Hi. I just made a simple LM1875 amplifier (for my headphones), but it didn't quite work out as expected. I've made an amp like this before and it worked great (except for some minor humming issues at first), but this one starts to output some strange and loud static noise at higher volumes. It's hard to scope it, but freezing the scope gave me this image:
http://i40.tinypic.com/att180.jpg
Perhaps the amp falls into oscillations at that volume? I don't know why though, I have 470nF and 100uF caps going from each supply pin to ground and this noise is heard from both amplifiers at the same time. It's powered from a +-12V 500mA center tapped transformer. Here's the board layout and schematic:
http://i43.tinypic.com/dwe5up.png
http://i39.tinypic.com/wixds.png
I also have an issue with humming on this amplifier as well (I'm bound to get some kind of noise regardless of what type of amp I'm making). I have a separate trace going to ground from the output jack, input jack and pot, but perhaps I should have had that for the feedback loop as well? Or perhaps the PSRR is too low for the voltage ripple on the supply?
http://i40.tinypic.com/att180.jpg
Perhaps the amp falls into oscillations at that volume? I don't know why though, I have 470nF and 100uF caps going from each supply pin to ground and this noise is heard from both amplifiers at the same time. It's powered from a +-12V 500mA center tapped transformer. Here's the board layout and schematic:
http://i43.tinypic.com/dwe5up.png
http://i39.tinypic.com/wixds.png
I also have an issue with humming on this amplifier as well (I'm bound to get some kind of noise regardless of what type of amp I'm making). I have a separate trace going to ground from the output jack, input jack and pot, but perhaps I should have had that for the feedback loop as well? Or perhaps the PSRR is too low for the voltage ripple on the supply?
So my amp starts to oscillate or what? Should I have a low-pass filter on the input to avoid amplifying RF? Can RF cause instability? I think I have plenty of decoupling caps, 470nF and 100uF right by the power pins and 2x1000uF in the middle of the board. Are you saying that the gain is very low or very high? I should perhaps implement an ability to switch between different gains for different headset impedance's.
Well... You seem to have the input GND runing directly to the chips decoupling caps...
Is that OK?
I would trace all input circuit gnd:s and the feedback to the same place, including the volume pot... And far away from the chips decoupling.
I'd say this is a PCB issue...
But I might be wrong.
Is that OK?
I would trace all input circuit gnd:s and the feedback to the same place, including the volume pot... And far away from the chips decoupling.
I'd say this is a PCB issue...
But I might be wrong.
The chip is specified for a minimum gain of 10 times.
At that minimum gain the phase and gain margins are very low.
Any non resistive impedance applied to the output could set up oscillation in the chipamp.
Some builders/designers report best sound from a gain of 20times to 40times. When you look at phase margins for this range of gains you see around 70degrees to 80degrees of phase margin.
A headphone amp needs very little gain. In many, just a 1times buffer is sufficient.
In my view a chipamp, with a usable gain around 20times, is not suitable for headphone driving. Look at the noise performance if you need further evidence of unsuitability.
And finally, fast input signals can cause instability in improperly designed amplifiers.
At that minimum gain the phase and gain margins are very low.
Any non resistive impedance applied to the output could set up oscillation in the chipamp.
Some builders/designers report best sound from a gain of 20times to 40times. When you look at phase margins for this range of gains you see around 70degrees to 80degrees of phase margin.
A headphone amp needs very little gain. In many, just a 1times buffer is sufficient.
In my view a chipamp, with a usable gain around 20times, is not suitable for headphone driving. Look at the noise performance if you need further evidence of unsuitability.
And finally, fast input signals can cause instability in improperly designed amplifiers.
Thanks for the replies. I understand what you mean AndrewT and I see now that this is not ideal for a headphone amplifier. One thing I could do though is to increase the gain until it becomes stable and "reduce" gain by having output resistors. The output resistors will protect the headphones, reduce noise and protect the amp from load inductance/capacitance, right?
JonSnell: 220p in parallel with R7+R9 to avoid high frequency oscillations? The oscillations I'm having are audible though, but perhaps avoiding high frequency oscillations will help still? I don't understand what an input resistor will do though, I didn't have that on my previous design. I was supposed to have a cap in series with R1+R10, but I somehow forgot :S Is this just to avoid the amplifier from amplifying DC? Won't the high-pass filter on the input prevent this anyway?
JonSnell: 220p in parallel with R7+R9 to avoid high frequency oscillations? The oscillations I'm having are audible though, but perhaps avoiding high frequency oscillations will help still? I don't understand what an input resistor will do though, I didn't have that on my previous design. I was supposed to have a cap in series with R1+R10, but I somehow forgot :S Is this just to avoid the amplifier from amplifying DC? Won't the high-pass filter on the input prevent this anyway?
The 220p will reduce the slew rate of the amplifier and stop the RF take off that possibly is causing the odd noise you are experiencing.
The input resistor acts as an anti parasitic device to stop the very low impedance from upsetting the input stage.
The noise you show in your screen shot looks like the amp is taking off and those were two typical ways of avoiding this happening.
The cap in series with the feedback resistor is to stop DC on your headphones as you quite rightly describe.
The input resistor acts as an anti parasitic device to stop the very low impedance from upsetting the input stage.
The noise you show in your screen shot looks like the amp is taking off and those were two typical ways of avoiding this happening.
The cap in series with the feedback resistor is to stop DC on your headphones as you quite rightly describe.
I've increased the gain to 20 and added a 100Ohm output resistor. The amp works fine now and the humming is almost inaudible. Is there any point in adding input resistors and feedback caps now that the amp is stable? Is being able to switch between different output resistors an acceptable way of adjusting for different headset impedance's?
JonSnell: Why would I need a cap in series with the feedback resistor when I have a high-pass filter on the input?
I'm also a little curious about the zobel network on the output. I'm not quite sure what a zobel network is, but I see that it will carry a lot of current at higher frequencies. Couldn't this be kind of dangerous if we are talking ultra sonic oscillations? Couldn't the zobel network be shorting the output to ground?
JonSnell: Why would I need a cap in series with the feedback resistor when I have a high-pass filter on the input?
I'm also a little curious about the zobel network on the output. I'm not quite sure what a zobel network is, but I see that it will carry a lot of current at higher frequencies. Couldn't this be kind of dangerous if we are talking ultra sonic oscillations? Couldn't the zobel network be shorting the output to ground?
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.