Hello, in this thread on eevblog: http://www.eevblog.com/forum/projects-designs-and-technical-stuff/does-anyone-know-what-this-diode-is-for/ a guy named DiligentMinds.com asked the question about why there is a 1N4148 diode on the output of the opamp in the LTZ1000 reference circuit, the one designed by Jim William in the appnotes, that can also be seen on the first page of the datasheet: http://cds.linear.com/images/product/1203_app_1.JPG Now I'm curious, anybody knows the answer? Thanks, Fabio. ---------------------------------------------------------------- This message was sent using IMP, the Internet Messaging Program.

Positive input voltage range for an LT1006 is typically 1.2V below the positive rail. mitch ----- Original Message ----- From: "Fabio Eboli" <fabioeb@quipo.it> To: <volt-nuts@febo.com> Sent: Wednesday, November 07, 2012 7:02 PM Subject: [volt-nuts] LTZ1000 reference appnote question > Hello, in this thread on eevblog: > > http://www.eevblog.com/forum/projects-designs-and-technical-stuff/does-anyone-know-what-this-diode-is-for/ > > a guy named DiligentMinds.com asked the question > about why there is a 1N4148 diode on the output of > the opamp in the LTZ1000 reference circuit, the one > designed by Jim William in the appnotes, that can > also be seen on the first page of the datasheet: > > http://cds.linear.com/images/product/1203_app_1.JPG > > Now I'm curious, > anybody knows the answer? > > Thanks, > Fabio. > > ---------------------------------------------------------------- > This message was sent using IMP, the Internet Messaging Program. > > > > _______________________________________________ > volt-nuts mailing list -- volt-nuts@febo.com > To unsubscribe, go to > https://www.febo.com/cgi-bin/mailman/listinfo/volt-nuts > and follow the instructions there.

On 11/07/2012 06:02 PM, Fabio Eboli wrote: > Hello, in this thread on eevblog: > > http://www.eevblog.com/forum/projects-designs-and-technical-stuff/does-anyone-know-what-this-diode-is-for/ > > > a guy named DiligentMinds.com asked the question > about why there is a 1N4148 diode on the output of > the opamp in the LTZ1000 reference circuit, the one > designed by Jim William in the appnotes, that can > also be seen on the first page of the datasheet: > > http://cds.linear.com/images/product/1203_app_1.JPG > > Now I'm curious, > anybody knows the answer? I believe it it to prevent negative lockup of the circuit output upon start. -- Bob Smither, PhD Circuit Concepts, Inc. ========================================================================= "If you want peace, work for justice. If you want justice, work for freedom." ========================================================================= Smither@C-C-I.Com http://www.C-C-I.Com 281-331-2744(office) -4616(fax)

Mitch,are you referring to startup behaviour, like Bob Smither suggests? Just now I was trying to imagine the startup (it's too late for firing up ltspice, I have to sleep :) ) and I was thinking that on startup both inputs are zero and output can take both ways: go down (remain approx zero) or go up, depending on input offset. Is this correct? If so what determines the correct startup of the circuit? Until the output diode is not forward biased the opamp has no feedback and the output should amplify any input noise some million times (open loop gain), so I imagine that is only waiting that the noise on inputs will steer the output the rignt way to forward-bias the diode and start the reference. Does all this makes sense? Fabio. Mitch Van Ochten <mitch@vincentelectronics.com> ha scritto: > Positive input voltage range for an LT1006 is typically 1.2V below > the positive rail. > > mitch > > > ----- Original Message ----- From: "Fabio Eboli" <fabioeb@quipo.it> > To: <volt-nuts@febo.com> > Sent: Wednesday, November 07, 2012 7:02 PM > Subject: [volt-nuts] LTZ1000 reference appnote question > > >> Hello, in this thread on eevblog: >> >> http://www.eevblog.com/forum/projects-designs-and-technical-stuff/does-anyone-know-what-this-diode-is-for/ >> >> a guy named DiligentMinds.com asked the question >> about why there is a 1N4148 diode on the output of >> the opamp in the LTZ1000 reference circuit, the one >> designed by Jim William in the appnotes, that can >> also be seen on the first page of the datasheet: >> >> http://cds.linear.com/images/product/1203_app_1.JPG >> >> Now I'm curious, >> anybody knows the answer? >> >> Thanks, >> Fabio. >> >> ---------------------------------------------------------------- >> This message was sent using IMP, the Internet Messaging Program. >> >> >> >> _______________________________________________ >> volt-nuts mailing list -- volt-nuts@febo.com >> To unsubscribe, go to >> https://www.febo.com/cgi-bin/mailman/listinfo/volt-nuts >> and follow the instructions there. > > > _______________________________________________ > volt-nuts mailing list -- volt-nuts@febo.com > To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/volt-nuts > and follow the instructions there. > ---------------------------------------------------------------- This message was sent using IMP, the Internet Messaging Program.

