I've played enough with slow op-amps circuits (<1 MHz) and now want to upgrade to faster toys ...
(which means I have zero experience with fast op-amps)
So my requirements were quite simple :
- Bandwidth ~100 MHz
- Gain bandwidth ~200 MHz+
- DIP package (then I can use my prototype boards and postpone switch to smd protoboards)
- Do not really care about capacitive load
- Not expensive
After lurking couple of hours on digikey I was amazed that op-amp with 12 GHz bandwidth is cheaper then most ~100 MHz class op-amps in DIP package (2.1 eur vs ~2.5 eur)
In the end I found some of them :
And that's it ... What other models you can recommend ?
DIP isn't a particularly good idea for those. Quite a bit of parasitic L and C - having used LM7171 in DIP before, I found that it wouldn't even work in a socket - oscillated like crazy, I had to desolder the socket and put the chip in directly, at which point it worked. This was with a relatively good layout, too. You'll have a hell of a time with it on protoboard without a ground plane.
Few more impressive ones I've used, though:
ADA4817, 1 GHz, JFET - 2pA leakage at DC!
AD8000 and close relative AD8003, 1.5 GHz, 4100 V/µs CFA
THS3001, only 420 MHz but a whopping 6500 V/µs, high output current with thermal pad, up to ±16V!, CFA, costs a fortune, nice for things like high-frequency function generator output stages as it can do the full usual voltage and current range with proper thermal management
Note that almost all amps in that range will be CFB (current feedback), not voltage type. This usually amounts to adding a series resistor or two on the inputs, turning voltage into current. A consequence of this is, the inputs will draw current from whatever the source is. Which makes sense and is unavoidable at RF, but may seem unfamiliar from LF because LF circuits have the luxury of arbitrarily high impedances.
Other consequences are, the bandwidth is more-or-less independent of gain, so you don't have a GBWP as such.
As mentioned, parasitic L and pin/trace delay are a big deal, and even SOIC are looking ponderously large in the 1GHz+ fT range.
Tim
and now want to upgrade to faster toys
And that's it ... What other models you can recommend ?
this is an open question thread. the real question is, why upgrade? there are gazzillion of high speed opamps out there each one with there own weakness and strength. if you really dont care about that, you already found the answer...
Got AD8055A, apparently it's a bit tricky to use them on breadboard - decoupling is just to much of a hassle. And yes, it oscillated at first, but not that bad. In the end managed to get G=2 at 66 MHz on breadboard, not bad.
AD830 ...Not that fast, but interesting as differential amplifier...360V/µs, 85Mhz unity gain bandwith.
I notice you're using the Rigol DS1054. Is it hacked to give 100MHz bandwidth? If not then the displayed amplitude will be lower than the actual value and you won't be able to tell if there's any distortion.
At 66 MHz that'll be true even with 100 MHz bandwidth.
I have a reel of AD8052 SO8 in the mail.
I am an "opamp virgin"
, and bought those for playing.
So i hope they are ok for some opamp fun , i have only played a bit with some LM324 before.
Anyone cares to comment on those ?
I'll prob try to sell some, as i never think i can use a reel.
/Bingo
A whole reel?
Yupp
I hope they can be used as general opamps
/Bingo
Well ... I still have 800+ LM324 so14 left, but they were not rail-to-rail.
I notice you're using the Rigol DS1054. Is it hacked to give 100MHz bandwidth? If not then the displayed amplitude will be lower than the actual value and you won't be able to tell if there's any distortion.
Yeah, it's hacked.
At 66 MHz that'll be true even with 100 MHz bandwidth.
Not really, indeed I cannot really say anything about distortion, but amplitude should be sane.
Here's frequency response chart through breadboard (light blue), coax from signal generator to oscilloscope directly (dark blue), sync signal (pink). It's a linear sweep from 1 to 100 MHz.
So it's almost linear up to 20 MHz, then it goes up with peak around 45 MHz, and goes down with a peak at 50 MHz. Then another peak around 60 MHz, and goes down ...
A whole reel?
Yupp
Really? That would be a couple hundred euros, at least...
Um, do you have any termination on that cable?
800+ LM324 you are nuts, a reel of AD8052 triples that. next order... a reel of high slewrate opamp...
A whole reel?
Yupp
I hope they can be used as general opamps
/Bingo
Well ... I still have 800+ LM324 so14 left, but they were not rail-to-rail.
Well, you can sell some of them to me - I'm in Denmark too.
800+ LM324 you are nuts, a reel of AD8052 triples that. next order... a reel of high slewrate opamp...
Well i got the LM324's on *Bay back when "EU-land" dumped the NON ROHS comp's for next to nothing.
The AD8052's are also NON ROHS , but as a hobbyist it doesn't matter.
@OP Sorry for "the noise"
/Bingo
Well i got my reel of AD8052's
What's a cheap and easy way to generate a usable (silent) +/- 5v from a 5v or 9..12v psu ?
Not an expensive chip, for +5$ if possible
/Bingo
Well i got my reel of AS8052's
What's a cheap and easy way to generate a usable (silent) +/- 5v from a 5v or 9..12v psu ?
Not an expensive chip, for +5$ if possible
/Bingo
The most simple way is to use 10v psu, threat 0 of psu as -5v and 10 as 5v then use voltage divider to get -5, 0, +5, and put a voltage buffer on 0v (you already got a reel of opamps)
Well i got my reel of AS8052's
What's a cheap and easy way to generate a usable (silent) +/- 5v from a 5v or 9..12v psu ?
Not an expensive chip, for +5$ if possible
/Bingo
The most simple way is to use 10v psu, threat 0 of psu as -5v and 10 as 5v then use voltage divider to get -5, 0, +5, and put a voltage buffer on 0v (you already got a reel of opamps)
Something like this ? - I'm an analog noob
https://www.eevblog.com/forum/beginners/op-amp-powered-by-voltage-divider//Bingo
This one should work just fine for non precision stuff.
PS. You probably cannot use AD8052 in this configuration because it's current controlled, you need voltage controlled op-amp for buffer.