Hi,
For the last few weeks, TI's site lists the 4562 and 49720 as out of stock. On top of that, the prices for both have increased by around 80% since I last saw them in stock. They are suggesting the OPA2891 as a replacement (it's a very different OA and my limited experiments suggest that although it's quiet with low source Z, its THD performance is nowhere near the 4562 in most situations).
They made the 4562 / 49720 EOL a few years back, but reintroduced them. Are they gone for good this time? I noticed that they had a few thousand of each just before the price increase... They were snapped up. Maybe someone knew the increase / EOL was coming?
For the last few weeks, TI's site lists the 4562 and 49720 as out of stock. On top of that, the prices for both have increased by around 80% since I last saw them in stock. They are suggesting the OPA2891 as a replacement (it's a very different OA and my limited experiments suggest that although it's quiet with low source Z, its THD performance is nowhere near the 4562 in most situations).
They made the 4562 / 49720 EOL a few years back, but reintroduced them. Are they gone for good this time? I noticed that they had a few thousand of each just before the price increase... They were snapped up. Maybe someone knew the increase / EOL was coming?
LM4562 easily available from major suppliers. Prices have gone up for many many things. The OPA2891 has a lot more supply current (upto 10mA per amp - it won't run cool enough at +/-18V supplies, basically), and has even more current noise than the LM4562, so for some audio applications its worse than LM4562, not better. The OPA2891 is more like AD797 I think, its designed for higher bandwidth than audio AFAICT.
Given the long lead times on some chips you get occasional shortages sometimes if they've sold faster than expected.
Given the long lead times on some chips you get occasional shortages sometimes if they've sold faster than expected.
On 11th of December, I ordered 1,000 49720s from Digikey @ £1.48 each
Today, they are £3.78 for a 1,000-off: https://www.digikey.co.uk/en/produc...8?s=N4IgTCBcDa4JwDYC0AWAzHA7GJBWJAcgCIgC6AvkA
That's 155% increase in under a month. TI's site shows similar increases. It occurred very recently, around the same time that the stock disappeared (either they pulled the stock or word got out to an OEM and they snapped up the whole lot - but it's weird that both types went around the same time). Something is happening and TI are not being open about it. The 2891 only replaces the 4562 / 49720 in a very narrow range of applications. It's nowhere near a real alternative.
edit: Digikey have the SOIC version at a much better price. For how long? Are TI pulling the DIP versions? Why aren't they communicating what's happening?
Today, they are £3.78 for a 1,000-off: https://www.digikey.co.uk/en/produc...8?s=N4IgTCBcDa4JwDYC0AWAzHA7GJBWJAcgCIgC6AvkA
That's 155% increase in under a month. TI's site shows similar increases. It occurred very recently, around the same time that the stock disappeared (either they pulled the stock or word got out to an OEM and they snapped up the whole lot - but it's weird that both types went around the same time). Something is happening and TI are not being open about it. The 2891 only replaces the 4562 / 49720 in a very narrow range of applications. It's nowhere near a real alternative.
edit: Digikey have the SOIC version at a much better price. For how long? Are TI pulling the DIP versions? Why aren't they communicating what's happening?
Last edited:
Is there any chance you can use some other device? While the PSRR on the 4562 family is outstanding, it seems in the intervening 15 yrs since it was introduced, many fine opamps have come along. The OPA1641/2 and 1651/2 come to mind for example.
There are many other good op amps. But ..... for us through hole hobbyists the 49720/4562 in a DIP package is a huge convenience. I sincerely hope it is not going EOL
I will have no tears when the 4562 goes away as it cost our company a lot of money because of uncontrolled popcorn noise. Also it has almost zero RF immunity.
There a many alternatives and most of them have quite similar or even much better specs
There a many alternatives and most of them have quite similar or even much better specs
IHMO 75dB at 10kHz for PSRR- is nothing to write home about. The 100dB you can have with a OPA1622 is.
LM4562 is the proof that a part needs no special features (and even may have quite negative aspects) to become famous in audio.
OPA1642 really is a better alternative.
OPA1642 really is a better alternative.
For use in todays RF ridden world audio opamps should be immune to RF.
