Quick Question Regarding Chipamp Kits

[CognitiveDiscord]

Hey guys! I've been lurking around the DIY audio scene for a few years, and decided to finally start a PCB based chipamp after having toyed around with a few breadboard headphone amps. I have a spare set of Advents that need some love.

I was all set to order the LM3886 kit from chipamp.com, when I stumbled upon this thread....

http://www.diyaudio.com/forums/chipamp/173382-lm-3886-volume-control.html

Specifically:

do not use this version of the chipamp. It has problems that beginners are unaware of and have to ask why their amp blew up or their speaker was damaged.

Now, I doubt there is anything terribly wrong with this kit, but I wanted to know for sure before I dropped the money on it and a toroid. I plan on using this kit as a starting base for future tweaks and experimentation. If there is a more recent "version" of the CarlosFM circuit out there for a PCB project, could you point me in the right direction? Otherwise, is this kit as awesome as it sounds, as far as a starting point?

Thanks again!

 

[benb]

I recall that thread, and another related thread. It appears the design in question is set up for frequency response down to DC. This is intended to have "no capacitors in the signal path" which may have slight sonic benefits over the use of capacitors, but there's a definite possible problem with many input devices.

There was another thread where someone asked about his amp of this design failing. It turns out he had an iPod earphone output connected to the amp's input. which for 99 percent of audio amplifiers causes no problem whatsoever, but the iPod output has a slight DC leakage current which the DC-coupled amp using the LM3886 dutifully amplifies until it overheats and/or damages the speaker it's connected to. A slightly different design that uses capacitor coupling, both in the input and between the gain-setting resistor and ground, will prevent this problem. It will also sound exactly the same as the DC couple amp to 99.5 percent of listeners.

(all statistics used in this post are approximate, but should give you the general idea)

 

[Atilla]

The kit from chipamp.com doesn't have a DC-filter capacitor at the input. If you want be safe in all cases, you can add a speaker protection circuit after it, to guarantee that if anything bad happens, for whatever the reason, your speakers will be safe. In fact, there's no good reason not to add that protection in any case. You can find them fairly cheap at ebay or electronic stores.

 

[jamal]

hi, I did wire up the LM3886 according to CarlosFM. it did work fine for me.

 

[Atilla]

Oh and the chipamp.com will not amplify any DC presented to it if you install the feedback capacitor. IIRC, it's labelled "optional" in the manual. Without it, what benb replied could indeed happen.

 

[AndrewT]

I doubt there is anything terribly wrong with this kit,

There is nothing wrong with the kit.
It does exactly what the very experienced designer wants it to do.
The problem is that the designer does not bring to the attention of beginners that there is a risk of damaged equipment solely due to the design decisions that he chose. It is now a while since I last read his literature supporting his product, I may now be out of date if that warning has now been appended.

DC coupled amplifiers are generally chosen because they may sound better.
AC coupled amplifiers tend to do less damage.

You do the research and then make an informed decision on whether to accept the risks or not.

 

[CognitiveDiscord]

Thanks for the quick answers!
Given the responses above, I will go ahead and order the kit over the next few days.


Correct me if I'm wrong, but these are a few changes I'll be making initially:


I'll be adding a speaker protection circuit post-amp, which couldn't hurt.
I'll be adding the "optional" feedback capacitor per the manual.
I'll be adding coupling capacitors in both the input stage and between the gain setting resistor and ground. Is there a recommended size?


I will be using an iPod Touch (most recent generation) for input initially, but I will be setting up a preamp in the next few months as well. Am I correct in assuming that this will eliminate the problem, or is there a specific type of preamp I should be researching?


Thanks again guys, you're a lot of help.

 

[AndrewT]

Hi,
if you add capacitors or add an output switch you are defeating the raison d'etre of this design. Omit everything that might interfere with the quality of the output signal. I could add irrespective of how much it increases the risk of broken equipment.

You decide if you want DC or AC coupled amplifier.
If you go AC coupled there is little point in adopting a design that does the opposite for a design route and then modifying it.

If you go DC coupled you have to decide whether to accept the risks or put in place some protection scheme.
For a DC coupled amplifier I recommend a DC servo to minimise the "normal" output offset with changes in operational temperatures + an input "mute" to remove input signal that has DC superimposed + DC detect and trigger to activate a speaker protection circuit. This does not need to be extra switching in the output feed to the speaker.

I'll be adding the "optional" feedback capacitor per the manual.
I'll be adding coupling capacitors ................. between the gain setting resistor and ground.

these are the same component.
If you go to AC coupled then yes, you have to design the filters and feedback to suit your choice of bandwidth.

 

[Atilla]

Thanks for the quick answers!
Given the responses above, I will go ahead and order the kit over the next few days.


Correct me if I'm wrong, but these are a few changes I'll be making initially:


I'll be adding a speaker protection circuit post-amp, which couldn't hurt.
I'll be adding the "optional" feedback capacitor per the manual.
I'll be adding coupling capacitors in both the input stage and between the gain setting resistor and ground. Is there a recommended size?


I will be using an iPod Touch (most recent generation) for input initially, but I will be setting up a preamp in the next few months as well. Am I correct in assuming that this will eliminate the problem, or is there a specific type of preamp I should be researching?


Thanks again guys, you're a lot of help.

In this design you should be just fine with the feedback cap (no DC-gain) and a protection circuit. Since you've got the boards already this is the fastest, cheapest way to make it nice and safe.

This is exactly what I'm running and the DC-offset is negligible. I get about 2mV.

The ipod touch shouldn't have dc-offest to trouble you.

For AC-coupled design and/or DC-servo you can find different designs/boards on this forum.

 

[benb]

Oh and the chipamp.com will not amplify any DC presented to it if you install the feedback capacitor. IIRC, it's labelled "optional" in the manual. Without it, what benb replied could indeed happen.

Just a note about the cap in the feedback circuit, it does not eliminate passing DC offset to the output, it only reduces the DC gain from that set by the gain resistors (their ratio + 1) to unity. If there's a 2-volt offset in the input (and the volume control is turned up all the way) there will also be a 2-volt output offset in the output. This is an improvement over full gain and the output putting the full power supply voltage across the output, but that's not the only reason for putting the capacitor in there. Like any opamp, the LM3886 has input offset error which gets amplified by the gain, and this capacitor reduces the gain factor to 1, reducing the offset that shows up at the output during normal operation.