What Makes An Audio Op Amp An Audio Op Amp?

"Rich Grise"

... Phil

;-)

=A0 =A0 =A0 =A0...Jim Thompson

=A0 =A0 =A0| =A0 =A0mens =A0 =A0 |

=A0 =A0 | =A0 =A0 et =A0 =A0 =A0|

=A0 =A0|

=A0 =A0 =A0 =A0 |

=A0|

...

Hi!

Is 74HC9046 a lot better?:

Do not know how much better 74HC7046 is to 74HC4046:

74HC9046 should be better than:

Glenn

look

else is

part.)

from

That is a bit evasive. How about some part numbers?

??-/

The part number is on the schematic, in plain sight.

John

Everyone makes the HC4046, though, and they're cheap. I regard the flat spot as a minor wart (and a pitfall for the unwary), but it doesn't get in my way. It isn't very big, for one thing, only a couple or three nanoseconds wide; the problem is that the loop wants to servo right on top of it, where it causes all sorts of evil behaviour. You add a resistor to ground to make the loop servo with a very small phase offset, and life gets better really fast.

The real issue with the HC4046 is the VCO : it stalls when the tuning voltage gets below half a volt, and it's fairly horribly nonlinear: df/dV varies as much as 5:1 with voltage, which will make frequency compensating a PLL significantly harder.

Judging by Fig 23 on P.24 of the datasheet, the HCT9046 has the same issues. According to Figures 18 through 23, the CD74HC7046 performs the astounding feat of having a much worse VCO than the HC4046--it dies at

So for my money, there's no point in paying $3 for an 'improved' part rather than 50 cents for the old one, when the flat spot can be fixed with a $0.002 resistor to ground. I'd much prefer to spend the dough on a decent oscillator.

Cheers

Phil Hobbs

ONLY a couple of nanoseconds wide?!!!

It's ONLY Niagra Falls.

It's ONLY the Grand Canyon.

John

Sure, if one cycle corresponds with the continental US, that's about right. You just have to set your PLL to run in Cincinnati instead of Niagara. ;)

Cheers

Phil Hobbs

Haven't you realized yet that Larkin is all bull-s*it and no hat ;-) ...Jim Thompson

This is such a simple, obvious circuit, and JT *still* doesn't get it.

Sad.

John

He doesn't want to cover his clown's fright wig with a hat?

ote:

else is

from

what's wrong with the part number on the schematic? moc8204s

-Lasse

Jim, Jim, Jim. You really need that extra glass of wine tonight with your meds.

Jim

is

from

JT has killfiled me, so you'd think he doesn't want to see my posts. But he follows me around, peeping by proxy so he won't miss anything, so naturally he gets confused.

Weird old git.

And he still doesn't understand a circuit with one opamp and two optoisolators. I suppose you don't understand it either.

John

BTW I just got an HP5372A Frequency and Time Interval Analyzer on eBay for $389 plus $75 shipping. (Including $3500 for the jitter FFT option, it was $34,000 in 1991, and it's still very useful.) The 5372A is just the thing for spotting PLL hitter problems, among many other jobs, and the FFT option will tell you right away if it's something like hum from a too-small filter cap, or AM interference, or what. A remarkable gizmo altogether, especially since it's basically a glorified time interval counter.

The guy selling it showed screen shots of it failing POST, but to me it looked like a backup battery failure, so I took a flyer on it. When it arrived, it turned out that once I cleared the first POST error, everything else worked flawlessly, and a new lithium-thionyl chloride backup battery fixed it right up.

I used it today to measure the jitter of an HP 3325A, which was much worse than I expected it to be--a few hundred picoseconds at 150 kHz, even with a 7th order lowpass to get rid of the DDS jaggies.

This really is an amazing time to be buying test equipment.

Cheers

Phil Hobbs

We still use 5370Bs in some of our test stands. It's a TIC with 20 ps LSB resolution. It's a remarkable design, and uses an 8-bit 6800 processor to do a lot of amazing math.

I have a herd of Tek 11801 scopes, the 40 GHz sampler, bought off ebay. Often they are sold cheap, because they won't start up, and all they need is batteries. Various models have jitters in the 1.2 to 3 ps RMS range.

I'd love to do a 5370 TIC clone some day. SRS has one, but it's mediocre. But it's hard to compete with ebay.

John

I'd love a better two-channel FFT spectrum analyzer. I have an HP

Something similar that worked up to, say, 10 or 50 MHz would be one of my favourite two or three instruments. (It would need to have much better ground loop immunity than the 35665A, too.)

Cheers

Phil Hobbs

Seems to me that all one would need these days is a good filter/digitizer, and let your PC do all the math. A small box could be fully floating.

John

Something like that would be terrific, except that almost any vendor I could think of would insist on doing it in LabView, which sucks so big it makes Washington DC look like a source term. ;)

Cheers

Phil Hobbs