Preamp with balanced outputs

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

I have a preamp based on the OPA627/BUF634 as described in another thread. My problem is that the GC I use is not giving me satisfying results, so I will move to a parallel/bridge configuration with OPA549's in inverting configuration (I am using now the LM3875).
For that I need to modify the preamp, in order to have balanced outputs to feed each side of the bridging amp. So I though of keeping the OPA627's, get rid of the BUF634 and put a DRV134 instead. The feedback loop of the OPA627 will have to move from the output of the BUF634 to the output of the OPA627. From here I feed the DRV134, which I guess have sufficient output to drive 3 OPA549's in parallel.
Any ideas on how this is going to work and any advices?

Miguel
 
Why not just take the internal circuitry of the DRV134 as by data sheet and duplicate it with 3 op amps and a few closely matched resistors? That gives you the best of both worlds. The only drawback is the cost of the additional 2 OPA627 and 1 BUF134 per side.
 
Ex-Moderator
Joined 2002
The problem is matching the resistors... Get it wrong just by a little and the noise rejection goes way down. Still, if you can test all your own resistors properly then it should be achievable. I am currently working on a little pre like this using OPA2604 and SSM1224 for the outputs
 
As far as I understand the crucial resistors to match are the output resistors (low R, usually 50 Ohms). Those will determine the CMRR of the following input stage. And there's only 2 of them ;). But apparently the unequal contact resistances in connectors, cables etc. will naturally put a limit on CMRR even with closely matched output resistors. In other words, match them as good as you can and live with it ;)

But wait. If you use the bal. input to drive 2 equal inverted gainclones that drive a load in bridge mode (something I also work on), shouldn't you get good CMRR even with loosely matched resistors? This seems to be close to the transformer input case, where even unbalanced inputs achieve high CMRR. The precise resistor matching is necessary solely due to the usual balanced input differential amp stage, which has unequal impedance to ground. In your case however you are feeding a topology akin to an instrumentation amp.
 
Which real advantages would an all opamp based balanced stage have? The DRV134 has good output drive capability according to the datasheet and I think they will not be stressed by the load of 3 OPA549's on each side.
The DRV134 datasheet mentions the use of 10uF caps on the output when offset is present. Also, I want to put some caps on the input of the bridging amp, as a safety measure, so only one cap is necessary on each output of the DRV134. Isn't 10u a bit too high, at least for a film cap?

I really don't thrust much on my DMM to match resistors... I am thinking of a way to match the 0.2 to 0.5 ohms resistors I will be using on the output of the amp, as I can get only 5% resistors. Any ideas?

Miguel
 
Miguel,

I had DRV134 and replaced it with your op-amp+BUF623 combo (pseudo-balanced, that means, only the + carries signal, the - is ground with equal output impedance as the buffer network). There was a clear improvement in detail and some in dynamics. (Test: I told wife if anything had changed, and she had these comments). Trouble, if you need a fully balanced output (signal on both + and - pins) you need 2 op amps with buffers, and a third to drive them. I didn't need signal on the - pin, so the pseudo-balanced suited me.
 
If you have a good bench supply regulate the voltage down .5v or so and see how much current is going thru.You should be able to match to about 0.1% which is precision resistor territory.
The drv134 is quite good sounding and will drive the opa549s easily.For something different you can implement a simple differential mosfet circuit ala PASS (see Pass bls or input stage of Aleph with ccs) and also try super-symmetry feedback(recent article by Pass on his site)
 
For something different you can implement...

I have already the OPA627/BUF634 pre so to change it to a DRV134 is just a plug and play operation (well almost that).

As for the resistor matching a wheastone bridge is something I had not thought of yet. But I don't have a bench supply. Can I use the supply of the preamp, with my DMM precision on measuring the voltage, and get good results on measuring the voltage on the other exteremes of the bridge to get the resistor value or is this not good?

Miguel
 
I have been thinking about the volume control for the preamp.
Something like this:

input -> DRV134 -> DS1802 - > 2x OPA627 -> bridge OPA549

The input level in the DS1802 must be low so a fixed attenuator (voltage divider) must be present there.
In this way the signal is splitted at the input and all the not so good effects of the chips after will be attenuated bay the balancing signal. Comments?

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