On Fri, 11 Dec 2015 11:50:08 -0500, Phil Hobbs <pcdhSpamMeSenseless@electrooptical.net> wrote:>On 12/10/2015 07:25 PM, Phil Hobbs wrote: >> On 12/10/2015 03:51 PM, John Larkin wrote: >>> On Thu, 10 Dec 2015 13:46:24 -0500, Phil Hobbs >>> <pcdhSpamMeSenseless@electrooptical.net> wrote: >>> >>>> On 12/10/2015 01:41 PM, John Larkin wrote: >>>>> On Thu, 10 Dec 2015 05:11:41 -0800 (PST), Phil Hobbs >>>>> <pcdhobbs@gmail.com> wrote: >>>>> >>>>>>> Teflon and some ceramics are the best capacitor types; 100 pF >>>>>>> gives you >>>>>> Iout = (1e-10) dV/dt >>>>>>> and I've made ramp generators that did about 10 mv/hour, and can keep >>>>>>> it up all day. that'd be Iout in the 3e-16 amps range... >>>>>> >>>>>> HP 3325As make nice ramps down to 1 uHz, and they're nearly as >>>>>> cheap as 610Cs. It's not a modern DDS/arb architecture, so I don't >>>>>> think its output is a staircase. >>>>>> >>>>>> >>>>>> Cheers >>>>>> >>>>>> Phil Hobbs >>>>> >>>>> You could charge a 10 uF cap through a 1G resistor, and couple that >>>>> into a picoammeter through a 10 pF cap. 10 volts makes 10 nA which >>>>> becomes 10 fA. >>>>> >>>>> >>>>> >>>>> >>>> >>>> Ceramic caps that big are pretty nonlinear to use for calibration, no? >>>> I have some 2-uf 200V dogbone mylars. When my 610C gets here this >>>> evening I'll test their insulation resistance. ;) >>> >>> A 10 pF NPO should be pretty good on the output side. I'd use a film >>> cap for the big one. >>> >>> >>>> >>>> An LM662 triangle wave generator with one of your patented >>>> bootstrapped-voltage-references could go pretty slowly. >>>> >>> >>> Yeah, if you want a linear ramp a bootstrap current source is a good >>> way to go, almost the only way to go if you want, say, 1 nA. >>> >>> Film cap, cmos opamp follower, bandgap level shifter, 1G resistor back >>> to the cap. Reset it with a paper clip. >> >> >> The $125 610C arrived just now, by UPS ground all the way from SoCal. At >> first blush, it looks like it works pretty well. It zeroes fine, which >> is one of the main indications of health. No trace of stickiness in the >> meter movement, which is another important indication. >> >> On the 10 fA FS range, with the input disconnected but with the cap off >> the connector, waving my hand a foot away from the front panel makes the >> meter move all over the place. >> >> Works on both the slow and fast settings. >> >> Very pretty! >> > >Resistance reads about 2% low up to 1E12 ohms, calibrated using a 5-1/2 >digit, 10M DVM and a 10V HP precision power supply. A Russian 1 teraohm >10% resistor reads about 0.86 Tohms on the 610C and about 0.89 Tohms >using the digital voltmeter trick. The fast setting pegs the meter in >the resistance mode. (I didn't try re-zeroing it.) > >Voltage is within 1% where I looked. > >Stability is excellent--on the 10 fA FS range, it stays zeroed to within >1% of FS over several minutes, and that's after only about half an hour >of warm-up. > >With a 100 pF polystyrene cap across the input (just so I know roughly >what the capacitance is there), and measuring the 0-3V output on the >back near full scale (2.986V), the voltage decay is less than 100 >uV/min and is consistent with zero. > >That means the input current on the 10V range is less than >100uV * 10V/3V * 100 pF /60 s = 0.5 fA. Amazing. (Not bad on the >polystyrene, either.) > >I measured a 2.2 uF 600V wound mylar, charged to 10V for a few seconds. > On the 10-V range, it produced nearly full scale on the 0-3V DC output. > This output dropped about 220uV out of 3V in 3 minutes, for a >self-discharge time constant of about 2.4 million seconds (about a >month). So with a 47 G resistor and an LM662, you could actually make a >decent 10 microhertz sawtooth out of it. > >Not a bad gizmo for $125 plus $25 shipping! > >Cheers > >Phil HobbsIt's a beautiful instrument, as long as you don't look inside. It's sort of a rat's nest. The change you are seeing on the mylar cap may be dielectric absorption, maybe not leakage. They are awful that way. Someone here once posted a long-term measurement of film cap leakage, and as I recall he saw essentially zero voltage change per year for some.