Hello together, what about another theory: J.W. first developped the negative reference cirquit in the datasheet. In the negative cirquit a diode is essential that on no circumstance the diode between reference and heater within the LTZ is not forward biased. Later on J.W. gave the order to his apprentice to develop a positive cirquit. The apprentice did not dare to remove a "holy part" from the cirquit. 3rd theory: The diode helps to limit inrush current on startup through the zener which is not specified above 5mA. I have a modified cirquit with a J-FET after the LT1013 and had to add a limiting resistor to limit to below 14 mA. 4th theory: the startup pull up resistor after the diode has been forgotten in the final cirquit. (like the ground connections and the "star" point). On the other side: the cirquit in my opinion lacks on stability. (at least for large signal behaviour). The 22nF is charged by a high ohmic resistor and discharged by a low ohmic transistor. The DATRON reference cirquit is using some kind of integrator for the current regulation. With best regards Andreas ----- Original Message ----- From: "Fabio Eboli" <fabioeb@quipo.it> To: "Discussion of precise voltage measurement" <volt-nuts@febo.com> Sent: Thursday, November 08, 2012 2:41 AM Subject: Re: [volt-nuts] LTZ1000 reference appnote question > Mitch,are you referring to startup behaviour, > like Bob Smither suggests? > > Just now I was trying to imagine the startup > (it's too late for firing up ltspice, I have to sleep :) ) > and I was thinking that on startup both inputs > are zero and output can take both ways: go down (remain approx zero) > or go up, depending on input offset. > Is this correct? > > If so what determines the correct startup > of the circuit? > > Until the output diode is not forward biased > the opamp has no feedback and the output should amplify > any input noise some million times (open loop gain), > so I imagine that is only waiting that the noise on > inputs will steer the output the rignt way to forward-bias > the diode and start the reference. > Does all this makes sense? > > Fabio. > > Mitch Van Ochten <mitch@vincentelectronics.com> ha scritto: > >> Positive input voltage range for an LT1006 is typically 1.2V below the >> positive rail. >> >> mitch >> >> >> ----- Original Message ----- From: "Fabio Eboli" <fabioeb@quipo.it> >> To: <volt-nuts@febo.com> >> Sent: Wednesday, November 07, 2012 7:02 PM >> Subject: [volt-nuts] LTZ1000 reference appnote question >> >> >>> Hello, in this thread on eevblog: >>> >>> http://www.eevblog.com/forum/projects-designs-and-technical-stuff/does-anyone-know-what-this-diode-is-for/ >>> >>> a guy named DiligentMinds.com asked the question >>> about why there is a 1N4148 diode on the output of >>> the opamp in the LTZ1000 reference circuit, the one >>> designed by Jim William in the appnotes, that can >>> also be seen on the first page of the datasheet: >>> >>> http://cds.linear.com/images/product/1203_app_1.JPG >>> >>> Now I'm curious, >>> anybody knows the answer? >>> >>> Thanks, >>> Fabio. >>> >>> ---------------------------------------------------------------- >>> This message was sent using IMP, the Internet Messaging Program. >>> >>> >>> >>> _______________________________________________ >>> volt-nuts mailing list -- volt-nuts@febo.com >>> To unsubscribe, go to >>> https://www.febo.com/cgi-bin/mailman/listinfo/volt-nuts >>> and follow the instructions there. >> >> >> _______________________________________________ >> volt-nuts mailing list -- volt-nuts@febo.com >> To unsubscribe, go to >> https://www.febo.com/cgi-bin/mailman/listinfo/volt-nuts >> and follow the instructions there. >> > > > > ---------------------------------------------------------------- > This message was sent using IMP, the Internet Messaging Program. > > > > _______________________________________________ > volt-nuts mailing list -- volt-nuts@febo.com > To unsubscribe, go to > https://www.febo.com/cgi-bin/mailman/listinfo/volt-nuts > and follow the instructions there.

