need some help with making an amp
Aug 23, 2007 at 5:41 PM Thread Starter Post #1 of 22

spiftacu1ar

Head-Fier
Joined
Aug 7, 2007
Posts
58
Likes
10
HI, I am pretty new to all this stuff, but i am pretty good with my soldering iron. I just am not so good at circuit design (not really layout, but more just what to use). I was going to make a cmoy, but i found this:

Nationmal Semiconductor: LM1036: Dual DC Operated Tone/Volume/Balance Circuit

I am not sure if the specs are any good for headphones, but i was hoping some of you could help me out. I can get free samples, which is the main reason i liked it. Also, what opamp should i use. I have 6 LME4986 opamps (the same ones used in the "hotrodding the x-fi" thread). Are these any good for an amp, or is there somehting better that i could get (from national semiconductor preferably so that i can get them for free). I don't want to spend too much money, because this is more of an experiment which i will be doing for fun. I will also be buying the goVibe 7 as my main amp (for those of you who say its easier to buy, i am. But i want to make one too).

If these parts are any good, could someone please help me figure out how to make an amp out of this. (i.e. parts list and/or a basic diagram). I can figure out the board layout later on my own time, that is not a problem. I just don't have much knowledge about how analog products work.

Thanks guys.

extrenal links:
LME49860 specs

Hotrodding the X-fi
 
Aug 23, 2007 at 6:15 PM Post #2 of 22
if u plan on using LME4986 get +/- 25V supply. Regulate it at +/-22V. The bigger the supply voltage the better are the dynamics. If u drive hphones with impedance smaller than 600 put a class A single ended buffer after opamp or better make it a part of the feedback network.
 
Aug 23, 2007 at 7:03 PM Post #3 of 22
let me reword, i know nothing about circuits, i am mostly a mechanical person. I need a lot of help. I don't understand most of your last post. So, basically, the only thing i am good for is soldering, and the final layout. I am 16, and i want to learn about this stuff. I am intersested EE and ME, but i don't know much about EE yet.

I have a LOT of time to do this. I am in no rush, it is like a weekend project for me, that i can work on whenever i figure somehting out. You guys don't have to sit for hours and draw the whole diagram at once, but maybe a little at a time. I don't want to cause too much of a hassle.

Also, i found this too, maybe i could use it, but I don't know not, its a tad advanced. I think it is a DAC, but not 100% sure: (it would be nice if i could bypass the analog in my x-fi).
LM4935

Feel free to give any feedback, good or bad (i know that i don't know what i am dealing with), as long as you explain a little.

thanks
 
Aug 23, 2007 at 7:48 PM Post #5 of 22
I don't see the circuit you are referring to in the link. From the description it doesn't sound like an amplifier.

You could build a CMoy and if you can read schematics build it with a more open layout since the hard part is getting the components so close together to fit in an altoids tin. None of mine are in tins. Use quality components.

A Pimeta is very highly regarded (Tangent site) although I have not heard one.

I use a regulated wall transformer from Jameco for all my amps and don't bother with building a separate power supply.

Note: in the CMoy the wall xformer replaces the battery. You need the resistors and capacitors and virtual ground as described on the Tangent site.

Good luck.