Keithley 610C repair
Started by ●December 1, 2015
Reply by ●December 11, 20152015-12-11
Reply by ●December 11, 20152015-12-11
On 12/11/2015 11:56 AM, John Larkin wrote:> On Fri, 11 Dec 2015 11:50:08 -0500, Phil Hobbs > <pcdhSpamMeSenseless@electrooptical.net> wrote: > >> On 12/10/2015 07:25 PM, Phil Hobbs wrote: >>> On 12/10/2015 03:51 PM, John Larkin wrote: >>>> On Thu, 10 Dec 2015 13:46:24 -0500, Phil Hobbs >>>> <pcdhSpamMeSenseless@electrooptical.net> wrote: >>>> >>>>> On 12/10/2015 01:41 PM, John Larkin wrote: >>>>>> On Thu, 10 Dec 2015 05:11:41 -0800 (PST), Phil Hobbs >>>>>> <pcdhobbs@gmail.com> wrote: >>>>>> >>>>>>>> Teflon and some ceramics are the best capacitor types; 100 pF >>>>>>>> gives you >>>>>>> Iout = (1e-10) dV/dt >>>>>>>> and I've made ramp generators that did about 10 mv/hour, and can keep >>>>>>>> it up all day. that'd be Iout in the 3e-16 amps range... >>>>>>> >>>>>>> HP 3325As make nice ramps down to 1 uHz, and they're nearly as >>>>>>> cheap as 610Cs. It's not a modern DDS/arb architecture, so I don't >>>>>>> think its output is a staircase. >>>>>>> >>>>>>> >>>>>>> Cheers >>>>>>> >>>>>>> Phil Hobbs >>>>>> >>>>>> You could charge a 10 uF cap through a 1G resistor, and couple that >>>>>> into a picoammeter through a 10 pF cap. 10 volts makes 10 nA which >>>>>> becomes 10 fA. >>>>>> >>>>>> >>>>>> >>>>>> >>>>> >>>>> Ceramic caps that big are pretty nonlinear to use for calibration, no? >>>>> I have some 2-uf 200V dogbone mylars. When my 610C gets here this >>>>> evening I'll test their insulation resistance. ;) >>>> >>>> A 10 pF NPO should be pretty good on the output side. I'd use a film >>>> cap for the big one. >>>> >>>> >>>>> >>>>> An LM662 triangle wave generator with one of your patented >>>>> bootstrapped-voltage-references could go pretty slowly. >>>>> >>>> >>>> Yeah, if you want a linear ramp a bootstrap current source is a good >>>> way to go, almost the only way to go if you want, say, 1 nA. >>>> >>>> Film cap, cmos opamp follower, bandgap level shifter, 1G resistor back >>>> to the cap. Reset it with a paper clip. >>> >>> >>> The $125 610C arrived just now, by UPS ground all the way from SoCal. At >>> first blush, it looks like it works pretty well. It zeroes fine, which >>> is one of the main indications of health. No trace of stickiness in the >>> meter movement, which is another important indication. >>> >>> On the 10 fA FS range, with the input disconnected but with the cap off >>> the connector, waving my hand a foot away from the front panel makes the >>> meter move all over the place. >>> >>> Works on both the slow and fast settings. >>> >>> Very pretty! >>> >> >> Resistance reads about 2% low up to 1E12 ohms, calibrated using a 5-1/2 >> digit, 10M DVM and a 10V HP precision power supply. A Russian 1 teraohm >> 10% resistor reads about 0.86 Tohms on the 610C and about 0.89 Tohms >> using the digital voltmeter trick. The fast setting pegs the meter in >> the resistance mode. (I didn't try re-zeroing it.) >> >> Voltage is within 1% where I looked. >> >> Stability is excellent--on the 10 fA FS range, it stays zeroed to within >> 1% of FS over several minutes, and that's after only about half an hour >> of warm-up. >> >> With a 100 pF polystyrene cap across the input (just so I know roughly >> what the capacitance is there), and measuring the 0-3V output on the >> back near full scale (2.986V), the voltage decay is less than 100 >> uV/min and is consistent with zero. >> >> That means the input current on the 10V range is less than >> 100uV * 10V/3V * 100 pF /60 s = 0.5 fA. Amazing. (Not bad on the >> polystyrene, either.) >> >> I measured a 2.2 uF 600V wound mylar, charged to 10V for a few seconds. >> On the 10-V range, it produced nearly full scale on the 0-3V DC output. >> This output dropped about 220uV out of 3V in 3 minutes, for a >> self-discharge time constant of about 2.4 million seconds (about a >> month). So with a 47 G resistor and an LM662, you could actually make a >> decent 10 microhertz sawtooth out of it. >> >> Not a bad gizmo for $125 plus $25 shipping! >> >> Cheers >> >> Phil Hobbs > > It's a beautiful instrument, as long as you don't look inside. It's > sort of a rat's nest. > > The change you are seeing on the mylar cap may be dielectric > absorption, maybe not leakage. They are awful that way. > > Someone here once posted a long-term measurement of film cap leakage, > and as I recall he saw essentially zero voltage change per year for > some. > > >Interestingly the major source of instability on the 10 fA range appears to be slow capacitive pickup from things moving around inside as it warms up. You can make it move > 10 fA in each direction by pressing gently on the front panel. Pushing on the range switch knob is about 10x more sensitive. When you're down in the hundreds of electrons, these things matter! Cheers Phil Hobbs -- Dr Philip C D Hobbs Principal Consultant ElectroOptical Innovations LLC Optics, Electro-optics, Photonics, Analog Electronics 160 North State Road #203 Briarcliff Manor NY 10510 hobbs at electrooptical dot net http://electrooptical.net
