Calcs for op amp loading vs THD?

[user 133392]

When do you use the Birt integrator vs your Deboo integrator?

I won't be using the DeBoo in the future: The DC accuracy is very poor without extremely precise resistors.

The DeBoo has a common mode gain of "1" so it floats to whatever the common mode voltage of its' inputs are. That's why I initially chose it for the Flying Rail Preamp.

I use both the Birt and the cross-coupled 124X/128X. The advantage of the cross-coupled 124X/128X is the precise resistor matching and no external resistors. The Birt has the advantage of looking like an FDA, has access to circuit nodes that are internal to a 124X/128X and can use lower-value resistors. Also the inverting op amp's noise contribution appears in the output in common mode. The Birt lends itself to being an A/D driver or FDA servo.

 

[thor.zmt]

I forgot to multiply by a hundred to get it in percentage. Your plot says -90 for the 2nd harmonic and 20 for the fundamental, I am guessing those are decibels?

It's dBV. That's how TINA-TI displays it.

So yes, it's -110dB HD or 0.0003%.

The other one was a brain fade by me.

Thor

 

[thor.zmt]

I've posted an article here on the many uses of the versatile Birt FDA: Visualizing the Birt Line Receiver As A Fully Differential Amplifier - Pro Audio Design Forum

Excellent work.

Only thing that is missing is the direct coupled, phantom powered microphone preamp.

And this one:

I derived a version of this myself, with additional circuitry to make it switchable for line, microphone (direct coupled, phantom powered) and Instrument. Sadly the product got cancelled during development.

Thor

 

[thor.zmt]

Note that that 450mW I got is an instantaneous power under the specific condition that occurs when the pair of opamps are pushing +-10V into the load resistor.

The power dissipation actually increases as you LOWER the signal amplitude peaking at 1/2 the supply voltage (so 7.5V) before falling again, and the pathological case is a square wave of 15V Pk differential. The REALLY pathological case would be to make the load phase angle 60 degrees or so (But that is just mean)....

Not really that relevant your pathological cases, if we are dealing with audio.

I quickly did the numbers for 10.6V into 600 Ohm, 460mW.

I have seen a D/A card (Drake?) done with a pair of TDA2030 driving a fairly large number of Lundhal transformers for the outputs, which I thought was a fun, if expensive approach.

In 2019BC (before Covi d/Chipageddon) I'd have recommended a Jung Compound using OPA1679 & 2 x THS6120. Can easily drive 20 parallel 600 Ohm loads to high levels, low noise and distortion.

Today? My little discrete can be amended to give more load drive capabilities. Change the diamond output to diamond plus compound feedback pair using TO-220 on a suitable heatsink.

Thor

 

[Dan Mills]

Not really that relevant your pathological cases, if we are dealing with audio.

Or say rather, only relevant to conservative thermal design, and mainly power amps at that.

In 2019BC (before Covi d/Chipageddon) I'd have recommended a Jung Compound using OPA1679 & 2 x THS6120. Can easily drive 20 parallel 600 Ohm loads to high levels, low noise and distortion.

Indeed you can do much better, but Telly broadcast was generally about simple, low parts count and utterly reliable, so that any idiot can fix the thing.
You can do appreciably better then old junk like the 2030, but they were once widely available and most techs understood them, something to be said for that!

In an age where most tellys had at best about a 4 inch internal speaker, the 2030 pushing the transformer in the DA was NOT the limiting factor, now the radio guys sometimes actually cared, but even there those DA were widely used and were typically not a problem, topology usually trumps parts choice.

 

[atavacron]

Only thing that is missing is the direct coupled, phantom powered microphone preamp.

With what values? Single gang gain?

 

[atavacron]

Or say rather, only relevant to conservative thermal design, and mainly power amps at that.

There are square wave patches on synths, and some pedals put out severely clipped waveforms by design. I could see it happening at 15Vpk / 10.6Vrms / 22.7dBu. But probably only for a section of a song, which isn’t long enough to heat up an amp. And even if it was struggling, you’d be hard-pressed to tell the difference between the elevated THD and that wild sound.

 

[user 133392]

Excellent work.

Only thing that is missing is the direct coupled, phantom powered microphone preamp.

And this one:

I derived a version of this myself, with additional circuitry to make it switchable for line, microphone (direct coupled, phantom powered) and Instrument. Sadly the product got cancelled during development.

Thor

Thanks!
I have an article on the evolution of the SuperBal here: Balanced Inputs - Pro Audio Design Forum
The Birt and SuperBal are predated by Dobkin in 1970 while at National published in AN-31.
I've also done work on an input-capacitorless phantom-capable mic preamp beginning around 2007.
It's on my list to get back to...

FWIW I ran some tests comparing various op amp's output short circuit currents and the ranges were large: Dual Op Amp Output Current Comparison - Pro Audio Design Forum
It was only one sample of each so "your mileage may vary."
I'll readily acknowledge that Isc doesn't correlate with THD at any given load within linear operating range.

 

[atavacron]

I have an article on the evolution of the SuperBal here: Balanced Inputs - Pro Audio Design Forum

Chris Mackensen came up with what he termed the Gainuator around 2010 (PDF attachment to the linked post). Same gain concept as the variable Birt/SuperBal/etc, applied to a regular FDA that uses equal values in the input and feedback networks, so that the single variable resistance across the nodes could be switched with a relay to perform attenuation or gain, respectively, using the same value. You could alternatively use a 2P8T Grayhill/Elma, a 2P12T Grayhill, or get spendy and use a 2P24T Elma. I think there’s no current across the nodes, so it should be mux-friendly. Its best and highest use (IMHO) is if preceded by a differential buffer at the input, which would enable one to use an OSOP or similar to do it with all 1K resistors and get pretty good CMR. You could also use it as a variable output, and dispense with the matched eight-resistor package if CMR is addressed upstream. More like all-1K5 for that. I checked in with Chris about it because he hasn’t logged in in seven years, and the Gainuator did indeed make it into a product and performs as advertised.

