But holy grail... Lets keep our fingers crossed...
No, only the most hardest of the hard hardcore, THS4021, AD8067, LM7171, LT1222.
There are no way to override their huge feedback depth while you are familiar with proper compensation techniques.
*attached*
The winter wasn't cold this year, so, yes, we could talk about push-pull or even AB-output.
Ok, EF-1 could be easily changed to BUF634 or even, with some tweaks, to LT1210. Supply rails could be decoupled with solid-polymer caps.
Moreover, we really need at least 4 differently optimized designs.
1. For ultrahigh sensitivity armature headphones with low impedance (10-32 Ohm)
2. For high sensitivity and moderate impedance common dynamic (16-64)
3. For high sensitivity and high impedance dynamic (300-600)
4. For low sensitivity and low impedance (16-64)
First pair could run at moderate 8-12 V rails while first could demand something like -20 dB gain instead of +6dB of second.
Second pair are better to have bridged output and something like THS4131 as input while third needs +12+18 dB instead of last fourth could demand even +26 dB.
Attachments
Yes that error (don't know when it appeared) would not have been deliberate because the schematic is already very simplified. The data sheets were done back in the day with stick on symbols, tape, and stat text.
BTW it's sad that most places no longer appreciate all the access to the innards.
Last edited:
50 pf is a really good joke!
Let's compete?
The goal will be to remove that ugly cap and achieve as much loopgain as possible at 20 kHz while keeping amp schematic and biases.
Sorry, I don't get it.
There are a bazillion really, really, really good chip-based op-amps out there; moreover, there are a few testers that have taken at least some of these op-amps and tested the hêll out of them in 'ladder-of-them-to-see-how-bad-they-are' layouts.
Sure, there is an intellectual challenge in designing an operational amplifier from discrete components; then another challenge to cherry-pick components to exacting matching standards, then another to gin up eensy little circuit boards, and solder everything (super-competently!) to them, capped off with the intellectual challenge of then testing a raft-load of the things, to measure the supposedly-superior characteristics.
Heck, except for the FT degradation, noise pickup and capricious device-heating-instabilities of laying everything out on a standard engineering socket-board strip, one really can cobble together something in an hour or two, and run up tests immediately. This I've done some decades back, driven by the same intellectual curiosity.
With all the blessings of a 50 MHz quad-channel oscilloscope, and having ginned up 10 well-matched parts discrete op-amp eensy-weensy boards, I found that none of them had characteristics superior to any of the handful of “better chip amps” I purchased some time before the enterprise.
So… again, as my uncle would say, “what's the poin?” (sic: no T)
⋅-=≡ GoatGuy ✓ ≡=-⋅
There are a bazillion really, really, really good chip-based op-amps out there; moreover, there are a few testers that have taken at least some of these op-amps and tested the hêll out of them in 'ladder-of-them-to-see-how-bad-they-are' layouts.
Sure, there is an intellectual challenge in designing an operational amplifier from discrete components; then another challenge to cherry-pick components to exacting matching standards, then another to gin up eensy little circuit boards, and solder everything (super-competently!) to them, capped off with the intellectual challenge of then testing a raft-load of the things, to measure the supposedly-superior characteristics.
Heck, except for the FT degradation, noise pickup and capricious device-heating-instabilities of laying everything out on a standard engineering socket-board strip, one really can cobble together something in an hour or two, and run up tests immediately. This I've done some decades back, driven by the same intellectual curiosity.
With all the blessings of a 50 MHz quad-channel oscilloscope, and having ginned up 10 well-matched parts discrete op-amp eensy-weensy boards, I found that none of them had characteristics superior to any of the handful of “better chip amps” I purchased some time before the enterprise.
So… again, as my uncle would say, “what's the poin?” (sic: no T)
⋅-=≡ GoatGuy ✓ ≡=-⋅
Yes, let it be as challenge at unity gain stable and inverting connection!
No, not without any caps, but without any cap at the C1 position of yours schematic.
Only external circuitry relative to the shown opamp structure.
Just a simulator compete for curiosity how and how much...
PS
Oh, yes, and let's add aperiodic step response at, say, 2 Volts input square wave.
PPS
Also i'ld set R1&R2 smaller, say 22 Ohm and R5 say 1k for making things even complicated.
Attachments
Last edited:
No-no, nothing about "universally" applicable.
That's nonsense with something having at least 120 dB gain at 10 kHz.
This is really beast to be tamed anyway.
So, let's try? For curiosity, of course!
Well, if one has a goal of applying to existing circuits as a drop-in “better” (or at least more-easily-talked-up) replacement, then by definition, it can not be engineered to not need compensation.
Because … many an existing circuit, having used op-amps with parametric compensation will have chosen the best compensation values, and have them in-circuit.
Dropping in a device with 'universal' compensation changes the existing design's criteria-base!
⋅-⋅-⋅ Just saying, ⋅-⋅-⋅
⋅-=≡ GoatGuy ✓ ≡=-⋅
I agree wholeheartedly with № 1.
And № 2 is equally defensible.
One thing's for sure — and that is that the string-of-modest-gain op-amp testing methods (along with a quad channel scope's marvelous comparative functions) quickly and visually identifies op-amp non-linearities. That, and a very, very pure sine-wave signal generator.
Best of The Art to you, friend.
⋅-=≡ GoatGuy ✓ ≡=-⋅
How would a discrete output stage be done? For example the Topping A30 has an LM337D2 going to the output RCA's. Would that be considered a discrete output?
Also, can an OPA627 with a BUF634 be replaced with an AD744 for better noise, distortion while still using a discrete output to drive big loads like 300 ohm easily? I've been thinking about paralleling 4 of them but it's not exactly cost effective.
Thanks guys
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.