Reply by ●December 11, 20152015-12-11
On Fri, 11 Dec 2015 18:18:05 -0500, Phil Hobbs <pcdhSpamMeSenseless@electrooptical.net> wrote:>On 12/11/2015 11:56 AM, John Larkin wrote: >> On Fri, 11 Dec 2015 11:50:08 -0500, Phil Hobbs >> <pcdhSpamMeSenseless@electrooptical.net> wrote: >> >>> On 12/10/2015 07:25 PM, Phil Hobbs wrote: >>>> On 12/10/2015 03:51 PM, John Larkin wrote: >>>>> On Thu, 10 Dec 2015 13:46:24 -0500, Phil Hobbs >>>>> <pcdhSpamMeSenseless@electrooptical.net> wrote: >>>>> >>>>>> On 12/10/2015 01:41 PM, John Larkin wrote: >>>>>>> On Thu, 10 Dec 2015 05:11:41 -0800 (PST), Phil Hobbs >>>>>>> <pcdhobbs@gmail.com> wrote: >>>>>>> >>>>>>>>> Teflon and some ceramics are the best capacitor types; 100 pF >>>>>>>>> gives you >>>>>>>> Iout = (1e-10) dV/dt >>>>>>>>> and I've made ramp generators that did about 10 mv/hour, and can keep >>>>>>>>> it up all day. that'd be Iout in the 3e-16 amps range... >>>>>>>> >>>>>>>> HP 3325As make nice ramps down to 1 uHz, and they're nearly as >>>>>>>> cheap as 610Cs. It's not a modern DDS/arb architecture, so I don't >>>>>>>> think its output is a staircase. >>>>>>>> >>>>>>>> >>>>>>>> Cheers >>>>>>>> >>>>>>>> Phil Hobbs >>>>>>> >>>>>>> You could charge a 10 uF cap through a 1G resistor, and couple that >>>>>>> into a picoammeter through a 10 pF cap. 10 volts makes 10 nA which >>>>>>> becomes 10 fA. >>>>>>> >>>>>>> >>>>>>> >>>>>>> >>>>>> >>>>>> Ceramic caps that big are pretty nonlinear to use for calibration, no? >>>>>> I have some 2-uf 200V dogbone mylars. When my 610C gets here this >>>>>> evening I'll test their insulation resistance. ;) >>>>> >>>>> A 10 pF NPO should be pretty good on the output side. I'd use a film >>>>> cap for the big one. >>>>> >>>>> >>>>>> >>>>>> An LM662 triangle wave generator with one of your patented >>>>>> bootstrapped-voltage-references could go pretty slowly. >>>>>> >>>>> >>>>> Yeah, if you want a linear ramp a bootstrap current source is a good >>>>> way to go, almost the only way to go if you want, say, 1 nA. >>>>> >>>>> Film cap, cmos opamp follower, bandgap level shifter, 1G resistor back >>>>> to the cap. Reset it with a paper clip. >>>> >>>> >>>> The $125 610C arrived just now, by UPS ground all the way from SoCal. At >>>> first blush, it looks like it works pretty well. It zeroes fine, which >>>> is one of the main indications of health. No trace of stickiness in the >>>> meter movement, which is another important indication. >>>> >>>> On the 10 fA FS range, with the input disconnected but with the cap off >>>> the connector, waving my hand a foot away from the front panel makes the >>>> meter move all over the place. >>>> >>>> Works on both the slow and fast settings. >>>> >>>> Very pretty! >>>> >>> >>> Resistance reads about 2% low up to 1E12 ohms, calibrated using a 5-1/2 >>> digit, 10M DVM and a 10V HP precision power supply. A Russian 1 teraohm >>> 10% resistor reads about 0.86 Tohms on the 610C and about 0.89 Tohms >>> using the digital voltmeter trick. The fast setting pegs the meter in >>> the resistance mode. (I didn't try re-zeroing it.) >>> >>> Voltage is within 1% where I looked. >>> >>> Stability is excellent--on the 10 fA FS range, it stays zeroed to within >>> 1% of FS over several minutes, and that's after only about half an hour >>> of warm-up. >>> >>> With a 100 pF polystyrene cap across the input (just so I know roughly >>> what the capacitance is there), and measuring the 0-3V output on the >>> back near full scale (2.986V), the voltage decay is less than 100 >>> uV/min and is consistent with zero. >>> >>> That means the input current on the 10V range is less than >>> 100uV * 10V/3V * 100 pF /60 s = 0.5 fA. Amazing. (Not bad on the >>> polystyrene, either.) >>> >>> I measured a 2.2 uF 600V wound mylar, charged to 10V for a few seconds. >>> On the 10-V range, it produced nearly full scale on the 0-3V DC output. >>> This output dropped about 220uV out of 3V in 3 minutes, for a >>> self-discharge time constant of about 2.4 million seconds (about a >>> month). So with a 47 G resistor and an LM662, you could actually make a >>> decent 10 microhertz sawtooth out of it. >>> >>> Not a bad gizmo for $125 plus $25 shipping! >>> >>> Cheers >>> >>> Phil Hobbs >> >> It's a beautiful instrument, as long as you don't look inside. It's >> sort of a rat's nest. >> >> The change you are seeing on the mylar cap may be dielectric >> absorption, maybe not leakage. They are awful that way. >> >> Someone here once posted a long-term measurement of film cap leakage, >> and as I recall he saw essentially zero voltage change per year for >> some. >> >> >> > >Interestingly the major source of instability on the 10 fA range appears >to be slow capacitive pickup from things moving around inside as it >warms up. You can make it move > 10 fA in each direction by pressing >gently on the front panel. Pushing on the range switch knob is about >10x more sensitive. > >When you're down in the hundreds of electrons, these things matter! > >Cheers > >Phil Hobbs1 fA is 6000 of them per second. Hey, back to the flame detector. What does a typical flame sound like? On o/e converter and some headphones might be interesting. Burning cotton must have features. Imagine listening to a fireworks display, or a sparkler, or a gunshot, or a fountain. Lightning. When I was a kid, I used to walk around listening to magnetic fields and electric fields and light. I was a weird kid. -- John Larkin Highland Technology, Inc picosecond timing precision measurement jlarkin att highlandtechnology dott com http://www.highlandtechnology.com
Reply by ●December 11, 20152015-12-11
On 12/11/2015 06:51 PM, John Larkin wrote:> On Fri, 11 Dec 2015 18:18:05 -0500, Phil Hobbs > <pcdhSpamMeSenseless@electrooptical.net> wrote: > >> On 12/11/2015 11:56 AM, John Larkin wrote: >>> On Fri, 11 Dec 2015 11:50:08 -0500, Phil Hobbs >>> <pcdhSpamMeSenseless@electrooptical.net> wrote: >>> >>>> On 12/10/2015 07:25 PM, Phil Hobbs wrote: >>>>> On 12/10/2015 03:51 PM, John Larkin wrote: >>>>>> On Thu, 10 Dec 2015 13:46:24 -0500, Phil Hobbs >>>>>> <pcdhSpamMeSenseless@electrooptical.net> wrote: >>>>>> >>>>>>> On 12/10/2015 01:41 PM, John Larkin wrote: >>>>>>>> On Thu, 10 Dec 2015 05:11:41 -0800 (PST), Phil Hobbs >>>>>>>> <pcdhobbs@gmail.com> wrote: >>>>>>>> >>>>>>>>>> Teflon and some ceramics are the best capacitor types; 100 pF >>>>>>>>>> gives you >>>>>>>>> Iout = (1e-10) dV/dt >>>>>>>>>> and I've made ramp generators that did about 10 mv/hour, and can keep >>>>>>>>>> it up all day. that'd be Iout in the 3e-16 amps range... >>>>>>>>> >>>>>>>>> HP 3325As make nice ramps down to 1 uHz, and they're nearly as >>>>>>>>> cheap as 610Cs. It's not a modern DDS/arb architecture, so I don't >>>>>>>>> think its output is a staircase. >>>>>>>>> >>>>>>>>> >>>>>>>>> Cheers >>>>>>>>> >>>>>>>>> Phil Hobbs >>>>>>>> >>>>>>>> You could charge a 10 uF cap through a 1G resistor, and couple that >>>>>>>> into a picoammeter through a 10 pF cap. 10 volts makes 10 nA which >>>>>>>> becomes 10 fA. >>>>>>>> >>>>>>>> >>>>>>>> >>>>>>>> >>>>>>> >>>>>>> Ceramic caps that big are pretty nonlinear to use for calibration, no? >>>>>>> I have some 2-uf 200V dogbone mylars. When my 610C gets here this >>>>>>> evening I'll test their insulation resistance. ;) >>>>>> >>>>>> A 10 pF NPO should be pretty good on the output side. I'd use a film >>>>>> cap for the big one. >>>>>> >>>>>> >>>>>>> >>>>>>> An LM662 triangle wave generator with one of your patented >>>>>>> bootstrapped-voltage-references could go pretty slowly. >>>>>>> >>>>>> >>>>>> Yeah, if you want a linear ramp a bootstrap current source is a good >>>>>> way to go, almost the only way to go if you want, say, 1 nA. >>>>>> >>>>>> Film cap, cmos opamp follower, bandgap level shifter, 1G resistor back >>>>>> to the cap. Reset it with a paper clip. >>>>> >>>>> >>>>> The $125 610C arrived just now, by UPS ground all the way from SoCal. At >>>>> first blush, it looks like it works pretty well. It zeroes fine, which >>>>> is one of the main indications of health. No trace of stickiness in the >>>>> meter movement, which is another important indication. >>>>> >>>>> On the 10 fA FS range, with the input disconnected but with the cap off >>>>> the connector, waving my hand a foot away from the front panel makes the >>>>> meter move all over the place. >>>>> >>>>> Works on both the slow and fast settings. >>>>> >>>>> Very pretty! >>>>> >>>> >>>> Resistance reads about 2% low up to 1E12 ohms, calibrated using a 5-1/2 >>>> digit, 10M DVM and a 10V HP precision power supply. A Russian 1 teraohm >>>> 10% resistor reads about 0.86 Tohms on the 610C and about 0.89 Tohms >>>> using the digital voltmeter trick. The fast setting pegs the meter in >>>> the resistance mode. (I didn't try re-zeroing it.) >>>> >>>> Voltage