Hi all, first post here. After reading few of other's posts and some document. I copy and combine them to have the design like the one in the graph. I know it is very simple to most of you. I would like to have some suggestions and corrections. I want to try the opa637 and buf634 combination and also parallel the buffer. I will power them with +-12V inside a cmoy like candy can.
An externally hosted image should be here but it was not working when we last tested it.
First of all, better persuade your CAD software to draw some dots where lines are connected, as it's usually done. Anything else gets confusing.
What do you intend to use in terms of cans and power source to warrant parallel BUF634s? You'd need <30 ohm loads driven at full swing for a single one to break into sweat.
Also, wouldn't that potentially get a little warm under serious loads? And are you sure that you could fit 2 channels of that onto a board small enough to fit something the size of a classic Altoids tin?
You have a single stage concept with a 11x (21 dB) gain there, hmm. First of all the gain setting resistors R8/R5 seem a bit high in value (I'd drop these to 1k/10k or so). I would also consider ditching the 2nd buffer in favor of a two-stage concept (10 dB before volume control, 10 dB after), for lower noise and better distortion performance.
In any case you should mention the intended performance goals and signal sources.
What do you intend to use in terms of cans and power source to warrant parallel BUF634s? You'd need <30 ohm loads driven at full swing for a single one to break into sweat.
Also, wouldn't that potentially get a little warm under serious loads? And are you sure that you could fit 2 channels of that onto a board small enough to fit something the size of a classic Altoids tin?
You have a single stage concept with a 11x (21 dB) gain there, hmm. First of all the gain setting resistors R8/R5 seem a bit high in value (I'd drop these to 1k/10k or so). I would also consider ditching the 2nd buffer in favor of a two-stage concept (10 dB before volume control, 10 dB after), for lower noise and better distortion performance.
In any case you should mention the intended performance goals and signal sources.
I am using EAGLE under mac osx, I have no idea how to turn the dot thing one. Can someone help. Thanks sgrossklass, here is an update from your suggestions.
An externally hosted image should be here but it was not working when we last tested it.
You have to place the dots one-by-one, yourself. Look for it on the toolbar in Eagle, somewhere close to the line tool.
Your circuit is basically this:
PIMETA v1 Headphone Amplifier
(don't go with the 3-channel/virtual ground idea though)
I would leave space to put a 50 to 100 pF local feedback cap on the op-amp in case of instability, like shown in the schematic in this relevant article:
http://waltjung.org/PDFs/WTnT_Op_Amp_Audio_1.pdf
I would also add a zobel network on the output, maybe 47 ohm in series with 100 nF, output to ground.
For the buffer you can use somewhat bigger bypass capacitors, maybe 1uF//470uF.
Last edited:
Thanks for the tip on eagle.
Here is an update from your suggestions. One thing I don't understand is the bypass capacitor. Do you mean C6, C7, C8 and C9?
An externally hosted image should be here but it was not working when we last tested it.
Those are the ones.
Bigger is better, but with rapidly diminishing returns above some level, depending on your power supply, power feed layout and headphone impedance.
There is no absolute right and wrong here.
An electroylitic should be used of around 10 to 47uf (these are bypass, and local decoupling remember, not reservoir caps) and these should have a small film/polyester type in parallel with them of around 0.1uf
There are some circumstances where that is actually the wrong thing to do and the film cap should also have a small resistance of around 1 ohm (carbon/metal non inductive) in series with it.
This is because if there is a specific problem such as spikes from an SMPS then just paralleling the caps can actually make the problem worse if the frequency sets up some resonance with other inductive components on the PCB such as print or component leadouts.
In your case you won't go far wrong just paralling a film cap with the electros.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.