variable-gain balanced amp

[hobiesound]

Hi All,

i'm building a api 553/opamps labs eq at the moment and i was wondering what the pro's and cons would be of slapping this variable-gain stage(from the d self page) in front of it so i could just use a 1:1 or 1:2 transformer on the output instead of the normal 1:3 output. Also is it ok to do the gain in front of the passive network or after with a separate gain stage.

thanx for looking!

greetings,

Thomas

 

[abbey road d enfer]

The use of a step-up output xfmr is justified for two reasons, almost similar and complimentary:
a) increasing headroom
b) operating the o/p stage at reduced level
With +/- 15V rails, the max o/p level is ca. +20dBu. API wanted to boast 20dB headroom i.e. +24dBu, so they opted for a 1:2 step-up OT. That means the nominal internal operating level will be -2dBu (for a nominal outside world +4).
Still, I don't know why anyone in his right mind would want to use a 1:3 OT! ???

Also is it ok to do the gain in front of the passive network or after with a separate gain stage. 

You have to analyse the level dgm. If you reduce the post-network gain (which would be the case if you put a 1:1 xfmr) and put gain before the EQ network, you run the risk of running it too hot. So, a strict answer is, if you put a 1:1 xfmr, you should increase the gain of the o/p stage. A more elaborate answer is: the overall attenuation of the EQ network is so variable that no perfect compromise exists. The ideal would be an input gain control (such as the D. Self circuit - very interesting because it can attenuate and boost) AND an output gain control. The combination would allow one to optimise the operating level at each stage. But it would also take some serious metering to really take advantage of it.

 

[hobiesound]

Hi Abbey road,

thanx for the reply, i didn't even think about the headroom that the OT would bring. In the original schematic they use all 3 secondary windings for a 1:3 ratio thus a 9.5 dB boost. The passive network has a volume drop of a little bit less
(don't know how much don't have my simulation data at hand) The DOA in the 553 is a differential amp without gain so all of the make-up gain has to come from the Xformer. I want the circuit to behave at unity so i need a little gain and this circuit could provide the extra 3dB (+- something) i want. Since i already have a few OT's lying around with 1:2 ratio this will probably be perfect. Ok it will have a little less headroom but i think for most applications i'll be fine.

greetings,

Thomas

 

[abbey road d enfer]

Seems like a weird design to me... Using an opamp in unity gain is probably the worst configuration, and a 1:3 xfmr may put serious stress on the output devices.
Can you post the schematic?

 

[hobiesound]

schematic is in this post: http://www.groupdiy.com/index.php?topic=35860.0
at first i thought i could just lower the 27k resistors but then i get less boost and cut from the eq.

 

[hobiesound]

Ok i have learned something today ..... if i raise the two 10K resistors around the DOA i get more gain.......  :-[

 

[abbey road d enfer]

OK, I've looked at the schemo; it's the "swinging input" type circuit. Why they have chosen to operate it at -10 is beyond me...
To be picky, the proper boost/cut amount and BW depend on the actual R1//R4 (or R2//R3) value. At the moment it's 7.3k. If you want to run it at unity gain, you should replace R1-R4 with 15k.

 

[ppa]

Seems like a weird design to me... Using an opamp in unity gain is probably the worst configuration, and a 1:3 xfmr may put serious stress on the output devices.
Can you post the schematic?

I agree with you that a 1:2 xfmr is the good choise, I think, and that an opamp in unity gain is probably the worst configuration.

The reason for the unity gain configuration of the opamp in the feedback network, I suppose, at this point, It was used maybe to increase as well as possible the feedback network bandwith to giving the better stability possible to the stage, considering that the old 2520's stability isn't very great (even if however a very good opamp), since , moreover, that stage works at -10dB of gain






 

 

[abbey road d enfer]

an opamp in unity gain is probably the worst configuration.

Actually, the noise gain is about 17dB, which gives good stability, but OTOH, the large-signal gain modulation is still there.

The reason for the unity gain configuration of the opamp in the feedback network, I suppose, at this point, It was used maybe to increase as well as possible the feedback network bandwith to giving the better stability possible to the stage, considering that the old 2520's stability isn't very great (even if however a very good opamp), since , moreover, that stage works at -10dB of gain

Honestly, I don't know; Spice sim shows that it would be very easy to operate at unity gain. Only the designer knows his motivations.

 

[ppa]

Spice sim shows that it would be very easy to operate at unity gain. Only the designer knows his motivations.

OK, the 2520 in unity gain works good.

but in this case, I wanted to say (I have explained myself badly) that, since the A1 op amp , that is a 2520,  has an active feedback network, the stability of the amplifier stage (all circuit A1 op amp + A2 op amp, etc...) depend by  this active feedback network. So, I suppose, that maybe to have the best stability possible the designer has configured the A2 op amp (another 2520) at unity gain for having the maximum bandwith possible from this active feedback network, considering that A1 (2520) stability isn't so great. I've done no simulations of this circuit, so this is only a my supposition.