Opamp input buffer choices

[tfboy]

I'll carry on with more searching and reading up, but in the meantime...

I'm trying out Hypex's new Ncore500 units. They come without any buffering so I want to add my own. Their test rig (which I've also picked up) comes with a LM4562.

But it looks like there are as many solutions as you could dream of, and some of the suggestions in their datasheets are rather juicy, particularly the LME49724.

So has anyone integrated one of these (or other opamps) with an NC500?

Any ideas or thoughts to share?

 

[AndrewT]

What is the input impedance of the Ncore500?

What about that gives you reason to think you need some "buffering"?

The more usual problem with interconnecting audio equipment is the inability of the Source equipment to drive the cables.
It's generally the Source that may need a Buffer and this Buffer is inside the Source equipment.

It is very rare that Receiving equipment would need a Buffer.
an example is converting a single ended or unbalanced input to balanced. In this situation the Cold feed to the -IN pin can see a very low impedance.
R.Cordell designed and posted details of the jFET Buffer to meet, exactly, this duty.

 

[tfboy]

Hi Andrew.

The amp is a differential / balanced audio one and the input impedance is relatively low; 25Ω per leg, i.e. 50Ω in total.

I think most modern electronics can drive that without too much difficulty, it's more trying to cover "weaker" pre-amps which would have too much voltage drop.

I've not dabbled with pre-amp side of things before so it's a learning experience for me 🙂

 

[AndrewT]

I don't believe this:

the input impedance is relatively low; 25Ω per leg, i.e. 50Ω in total.

Check what you have read and ensure you understand what the specification means.
Can you post an extract from the spec here?

 

[tfboy]

Oops. Sorry, before the compulsory morning coffee.

It's 1.8kΩ

Additional note: The INH/INC inputs form a differential pair. Note that the input impedance is fairly low meaning that minimalist discrete circuits or valve input stages won't work. All op amps commonly used in audio can handle them though.

 

[AndrewT]

2.2Vac = 3.11Vpk.
that voltage into 1k8 is ~1.7mApk
Any Source should easily drive 1.7mApk into a resistive load. Even an iPad can do this.

It's the cable capacitance that can be the far bigger current demand.

edit,
Av is +12.4dB requiring a maximum input of 80Vpk/4.17=19.2Vpk

The input will draw ~10.6mApk if driven to near clipping.
That will require a good pre-amp capable of ~14Vac & ~14mApk

It's a mini power amp rather than a pre-amp and will probably operate on +-24Vdc to +-30Vdc supply rails. Opamps cannot look at this duty.

 

[EssB]

Are you sure? 25/50Ω is ludicrously low input impedance. More likely its 25/50 KΩ.

Can't find any data on the NC500 as it appears to be an OEM unit, can you link to a datasheet?

 

[EssB]

OK , that's more like it.

 

[abraxalito]

2.2Vac = 3.11Vpk.

Can't see a reference to 2.2V in the DS included in post 5.

I do see though that the output power into 8R is typically 400W. This requires 56VRMS, 80V peak. The voltage gain is stipulated as 12.4dB so *4.16. Dividing the output voltage by this gain gets us to 19.2V peak. Probably a bit less in practice coz the 400W figure was clipping slightly (1% distortion). So it looks like any buffer at the front end needs to do 10mA.

There must be a mistake somewhere because an LM4562 is only rated to 17V on each rail so won't drive this into clipping. Unless of course the input is balanced which would mean each side only needs to swing up to 10V or so.

<edit> Ah I see AndrewT and I have eventually reached pretty much the same conclusion. I am pretty sure Bruno's a fan of balanced inputs so opamps will handle it but you'll certainly want to bias the outputs into classA with CCSs to VEE.

 

[AndrewT]

Can't see a reference to 2.2V in the DS included in post 5...................

He had not included any Ncore 500 spec details.
I took a guess at a typical CDP maximum voltage and showed an example of how to calculate the current requirement of the source.

Picking up from Abrax comment:

Unless of course the input is balanced which would mean each side only needs to swing up to 10V or so.

two channels of phase inverted output for an effective balanced impedance source would require ~10Vpk and ~14mApk from each of the phases (hot to pin2 and cold to pin3).
A pair of Opamps operating on +-15Vdc to +-18Vdc could manage this duty.
Bruno Putzeys' balanced vol pot could probably be used if modified to have an actively driven second phase.
I think the second phase is balanced impedance passive without a signal voltage.

 

[tfboy]

Andrew, I presume you initially chose 2.2V as an arbitrary value of input voltage to the Amp?

In my case, I'm using balanced pre-outs: a Marantz AV8801 processor which gives 2.4V output pk.

However reading up on the datasheet of some other highly rated opamps like the LME49724 with +/-18V gives sufficient voltage swing on the output of over 50V.