F
 
Aug 23, 2007 at 9:26 PM Post #7 of 22
Ok so u want to build a power amplifier. I think in order to understand some things we should start from the end and work the way to the begining. So the last element in ur chain are the headphones. In this device u will deliver power. Now the amp doesnt care if its headphone or speaker or another amp or a very long cable. He doesnt care about the nature of the thing in which he will deliver power for all that the amp "feels" is a resistance. The resistance of ur headphones in ur case. Hphones dont have a linear resistance throughout the audio freq range therefore we will refer it as impedance which is a real value. U want to deliver power into this impedance. Because of this non-linear nature of headphone's resistance the voltage applied across it and the current flowing through it will rarely be in phase therefore u cant apply the famous P=U*I. but u can apply Prms=Vrms^2/impedance. Prms is the rms value of power Vrms is rms value of voltage across the load. ok now we look back to ur hphones. for example lets say it rated at 300mW with 300ohm impedance. So ur amp should deliver maximum 300mW continuously. Doing the math results that ur amp should output about 9.4V rms that means about 13.2V peak. So u should have at least 30V supply. Will come back on this one later. So ur voltage peak for 0.3W is 13.2V that gives u a current peak of 44mA. So ur amplifier must be able to deliver 44mA into ur hphone with no problems. Looking in the LME4986 datasheet u'll see that the opamp cant deliver that current so u will have to boost it somehow. U can do that by adding a current amplifier aka buffer after the opamp voltage amplification. The current amplifier has a voltage gain of 1 or a little less and if it runs in class A adds little to none degradation of signal. Now we only talked about continuous levels of power. But music doesnt have a constant intensity level. Its what i was refering to as dynamics. Sometimes the intensity increases for shorts periods of time or longer and then falls back again and so on. We said that the hphone is rated at 0.3W rms. So its safe to say that if u apply more power at continuous levels u will melt the coil. But for shorts periods of time that coil can stand to more power. Sometime alot more. So the hphone itself can handle dynamics pretty good. (ofc depends on many more but lets leave it like this for now) The problem is can ur amp deal with dynamics? We saw that for 0.3W u need 13.2V peak. So that resulted in about 30V supply. But when the music gets louder the signal will be cliped by supply. So in order to deal with this problem u need as large supply voltages as possible. In ur case about 40V mb 44V... thats the limit of ur LME4986. I usually like to have at least 10db headroom so that will put me with 60V supply. Hahaha im sure some of you will say thats ridiculous. And u may be right since its doubtfull u'll be listening at 0.3W rms continuously. But u get the idea. The next thing u must be sure of is that the current amplifier can deal with the peak current that will flow while playing the loud passages. Ok if this is of any help to you i will continue at another time. Good luck!
 
Aug 23, 2007 at 9:53 PM Post #8 of 22
THanks, that helped a lot. A few things, first, i will be using this amp with the sennheiser hd-595, which i think is 50ohms, but not sure.

Also, say i want to add volume control, where does that go? Or, even better, how could i implement (if its even possible) the LM1036 (link in first post) to achieve bass and treble control as well?

Lastly, since the lme4986 cannot take enough voltage, which opamp should i use (from national semiconductor)?
 
Aug 24, 2007 at 11:40 AM Post #9 of 22
okz it seems HD595 has an impedance of 120 ohms. So for 0.3W u need about 6V rms thus 8.46 peak. The current requirment tho will be higher. About 50mA.

About LM1036, personally i wouldnt use it. it has big THD figures and operates on low single supply voltages. The only advantage is that u can have a relatively small component count for the functions it provides. IF u really want to use it this stage goes betwen input circuitry and power amplification.

lme4986 has very good supply voltage rating for a low power opamp! And since ur hphone has an imp of 120 ohms. +/-20V supply will achive that nice 10db headroom even with the amp delivering 300mW continuously! So ur in good shape. Now for the current amplifier... Im building a hphone amp right now which uses an opamp. Take a look at the picture. the bottom transistors are constant current sources for the upper transistors wich provide current amplification. The gain is set as with any opamp by the ratio of R1&R2. Good luck.
 
Aug 24, 2007 at 6:01 PM Post #11 of 22
I didnt put the resistor value for they were calculated to suite my requirments and not urs. But i see u dont know all the electronic symbols?..... If so u must first find out more about that and some basic principles in electronics.....
 
Aug 25, 2007 at 6:43 PM Post #13 of 22
They just come normally in 8 pin packages. Dual opamp: 2x2 inputs, 2x1 outputs, V+ and V-. Makes 8.
With single opamps, you sometimes have a DC offset correction pin, or a compensation pin. But sometimes there are just unconnected pins
 
Aug 25, 2007 at 6:51 PM Post #14 of 22
well i omitted the offset nulling part... basicaly u have to calculate the current that will flow in idle state through the output transistors. When calculating the gain its good to know with how much "level" u will attack the amp so u will not overdrive it.
 
Aug 26, 2007 at 6:17 PM Post #15 of 22
So i have been looking at a few opamps, and i noticed that most of the ones (that i could find) actually have a very low max voltage. I think the best one from national semiconductor is LME49720, but it only takes 34v. If I were to use somehting like this, would it be possible to add more voltage after the opamp. If not, what is the use of all these opamps?

Also, is the 12v power supply in a computer good enough for audio, or will there be too much noise added? Is there a way to shield it form the noise in the copmuter. It would be nice if I could use this amp permanently with my computer (connecting to the x-fi). If there is no way without adding lots of noise, then forget it.
 

Users who are viewing this thread

Back
Top