Single-ended input to balanced output, no DC offset

Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
Has anyone designed a circuit that can translate the single-ended outputs from my active crossover into a balanced output with no +/- DC separation to drive my power amp. I'm trying not to use blocking capacitors in this signal path. The crossover circuit is on the same PCB as the power amp driver circuitry. The power amp has a DC servo, but this cannot null the outputs from two opamps if the +IN and -IN have a fixed DC separation plus drift. My search in diyAudio has yielded typical two opamp buffers which generate a few millivolts of non-symmetric and variable drift voltage separation which translates to 100ma of bias current at the amp's bridged output.
 
Hi LineSource,

Your quest of no signal path C's is commendable. However, I may be missing the 'drift' here. You are wanting to run a bridge amp sub, correct? You say the amp servo can only track one, but each power amp channel should have it's own dedicated servo so nulling it's output. So if you run an inverted signal from your LF x'over out it will invert the DC offset and drift being fed to the other cannel which is on side of the bridge out while the other goes the opposite direction, amounting, as you say, to 100's of mA Iq. But each output is independantly servoed for NULL at it's output so there should be no DC across the bridged load because both ends are corrected to DC=0.

Am I missing something, or is this a non-issue. :D

Cheers,
Greg
 
xplod, thanks for the BB pointer. I simulated a similar circuit using a discrete opamp (diff JFET Pass crossover clone ckt) and noted about +/- 12 mv from the sum of the balanced outputs(attached). Maybe discrete circuits cannot track as well as the Burr Brown IC with laser trimmed resistors and single die thermal tracking. I keep wondering if a good discrete transistor solution is possible.


ampguru, the amp uses one complementary differential folded cascode input stage with balanced inputs to drive both +/- bipolar outputs, so there is only one servo amp. In this topology, if the +/- balanced inputs have a DC sum on top of the signal, instead of just a balanced tracking DC offset, the outputs drive a DC current into the speaker.

Are DC blocking capacitors always required on balanced input amps?
 

Attachments

  • singlebalanced.jpg
    singlebalanced.jpg
    46.7 KB · Views: 719
Hi,
you could try a pseudo balanced output from the pre to feed the balanced input stage of the power.
The pre single ended output is normal but you add a complementary shunting resistor from cold output to ground after the extra line source resistor.
The balanced input sees identical source impedances on both lines and thinks a balanced output is driving it.
Look up the Jensen website to see the circuit. It is two extra components and an XLR socket. They show the in line caps to block DC but you could try without if you are confident that the pre is NOT putting out a DC offset and that your power can correct itself for its own DC offset drift.
By the way you will need to select the balanced source resistors very accurately try for +-0.1% or better. A little bridge measuring circuit using your DMM will easily achieve this. A DMM on it's own will be right at it's limit of resolution
 
DC blocking caps

Hi,
DC blocking between pre and power is always required unless the designer has the chosen to use another method of eliminating the output offset. This would be rare and could only apply to a matched pair of pre & power.
Either the pre must have a DC block or the power must have a DC block. Most equipment has DC block at both the input and output resulting in the two RC time constants compounding themselves. In the worst case if both had the same turnover frequency then the overall Bass turnover would move up one octave and all the phase angle errors would also move up one octave.
There is one thing to look out for that is often omitted;- A shunting resistor (1M or higher is generally OK) at the RCA connector to bleed DC on the blocking cap to earth (on both the pre and the power) to help eliminate that loud pop (or worse) if you ever try to hot plug in RCAs without switching off.
 
Attached the Burr Brown drv134 buffer internal ckt diagram.

My design is all discrete, with JFET opamps used in the crossover.

A discrete circuit for single-ended-output from crossover to balanced-input on amplifier that nulled any DC offset created in the generation of the +/- balanced drive would help me and all of the SuperSym_X amp designers as well.

Let the transistors fly.........
 

Attachments

  • burrb.jpg
    burrb.jpg
    41.7 KB · Views: 680
I thought is was standard practise for at least one capacitor before the power amp input to prevent damage to the loudspeaker....try a Auricap.

Remember virtually all recordings known to man pass through countless capacitor coupled stages..so one more...the most important can't hurt.

macka
 
Hi Sy,
your suggestion for an input transformer will not avoid the DC offset from the Pre. I think the DC current will upset the transformer core with some effect on the electrical output that feeds the power amp (early saturation and probably other distortions).
The tranny only provides a balanced input which he already has.
If I have misunderstood the workings of an electrically balanced input and a transformer coupled balanced input then please correct me.
 
THIS IS WHY DIGITAL WILL WIN!


It takes 23 opamps, 115 transistors and about 200 capacitors and resistors, to execute a steep slope 3-way analog active crossover. The long chain of amplifiers adds noise, distortion, and phase shifts. I'm laying out a 50 sq. in. PCB for the Xover and Class A amp, and budgeting 180 sq. in. of heat sinks. I using capacitors to block DC into the amp because I can't find a good circuit to convert the Xover single-ended output into DC-Free differential-inputs for the amp.


IN DIGITAL THE CROSSOVER AND AMP WOULD BE ONE CHIP, DRIVING POWER SUPPLY MOSFETS THOUGH INDUCTOR FILTERS.
 

Attachments

  • crossover_midamp.jpg
    crossover_midamp.jpg
    59.6 KB · Views: 344
> I can't find a good circuit to convert the Xover
> single-ended output into DC-Free differential-inputs for the amp.

Is it line-level? DRV134? http://focus.ti.com/lit/ds/symlink/drv134.pdf

(Note: no practical experience, but I plan to have some by Christmas...)

BTW, I agree with your digital comment. It will also make those Lynkwitz thingamajigs a lot more interesting. And room balancing!

Wes
 
Wes,


I mimiced the Buffer134 using my discrete opamps and got the +/- 12mv offset shown in post #4. Burr Brown specs laser trimmed resistors and special on-die thermal tracking for their buff134. The Burr Brown app note shows DC blocking caps.
 

Attachments

  • buff134.jpg
    buff134.jpg
    54.8 KB · Views: 328
Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.