Thank you Andreas, very interesting points you made. In the meantime the original OP on eevblog thread reported that Linear answer (from Bob Dobkin) was about startup behaviour: "It is for start up. If the amp has a negative supply, it can start in the wrong direction. Dobby" Remains the doubt if he referring to negative offset, or psu negative transient. Fabio. Andreas Jahn <Andreas_-_Jahn@t-online.de> ha scritto: > Hello together, > > what about another theory: > J.W. first developped the negative reference cirquit in the datasheet. > In the negative cirquit a diode is essential that on no circumstance > the diode between reference and heater within the LTZ is not forward > biased. > Later on J.W. gave the order to his apprentice to develop a positive cirquit. > The apprentice did not dare to remove a "holy part" from the cirquit. > > 3rd theory: > The diode helps to limit inrush current on startup through the zener > which is not specified above 5mA. > I have a modified cirquit with a J-FET after the LT1013 and had to > add a limiting resistor to limit to below 14 mA. > > 4th theory: > the startup pull up resistor after the diode has been forgotten in > the final cirquit. > (like the ground connections and the "star" point). > > On the other side: the cirquit in my opinion lacks on stability. (at > least for large signal behaviour). > The 22nF is charged by a high ohmic resistor and discharged by a low > ohmic transistor. > > The DATRON reference cirquit is using some kind of integrator for > the current regulation. > > With best regards > > Andreas > > > ----- Original Message ----- From: "Fabio Eboli" <fabioeb@quipo.it> > To: "Discussion of precise voltage measurement" <volt-nuts@febo.com> > Sent: Thursday, November 08, 2012 2:41 AM > Subject: Re: [volt-nuts] LTZ1000 reference appnote question > > >> Mitch,are you referring to startup behaviour, >> like Bob Smither suggests? >> >> Just now I was trying to imagine the startup >> (it's too late for firing up ltspice, I have to sleep :) ) >> and I was thinking that on startup both inputs >> are zero and output can take both ways: go down (remain approx zero) >> or go up, depending on input offset. >> Is this correct? >> >> If so what determines the correct startup >> of the circuit? >> >> Until the output diode is not forward biased >> the opamp has no feedback and the output should amplify >> any input noise some million times (open loop gain), >> so I imagine that is only waiting that the noise on >> inputs will steer the output the rignt way to forward-bias >> the diode and start the reference. >> Does all this makes sense? >> >> Fabio. >> >> Mitch Van Ochten <mitch@vincentelectronics.com> ha scritto: >> >>> Positive input voltage range for an LT1006 is typically 1.2V below >>> the positive rail. >>> >>> mitch >>> >>> >>> ----- Original Message ----- From: "Fabio Eboli" <fabioeb@quipo.it> >>> To: <volt-nuts@febo.com> >>> Sent: Wednesday, November 07, 2012 7:02 PM >>> Subject: [volt-nuts] LTZ1000 reference appnote question >>> >>> >>>> Hello, in this thread on eevblog: >>>> >>>> http://www.eevblog.com/forum/projects-designs-and-technical-stuff/does-anyone-know-what-this-diode-is-for/ >>>> >>>> a guy named DiligentMinds.com asked the question >>>> about why there is a 1N4148 diode on the output of >>>> the opamp in the LTZ1000 reference circuit, the one >>>> designed by Jim William in the appnotes, that can >>>> also be seen on the first page of the datasheet: >>>> >>>> http://cds.linear.com/images/product/1203_app_1.JPG >>>> >>>> Now I'm curious, >>>> anybody knows the answer? >>>> >>>> Thanks, >>>> Fabio. >>>> >>>> ---------------------------------------------------------------- >>>> This message was sent using IMP, the Internet Messaging Program. >>>> >>>> >>>> >>>> _______________________________________________ >>>> volt-nuts mailing list -- volt-nuts@febo.com >>>> To unsubscribe, go to >>>> https://www.febo.com/cgi-bin/mailman/listinfo/volt-nuts >>>> and follow the instructions there. >>> >>> >>> _______________________________________________ >>> volt-nuts mailing list -- volt-nuts@febo.com >>> To unsubscribe, go to >>> https://www.febo.com/cgi-bin/mailman/listinfo/volt-nuts >>> and follow the instructions there. >>> >> >> >> >> ---------------------------------------------------------------- >> This message was sent using IMP, the Internet Messaging Program. >> >> >> >> _______________________________________________ >> volt-nuts mailing list -- volt-nuts@febo.com >> To unsubscribe, go to >> https://www.febo.com/cgi-bin/mailman/listinfo/volt-nuts >> and follow the instructions there. > > > _______________________________________________ > volt-nuts mailing list -- volt-nuts@febo.com > To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/volt-nuts > and follow the instructions there. > ---------------------------------------------------------------- This message was sent using IMP, the Internet Messaging Program.