Also DIL has had its time. We are in an SMD world for at least 25 years and with the high quality of recent opamps and pretty exact “designing by simulation” opamp rolling will likely also become something of the past. We had fun with exact methods like empirical testing, gut feeling and measuring by ear but time goes on! 😉
Also DIL has had its time. We are in an SMD world for at least 25 years and with the high quality of recent opamps and pretty exact “designing by simulation” opamp rolling will likely also become something of the past. We had fun with exact methods like empirical testing, gut feeling and measuring by ear but time goes on! 😉
Last edited:
I've been using the 4562 for over 10 years. It has caveats: RFI (you need to work with someone who understands EMC, although it can be annoying when testing assemblies with the enclosure removed) and popcorn (I've written about it here several times before - it's a total PITA and I use a jig that has ZIF sockets in order to screen for it). However, the 4562 / 49720 works better (lower THD+N) than the 1612 and 1642 in many areas for the designs that I work with. It all depends on the application. The 2891 offers better noise (if source Z is ultra low), but I haven't been able to get near the 4562's THD performance in any of the circuits I've tried it on. In some applications I've seen wild distortion from the 2891, which is odd as it's definitely stable and passes a perfect square. The 4562 is also rugged. I haven't seen one fail outside of user error. The same cannot be said for some of the newer OAs...particularly the CMOS-based ones, which need very careful handling and external TVS diodes. If you consider the price before the increase + all-round performance, the 4562 was a strong contender in many applications. The vicious price increase changes this. My complaint is the opaque nature of how TI are handling this. We've been using the 4562 for over 10 years. There have been no mail-outs discussing whether it's being phased out or why the price has increased so drastically. Let's say you make a product that has 50 of these ICs in it. TI just added around $75 to the cost of your product, with no warning. TI are behaving in an arrogant, laissez-faire manner IMO. The 2891 really isn't a substitute by any stretch.
Thr OPA1612 is s great product for low source impedance designs - 0.9nV/rt Hz. I used it on a MC head amp with more than passable performance. The 1642 is my go to GP opamp. Low distortion (300 ppb quoted) and all the benefits of JFET IPS. For MM work and line stages it’s hard to beat and is easily up to the AD797 which I have also used.
It might be with the 4562 TI are trying to shift users to more profitable parts. That said, having worked in semis for 2 decades, it could also be TI forward built die stock and it’s coming to an end so they’ve upped prices. Typically this would be done because a process in the fab where these are made is going EOL. Semiconductors is an exceedingly demanding business.
It might be with the 4562 TI are trying to shift users to more profitable parts. That said, having worked in semis for 2 decades, it could also be TI forward built die stock and it’s coming to an end so they’ve upped prices. Typically this would be done because a process in the fab where these are made is going EOL. Semiconductors is an exceedingly demanding business.
I used it as a low power regulator (5-10 mA) because of its outstanding PSRR. The picture below is a partial snapshot of the data sheet. Most opamps have unbalanced PSRR performance so it’s not unique to the 4562.
EOL is a pain. Customers used to roast us for doing it but if the process goes the parts will only be transferred to a new process if it’s economically viable. Ditto parts in a portfolio that aren’t selling enough volume, or just aren’t making margin targets.
Don't shoot me and call me a luddite. But this price increase has me a on a new hunt for a PDIP part. Is there a good low noise PDIP op-amp suitable for MM/MC pre use that can replace the 4562/49720? That one WAS really great in that slot at a decent price. I've built a few pre-amps with it and it sounds remarkably good. Not perfect but really, really good, in fact. I know better SM devices can raise the bar. I've done that, too. But.... I need something easy for anyone to hand assemble that won't break the bank. So a good PDIP is needed. Does anyone have any good suggestions?
Redesign the PCB for contemporary SOIC and contemporary opamps (= younger than 10 years) and you have many good opamps to choose from. Time goes on, there is no stopping that. Many new opamps do not exist at all in PDIP/DIL or they won't exist anymore in a while so in fact you are missing out or will be missing out. Unless you buy thousands of them right now, that is an option. It is either live in the now or in the future but not in the past.
What about OPA828? It is SOIC and very good. Probably in the top 3 of all audio opamps even so probably the devices will perform even better than expected. Yes they're singles, you can sell that as better channel separation 🙂
What about OPA828? It is SOIC and very good. Probably in the top 3 of all audio opamps even so probably the devices will perform even better than expected. Yes they're singles, you can sell that as better channel separation 🙂
Last edited:
@jean-paul I understand that and would do it if just for me or for a commercial build. But, unfortunately, SOIC is not an option under current circumstances.
Luckily, I have 50x 4562 and 70x 49720 on the way, that should last me for quite a while. "Old price", of course.