Reply by February 24, 20182018-02-24
On Sunday, February 25, 2018 at 4:26:37 AM UTC+11, dagmarg...@yahoo.com wrote:
> On Saturday, February 24, 2018 at 11:49:58 AM UTC-5, John Larkin wrote: > > Phil might like this one: > > > > http://www.linear.com/product/LT6018 > > > > This is a 1 nv-noise-class opamp with a 1 Hz 1/f corner. It doesn't > > seem to be a chopper, but that may be hidden inside. Input noise > > current is fairly high. > > Chopper. The noise density plot gives it away @ 100kHz-10MHz. But still > pretty impressive to my somewhat chopper-naive eyeballs. (I've never had > to fool with truly tiny signals.)
Seems unlikely. None of the noise plots on page 7 of the data sheet suggest that to me - pseudo-random chopping waveforms can spread out chopper noise, but those curves are pretty smooth. The simplified schematic on page 11 certainly doesn't suggest any chopper element. -- Bill Sloman, Sydney
Reply by February 24, 20182018-02-24
On Saturday, February 24, 2018 at 2:42:07 PM UTC-5, Gerhard Hoffmann wrote:
> Am 24.02.2018 um 18:26 schrieb dagmargoodboat@yahoo.com: > > On Saturday, February 24, 2018 at 11:49:58 AM UTC-5, John Larkin wrote: > >> Phil might like this one: > >> > >> http://www.linear.com/product/LT6018 > >> > >> This is a 1 nv-noise-class opamp with a 1 Hz 1/f corner. It doesn't > >> seem to be a chopper, but that may be hidden inside. Input noise > >> current is fairly high. > > > > Chopper. The noise density plot gives it away @ 100kHz-10MHz. But still > > pretty impressive to my somewhat chopper-naive eyeballs. (I've never had > > to fool with truly tiny signals.) > > That does not look like a chopper, more like the usual noise peak > that plagues 3 stage opamps. The LT1028 has it at 300 KHz. > ADA4898 and AD797 do not have it.
I said "chopper" because somehow I thought I saw no 1/f rise, and thought the 200kHz-15MHz band was 1/f heterodyned up with some fancy hardware, maybe spread spectrum-ish. But, looking again there it is, 1/f right where it should be. Rats. I am duly chastened.
> The input stage must be run at high current to get the low voltage > noise. That brings high current noise automatically. > > I have made a low voltage noise amplifier (220pV/rtHz) by paralleling > 20 opamps (10 ADA4898-2 pairs). The current noise is awful. But that's > not bad if the input source is low impedance.
Yes, and thanks for posting those. Cheers, James Arthur
Reply by Winfield Hill February 24, 20182018-02-24
John Larkin wrote...
> > Winfield Hill wrote: >> George Herold wrote... >>> On February 24, 2018, dagmarg...@yahoo.com wrote: >>>> On February 24, 2018, John Larkin wrote: >>>>> >>>>> Phil might like this one: >>>>> http://www.linear.com/product/LT6018 >>>>> >>>>> This is a 1 nv-noise-class opamp with a 1 Hz 1/f corner. >>>>> It doesn't seem to be a chopper, but that may be hidden >>>>> inside. Input noise current is fairly high. >>>> >>>> Chopper. The noise density plot gives it away @ 100kHz-10MHz. >>>> But still pretty impressive to my somewhat chopper-naive >>>> eyeballs. (I've never had to fool with truly tiny signals.) >>> >>> Huh, wouldn't a chopper be peakier? I was thinking it >>> could be shot noise from some gain stage * some resistance. >>> It would be annoying if you wanted to use it past 100k Hz. >> >> It must have a bias-current-cancelling input stage, like >> the LT1028, which has 25nA input-bias current, but likely >> 250nA actual base current, making about 1pA/rt-Hz noise >> (see AoE pg 327). Adding a trace for the L6018 on the >> integrated voltage noise graph (Fig. 5.54 pg 338), shows >> that it's pretty close to the LT1028, and not much like >> the other choppers. Also, adding pencil traces on the >> voltage- and current-noise density graphs (Fig 8.60, >> pg 528), shows it coming close to matching the LT1028. >> Yes, all three plots do show better 1/f performance. >> But I suspect it's not a chopper. > > It looks like giant NPNs in the input stage, with very > good Ib cancellation and trimming. Input capacitances > are high. The noise hump does resemble the LT1028.
Yes, input capacitances are higher than the LT1028. Lowering the current density helps to move the 1/f break point lower. Too bad it doesn't come in a TO-5 package. I sawed the top off a metal-case LT1028 to get the die photo in Fig 8.59 page 527, look how big the input transistors are there. -- Thanks, - Win
Reply by Phil Hobbs February 24, 20182018-02-24
On 02/24/2018 11:49 AM, John Larkin wrote:
> > > Phil might like this one: > > http://www.linear.com/product/LT6018 > > This is a 1 nv-noise-class opamp with a 1 Hz 1/f corner. It doesn't > seem to be a chopper, but that may be hidden inside. Input noise > current is fairly high. > >
Looks awesome. I'll try putting it in a compound-amp loop with an LM6171 and see how it works as a 100X instrument amp. My HP89441A is supposed to be here Tuesday, so that'll come in handy. Cheers Phil Hobbs -- Dr Philip C D Hobbs Principal Consultant ElectroOptical Innovations LLC / Hobbs ElectroOptics Optics, Electro-optics, Photonics, Analog Electronics Briarcliff Manor NY 10510 http://electrooptical.net https://hobbs-eo.com
Reply by John Larkin February 24, 20182018-02-24
On 24 Feb 2018 12:03:08 -0800, Winfield Hill
<hill@rowland.harvard.edu> wrote:

>George Herold wrote... >> >> On February 24, 2018, dagmarg...@yahoo.com wrote: >>> On February 24, 2018, John Larkin wrote: >>>> >>>> Phil might like this one: >>>> >>>> http://www.linear.com/product/LT6018 >>>> >>>> This is a 1 nv-noise-class opamp with a 1 Hz 1/f corner. >>>> It doesn't seem to be a chopper, but that may be hidden >>>> inside. Input noise current is fairly high. >>> >>> Chopper. The noise density plot gives it away @ 100kHz-10MHz. >>> But still pretty impressive to my somewhat chopper-naive >>> eyeballs. (I've never had to fool with truly tiny signals.) >> >> Huh, wouldn't a chopper be peakier? I was thinking it >> could be shot noise from some gain stage * some resistance. >> It would be annoying if you wanted to use it past 100k Hz. > > It must have a bias-current-cancelling input stage, like > the LT1028, which has 25nA input-bias current, but likely > 250nA actual base current, making about 1pA/rt-Hz noise > (see AoE pg 327). Adding a trace for the L6018 on the > integrated voltage noise graph (Fig. 5.54 pg 338), shows > that it's pretty close to the LT1028, and not much like > the other choppers. Also, adding pencil traces on the > voltage- and current-noise density graphs (Fig 8.60, > pg 528), shows it coming close to matching the LT1028. > Yes, all three plots do show better 1/f performance. > But I suspect it's not a chopper.
It looks like giant NPNs in the input stage, with very good Ib cancellation and trimming. Input capacitances are high. The noise hump does resemble the LT1028. -- John Larkin Highland Technology, Inc lunatic fringe electronics
Reply by Winfield Hill February 24, 20182018-02-24
George Herold wrote...
> > On February 24, 2018, dagmarg...@yahoo.com wrote: >> On February 24, 2018, John Larkin wrote: >>> >>> Phil might like this one: >>> >>> http://www.linear.com/product/LT6018 >>> >>> This is a 1 nv-noise-class opamp with a 1 Hz 1/f corner. >>> It doesn't seem to be a chopper, but that may be hidden >>> inside. Input noise current is fairly high. >> >> Chopper. The noise density plot gives it away @ 100kHz-10MHz. >> But still pretty impressive to my somewhat chopper-naive >> eyeballs. (I've never had to fool with truly tiny signals.) > > Huh, wouldn't a chopper be peakier? I was thinking it > could be shot noise from some gain stage * some resistance. > It would be annoying if you wanted to use it past 100k Hz.
It must have a bias-current-cancelling input stage, like the LT1028, which has 25nA input-bias current, but likely 250nA actual base current, making about 1pA/rt-Hz noise (see AoE pg 327). Adding a trace for the L6018 on the integrated voltage noise graph (Fig. 5.54 pg 338), shows that it's pretty close to the LT1028, and not much like the other choppers. Also, adding pencil traces on the voltage- and current-noise density graphs (Fig 8.60, pg 528), shows it coming close to matching the LT1028. Yes, all three plots do show better 1/f performance. But I suspect it's not a chopper. -- Thanks, - Win
Reply by Gerhard Hoffmann February 24, 20182018-02-24
Am 24.02.2018 um 18:26 schrieb dagmargoodboat@yahoo.com:
> On Saturday, February 24, 2018 at 11:49:58 AM UTC-5, John Larkin wrote: >> Phil might like this one: >> >> http://www.linear.com/product/LT6018 >> >> This is a 1 nv-noise-class opamp with a 1 Hz 1/f corner. It doesn't >> seem to be a chopper, but that may be hidden inside. Input noise >> current is fairly high. > > Chopper. The noise density plot gives it away @ 100kHz-10MHz. But still > pretty impressive to my somewhat chopper-naive eyeballs. (I've never had > to fool with truly tiny signals.)
That does not look like a chopper, more like the usual noise peak that plagues 3 stage opamps. The LT1028 has it at 300 KHz. ADA4898 and AD797 do not have it. The input stage must be run at high current to get the low voltage noise. That brings high current noise automatically. I have made a low voltage noise amplifier (220pV/rtHz) by paralleling 20 opamps (10 ADA4898-2 pairs). The current noise is awful. But that's not bad if the input source is low impedance. cheers, Gerhard
Reply by George Herold February 24, 20182018-02-24
On Saturday, February 24, 2018 at 12:26:37 PM UTC-5, dagmarg...@yahoo.com wrote:
> On Saturday, February 24, 2018 at 11:49:58 AM UTC-5, John Larkin wrote: > > Phil might like this one: > > > > http://www.linear.com/product/LT6018 > > > > This is a 1 nv-noise-class opamp with a 1 Hz 1/f corner. It doesn't > > seem to be a chopper, but that may be hidden inside. Input noise > > current is fairly high. > > Chopper. The noise density plot gives it away @ 100kHz-10MHz. But still > pretty impressive to my somewhat chopper-naive eyeballs. (I've never had > to fool with truly tiny signals.)
Huh, wouldn't a chopper be peakier? I was thinking it could be shot noise from some gain stage * some resistance. It would be annoying if you wanted to use it past 100k Hz. George H.
> > Cheers, > James Arthur
Reply by February 24, 20182018-02-24
On Saturday, February 24, 2018 at 11:49:58 AM UTC-5, John Larkin wrote:
> Phil might like this one: > > http://www.linear.com/product/LT6018 > > This is a 1 nv-noise-class opamp with a 1 Hz 1/f corner. It doesn't > seem to be a chopper, but that may be hidden inside. Input noise > current is fairly high.
Chopper. The noise density plot gives it away @ 100kHz-10MHz. But still pretty impressive to my somewhat chopper-naive eyeballs. (I've never had to fool with truly tiny signals.) Cheers, James Arthur
Reply by John Larkin February 24, 20182018-02-24

Phil might like this one:

http://www.linear.com/product/LT6018

This is a 1 nv-noise-class opamp with a 1 Hz 1/f corner. It doesn't
seem to be a chopper, but that may be hidden inside. Input noise
current is fairly high.


-- 

John Larkin         Highland Technology, Inc

lunatic fringe electronics