 

[atavacron]

I've also done work on an input-capacitorless phantom-capable mic preamp beginning around 2007.
It's on my list to get back to...

You and Thor are the only designers I’ve seen talking with experience about a flying rail preamp. It would be very cool to see a collab on that using modern parts, HINT HINT

 

[JohnRoberts]

You and Thor are the only designers I’ve seen talking with experience about a flying rail preamp. It would be very cool to see a collab on that using modern parts, HINT HINT

it's kind of a solution searching for a hypothetical problem.
===
The parts in Wayne's design are modern enough (IMO).

JR

 

[atavacron]

it's kind of a solution searching for a hypothetical problem.
===
The parts in Wayne's design are modern enough (IMO).

JR

Right but the INA1620 exists now and would make a simple and elegant back end, double balanced, +6dB, no need for the AC coupled LME49860 (IIRC, haven’t looked at the thread in awhile). And the JFE2140 is a real front end asset if applicable, could dispense with the reliance on a THAT part.

 

[user 133392]

You and Thor are the only designers I’ve seen talking with experience about a flying rail preamp. It would be very cool to see a collab on that using modern parts, HINT HINT

There was actually a patent issued on one that "borrowed" from posts originally made here and elsewhere. So we can count three.

 

[atavacron]

There was actually a patent issued on one that "borrowed" from posts originally made here and elsewhere. So we can count three.

Ballsy. Link?

 

[user 133392]

Chris Mackensen came up with what he termed the Gainuator around 2010 (PDF attachment to the linked post). Same gain concept as the variable Birt/SuperBal/etc, applied to a regular FDA that uses equal values in the input and feedback networks, so that the single variable resistance across the nodes could be switched with a relay to perform attenuation or gain, respectively, using the same value. You could alternatively use a 2P8T Grayhill/Elma, a 2P12T Grayhill, or get spendy and use a 2P24T Elma. I think there’s no current across the nodes, so it should be mux-friendly. Its best and highest use (IMHO) is if preceded by a differential buffer at the input, which would enable one to use an OSOP or similar to do it with all 1K resistors and get pretty good CMR. You could also use it as a variable output, and dispense with the matched eight-resistor package if CMR is addressed upstream. More like all-1K5 for that. I checked in with Chris about it because he hasn’t logged in in seven years, and the Gainuator did indeed make it into a product and performs as advertised.

I did a "Universal Input" using a THAT 1570/5171 combo which under software control flipped the position of the 5171 U-pad so the 1570-based input could span line to mic level gain. It's similar to the circuit you posted with the FDA.

 

[user 133392]

Markey US US 8,233,643
Also forgot about Chalupa who published an AES paper. So that makes at least 4 of us.
Both Chalupa and Markey burn off the DC CM voltage whereas I AC-coupled it with films after the floating stage.

I have a recent Idea For Design somewhere I need to find that shows an INA with both variable gain and attenuation which, IIRC, used a combination of positive and negative feedback or was cross-coupled somehow and used only one variable resistor. It actively attenuated I recall.

 

[atavacron]

I have a recent Idea For Design somewhere I need to find that shows an INA with both variable gain and attenuation which, IIRC, used a combination of positive and negative feedback or was cross-coupled somehow and used only one variable resistor. It actively attenuated I recall.

This one?

I had an email convo with Rusty, who had no idea that he’d just invented a siiiiick line trim.

 

[JohnRoberts]

Markey US US 8,233,643
Also forgot about Chalupa who published an AES paper. So that makes at least 4 of us.
Both Chalupa and Markey burn off the DC CM voltage whereas I AC-coupled it with films after the floating stage.

I have a recent Idea For Design somewhere I need to find that shows an INA with both variable gain and attenuation which, IIRC, used a combination of positive and negative feedback or was cross-coupled somehow and used only one variable resistor. It actively attenuated I recall.

In one of my late night scribbles I designed a 3 transistor long tail pair triple input differential stage (3-put) that allowed me to DC couple and level shift the audio down to normal rail voltages (the common mode DC voltage cancelled out). I'm sure my rants about that still exist over there. I never melted solder on my approach so this is just mental masturbation on my part. I suspect there is probably a way to cobble some op amps together to look like my 3-put, but I never wasted any brain cells on that, my discrete design didn't worry about above average rail voltages.

JR

 

[user 133392]

This one?

I had an email convo with Rusty, who had no idea that he’d just invented a siiiiick line trim.

That jogged my memory. I recall emailing Rusty at the time as well when I saw that one.

The circuit I think I have in mind was a three op amp INA published as an IFD. It's been on my mind for a few weeks and I've been digging for it.

 

[user 133392]

In one of my late night scribbles I designed a 3 transistor long tail pair triple input differential stage (3-put) that allowed me to DC couple and level shift the audio down to normal rail voltages (the common mode DC voltage cancelled out). I'm sure my rants about that still exist over there. I never melted solder on my approach so this is just mental masturbation on my part. I suspect there is probably a way to cobble some op amps together to look like my 3-put, but I never wasted any brain cells on that, my discrete design didn't worry about above average rail voltages.

JR

Yours JR is/was more elegant. In Chalupa, Markey, Kirkwood et al and possibly thor we cheated by bracketing supplies around phantom Vcm.