CL-D Bi and Tri-amping (lots of photos and plots)

[John Warren]

On ‎11‎/‎22‎/‎2017 at 9:28 AM, CECAA850 said:

The advantages of using this over, say, a couple of RCA splitters?

 

Some advantages-

 

The three outputs from each channel are buffered with respect to the input signal and to its neighbors.  With the LME49720 each output is capable of driving loads with input impedances as low as 600 Ohms to provide an output signal "essentially" equivalent in both bandwidth and amplitude to the input signal across a wide range of loads.  

 

Some chip amps require buffers at the input terminals. 

 

The popular active filter architectures like Sallen-Key, 12 and 24dB/oct Linkwitz-Riley need input buffer front ends.  

 

XLR combo allows advantages of balanced signal noise immunity to be exploited.

 

RFI/EMI filters can be applied and gain adjusted (via feedback) to provide near unity gain output while filtering out SMPS noise and other unwanted signals.

 

Provides proper front end for XLR connectors to interface with chip type amps capable of differential input interconnecting.

[CECAA850]

Thanks John.

[John Warren]

12 hours ago, CECAA850 said:

Thanks John.

My pleasure!  Great question too, btw.

 

 

 

[John Warren]

Common mode rejection.

 

Red FFT is both the inverting and non-inverting inputs with 5VRMS, 1kHz sine wave with 0.002% THD.  Green in non-inverting with same signal.  CMR at 1kHz is over 90dB.  Op-amp is the LME49720.  There's entitlement in the design but the result is very respectable none the less.  Need to match resistors to about +/-0.05% or better to achieve high CMR levels.

 

With this test, the input isn't in-step with the sampling clock of the analyzer so need to window.  The Blackman provides decent side lobe attenuation for good resolution of the fundamental.  

 

[babadono]

I do love me them LME49720 opamps. Bob Pease et al @National Semi knocked it out of the park with these.

Sorry i missed this thread until now

[John Warren]

Buffer board being supplied power from a +/-100VDC supply.  The capacitor board consists of 40000uF of filtering.   The voltage shown is voltage between the + and - power supply rails.   Under these conditions the power transistors reach a temp of about 140C which is close to upper end of the recommended range of 150C.   Maximum dissipation on the resistors is about 800mW.  

 

 

[John Warren]

The XLR shield drains are connected to  on-board RC hybrid filters with traces that back to a dedicated chassis ground circuit location for both XLR connectors.   This keeps all shield garbage away for power supply reference and star grounds it back to the IEC plug.   The green/black lead from "chassis" (in this case the Plitron mount) is attached to the RC filter on the buffer.  The green lead with Faston female spade connectors to the IEC plug.

[John Warren]

The R/W/B leads source +/-100VDC to the buffer.  The Fastons source to the amplifier.  The buffer current draw is negligible (about 40mA).

 

[John Warren]

The Plitron, the capacitor board and the buffer waiting for CL-D amps.

[John Warren]

FFT of CH1 buffer out with 560 Ohm load, +/-100VDC supply, 1VRMS balanced input at 1kHz.  

 

[John Warren]

FFT between 500-20,000Hz.

 

Nearly 100dB of distortion free dynamic range(!).

 

[John Warren]

So, having recently had a chance to break the redesigned buffer board out as a 2-CH separate, I've order a bunch of boards to play with. Compare it to the board above.  

 

The front end can now accept up to +/-100VDC input which allows it to front end a pretty good sized amplifier.  Of course can take XLR/TRS balanced or unbalanced.  Can source rail power thru the terminal block to three additional devices (additional front ends or meters or active filters).  +/100VRMS is about the limit given the MJE350/340 compliments are running near maximum dissipation at 145C when all three op-amps are outputting to full dissipation into 600 Ohm 100pF dummy loads (x6).  The XLR inputs are both common-mode and differential mode filtered so even stupid long balanced cables are free of EMI/RFI garbage as is shown in the bandwidth plots over a range of VRMS levels (0.5, 2, 4 and 5VRMS) into a 600 Ohm, 100pF load.  Note how the response drops hard above about 30kHz (it's by design).  The response is about 0.5dB down at 20kHz. 

 

CMR at 10kKz is a pretty outstanding at 94dB thanks to matching all RC pairs on the balanced side with within about 0.05%.  Considering the board is old school, radial leads and DIPs that's actually quite good.  Chassis and reference grounds are traced individually and I've gone to extremes to hand layout most of the traces for grounding and return circuits.  

 

I've completed a 2-CH Sallen-Key active filter design that will be used in conjunction with this module to assemble a 2-CH, actively filtered, triamplifer front end for my three way setup at our new place.  The Sallen-Key modules I've designed are beautifully laid out and share most of the utilities designed into the front end (rail distribution, on-board power, socketed op-amps, terminal block outputs, dedicated grounds for reference and chassis).  They allow for a wide range of capacitor sizes and types to accommodate very low frequency, 4th order band-pass filters. 

 

Distortion plot below is three runs at 1kHz (bottom) and three at 20kHz (top).   %THD+N at 20KHz is <0.04% maximum.

 

I've also added a photo of the front end sourcing a signal to a filter board I designed a few years ago (just so happens to be for a Klipschorn bass unit).  It's a S-K topology but not quite right for what I've got in mind.  The power supply is a pretty stout +/100VRMS unit with enough filter capacity to put a hole in something!

 

 

[John Warren]

On ‎8‎/‎21‎/‎2018 at 8:02 PM, John Warren said:

I've completed a 2-CH Sallen-Key active filter design that will be used in conjunction with this module to assemble a 2-CH, actively filtered, triamplifer front end for my three way setup at our new place.  The Sallen-Key modules I've designed are beautifully laid out and share most of the utilities designed into the front end (rail distribution, on-board power, socketed op-amps, terminal block outputs, dedicated grounds for reference and chassis).  They allow for a wide range of capacitor sizes and types to accommodate very low frequency, 4th order band-pass filters. 

 

Here's the module, very easy to build and make connections to power and amplifier.   Requires a center-tapped power supply greater than about 20VDC.  All tests of the module were using a +/-100VDC.  There's on-board power supplies for the op-amp packages.  Here using two AD713 quad for each channel.  Useful for putting a bass horn on a low-frequency bandpass with dedicated amplifier.  I'll put a webpage together showing how to use it and what it's capable of.  

 

The low and high frequency -3dB corner frequencies can be shifted left or right by changing a few caps and resistors.  Sallen-Key makes that possible.