is within 1% where I looked. >>>> >>>> Stability is excellent--on the 10 fA FS range, it stays zeroed to within >>>> 1% of FS over several minutes, and that's after only about half an hour >>>> of warm-up. >>>> >>>> With a 100 pF polystyrene cap across the input (just so I know roughly >>>> what the capacitance is there), and measuring the 0-3V output on the >>>> back near full scale (2.986V), the voltage decay is less than 100 >>>> uV/min and is consistent with zero. >>>> >>>> That means the input current on the 10V range is less than >>>> 100uV * 10V/3V * 100 pF /60 s = 0.5 fA. Amazing. (Not bad on the >>>> polystyrene, either.) >>>> >>>> I measured a 2.2 uF 600V wound mylar, charged to 10V for a few seconds. >>>> On the 10-V range, it produced nearly full scale on the 0-3V DC output. >>>> This output dropped about 220uV out of 3V in 3 minutes, for a >>>> self-discharge time constant of about 2.4 million seconds (about a >>>> month). So with a 47 G resistor and an LM662, you could actually make a >>>> decent 10 microhertz sawtooth out of it. >>>> >>>> Not a bad gizmo for $125 plus $25 shipping! >>>> >>>> Cheers >>>> >>>> Phil Hobbs >>> >>> It's a beautiful instrument, as long as you don't look inside. It's >>> sort of a rat's nest. >>> >>> The change you are seeing on the mylar cap may be dielectric >>> absorption, maybe not leakage. They are awful that way. >>> >>> Someone here once posted a long-term measurement of film cap leakage, >>> and as I recall he saw essentially zero voltage change per year for >>> some. >>> >>> >>> >> >> Interestingly the major source of instability on the 10 fA range appears >> to be slow capacitive pickup from things moving around inside as it >> warms up. You can make it move > 10 fA in each direction by pressing >> gently on the front panel. Pushing on the range switch knob is about >> 10x more sensitive. >> >> When you're down in the hundreds of electrons, these things matter! >> >> Cheers >> >> Phil Hobbs > > 1 fA is 6000 of them per second.Yup. The meter is a bit faster than that, of course, and the fluctuations are smaller than 1/10 of full scale (fun).> > Hey, back to the flame detector. What does a typical flame sound like? > On o/e converter and some headphones might be interesting. Burning > cotton must have features.Diffusion flames such as candles and small sparks are pretty quiet, I expect. Certainly quiet compared with all the vibrating duct work and reciprocating machinery and cussing workers. ;)> > Imagine listening to a fireworks display, or a sparkler, or a gunshot, > or a fountain. Lightning. > > When I was a kid, I used to walk around listening to magnetic fields > and electric fields and light. I was a weird kid.When did that end? Cheers Phil Hobbs -- Dr Philip C D Hobbs Principal Consultant ElectroOptical Innovations LLC Optics, Electro-optics, Photonics, Analog Electronics 160 North State Road #203 Briarcliff Manor NY 10510 hobbs at electrooptical dot net http://electrooptical.net
Reply by ●December 11, 20152015-12-11
On Friday, December 11, 2015 at 6:18:15 PM UTC-5, Phil Hobbs wrote:> On 12/11/2015 11:56 AM, John Larkin wrote: > > On Fri, 11 Dec 2015 11:50:08 -0500, Phil Hobbs > > <pcdhSpamMeSenseless@electrooptical.net> wrote: > > > >> On 12/10/2015 07:25 PM, Phil Hobbs wrote: > >>> On 12/10/2015 03:51 PM, John Larkin wrote: > >>>> On Thu, 10 Dec 2015 13:46:24 -0500, Phil Hobbs > >>>> <pcdhSpamMeSenseless@electrooptical.net> wrote: > >>>> > >>>>> On 12/10/2015 01:41 PM, John Larkin wrote: > >>>>>> On Thu, 10 Dec 2015 05:11:41 -0800 (PST), Phil Hobbs > >>>>>> <pcdhobbs@gmail.com> wrote: > >>>>>> > >>>>>>>> Teflon and some ceramics are the best capacitor types; 100 pF > >>>>>>>> gives you > >>>>>>> Iout = (1e-10) dV/dt > >>>>>>>> and I've made ramp generators that did about 10 mv/hour, and can keep > >>>>>>>> it up all day. that'd be Iout in the 3e-16 amps range... > >>>>>>> > >>>>>>> HP 3325As make nice ramps down to 1 uHz, and they're nearly as > >>>>>>> cheap as 610Cs. It's not a modern DDS/arb architecture, so I don't > >>>>>>> think its output is a staircase. > >>>>>>> > >>>>>>> > >>>>>>> Cheers > >>>>>>> > >>>>>>> Phil Hobbs > >>>>>> > >>>>>> You could charge a 10 uF cap through a 1G resistor, and couple that > >>>>>> into a picoammeter through a 10 pF cap. 10 volts makes 10 nA which > >>>>>> becomes 10 fA. > >>>>>> > >>>>>> > >>>>>> > >>>>>> > >>>>> > >>>>> Ceramic caps that big are pretty nonlinear to use for calibration, no? > >>>>> I have some 2-uf 200V dogbone mylars. When my 610C gets here this > >>>>> evening I'll test their insulation resistance. ;) > >>>> > >>>> A 10 pF NPO should