So I wanted to design my own portable headphone amplifier (I broke my FiiO). Now I've built some Cmoy's and PPAs in the past, and I've completed some decently big electronics projects before - so I'm confident in my ability to actually build it, but I don't know how to design one. I know the theory in terms of the electronics : For example I know in theory how an opamp works. But I'm not able to make the jump into the actual designing something myself part. so how do I decide on topology's? should I power it off've a recharable battery and include a charging circuit or just use pencil cells? Can someone give me a general direction? b
First major DIY project : designing a portable headphone amp. Where to begin?
Head-Fi's Best Sellers
You need some gain, not very much, 4x to 10x is typical.
You need some current output capacity, 250mA would be nice.
You need some voltage swing, typical opamp voltage, +/- 18V, will be more than enough. +/- 12V will be OK if you get rail-to-rail output swing. You can get away with much less voltage swing, but some HP require more than others, so the amplifier is more versatile if the voltage out is good.
A rechargeable battery is desirable, but adds to the complication, especially if you want to run while charging. You can recharge offboard though. A dual-rail supply is desirable, because you can get rid of a lot of DC blocking caps. Then you have to worry about output DC offset though, which can destroy HP if excessive. Stay clear of rail-splitters, they're more trouble than they're worth. If you want to run from a wallwart, you'll have to come to terms with using an AC one, with a dual-rail voltage doubler for rectification, and you need to work at keeping the noise (spikes) down, so regulators definitely and maybe some ferrites or other measures to prevent EMI (radiation) might be desirable. Forget switch-mode.
Many people go for 2 * 9V PP3 batteries, rechargeable or otherwise.
In most people's book, this adds up to a 2-stage amplifier, with the first stage providing the majority or all of the gain, and the second having some current drive capacity. The feedback loop wraps around both stages for the best distortion. You can use discretes for an output buffer, but why? Google Walt Jung and read his stuff on multiloop configurations.
Both stages will be non-inverting, since you want to preserve polarity. This isn't considered an absolute necessity, but its easy to do, and consequently most people do it now. You could have 2 inverting stages, but why? It'll just add to the complication. There are slight differences in performance of inverting and non-inverting configurations, but they're best ignored for the moment.
A 10k pot is typical at the input. It'll be inaudible in terms of noise unless you really mess things up, so no real need to put it in between the opamps.
Pick an output opamp with short-circuit protection. You can parallel opamps to get more current drive, load-sharing resistors are desirable in that case. Consider some measures to prevent transients (popping) at switch-on and -off. Put in a shunt cap at the input to prevent the ingress of RF. Read up on star grounding; the layout of the grounds, power and signal traces contribute significantly to the final performance and can degrade an otherwise excellent design.
You can see a few HP amp designs including schematics here:- http://wakibaki.com/audio.php#t6
- 15 Posts. Joined 1/2012
- Select All Posts By This User
Thank you so much, just a couple of questions - wouldn't 2 9V batteries be very bulky for a portable headphone amp? I don't intend on using a wall adapter or anything like that so plain DC sounds fine to me.
If I'm reading this right, I would need 2 stages, the first stage that has an opamp for each channel giving me gain, connected to a final current driver that acts as an output?
- 8,765 Posts. Joined 1/2005
- Location: northern NJ
- Select All Posts By This User
2x9V is bulkier than 1x9V but it usually buys you some advantages with regards to distortion, op amp stability, and simple voltage drive.
The majority of DIY amps are built a lot like the PPA or Pimeta with separate "input" and "output" sections. The Pimeta/PPA with the Jung multiloop are pretty slick designs. I would not discount them.
Inverting op amp configuration has its merits - the way feedback is applied normally results in lower THD than the same op amp in non-inverting mode. The (greatly) reduced input impedance may be an obstacle, but if we wanted simple straight paths we would just build the O2 and say F*** this hobby! That is kind of boring though.
Have you read Tangent's website?
Have you read EVERYTHING about the Cmoy?
Start off small. Read all about the Cmoy until you can walk through the circuit and tell a story about every part of the amp's opperation - like walking through a park and telling why each path goes each way and how changing the path would change the whole park... Once you can do that adding some swings and a slide to the park (or a buffer to your cmoy) is easy as cake.
Obviously, if you look at a Sansa Clip+, it's tiny and it drives a lot of phones pretty good. The battery life is 15 hours I think, it has an oled display, spare computing power, storage, decompression, the equivalent of a DAC and amplification and even FM radio and recording facilities. The whole thing probably occupies less volume than a 9V battery, it's certainly lighter. It's SOC-based, System-On-a-Chip.
In order to achieve this kind of performance, the Clip makes use of SMT devices and a lithium battery. It's possible to compete with it as a DIY constructor, but lithium batteries alone present implementation problems that Sansa probably have an engineer or even a team of engineers tasked on.
You have to start somewhere, I'm suggesting a starting point that many experienced home constructors would adopt as a balance between many factors, but with a focus on performance ultimately. Not everybody would agree with my recommendations, but most would see their point. Although the device you will end up with will not be the ultimate in miniaturization, with care it could achieve a performance (in measurable terms such as noise and distortion) which will be very difficult to exceed at any price, and this is what most home constructors prize above all else.
Yes, you understood my recommendations about topology.