[DolbyR]

Hi,
I was looking at this http://www.raymondaudio.nl/projects/.../project7.html amp and some things caught my attention.
The author uses a 2x12v/8VA transformer for it.
Wouldnt it be better to use something in the range of 15-18v?
And the main question, if i'd use higher voltage (15-18v), what kind of mA levels should i ned for it? I believe the OPAs dont really need that much. But am I wrong?

 

[heatmizer]

The rectifier bridge is going to bring the voltage to roughly 18v
Hmm something like an A47
It might be good to have tantalums on the outputs of the the regulators and just make them lm7915 and lm7815

 

[Sinbios]

This the reason why the datasheet exists

Max supply voltage: 36V
Max quiescent current draw: 5mA/amp, so for the 2134 it'd be 10mA total.

 

[DolbyR]

Thanks Sinbios.
I did look at the datasheet but wasnt sure what is quiescent current

According to wikipedia, Quiescent current is the current that flows in an electrical circuit when no load is present.
So, I thought I'd need more than that.

 

[balou]

max output current is 40mA per chan, so peak current will be 90mA. I think that with normal sized psu caps, you could get away with a 20mA power supply. I did some tests with a 2134 in a cmoy, max current was irrespective of listening volume (my yamaha headphones can bear 3w[!]) in the range of 12mA (10mA opa2132, 2mA led)

about the voltage: you only need high voltages if you want to listen really, really loud with extremely low-volume headphones.

 

[jcx]

you also have quiescent current in the reg (~6mA bias current in the zeners)

if you want to drive 300 Ohm 98 dB/mW HD600/650 to 120 dB spl you need 9.8V peak and 32/channel mA peak

the opa134 common mode input range and the output swing limit at about 2.5 V from either ps rail

with 13.25 Vs the supply dropout budget for 9.8 V output into 300 Ohms:
13.25 -2.5 - (50/2)*.032 = 10 V
which is just OK

4*5 mA + 6mA + 2*32 mA = 90 mA peak load current driving 300 Ohm loads to 9.5 V pk

24 V * 90 mA = 2.2 VA load on transformer

upping the xfmr sec to 2x15 Vrms at the same VA (or 1/2 size) and increasing the zeners to 15 V would give a +/-16.25 V supply which would have a little more headroom but the design is fine as is

I would drop R3 to 1K (and R5 accordingly to set your gain), the lower impedance feedback avoids some potential stability problems and could lower distortion and noise

I would also drop the volume pot R to 5-10K and short R2 (or drop value to ~50 Ohms):
http://www.head-fi.org/forums/showpo...7&postcount=24

 

[DolbyR]

Thanks a lot for all the help.
Just one more thing I cant figure out, whats the ACDC1 in the version of this finnish guy http://www.kolumbus.fi/antti.penttal...ponentsilk.pdf the schem is the same as Raymond's but I dont see someth like this in it.

 

[amb]

A bridge rectifier in a 4-pin DIP package.

 

[DolbyR]

Thanks AMB

 

[DolbyR]

Sh**, am I just stupid, blind, or someth else? There's no bridge rectifier in the original schem, or is there?

 

[amb]

The raymondaudio.nl schematic has four 1N4007 rectifier diodes instead of a single bridge rectifier.

 

[DolbyR]

Thanks again AMB.
Well, unfortunately theres few more questions I have. What kind of voltage rating should the bridge rectifier have? I found only 50v 1.5A in my local shop. I could source them from Farnell through them though. as thats how I'll have to get the OPAs too.
Then, I'm builduing this on veroboard, using Penttala's layout, just wondering what kind of wire i should use for the traces. I've been using componen leads before but no way to have that long ones as i'll need. I've also tried solid CAT5, works good for a CMOY, but it needs to be bit stiffer for this project so that it will look nice and be easier to work with. Copper also oxidizes easily so I dont think it would be such a good idea.
I'll be using 1% metal films as resistors, 1/4W and for the signal side, Im thinking about Wima MKP's. Does it sound like a good choice?
Btw, just coz im 100+ member, it doesnt mean I know much of this stuff yet

 

[amb]

While a 50V bridge would work fine I'd feel more comfortable using at least 100V rating in case of nasty voltage spikes. I suggest using four separate 1N400x diodes as in the first schematic. These are easy to find and cheap. 1N4007 (rated 1000V) is a bit extreme and a 1N4002 or 1N4004 would do fine, but most often the higher voltage variant aren't any more expensive, so why not.

Tinned #24 AWG (~0.5mm diameter) or slightly thicker solid hookup wire would work fine for your veroboard. The solder tinning keeps the copper inside from getting tarnished.

Your resistor and capacitor choices are fine.

 

[DolbyR]

Would there be an advantage in using a diode bridge over the other kind? Also, as i'm using Penttala's layout, I dont really know how i'd connetc the diodes. Just one end to the "pins" of his dip-4 place and the other ends to the trace? Or i'd need to change something more in the layout?

 

[amb]

Convenience... instead of 4 separate diodes, you have one package containing 4 diodes. Otherwise there is no real advantage, at least not for a small bridge. As for how to connect individual diodes to make a bridge, just look up the datasheet of a bridge rectifier and see for yourself.