EEVblog Electronics Community Forum
Electronics => Projects, Designs, and Technical Stuff => Topic started by: hanakp on November 17, 2020, 04:32:43 pm
-
Over the last few years, I've used TLC2272 in several projects, because it has
-low noise 9 nV/sqrt(Hz)
-rail-to-rail output
-decent GBW (2.2 MHz) and slew rate (3.6 V/us)
-tolerates high load capacitance (up to 10 nF)
-works in both single and dual supply environments (Vspan up to 16 V)
-yet is surprisingly cheap
Now I'm looking for a single opamp with as close parameters as possible, preferably in 5-pin package. So far, the closest one I've found is AD8665, but it can handle capacitive loads only to 1 nF. On TI side, OPA171 is close, but it has lower slew rate and again only 1 nF load. Faster TI alternative is OPA991, but it still can handle only up to 4 nF.
I know it's a long shot, but given TLC2272's popularity, maybe someone knows a better replacement.
-
Why do you insist on single channel? Presumably PCB space or power consumption? I doubt you'll find anything as cheap.
I can't suggest anything, other than the parametric search on manufacture and distributor websites.
-
TLC272? Should be pretty close. ;)
-
OP wanted SOT23-5 :P
-
Yes, there are power consumption and board space concerns. Moreover, I need to minimize crosstalk between individual amps, so I'd like to use 1-channel ones. The amps will feed Sallen-key filters which represent significant capacitive load. Well, I'll try OPA991 first and see how it fares.
-
I think SO8 duals could take less space than singles. Are you sure that crosstalk is a problem or merely paranoid? ;)
-
The Sallen-key filters usually don't represent a significant capacitive load, unless the resistors are chosen to be very small. There may be a capacitor at the output, but further down the circuit there is additional resistance. The low pass filter is usually quite forgiving, the high pass filter is more tricky / prone to oscillation. The low pass filter profit from a low open loop output impedance that often comes with good capacitive drive. So indirectly it makes sense to look for capacitive drive if the impedance values are not given, but it is not an essential parameter.
The TLC071 may be a possible alternative: however not SOT23-5, but only VSSOP8. It may not be high capacitive drive, but I don't think this is really needed. The higher GBW may make up for it in the filter circuit.
-
Yes, I already confirmed crosstalk is an issue at 50 kHz (measurement on Audio Precision APX525) and the circuit needs to work up to 150 kHz, so it will get only worse. Another concern is symmetry - the system will have several channels and I want to make them identical to have the same parasitics on all of them.
As for the filters, I unfortunately need high-pass ones. The resistors are already in the 40-70 k\$\Omega\$ range and I don't want to increase them anymore (so I could decrease capacitors), because it increases noise, too.
Oh well, if OPA991 doesn't pan out, I guess I'll have to ditch the SOT23-5 limitation, I remember I saw some good Analog Devices candidate in SO8.
-
The high pass is a little prone to oscillation / ringing, but it still does not need to drive capacitive load.
Using smaller resistors could likely reduce the cross talk. 40-70 K is a bit high to get good performance at 50 kHz.
Chances are it can work better with lower impedance, like a factor 10 down.
-
A quick search of low noise, low input bias current, and rail-to-rail output TI parts in the SOT23-5 package which operate at 16 volts and higher includes OPA145, OPA197, and OPA140 in order of gain-bandwidth product.
I like the TLC2272 for the same reasons you do and often used it as a rail-to-rail output replacement for the LM358/LM324. But beware of its 16 volt absolutely maximum supply voltage; they are not kidding about that on their LinCMOS parts.
-
@Kleinstein: Sallen-Key high-pass filter input has two capacitors in series that are connected to opamp's virtual ground. So in essence, they behave like a frequency-dependent capacitance connected to the ground. These problems are well known and you need to drive them properly:
https://www.eeweb.com/hidden-hazards-in-the-sallen-key-2nd-order-high-pass-active-filter/ (https://www.eeweb.com/hidden-hazards-in-the-sallen-key-2nd-order-high-pass-active-filter/)
In my project, the capacitors are in 6-10 nF range already and I can't increase them much, because I need to use NP0 dielectric. With TLC2272s, 0603 components and proper layout, the drivers and filters work surprisingly well up to 200 kHz and 1 Vrms.
@David Hess: none of them is specified up to 10 nF like TLC2272, the best is OPA140 with meager 1.5 nF. I already checked most AD and TI parts, but I thought someone may know about better alternative from the 15+ other opamp producers...
-
Following the article (since Michael Steffes knows his stuff!), note that resistor R1 will isolate the capacitances C1, C2 from the op-amp output. R1 is ~400 ohms or larger in the example designs he shows. This is usually enough resistance to not cause loading trouble for the op-amp, so you don't need a heavy C-load stable part. Accordingly, I'd start with the OPA197, but that's because I really like it (and its same-die? relatives 196/191/192) so I have it on hand and always start there.
Other things worth noting: a Sallen-Key filter may not be the preferred choice for you if lowest distortion is your goal. Operating the amplifier in inverting mode rather than non-inverting is always going to be better for distortion. (You can also bootstrap the supplies of a non-inverting amplifier if you have to, but that's enough work that you don't do it unless you have to.) That said, the most common inverting filter topology (MFB) is not great in high-pass form either....
If you want more than 10nF NP0, there are some 1206 100nF NP0 parts out there that are very, very useful, and high but bearable in cost at around US$0.25 each. Above that it gets ugly.
-
@David Hess: none of them is specified up to 10 nF like TLC2272, the best is OPA140 with meager 1.5 nF. I already checked most AD and TI parts, but I thought someone may know about better alternative from the 15+ other opamp producers...
I missed that but where does the high capacitive load driving requirement come from? It is not inherent or relevant to Sallen-Key filters because the impedances involved have considerable series resistance.
-
How about the TLV9151? It doest quite meet all of the criteria in the original post, but design is all about compromise and it's cheap.
https://www.ti.com/lit/ds/symlink/tlv9151.pdf (https://www.ti.com/lit/ds/symlink/tlv9151.pdf)