be pretty good on the output side. I'd use a film > >>>> cap for the big one. > >>>> > >>>> > >>>>> > >>>>> An LM662 triangle wave generator with one of your patented > >>>>> bootstrapped-voltage-references could go pretty slowly. > >>>>> > >>>> > >>>> Yeah, if you want a linear ramp a bootstrap current source is a good > >>>> way to go, almost the only way to go if you want, say, 1 nA. > >>>> > >>>> Film cap, cmos opamp follower, bandgap level shifter, 1G resistor back > >>>> to the cap. Reset it with a paper clip. > >>> > >>> > >>> The $125 610C arrived just now, by UPS ground all the way from SoCal. At > >>> first blush, it looks like it works pretty well. It zeroes fine, which > >>> is one of the main indications of health. No trace of stickiness in the > >>> meter movement, which is another important indication. > >>> > >>> On the 10 fA FS range, with the input disconnected but with the cap off > >>> the connector, waving my hand a foot away from the front panel makes the > >>> meter move all over the place. > >>> > >>> Works on both the slow and fast settings. > >>> > >>> Very pretty! > >>> > >> > >> Resistance reads about 2% low up to 1E12 ohms, calibrated using a 5-1/2 > >> digit, 10M DVM and a 10V HP precision power supply. A Russian 1 teraohm > >> 10% resistor reads about 0.86 Tohms on the 610C and about 0.89 Tohms > >> using the digital voltmeter trick. The fast setting pegs the meter in > >> the resistance mode. (I didn't try re-zeroing it.) > >> > >> Voltage is within 1% where I looked. > >> > >> Stability is excellent--on the 10 fA FS range, it stays zeroed to within > >> 1% of FS over several minutes, and that's after only about half an hour > >> of warm-up. > >> > >> With a 100 pF polystyrene cap across the input (just so I know roughly > >> what the capacitance is there), and measuring the 0-3V output on the > >> back near full scale (2.986V), the voltage decay is less than 100 > >> uV/min and is consistent with zero. > >> > >> That means the input current on the 10V range is less than > >> 100uV * 10V/3V * 100 pF /60 s = 0.5 fA. Amazing. (Not bad on the > >> polystyrene, either.) > >> > >> I measured a 2.2 uF 600V wound mylar, charged to 10V for a few seconds. > >> On the 10-V range, it produced nearly full scale on the 0-3V DC output. > >> This output dropped about 220uV out of 3V in 3 minutes, for a > >> self-discharge time constant of about 2.4 million seconds (about a > >> month). So with a 47 G resistor and an LM662, you could actually make a > >> decent 10 microhertz sawtooth out of it. > >> > >> Not a bad gizmo for $125 plus $25 shipping! > >> > >> Cheers > >> > >> Phil Hobbs > > > > It's a beautiful instrument, as long as you don't look inside. It's > > sort of a rat's nest. > > > > The change you are seeing on the mylar cap may be dielectric > > absorption, maybe not leakage. They are awful that way. > > > > Someone here once posted a long-term measurement of film cap leakage, > > and as I recall he saw essentially zero voltage change per year for > > some. > > > > > > > > Interestingly the major source of instability on the 10 fA range appears > to be slow capacitive pickup from things moving around inside as it > warms up. You can make it move > 10 fA in each direction by pressing > gently on the front panel. Pushing on the range switch knob is about > 10x more sensitive. > > When you're down in the hundreds of electrons, these things matter! > > Cheers > > Phil HobbsI assume you're keeping yourself ground during these proddings, (I grab the front panel ground... what did you grab?) Isn't there some bootstrap/feedback thing to get rid of the input C? That will depend the panel shape... and input knobs/switches are sensitive to pickup. (I just shipped this one/ two/ few of circuit in a plastic box. I had to put a ground strapped lug onto a toggle switch. Else when you touched it, you got 10's mV into 100 k ohm. well that's the pickup in my lab.) (Almost all my circuits suffer the cardinal sin of running the signal through control panel pots and switches... I tuck all the active elements right next to the switches/ pots, and that's worked so far. :^) George H.> > -- > Dr Philip C D Hobbs > Principal Consultant > ElectroOptical Innovations LLC > Optics, Electro-optics, Photonics, Analog Electronics > > 160 North State Road #203 > Briarcliff Manor NY 10510 > > hobbs at electrooptical dot net > http://electrooptical.net
Reply by ●December 11, 20152015-12-11