Assume that the diode isn't there, and the output of the opamp starts out at negative 0.1 volt. None of the semiconductor junctions will be active, so the 30K resistor will feed negative current into the "+" input of the opamp (positive feedback), and force the output of the opamp to the negative rail. The opamp will happily stay that way because there is nothing to prevent it. The diode keeps any negative voltages from the output pin from reaching the noninverting input, and allows the circuit to only start up in the positive direction. -Chuck Harris Fabio Eboli wrote: > Thank you Andreas, very interesting points you made. > > In the meantime the original OP on eevblog thread > reported that Linear answer (from Bob Dobkin) > was about startup behaviour: > > "It is for start up. If the amp has a negative supply, > it can start in the wrong direction. > > Dobby" > > Remains the doubt if he referring to negative offset, > or psu negative transient. > > Fabio. > > Andreas Jahn <Andreas_-_Jahn@t-online.de> ha scritto: > >> Hello together, >> >> what about another theory: >> J.W. first developped the negative reference cirquit in the datasheet. >> In the negative cirquit a diode is essential that on no circumstance the diode >> between reference and heater within the LTZ is not forward biased. >> Later on J.W. gave the order to his apprentice to develop a positive cirquit. >> The apprentice did not dare to remove a "holy part" from the cirquit. >> >> 3rd theory: >> The diode helps to limit inrush current on startup through the zener which is not >> specified above 5mA. >> I have a modified cirquit with a J-FET after the LT1013 and had to add a limiting >> resistor to limit to below 14 mA. >> >> 4th theory: >> the startup pull up resistor after the diode has been forgotten in the final cirquit. >> (like the ground connections and the "star" point). >> >> On the other side: the cirquit in my opinion lacks on stability. (at least for >> large signal behaviour). >> The 22nF is charged by a high ohmic resistor and discharged by a low ohmic transistor. >> >> The DATRON reference cirquit is using some kind of integrator for the current >> regulation. >> >> With best regards >> >> Andreas >> >> >> ----- Original Message ----- From: "Fabio Eboli" <fabioeb@quipo.it> >> To: "Discussion of precise voltage measurement" <volt-nuts@febo.com> >> Sent: Thursday, November 08, 2012 2:41 AM >> Subject: Re: [volt-nuts] LTZ1000 reference appnote question >> >> >>> Mitch,are you referring to startup behaviour, >>> like Bob Smither suggests? >>> >>> Just now I was trying to imagine the startup >>> (it's too late for firing up ltspice, I have to sleep :) ) >>> and I was thinking that on startup both inputs >>> are zero and output can take both ways: go down (remain approx zero) >>> or go up, depending on input offset. >>> Is this correct? >>> >>> If so what determines the correct startup >>> of the circuit? >>> >>> Until the output diode is not forward biased >>> the opamp has no feedback and the output should amplify >>> any input noise some million times (open loop gain), >>> so I imagine that is only waiting that the noise on >>> inputs will steer the output the rignt way to forward-bias >>> the diode and start the reference. >>> Does all this makes sense? >>> >>> Fabio. >>> >>> Mitch Van Ochten <mitch@vincentelectronics.com> ha scritto: >>> >>>> Positive input voltage range for an LT1006 is typically 1.2V below the positive >>>> rail. >>>> >>>> mitch >>>> >>>> >>>> ----- Original Message ----- From: "Fabio Eboli" <fabioeb@quipo.it> >>>> To: <volt-nuts@febo.com> >>>> Sent: Wednesday, November 07, 2012 7:02 PM >>>> Subject: [volt-nuts] LTZ1000 reference appnote question >>>> >>>> >>>>> Hello, in this thread on eevblog: >>>>> >>>>> http://www.eevblog.com/forum/projects-designs-and-technical-stuff/does-anyone-know-what-this-diode-is-for/ >>>>> >>>>> >>>>> a guy named DiligentMinds.com asked the question >>>>> about why there is a 1N4148 diode on the output of >>>>> the opamp in the LTZ1000 reference circuit, the one >>>>> designed by Jim William in the appnotes, that can >>>>> also be seen on the first page of the datasheet: >>>>> >>>>> http://cds.linear.com/images/product/1203_app_1.JPG >>>>> >>>>> Now I'm curious, >>>>> anybody knows the answer? >>>>> >>>>> Thanks, >>>>> Fabio. >>>>> >>>>> ---------------------------------------------------------------- >>>>> This message was sent using IMP, the Internet Messaging Program. >>>>> >>>>> >>>>> >>>>> _______________________________________________ >>>>> volt-nuts mailing list -- volt-nuts@febo.com >>>>> To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/volt-nuts >>>>> and follow the instructions there. >>>> >>>> >>>> _______________________________________________ >>>> volt-nuts mailing list -- volt-nuts@febo.com >>>> To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/volt-nuts >>>> and follow the instructions there. >>>> >>> >>> >>> >>> ---------------------------------------------------------------- >>> This message was sent using IMP, the Internet Messaging Program. >>> >>> >>> >>> _______________________________________________ >>> volt-nuts mailing list -- volt-nuts@febo.com >>> To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/volt-nuts >>> and follow the instructions there. >> >> >> _______________________________________________ >> volt-nuts mailing list -- volt-nuts@febo.com >> To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/volt-nuts >> and follow the instructions there. >> > > > ---------------------------------------------------------------- > This message was sent using IMP, the Internet Messaging Program. > > > > _______________________________________________ > volt-nuts mailing list -- volt-nuts@febo.com > To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/volt-nuts > and follow the instructions there. >

Hello Chuck, I understand that for dual supply, but is this valid also for single supply like that one? I was thinking that the lowest voltage in the circuit, at least after powerup transient, will be ground potential that is also negative rail's potential. I agree that, without the output diode, if the out is for some reason externally tied few mV below ground there is really no hope to start the reference. Fabio. Chuck Harris <cfharris@erols.com> ha scritto: > Assume that the diode isn't there, and the output of the > opamp starts out at negative 0.1 volt. None of the > semiconductor junctions will be active, so the 30K > resistor will feed negative current into the "+" input > of the opamp (positive feedback), and force the output > of the opamp to the negative rail. The opamp will happily > stay that way because there is nothing to prevent it. > > The diode keeps any negative voltages from the output > pin from reaching the noninverting input, and allows the > circuit to only start up in the positive direction. > > -Chuck Harris > > Fabio Eboli wrote: >> Thank you Andreas, very interesting points you made. >> >> In the meantime the original OP on eevblog thread >> reported that Linear answer (from Bob Dobkin) >> was about startup behaviour: >> >> "It is for start up. If the amp has a negative supply, >> it can start in the wrong direction. >> >> Dobby" >> >> Remains the doubt if he referring to negative offset, >> or psu negative transient. >> >> Fabio. >> >> Andreas Jahn <Andreas_-_Jahn@t-online.de> ha scritto: >> >>> Hello together, >>> >>> what about another theory: >>> J.W. first developped the negative reference cirquit in the datasheet. >>> In the negative cirquit a diode is essential that on no >>> circumstance the diode >>> between reference and heater within the LTZ is not forward biased. >>> Later on J.W. gave the order to his apprentice to develop a >>> positive cirquit. >>> The apprentice did not dare to remove a "holy part" from the cirquit. >>> >>> 3rd theory: >>> The diode helps to limit inrush current on startup through the >>> zener which is not >>> specified above 5mA. >>> I have a modified cirquit with a J-FET after the LT1013 and had to >>> add a limiting >>> resistor to limit to below 14 mA. >>> >>> 4th theory: >>> the startup pull up resistor after the diode has been forgotten in >>> the final cirquit. >>> (like the ground connections and the "star" point). >>> >>> On the other side: the cirquit in my opinion lacks on stability. >>> (at least for >>> large signal behaviour). >>> The 22nF is charged by a high ohmic resistor and discharged by a >>> low ohmic transistor. >>> >>> The DATRON reference cirquit is using some kind of integrator for >>> the current >>> regulation. >>> >>> With best regards >>> >>> Andreas >>> >>> >>> ----- Original Message ----- From: "Fabio Eboli" <fabioeb@quipo.it> >>> To: "Discussion of precise voltage measurement" <volt-nuts@febo.com> >>> Sent: Thursday, November 08, 2012 2:41 AM >>> Subject: Re: [volt-nuts] LTZ1000 reference appnote question >>> >>> >>>> Mitch,are you referring to startup behaviour, >>>> like Bob Smither suggests? >>>> >>>> Just now I was trying to imagine the startup >>>> (it's too late for firing up ltspice, I have to sleep :) ) >>>> and I was thinking that on startup both