Music. Anything with a wide dynamic range.
Scott,
Does that mean you are responsible for all competent designs? Must be a real burden!
I would like to be in on this as well. Let us know if there's a thread or such. The device is $100 on Mouser, and $5 for transformer or so.
Thermal distortion used to be almost the major distortion in op amps like the uA741. Yes! I measured it myself with an added plug-in to my TEK577 curve tracer back in 1974. The only thing that kept it in control was high amounts of negative feedback. Of course, over the decades, designers, perhaps embarrassed by it, worked hard to keep it down, but the close proximity of the internal parts inside the IC chip and the usual necessary Class B output, put plenty of pressure on thermal balance. I'm sure it is not as bad today, but who knows, IF we don't have a standard measurement for it? This is one main intrinsic advantages of discrete designs over IC's.
How about this thought experiment: I have this great amplifier with the usual Cdom around the Vas stage. It has an OL BW of 20kHz at an OL gain of say 60dB.
Now I am increasing the Vas stage load impedance, with the same Cdom. I believe the result would still be 60dB @ 20kHz, but increasing OL gain toward lower frequencies, like 72dB at 10kHz and 84dB at 5kHz etc, and I continue until the OL BW is reduced to 10Hz, where it has a very high gain.
I believe that the 2nd amp is much more linear and transparent in the main audio listening frequency range. Or is there something I overlook?
Jan
Man* on a DIY site makes an RIAA pre-amp out of bits he had lying around** and it both measures and sounds great. WNTL. Raise a glass to the fact that the spirit of DIY survives despite everything
*OK, he's a ringer, but it's still DIY
** Not all of us are luckily enough to have such a good stash, but there are affordable options even in through whole.
I'm in the middle of attempting to replicate an experiment performed by Richard Marsh and described on this website: purchase a boatload of complementary JFETs from a dozen different Chinese sellers on AliExpress, keep the batches separate*, and test them to find out whether any are legitimate or useable or both. It turns out this is inexpensive, 100 pcs for $15 with free shipping is about the median price. Why not give it a try; if it happens to work out you get difficult-to-source parts at low cost. If it doesn't, you haven't spent big money.
*so you know which vendors are worthy of a re-order
Music. Anything with a wide dynamic range.
Scott,
Does that mean you are responsible for all competent designs? Must be a real burden!
Strange conclusion. Dragging out the uA741 again does not teach much except that someone discovered something they could fix 40yr. ago. It is trivial now to demonstrate sub uV errors driving 600 Ohms from an SOIC op-amp. The poor 741 has to bear the burden of all that is wrong with IC op-amps.
Last edited:
Mark,
You could save your money (and time and effort) and donate them to me.
We have checked out multiple "reliable and respected" vendors on Chinese Ebay (TaoBao) known to us
for the usual 2SK170/2SJ74 of any grade, as well as 2SK369's.
All fakes, especially the P-JFET.
The odd real ones are matched with corresponding ebay prices.
If you are lucky you might still find genuine 2SK246/2SJ103 etc.
Maybe some other seldom used N devices.
But then don't let me stop you having fun in hunting.
Cheers,
Patrick
Digikey has 25k+. Interesting thing is that on a per opamp basis, it's twice as expensive as an NE5532! Is this the market for repair of legacy equipment?
Jan
Last edited:
Wayne,
I think you miscounted! They list the same device two or three ways as in reel, cut tape and individual. My count shows only about 69,000 or so pieces. Prices start just above 11 cents for a reel of them and 41 cents for a single piece. Only 27,000 of the NE5534 and at 2-3 X the price!
However the last time I measured a 741 it exceeded all of the specifications by just a bit for some like GBW but a lot for Vos. Enough that I suspect it has been improved while being shrunk to match current processes.
ES
Last edited:
Yes sure there are, but to pay extra for it compared to other run of the mill parts?
A classic. You can get today a 555 CMOS timer (TLC555) that takes less that 250uA standby current, is 10x faster and 50% cheaper than the original bipolar 555, still in production.
- Status
- Not open for further replies.