On 12/11/2015 07:24 PM, George Herold wrote:> On Friday, December 11, 2015 at 6:18:15 PM UTC-5, Phil Hobbs wrote: >> On 12/11/2015 11:56 AM, John Larkin wrote: >>> On Fri, 11 Dec 2015 11:50:08 -0500, Phil Hobbs >>> <pcdhSpamMeSenseless@electrooptical.net> wrote: >>> >>>> On 12/10/2015 07:25 PM, Phil Hobbs wrote: >>>>> On 12/10/2015 03:51 PM, John Larkin wrote: >>>>>> On Thu, 10 Dec 2015 13:46:24 -0500, Phil Hobbs >>>>>> <pcdhSpamMeSenseless@electrooptical.net> wrote: >>>>>> >>>>>>> On 12/10/2015 01:41 PM, John Larkin wrote: >>>>>>>> On Thu, 10 Dec 2015 05:11:41 -0800 (PST), Phil Hobbs >>>>>>>> <pcdhobbs@gmail.com> wrote: >>>>>>>> >>>>>>>>>> Teflon and some ceramics are the best capacitor types; 100 pF >>>>>>>>>> gives you >>>>>>>>> Iout = (1e-10) dV/dt >>>>>>>>>> and I've made ramp generators that did about 10 mv/hour, and can keep >>>>>>>>>> it up all day. that'd be Iout in the 3e-16 amps range... >>>>>>>>> >>>>>>>>> HP 3325As make nice ramps down to 1 uHz, and they're nearly as >>>>>>>>> cheap as 610Cs. It's not a modern DDS/arb architecture, so I don't >>>>>>>>> think its output is a staircase. >>>>>>>>> >>>>>>>>> >>>>>>>>> Cheers >>>>>>>>> >>>>>>>>> Phil Hobbs >>>>>>>> >>>>>>>> You could charge a 10 uF cap through a 1G resistor, and couple that >>>>>>>> into a picoammeter through a 10 pF cap. 10 volts makes 10 nA which >>>>>>>> becomes 10 fA. >>>>>>>> >>>>>>>> >>>>>>>> >>>>>>>> >>>>>>> >>>>>>> Ceramic caps that big are pretty nonlinear to use for calibration, no? >>>>>>> I have some 2-uf 200V dogbone mylars. When my 610C gets here this >>>>>>> evening I'll test their insulation resistance. ;) >>>>>> >>>>>> A 10 pF NPO should be pretty good on the output side. I'd use a film >>>>>> cap for the big one. >>>>>> >>>>>> >>>>>>> >>>>>>> An LM662 triangle wave generator with one of your patented >>>>>>> bootstrapped-voltage-references could go pretty slowly. >>>>>>> >>>>>> >>>>>> Yeah, if you want a linear ramp a bootstrap current source is a good >>>>>> way to go, almost the only way to go if you want, say, 1 nA. >>>>>> >>>>>> Film cap, cmos opamp follower, bandgap level shifter, 1G resistor back >>>>>> to the cap. Reset it with a paper clip. >>>>> >>>>> >>>>> The $125 610C arrived just now, by UPS ground all the way from SoCal. At >>>>> first blush, it looks like it works pretty well. It zeroes fine, which >>>>> is one of the main indications of health. No trace of stickiness in the >>>>> meter movement, which is another important indication. >>>>> >>>>> On the 10 fA FS range, with the input disconnected but with the cap off >>>>> the connector, waving my hand a foot away from the front panel makes the >>>>> meter move all over the place. >>>>> >>>>> Works on both the slow and fast settings. >>>>> >>>>> Very pretty! >>>>> >>>> >>>> Resistance reads about 2% low up to 1E12 ohms, calibrated using a 5-1/2 >>>> digit, 10M DVM and a 10V HP precision power supply. A Russian 1 teraohm >>>> 10% resistor reads about 0.86 Tohms on the 610C and about 0.89 Tohms >>>> using the digital voltmeter trick. The fast setting pegs the meter in >>>> the resistance mode. (I didn't try re-zeroing it.) >>>> >>>> Voltage is within 1% where I looked. >>>> >>>> Stability is excellent--on the 10 fA FS range, it stays zeroed to within >>>> 1% of FS over several minutes, and that's after only about half an hour >>>> of warm-up. >>>> >>>> With a 100 pF polystyrene cap across the input (just so I know roughly >>>> what the capacitance is there), and measuring the 0-3V output on the >>>> back near full scale (2.986V), the voltage decay is less than 100 >>>> uV/min and is consistent with zero. >>>> >>>> That means the input current on the 10V range is less than >>>> 100uV * 10V/3V * 100 pF /60 s = 0.5 fA. Amazing. (Not bad on the >>>> polystyrene, either.) >>>> >>>> I measured a 2.2 uF 600V wound mylar, charged to 10V for a few seconds. >>>> On the 10-V range, it produced nearly full scale on the 0-3V DC output. >>>> This output dropped about 220uV out of 3V in 3 minutes, for a >>>> self-discharge time constant of about 2.4 million seconds (about a >>>> month). So with a 47 G resistor and an LM662, you could actually make a >>>> decent 10 microhertz sawtooth out of it. >>>> >>>> Not a bad gizmo for $125 plus $25 shipping! >>>> >>>> Cheers >>>> >>>> Phil Hobbs >>> >>> It's a beautiful instrument, as long as you don't look inside. It's >>> sort of a rat's nest. >>> >>> The change you are seeing on the mylar cap may be dielectric >>> absorption, maybe not leakage. They are awful that way. >>> >>> Someone here once posted a long-term measurement of film cap leakage, >>> and as I recall he saw essentially zero voltage change per year for >>> some. >>> >>> >>> >> >> Interestingly the major source of instability on the 10 fA range appears >> to be slow