My discrete opamp

Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
Hello to all!!


After a while of being interested in mic preamplifiers, and after
hearing a lot of good stuff about opamps like the AD797, AD825, OPA627, etc.. and their performance in audio circuits, I'v decided
to built a mic preamp based on opamps. :)

I must say that the AD797 and OPA627 sound pretty good, but something was missing...The sound was a bit cold. :xeye:
I wasn't giving up and tried a few combinations of input
transformers and opamps but still I wasn't satisfied with
the results... :eek:
So I decided to give discrete opamps a chance.... and built
some...the first I'v built was the one in M.Leach article
and you can find it in this thread.
I plugged it in and... the sound was very good!! not cold as
the opamps' sound and it was also very quiet!! :cool:
The reason I decided to build M. Leach design was because
I'v seen that all the good discrete designs are using
this topology of second LTP stage (if this is how it called),
like Melcor, API 2520, Kaneda preamp, JC-2 preamp...
so I guess there is a reason for that...

Not to make a long story longer I'v decided to try the same topology (fet input, second LTP, and bjts output) but
with fully complementary output stage and not the quasi-comp.
I must say that I'm not an EE, so the design is not perfect...
I was aiming for a 'standart' +-15v supply, low noise, max of
40 db gain with good bandwidth, good output capabilities and
yet simple.


Any comments about the story/design are welcomed...

Best regards,

Udi.
 

Attachments

  • opamp.jpg
    opamp.jpg
    92.3 KB · Views: 4,279
Hi! if i turn volume maximum, my speaker has "windy sound" .I thought that my amp wasn't stable! Now i have just found the reason . That sound's from pre-amp ,TL084 (but only use 1 of total 4 )
your pre-amp has balanced -input (do i mistake in my sentence? ) .A balanced-output,why not? perhaps if you do that, the mirror can be removed.
 
Hi,

tmblack said:
I got 2 variations:
1) how about adding extra 1N4148 diode in string to get class A?

2) remove Q2 and Q9 parts and connect base of Q4 to anode of LED maybe thru 220 resistor.
Now that resembles my PS audio preamp.Tom

1) Adding extra 1N4148? The output transistors are biased at
17ma, isn't this already class A?

2) I guess it's possible..



tmblack said:
Which one is better: the AD797 or OPA627?
I'm not sure what you mean by cold, the best I've trried was only the 5534 and it was lacking in precision.

Tom

In my opinion the OPA627 is better than the AD797,
and the AD797 is better than the NE5532.


Regards,

Udi.
 
In order to calculate the current I may draw from the opamp,
can I use this:

I = V(rms) / R(load),

For example, let say the max voltage swing is 30V (I know that
in reallity i'ts less than that) so the rms voltage is about 10V
(30/2/1.41), and let say the load is 600R so the current will be app. 17ma. and beyond this it will operate at class AB?


Please correct me if I'm wrong.


Thanks,

Udi.
 
You need to do that calculation for peak current, not RMS. For a 600 Ohm load and 17mA your amp can provide a maximum of 10V peak or 7.2V rms before one output device switches off completely. Still probably enough for many applications, but personally I would increase quiescent current a bit to provide plenty of headroom, unless efficiency is a concern.
 
Mr Evil said:
For a 600 Ohm load and 17mA your amp can provide a maximum of 10V peak or 7.2V rms before one output device switches off completely.


moamps said:
Hi,

Use Vpeak instead of Vrms in your calculation.

Regards,
Milan


Hi,

So If I'll need more that 10V peak into 600R load, for example,
the opamp will switch to class B?
I mean, It won't just stop working, right?

According to this, a class A push-pull output will disapate twice
the power of a single-ended output, but still with same output
capabilities? Shouldn't push-pull be more efficient than
single-ended (even when both operating in class A)?

Forgive me for my ignorance in this area..



Thanks guys,

Udi.
 
The output devices must NEVER switch off if you want to stay in class-a, so like the others have said you must use the peak voltage swing. It will just switch to class-b once the quiescent current limit is exceeded, just like a normal power amp ;)

This will not dissipate twice the power of a single ended output, it will be half. Think about it. Single ended uses a constant current source in one half, which is always trying to pull the same current through it. PP the 'lazy' half is not trying to pull this same amount of current, just has to keep some current flowing, just a little.
 
Mr Evil said:
You need to do that calculation for peak current, not RMS. For a 600 Ohm load and 17mA your amp can provide a maximum of 10V peak or 7.2V rms before one output device switches off completely. Still probably enough for many applications, but personally I would increase quiescent current a bit to provide plenty of headroom, unless efficiency is a concern.


Efficiency is not a concern to me.
The BC639/640 are rated at 1W, and I want to stay with
TO-92 devices, and not having a heatsink on them.
Will 500-600mw be to much?

Udi.
 
Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.