My "audiophile" LM3886 approach

Thanks Unixman, I am using a high spec soundcard as source so am wanting to just use a the PC asio volume control to save money and relieve the complication of finding a volume pot. The only disadvantage of this, needing a very high gain to get a enough output for lightly compressed music or quiet tracks.

This sounds ridiculous but Ive searched this thread and the forum a bit but can't actually find any recommendations for an active class A transistor preamp for myref

I have 4 channels and want to keep the amp as high spec at as low cost as possible which is why I haven't just gone for 4 OPA627's or any opamps.

I've found plenty on opamps, buffers and passive attenuators but nothing other than a design posted on decibal dungeon for class A and that wasn't for myref.
Forgive me if I've missed the obvious but could someone point me in the direction of the best discrete class A transistor pre's, I'm guessing I'm missing a big discussion on the best transistor pairs to use somewhere?

Many thanks
 
Last edited:
I thought they had a maximum gain of 1 being a buffer?
I'm looking for 10db maybe 15 max.

Thanks
Hi,
a gain of two times (+6dB) and 4times (+12dB) is really useful for low output sources.
ONLY USE GAIN when it is needed. Two separate input channels, one feeding the volume control and buffer only and the other feeding the switchable +6/+12dB pre-amp then to the volume control and buffer.
 
Hi Andrew,

I do have a very low output source. I am using a computer soundcard as the output but to save having to but an expensive pot or make an attenuator I am not having any gain control on the amp, I am only going to vary the volume via the lossless soundcard volume control.

This means I need a very large gain to allow me to still get good volume on quiet songs and hence why I'm looking for a 10-15db preamp.
 
AFAIK, soundcard output levels usually are in the same ballpark as today typical line level outputs, that is about 2Vrms.

Thus you should not need any extra gain.

And, if you're not goin' to use an analog volume control, you do NOT need a buffer either.

Just directly connect the (low impedance) output of the soundcard to the my_ref input and you're done.
 
AFAIK, soundcard output levels usually are in the same ballpark as today typical line level outputs, that is about 2Vrms.

Yes, if I had it at maximum volume all of the time. This is why I want the extra gain to give the headroom to use the PC volume control.

unless you're using some "high-end" (pro or semi-pro) audio card which includes an analog attenuator for volume control, I'm afraid that's quite unlikely...

I am using a pro studio RME soundcard with software in ASIO which bypasses the windows mixer etc and is known for not adding any audio anomalies.

Is there really no discussion about transistor preamps with gain :confused:
 
Last edited:
Andrew, would you mind rewording that I'm not sure I understand what you mean?

I think you're saying I'm amplifying the noise carried along the cable from the soundcard rather than just amplifying the DAC noise more if I had a higher gain at the source, and having a greater SNR along the cable?
 
Recently, my RevC has developed a buzzing sound during power down that last for about a second. Nothing has changed in my setup so far. It's doesn't sound like it will damage the speakers, but it hadn't happened before.

Anyone has any idea what caused it (where I should start looking) and how I can solve it?
 
PJN,

Some ideas presented here had addressed many technical issues I had in mind. Additionally, the idea of having the whole thing floating from the chassis got me thinking that since the amp is at the end of the audio chain, it does make sense to let it float.

I have used switching power on gainclones with relatively good results, love the efficiency and the punch it delivers which was a problem that I had with transformers where ratings were the same. But with this PCB, I will first try them without snubbers and then try to design snubbers according to application note from Calex on "Understanding Power Impedance Supply for Optimum Decoupling". This will be done on the LM3886 power first.
 
Last edited:
Since I have started to put my RevC together (finally), I thought it would be a good idea to start posting.
Basically, I will be replacing the linear PS with switching PS. The PS switches at somewhere between 100KHz~86KHz (since I don't know the version of these).
Just started playing.
First problems. The LEDs that are supposed to light up I think, are not turning on. I'll have to open the tight chassis arrangement and do some measurements to see what's wrong.

The sound is amazing from first impression. Very detailed and dynamic. Since I have one burned super tweeter, I'm sure the results will be better than now. I will start turning the volume up after I get a new super tweeter in. I am currently only driving it from a PC headphone output and will be comparing it against the HAFLER 280XL that I originally had driving the same speakers. I already think it's going to beat the HAFLER by quite a margin.

The whole thing is in a 10cm X 10cm X 23cm enclosure. The power supplies are two 29W modules in parallel @24V spec but 27V measured.
 
Last edited: