Reply by Allan Herriman February 24, 20162016-02-24
On Wed, 24 Feb 2016 11:50:10 -0800, Winfield Hill wrote:

> Allan Herriman wrote... >> >> I used to use Fujitsu relays when I worked for Fujitsu. >> Here though, it's the choice of a relay vs a US$0.10 FET that takes no >> board space, withstands any amount of vibration and will last forever >> in a corrosive atmosphere. [hyperbole] > > Keep in mind the FET has about 100pF of capacitance at 0V.
Thanks Win. In general that is a very good point, however in this application the voltage doesn't change very quickly and C dv/dt won't come close to 100pA. Nice book, BTW. Regards, Allan
Reply by Allan Herriman February 24, 20162016-02-24
On Wed, 24 Feb 2016 08:15:30 -0700, Jim Thompson wrote:

> On 24 Feb 2016 09:33:39 GMT, Allan Herriman <allanherriman@hotmail.com> > wrote: > > [snip] >> >>Actually, I'm still hopeful of finding an opamp that does exactly what I >>want in one chip. >>Contemporary 74xx series CMOS logic gates have the sort of ESD >>protection I want (i.e. input voltages > Vcc without drawing a lot of >>current). Why can't an opamp? Are the processes so different? Does >>that sort of ESD protection have some bad effect on analog circuits >>(e.g. leakage current)? >> >>Regards, >>Allan > > Didn't your original post say even while unpowered? The only way to do > that is with specially-designed ESD, such as is found in many bus > transceivers. Off-hand I don't know of an OpAmp so constructed. > > ...Jim Thompson
Hi Jim, Yes, the requirement is |Iin| < 100pA under the conditions of 0 < Vin < 4 and 0 < Vsupply < 5V. BTW, The Microchip MCP6002 family has that sort of ESD protection. Its abs max specs mean that I can't use it (Vin max is VDD + 1V) but I'm hoping there are others out there that I can use. http://www.microchip.com/wwwproducts/en/MCP6002 Regards, Allan
Reply by Winfield Hill February 24, 20162016-02-24
Allan Herriman wrote...
> > I used to use Fujitsu relays when I worked for Fujitsu. > Here though, it's the choice of a relay vs a US$0.10 > FET that takes no board space, withstands any amount > of vibration and will last forever in a corrosive > atmosphere. [hyperbole]
Keep in mind the FET has about 100pF of capacitance at 0V. -- Thanks, - Win
Reply by Jim Thompson February 24, 20162016-02-24
On 24 Feb 2016 09:33:39 GMT, Allan Herriman
<allanherriman@hotmail.com> wrote:

[snip]
> >Actually, I'm still hopeful of finding an opamp that does exactly what I >want in one chip. >Contemporary 74xx series CMOS logic gates have the sort of ESD protection >I want (i.e. input voltages > Vcc without drawing a lot of current). Why >can't an opamp? Are the processes so different? Does that sort of ESD >protection have some bad effect on analog circuits (e.g. leakage current)? > >Regards, >Allan
Didn't your original post say even while unpowered? The only way to do that is with specially-designed ESD, such as is found in many bus transceivers. Off-hand I don't know of an OpAmp so constructed. ...Jim Thompson -- | James E.Thompson | mens | | Analog Innovations | et | | Analog/Mixed-Signal ASIC's and Discrete Systems | manus | | San Tan Valley, AZ 85142 Skype: skypeanalog | | | Voice:(480)460-2350 Fax: Available upon request | Brass Rat | | E-mail Icon at http://www.analog-innovations.com | 1962 | I love to cook with wine. Sometimes I even put it in the food.
Reply by Allan Herriman February 24, 20162016-02-24
On Tue, 23 Feb 2016 19:52:39 -0800, John Larkin wrote:

> On 23 Feb 2016 10:11:00 GMT, Allan Herriman <allanherriman@hotmail.com> > wrote: > >>On Mon, 22 Feb 2016 15:41:07 -0800, whit3rd wrote: >> >>> On Monday, February 22, 2016 at 1:27:53 AM UTC-8, Allan Herriman >>> wrote: >>> >>>> I'm looking for a single supply jellybean opamp to meet the following >>>> requirements: >>>> >>>> V supply = 5.0V Vin (e.g. Vicm) 0V to 4V. >>>> Iin +/- 100pA max. (prefer < 10pA) >>>> >>>> Other parameters are "normal" e.g. slow, noisy, 1mA drive, 0-70 >>>> degree C, >>>> up to $2 for a dual, etc.. >>>> >>>> Here's the hard part: I want the Iin spec to apply over the Vin range >>>> *even when the supply voltage is zero*. >>> >>> What about an input relay, coil powered from your 5.0V supply? >>> DPST-NO would lift one input from two op amps, or both inputs of a >>> single op amp, and some of 'em don't eat your whole budget. >>> >>> <http://www.digikey.com/product-detail/en/EE2-5NU-L/399-11017-1- >>ND/4506469> >> >> >>Thanks for the suggestion. That particular relay operates at 3.75V, so >>it would probably need to be driven by a brownout detector or similar. >> >>I wouldn't use a relay for a "dry" switching application like this >>though. >> >>Regards, >>Allan > > We use thousands of a tiny Fujitsu DPDT relay for low-level switching, > like thermocouple inputs, and they work fine.
I used to use Fujitsu relays when I worked for Fujitsu. Here though, it's the choice of a relay vs a US$0.10 FET that takes no board space, withstands any amount of vibration and will last forever in a corrosive atmosphere. [hyperbole] Actually, I'm still hopeful of finding an opamp that does exactly what I want in one chip. Contemporary 74xx series CMOS logic gates have the sort of ESD protection I want (i.e. input voltages > Vcc without drawing a lot of current). Why can't an opamp? Are the processes so different? Does that sort of ESD protection have some bad effect on analog circuits (e.g. leakage current)? Regards, Allan
Reply by Allan Herriman February 24, 20162016-02-24
On Tue, 23 Feb 2016 18:39:43 -0800, bill.sloman wrote:

> On Tuesday, February 23, 2016 at 9:11:07 PM UTC+11, Allan Herriman > wrote: >> On Mon, 22 Feb 2016 15:41:07 -0800, whit3rd wrote: >> >> > On Monday, February 22, 2016 at 1:27:53 AM UTC-8, Allan Herriman >> > wrote: >> > >> >> I'm looking for a single supply jellybean opamp to meet the >> >> following requirements: >> >> >> >> V supply = 5.0V Vin (e.g. Vicm) 0V to 4V. >> >> Iin +/- 100pA max. (prefer < 10pA) >> >> >> >> Other parameters are "normal" e.g. slow, noisy, 1mA drive, 0-70 >> >> degree C, >> >> up to $2 for a dual, etc.. >> >> >> >> Here's the hard part: I want the Iin spec to apply over the Vin >> >> range *even when the supply voltage is zero*. >> > >> > What about an input relay, coil powered from your 5.0V supply? >> > DPST-NO would lift one input from two op amps, or both inputs of a >> > single op amp, and some of 'em don't eat your whole budget. >> > >> > <http://www.digikey.com/product-detail/en/EE2-5NU-L/399-11017-1- >> ND/4506469> >> >> Thanks for the suggestion. That particular relay operates at 3.75V, so >> it would probably need to be driven by a brownout detector or similar. >> >> I wouldn't use a relay for a "dry" switching application like this >> though. > > Mercury-wetted relays are nice for "dry" switching. They do have to > mounted within 15 degrees of the vertical, which can be a problem, but > they don't bounce on closing or opening - the mercury film provides > mechanical damping.
I keep looking for an ROHS compliant Mercury-wetted relay but no luck so far ... Allan
Reply by John Larkin February 23, 20162016-02-23
On 23 Feb 2016 10:11:00 GMT, Allan Herriman
<allanherriman@hotmail.com> wrote:

>On Mon, 22 Feb 2016 15:41:07 -0800, whit3rd wrote: > >> On Monday, February 22, 2016 at 1:27:53 AM UTC-8, Allan Herriman wrote: >> >>> I'm looking for a single supply jellybean opamp to meet the following >>> requirements: >>> >>> V supply = 5.0V Vin (e.g. Vicm) 0V to 4V. >>> Iin +/- 100pA max. (prefer < 10pA) >>> >>> Other parameters are "normal" e.g. slow, noisy, 1mA drive, 0-70 degree >>> C, >>> up to $2 for a dual, etc.. >>> >>> Here's the hard part: I want the Iin spec to apply over the Vin range >>> *even when the supply voltage is zero*. >> >> What about an input relay, coil powered from your 5.0V supply? DPST-NO >> would lift one input from two op amps, or both inputs of a single op >> amp, and some of 'em don't eat your whole budget. >> >> <http://www.digikey.com/product-detail/en/EE2-5NU-L/399-11017-1- >ND/4506469> > > >Thanks for the suggestion. That particular relay operates at 3.75V, so >it would probably need to be driven by a brownout detector or similar. > >I wouldn't use a relay for a "dry" switching application like this though. > >Regards, >Allan
We use thousands of a tiny Fujitsu DPDT relay for low-level switching, like thermocouple inputs, and they work fine. -- John Larkin Highland Technology, Inc lunatic fringe electronics
Reply by February 23, 20162016-02-23
On Tuesday, February 23, 2016 at 9:11:07 PM UTC+11, Allan Herriman wrote:
> On Mon, 22 Feb 2016 15:41:07 -0800, whit3rd wrote: > > > On Monday, February 22, 2016 at 1:27:53 AM UTC-8, Allan Herriman wrote: > > > >> I'm looking for a single supply jellybean opamp to meet the following > >> requirements: > >> > >> V supply = 5.0V Vin (e.g. Vicm) 0V to 4V. > >> Iin +/- 100pA max. (prefer < 10pA) > >> > >> Other parameters are "normal" e.g. slow, noisy, 1mA drive, 0-70 degree > >> C, > >> up to $2 for a dual, etc.. > >> > >> Here's the hard part: I want the Iin spec to apply over the Vin range > >> *even when the supply voltage is zero*. > > > > What about an input relay, coil powered from your 5.0V supply? DPST-NO > > would lift one input from two op amps, or both inputs of a single op > > amp, and some of 'em don't eat your whole budget. > > > > <http://www.digikey.com/product-detail/en/EE2-5NU-L/399-11017-1- > ND/4506469> > > Thanks for the suggestion. That particular relay operates at 3.75V, so > it would probably need to be driven by a brownout detector or similar. > > I wouldn't use a relay for a "dry" switching application like this though.
Mercury-wetted relays are nice for "dry" switching. They do have to mounted within 15 degrees of the vertical, which can be a problem, but they don't bounce on closing or opening - the mercury film provides mechanical damping. -- Bill Sloman, Sydney
Reply by Hul Tytus February 23, 20162016-02-23
Allan - your interpretation's correct. My suggestion shouldn't work but, 
if I were in your shoes, I'd try it anyway.

Hul

Allan Herriman <allanherriman@hotmail.com> wrote:
> On Mon, 22 Feb 2016 23:04:44 +0000, Hul Tytus wrote:
> > Maybe a diode in the positive supply line. > > > > Hul
> Whilst that would protect the opamp input, it wouldn't meet the 100pA Iin > requirement, as the input would end up powering the opamp through the ESD > protection diode. > (Assuming I've interpreted your suggestion correctly.)
> Thanks, > Allan
Reply by Allan Herriman February 23, 20162016-02-23
On Tue, 23 Feb 2016 04:19:38 -0800, bloggs.fredbloggs.fred wrote:

> On Monday, February 22, 2016 at 4:29:37 PM UTC-5, Allan Herriman wrote: >> On Mon, 22 Feb 2016 05:47:56 -0800, bloggs.fredbloggs.fred wrote: >> >> > On Monday, February 22, 2016 at 4:27:53 AM UTC-5, Allan Herriman >> > wrote: >> >> Hi, >> >> >> >> I'm looking for a single supply jellybean opamp to meet the >> >> following requirements: >> >> >> >> V supply = 5.0V Vin (e.g. Vicm) 0V to 4V. >> >> Iin +/- 100pA max. (prefer < 10pA) >> >> >> >> Other parameters are "normal" e.g. slow, noisy, 1mA drive, 0-70 >> >> degree C, >> >> up to $2 for a dual, etc.. >> >> >> >> Here's the hard part: I want the Iin spec to apply over the Vin >> >> range *even when the supply voltage is zero*. >> >> This requires the absence of ESD protection diodes between the >> >> inputs and the positive rail. Newer opamps almost always have these >> >> diodes. >> >> >> >> There are plenty of cheap older opamps that can come close to >> >> meeting my requirement, e.g. the ubiquitous LM358A and its siblings >> >> would be fine except that Vicm only gets to Vsupply - 1.5V or so. >> >> >> >> Microchip have some interesting parts, e.g. MCP6L02 and MCP6002 are >> >> about US$0.25 in my quantities. They explicitly state in the >> >> datasheet that they don't have diodes from the inputs to the >> >> positive supply rail (good) >> >> but the absolute maximum ratings say that "Analog Inputs (V IN+, V >> >> IN-) >> >> VSS - 1.0V to VDD + 1.0V", which isn't so good. >> >> >> >> >> >> Does anyone have experience with those Microchip parts, and know >> >> whether the abs max ratings can be exceeded without problem? >> >> Normally I would never do that, but I suspect in this case that the >> >> datasheet might be overly conservative. >> >> >> >> Any other opamp suggestions? >> >> >> >> >> >> If I can't find an opamp that meets these requirements I can use >> >> other circuit topologies that aren't as elegant as the one I have in >> >> mind. >> >> >> >> Thanks, >> >> Allan >> > >> > Something like this is pretty compact, I'm sure you'll find a problem >> > with it, "max" leakage spec sucks e.g.: >> > http://datasheets.maximintegrated.com/en/ds/MAX4505.pdf >> >> >> Thanks. That part has a recommended minimum supply voltage of 16V >> though, so it doesn't meet the +5V requirement. >> >> Regards, >> Allan > > It is used at 5V but it loses rail-to-rail performance. There's the > MAX366 spec'd for LV but the channel starts going hi-z for Ain within > 1.5V of rail. And all these parts are pricey, in the $3-4 dollar range, > so not jellybean. In the way of OAs there's the OP191 and later > derivative generations, but pricey too.
I hadn't come across the OP191 before. They only specify the input current for voltages between the rails. Looking at the schematic (figure 61 in the OP191 datasheet) it seems that it actually does have clamp diodes to the positive supply and also back- to-back diodes between the inverting and noninverting inputs. There are 5k series resistors to limit the input current. Unfortunately this won't meet my Iin requirement. The FET I was thinking about earlier (in my reply to Piglet's post) was the DMG1012UW. It has a threshold voltage of 0.5 to 1V @ 250uA, meaning I could use it to isolate the opamp input and still meet the 4V input with 5V supply requirement. I suspect that would work better than the Maxim parts you mentioned for voltages close to the rail. The leakage current is ok if I only look at the typical parameters. They don't specify a max IDSS at anything other than 25 degree C (100nA). The typical curve looks good though (about 7nA at 85 degree C and 4V). I use a lot of these FETs for driving things from 1.8V logic, so they're already on the board and only cost a few cents each. Regards, Allan