rails my max is 30 Watts into 8ohms. So i am only half way there. But, if a added a transistor with a voltage gain of 2, i would hit the rails and clip.

BUT, I will be making a preamp for this project, as well there is a variable gain select on the chip that can be lowered. There is allot of tweaking that could be done to get the max out of the 18V rails.

If i could do all the tweaking, what kind of configuration could be possible? What about a simple push-pull with a voltage gain of 2?

My main concern is damaging the chip.
 
AX tech editor
Joined 2002
Paid Member
Number04 said:
Ahh, i see my math blunder.
[snip]now, for arguments sake, lets say that with 18V rails my max is 30 Watts into 8ohms. So i am only half way there. [snip]

So, your ONLY option to increase power from here is to lower the speaker impedance to 4 Ohm. That should give you 60Watts BUT you will hit the max current & max dissipation limit of the chip. Next step then is to parallel two chips with a current sharing resistor, which is detailed discussed in another thread here.

Jan Didden
 
Sorry to flog this topic to death. I have one more crazy idea.

So we all agree that my rails are 18V. If i keep the chip @ 15W into 8ohms i will get 15.4V peak. That's also 1.369 amps.

I plan on using voltage regulators for my power supply, with an Iout of 3A. So, my transformer secondary output will have to be above 18, most likely i will use 22V.

Now, if i want to use a Class B output stage on my amp, i could theoretically take my VCC from before the regulators and get a VCC of 22V, Allowing for a bigger swing on my output.

Just as long a my filter caps are big enough to take the strain.

Can anyone make sense out my this?
 
18v rails up to 22v is not a big difference... you won't perceive the slight power increase. That's the first hassle. The second is that you now need a different o/p stage configuration capable of voltage gain - ie common-emitter arrangement or CFP/Sziklai - to take advantage of the extended rail headroom.

Filter caps would probably be 25v rated and the least of your problems.
 
Sorry about they typo.

About the 22V. If i use a positive and a negative regulator, would i not have to have a center taped transformer of double the regulator input voltage?

IE; 44VP volts across the secondary so i can get +22VP and -22VP in reference to the common center? So that would give me 44VP for my output, double what the chip is getting.
 
Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.