I would just like to be able to put pretty much anything up front, valve pre-amp, ipad or modern SS stuff 🙂

 

[abraxalito]

To make the input more 'universal' you'll probably be wanting some gain - I'd suggest a two-opamp instrumentation amp configuration with 20dB gain. For a highly rated opamp, try HA-5147.

 

[AndrewT]

Andrew, I presume you initially chose 2.2V as an arbitrary value of input voltage to the Amp?

In my case, I'm using balanced pre-outs: a Marantz AV8801 processor which gives 2.4V output pk.

I always get confused by the balanced output voltages, I don't understand whether they are the differential voltage between Hot and Cold or the half voltage on either Hot or on Cold. Do you know if the 2.4V is differential, or two poles of opposite phase?

However reading up on the datasheet of some other highly rated opamps like the LME49724 with +/-18V gives sufficient voltage swing on the output of over 50V.

I would just like to be able to put pretty much anything up front, valve pre-amp, ipad or modern SS stuff 🙂

If all your Sources have maximum outputs around 2Vac to 2.3Vac then look at using some gain to increase the ~2Vac to ~13.5Vac (19.2Vpk) or +16.5dB
+6dB comes from the unbal to bal conversion. (or you need more depending on how that 2.4V pk was measured).
You need around +10dB to +11dB of gain between a balanced impedance source and the balanced impedance input of the Ncore 500.

Have a look at the B.Putzeys' paper.

 

[tfboy]

Thanks.

I think the 2.4V must be differential, purely because on single-ended unbalanced connections, they say it's 1.2V.

One of these days, I'll plug the scope in and actually measure it.

 

[tfboy]

That HA-5147 looks nice, I'll see if they do a differential version.

 

[AndrewT]

Then it looks like the Marantz has 1.2Vpk relative to audio ground and 2.4Vpk differential voltage.

The Ncore reuires 19.2Vpk differential (if the gain is +12.4dB)
You need extra gain to allow you to just not quite exceed your maximum power.

extra gain = 19.2/2.4 = 8times (+6dB = 2times, +12dB = 4times, +18dB = 8times)

You need a balanced impedance gain stage that has +18dB of gain, if you want to use the full capability of your 400W amplifier.

Does the Marantz have vol pot facility?

It would be worth checking with Ncore that we are reading their spec sheet correctly.
Get them to confirm the gain you require to get from the Marantz to the Ncore.

 

[tfboy]

The Marantz has some calibration settings where you can adjust the output to reach a certain reference level. I don't know what the upper limit is in terms of how much voltage it can supply.
Looking at the specs, it's clearly the difference between hot and cold signals. They state the same kind of stuff for the inputs: balanced has double voltage swing, double impedance: http://us.marantz.com/DocumentMaster/US/mz_av8801_u_eng_bg001.pdf

The dev rig uses the LM4562.

They also suggest other opamps, and values are to be tweaked to get maximum gain out of the unit.

 

[AndrewT]

Note the different symbols indicating Chassis and Audio Ground. This is VERY important.

The upper diag is the first half of an instrumentation differential amplifier.
Look at B.Putzeys. He uses the same. And so do very many others (W.Jung, D.Self, R.Cordell etc...). The gain is {3k3+3k3} / 1k5 =4.4times (+12.9dB)

The supply rails determine the maximum output. expect max out to be ~ supply rail minus 3V if the amp can easily drive the load impedance.
If you need 10Vpk, then +-13Vdc is just enough to get your 10Vpk from each chip.
i.e. 20Vpk differential.
The lower diagram shows a dif in and dif out chip. It too can drive to within a few volts of supply rail and the diff output is double the maximum peak value from either half.

The pdf download does NOT state 2.4V pk. I think it means 2.4Vac, or 2.4Vrms, even though it is ambiguous.

 

[tfboy]

Note the different symbols indicating Chassis and audio ground. This is VERY important.

Haha. I know. Bruno cracks me up with it and the whole "pin 1 / ground loop" problem. 🙂

And thanks for the details. I knew I shouldn't have dozed off in those opamp lectures at university all those years ago 😉

 

[AndrewT]

...............I knew I shouldn't have dozed off in those opamp lectures at university all those years ago 😉

I guess my lecturers did not know that SS opamps had even been invented back then.

There was a reference to FEA and a link to Wiki.
I was very surprised to see how recently that Wiki claimed it was invented.
I did get some input from our lecturers on FEA before 1970 and yet it must have been a very immature technique/method at that time.
Unless Wiki is wrong and it was around long before the early 1960s

Its development can be traced back to the work by A. Hrennikoff [2] and R. Courant.[3] In China, in the later 1950s and early 1960s,