inputs >>>> are zero and output can take both ways: go down (remain approx zero) >>>> or go up, depending on input offset. >>>> Is this correct? >>>> >>>> If so what determines the correct startup >>>> of the circuit? >>>> >>>> Until the output diode is not forward biased >>>> the opamp has no feedback and the output should amplify >>>> any input noise some million times (open loop gain), >>>> so I imagine that is only waiting that the noise on >>>> inputs will steer the output the rignt way to forward-bias >>>> the diode and start the reference. >>>> Does all this makes sense? >>>> >>>> Fabio. >>>> >>>> Mitch Van Ochten <mitch@vincentelectronics.com> ha scritto: >>>> >>>>> Positive input voltage range for an LT1006 is typically 1.2V >>>>> below the positive >>>>> rail. >>>>> >>>>> mitch >>>>> >>>>> >>>>> ----- Original Message ----- From: "Fabio Eboli" <fabioeb@quipo.it> >>>>> To: <volt-nuts@febo.com> >>>>> Sent: Wednesday, November 07, 2012 7:02 PM >>>>> Subject: [volt-nuts] LTZ1000 reference appnote question >>>>> >>>>> >>>>>> Hello, in this thread on eevblog: >>>>>> >>>>>> http://www.eevblog.com/forum/projects-designs-and-technical-stuff/does-anyone-know-what-this-diode-is-for/ >>>>>> >>>>>> >>>>>> a guy named DiligentMinds.com asked the question >>>>>> about why there is a 1N4148 diode on the output of >>>>>> the opamp in the LTZ1000 reference circuit, the one >>>>>> designed by Jim William in the appnotes, that can >>>>>> also be seen on the first page of the datasheet: >>>>>> >>>>>> http://cds.linear.com/images/product/1203_app_1.JPG >>>>>> >>>>>> Now I'm curious, >>>>>> anybody knows the answer? >>>>>> >>>>>> Thanks, >>>>>> Fabio. >>>>>> >>>>>> ---------------------------------------------------------------- >>>>>> This message was sent using IMP, the Internet Messaging Program. >>>>>> >>>>>> >>>>>> >>>>>> _______________________________________________ >>>>>> volt-nuts mailing list -- volt-nuts@febo.com >>>>>> To unsubscribe, go to >>>>>> https://www.febo.com/cgi-bin/mailman/listinfo/volt-nuts >>>>>> and follow the instructions there. >>>>> >>>>> >>>>> _______________________________________________ >>>>> volt-nuts mailing list -- volt-nuts@febo.com >>>>> To unsubscribe, go to >>>>> https://www.febo.com/cgi-bin/mailman/listinfo/volt-nuts >>>>> and follow the instructions there. >>>>> >>>> >>>> >>>> >>>> ---------------------------------------------------------------- >>>> This message was sent using IMP, the Internet Messaging Program. >>>> >>>> >>>> >>>> _______________________________________________ >>>> volt-nuts mailing list -- volt-nuts@febo.com >>>> To unsubscribe, go to >>>> https://www.febo.com/cgi-bin/mailman/listinfo/volt-nuts >>>> and follow the instructions there. >>> >>> >>> _______________________________________________ >>> volt-nuts mailing list -- volt-nuts@febo.com >>> To unsubscribe, go to >>> https://www.febo.com/cgi-bin/mailman/listinfo/volt-nuts >>> and follow the instructions there. >>> >> >> >> ---------------------------------------------------------------- >> This message was sent using IMP, the Internet Messaging Program. >> >> >> >> _______________________________________________ >> volt-nuts mailing list -- volt-nuts@febo.com >> To unsubscribe, go to >> https://www.febo.com/cgi-bin/mailman/listinfo/volt-nuts >> and follow the instructions there. >> > > _______________________________________________ > volt-nuts mailing list -- volt-nuts@febo.com > To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/volt-nuts > and follow the instructions there. > ---------------------------------------------------------------- This message was sent using IMP, the Internet Messaging Program.

Hi Fabio, I think that startup problem could occur if the opamp's input offset current happened to be such that the - input was above the + input. The 30K resistor, by virtue of its positive feedback, would enforce that condition, locking the circuit at near zero output. The diode, at least, prevents that extra positive feedback from contributing to the problem. I don't think the diode will have much effect most of the time. -Chuck Harris Fabio Eboli wrote: > Hello Chuck, I understand that for dual supply, but is > this valid also for single supply like that one? > I was thinking that the lowest voltage in the circuit, > at least after powerup transient, will be ground > potential that is also negative rail's potential. > > I agree that, without the output diode, if the out > is for some reason externally tied few mV below ground > there is really no hope to start the reference. > > Fabio. > >