Starting a headphone amp project, need general help.

[Bobbias]

I've been trying to start this project for a while. Long story short, I've got a few op-amps and enough stuff to begin putting something together, but I'd like some input and help. I'm limited by working with what I have on hand now. I haven't been able to order anything online for the project and everything I have was paid for by a friend.

Anyway, on to the specifics. I'm looking at creating an amp for my Grado SR80i's, which means I've got the fact that they're low impedance to deal with along with all the other design considerations.

I'm planning on using a 15 volt wall wart (has multiple output voltages actually) which is probably a SMPS.

I'm planning on creating a virtual ground and then creating a buffer circuit out of discrete elements.

I have 2 different op-amps on hand, and I really dont know which one would be better for the amplification process and which would be better suited for the virtual ground (or if there would even be a difference in the vground between them). My op-amp knowledge is pretty much limited to a very basic understanding along with being able to understand some of the datasheet information (I am an electrical engineering student, but it's been quite some time since I covered op-amps and what we did cover was more or less an overview of "here's what they do, here's some places they're used, on to the next subject")

I've got some TLC272ACP and some TL074CN op amps, and 2 2N3904's.

I'm aware this is likely not exactly optimal in any way shape or form, but I consider this a starting point. If I can get something that works out of this, I'll have learned a lot about what I'm working with and maybe in the future when I have money and the means to order parts online, I'll order some better parts to work with.

TL;DR
Got me some TLC272APC, TL074CN, and 2 2N3904's, help me design a working vground/buffer circuit and let me know which would be better in the audio line.

[counter culture]

Google cmoy. Click the link to tangentsoft.net

Build a basic cmoy with 2 * 9V batteries using the TLC272.

I'd omit the virtual ground and use the junction of the batteries as ground, but that's my personal preference, there's an arguable risk of damage to the 'phones if one battery is discharged, but the technical performance is better, just make sure you use 2 new batteries from a 2-pack (same sell-by date).

That's it.

[ChicagoJTW]

You can also create a dual power supply with the wall wart. Headwize shows simple plans for a NiCad trickle charger that will enable you to power off the wall and save your batteries when you are near an outlet. I have built this and it works well.

[Bobbias]

I was actually looking at the cmoy as a basis. Quick question: Why is the TLC272 better for the amp than the TL074?

I don't have a battery holder or any money to get a hold of one

I don't really plan on the amp being ultra-portable, which is why I'm fine with the wall wart.

I'm a broke-*** college student, I'm avoiding taking loans so my parents are helping pay for things and any work I can find helps, but I haven't found any work recently so yeah :/

If I'm going with the cmoy design should I add the load side resistor suggested in the addendum for dealing with low impedance phones?

[counter culture]

Originally Posted by Bobbias

If I'm going with the cmoy design should I add the load side resistor suggested in the addendum for dealing with low impedance phones?

This is a judgement call you have to make.

It's unlikely the DC offset resulting from a virtual ground and low impedance phones will result in damage to the phones, but the Grados are not cheap. You can measure the offset to help you make up your mind.

A large cap (470u) at each output is a better solution than a resistor IMO, any distortion is likely to be less audible than the effect on the frequency response of a resistor, but resistors are cheaper than caps.

You may be able to salvage some components from dead equipment.

[FREDtheDOG]

Would one of these be any good for you

Under $20.00 DIY Cmoy Hi-Fi Headphone amplifier KIT 18volt power Guaranteed ! | eBay

cheers

[godfrey]

Originally Posted by Bobbias

Quick question: Why is the TLC272 better for the amp than the TL074?

The TCL272 has a decent amount of output current available (+-30mA). I doubt the TL074 would get very far trying to drive your Grados.

The CMoy schematic I found has quite a high voltage gain (times 10), which seems excessive. IMHO, since you probably don't want (and can't get) more than about 1V output and most sources can give more than 1V input, a unity gain configuration might be suitable.

Since you mentioned having more than one TLC272 available, there's a couple of ways to put extra chips to use:
A) Using 2 opamps in parallel per channel gives you double the maximum output, and lower distortion at lower levels.
B) You could use one chip to make a better virtual ground.

I'm assuming you can get hold of a few resistors and capacitors?

Some ideas below.

[Bobbias]

Originally Posted by FREDtheDOG

Would one of these be any good for you

Under $20.00 DIY Cmoy Hi-Fi Headphone amplifier KIT 18volt power Guaranteed ! | eBay

cheers

I kinda wanted to make this as much DIY as possible, and I've already got pretty much everything I need, so that's not really much help at this point. I'm also currently dead broke (well, I got $20 from my grandmother today, but part of that is already going to go towards parking at the college :/)

Thanks for the heads up though. Also, I'm canadian, so that's another $5 shipping. I'll work with what I have now to get something up and runnign and look at improving things later when I have some income.

This is a judgement call you have to make.

It's unlikely the DC offset resulting from a virtual ground and low impedance phones will result in damage to the phones, but the Grados are not cheap. You can measure the offset to help you make up your mind.

A large cap (470u) at each output is a better solution than a resistor IMO, any distortion is likely to be less audible than the effect on the frequency response of a resistor, but resistors are cheaper than caps.

You may be able to salvage some components from dead equipment.

Now I wish I had grabbed a few of the 470uf caps my friend had in his collection just in case. Would paralleling a few smaller caps work instead?

The TCL272 has a decent amount of output current available (+-30mA). I doubt the TL074 would get very far trying to drive your Grados.

The CMoy schematic I found has quite a high voltage gain (times 10), which seems excessive. IMHO, since you probably don't want (and can't get) more than about 1V output and most sources can give more than 1V input, a unity gain configuration might be suitable.

Since you mentioned having more than one TLC272 available, there's a couple of ways to put extra chips to use:
A) Using 2 opamps in parallel per channel gives you double the maximum output, and lower distortion at lower levels.
B) You could use one chip to make a better virtual ground.

I'm assuming you can get hold of a few resistors and capacitors?

Some ideas below.

Thanks for the answer, I had actually been thinking about using two in parallel. Is there any reason not to use the two 2N3904s as a buffer stage? (My transistor knowledge is kinda spotty at best. I've forgotten a lot of what I learned in my AC fundamentals class about transistors because I've basically never had to use them since then).

[godfrey]

Originally Posted by Bobbias

I'm also currently dead broke...
Now I wish I had grabbed a few of the 470uf caps my friend had in his collection just in case. Would paralleling a few smaller caps work instead?

Paralleling smaller caps to make a larger capacitance works. I'm not a big fan of output capacitors though, so I'd probably be using them for power supply smoothing.

As for parts, you mentioned being an EE student. Isn't there an electronics lab on campus where you can try scrounge some components? (Bambi eyes, "Please Sir...") I once got a whole box of stuff from a local college when they were clearing out a store room and tossing the stuff out.

Plan B is to head down to the local garbage dump and look for an old bust radio or something you can strip.

Originally Posted by Bobbias

Is there any reason not to use the two 2N3904s as a buffer stage?

I was thinking about that too. It's a bit awkward making a decent buffer with just one transistor per channel, though. I suppose you could make a simple emitter follower loaded with a resistor to the negative rail, but I'm not convinced the results would be any better than just using the chips.

p.s. Power supplies from old desktop computers are a goldmine too. If it's still working, you can make a nice standalone 12V power supply. If it's dead, you can strip it for parts.

btw; How are you planning to build this - point to point wiring or Veroboard or...?

[Bobbias]

I've already tried getting stuff from the college, but it didn't really work out. The EE stuff there mostly focuses on either industrial automation or power generation and distribution, and there really isn't much in the way of parts. I did manage to get a hold of an old oscilloscope for free when they replaced the old ones with fancy new ones, which is kinda nice considering that's what I'm using as a multimeter for the time being (yeah, shoestring budget :P)

I might just see if there is an old power supply laying around I can scavenge. I'm not a stranger to scavenging for parts. I've got a microwave transformer rated at 120v:4200v sitting around here somewhere from my girlfriend's old microwave.

What effect does output capacitors have vs resistors?

I have a total of 4 100uf caps on me, 2 rated at 25v and 2 rated at 50. The 50V's are from the same company, the other 2 aren't.

I do have a number of small tantalum caps (3 of them read 0.1 35, one reads 105 35). I had a few more and they might still be laying around.

I plan on breadboarding this first and then once the design is finalized I've got some protoboard I can use.