capacitive pickup from things moving around inside as it >> warms up. You can make it move > 10 fA in each direction by pressing >> gently on the front panel. Pushing on the range switch knob is about >> 10x more sensitive. >> >> When you're down in the hundreds of electrons, these things matter! >> >> Cheers >> >> Phil Hobbs > > I assume you're keeping yourself ground during these proddings, > (I grab the front panel ground... what did you grab?) > Isn't there some bootstrap/feedback thing to get rid of the > input C? That will depend the panel shape... > and input knobs/switches are sensitive to pickup. > > (I just shipped this one/ two/ few of circuit in a plastic > box. I had to put a ground strapped lug onto a toggle switch. > Else when you touched it, you got 10's mV into 100 k ohm. > well that's the pickup in my lab.) > (Almost all my circuits suffer the cardinal sin of > running the signal through control panel pots and > switches... I tuck all the active elements right next > to the switches/ pots, and that's worked so far. :^) > > George H. >>This was done with the metal blank cap across the input jack (a Keithley special teflon SO-239). With the cap in place, it makes no difference if I ground myself to the chassis or not. Interestingly, with the cap off, if I grab hold of the connector body (ground) I can still make the meter move a whole lot by flexing my fingers. Cheers Phil Hobbs -- Dr Philip C D Hobbs Principal Consultant ElectroOptical Innovations LLC Optics, Electro-optics, Photonics, Analog Electronics 160 North State Road #203 Briarcliff Manor NY 10510 hobbs at electrooptical dot net http://electrooptical.net
Reply by ●December 11, 20152015-12-11
On Friday, December 11, 2015 at 6:51:37 PM UTC-5, John Larkin wrote:> On Fri, 11 Dec 2015 18:18:05 -0500, Phil Hobbs > <pcdhSpamMeSenseless@electrooptical.net> wrote: > > >On 12/11/2015 11:56 AM, John Larkin wrote: > >> On Fri, 11 Dec 2015 11:50:08 -0500, Phil Hobbs > >> <pcdhSpamMeSenseless@electrooptical.net> wrote: > >> > >>> On 12/10/2015 07:25 PM, Phil Hobbs wrote: > >>>> On 12/10/2015 03:51 PM, John Larkin wrote: > >>>>> On Thu, 10 Dec 2015 13:46:24 -0500, Phil Hobbs > >>>>> <pcdhSpamMeSenseless@electrooptical.net> wrote: > >>>>> > >>>>>> On 12/10/2015 01:41 PM, John Larkin wrote: > >>>>>>> On Thu, 10 Dec 2015 05:11:41 -0800 (PST), Phil Hobbs > >>>>>>> <pcdhobbs@gmail.com> wrote: > >>>>>>> > >>>>>>>>> Teflon and some ceramics are the best capacitor types; 100 pF > >>>>>>>>> gives you > >>>>>>>> Iout = (1e-10) dV/dt > >>>>>>>>> and I've made ramp generators that did about 10 mv/hour, and can keep > >>>>>>>>> it up all day. that'd be Iout in the 3e-16 amps range... > >>>>>>>> > >>>>>>>> HP 3325As make nice ramps down to 1 uHz, and they're nearly as > >>>>>>>> cheap as 610Cs. It's not a modern DDS/arb architecture, so I don't > >>>>>>>> think its output is a staircase. > >>>>>>>> > >>>>>>>> > >>>>>>>> Cheers > >>>>>>>> > >>>>>>>> Phil Hobbs > >>>>>>> > >>>>>>> You could charge a 10 uF cap through a 1G resistor, and couple that > >>>>>>> into a picoammeter through a 10 pF cap. 10 volts makes 10 nA which > >>>>>>> becomes 10 fA. > >>>>>>> > >>>>>>> > >>>>>>> > >>>>>>> > >>>>>> > >>>>>> Ceramic caps that big are pretty nonlinear to use for calibration, no? > >>>>>> I have some 2-uf 200V dogbone mylars. When my 610C gets here this > >>>>>> evening I'll test their insulation resistance. ;) > >>>>> > >>>>> A 10 pF NPO should be pretty good on the output side. I'd use a film > >>>>> cap for the big one. > >>>>> > >>>>> > >>>>>> > >>>>>> An LM662 triangle wave generator with one of your patented > >>>>>> bootstrapped-voltage-references could go pretty slowly. > >>>>>> > >>>>> > >>>>> Yeah, if you want a linear ramp a bootstrap current source is a good > >>>>> way to go, almost the only way to go if you want, say, 1 nA. > >>>>> > >>>>> Film cap, cmos opamp follower, bandgap level shifter, 1G resistor back > >>>>> to the cap. Reset it with a paper clip. > >>>> > >>>> > >>>> The $125 610C arrived just now, by UPS ground all the way from SoCal. At > >>>> first blush, it looks like it works pretty well. It zeroes fine, which > >>>> is one of the main indications of health. No trace of stickiness in the > >>>> meter movement, which is another important indication. > >>>> > >>>> On the 10 fA FS range, with the input disconnected but with the cap off > >>>> the connector, waving my hand a foot away from the front panel makes the > >>>> meter move all over the place. > >>>> > >>>> Works on both the slow and fast settings. > >>>> > >>>> Very pretty! > >>>> > >>> > >>> Resistance reads about 2% low up to 1E12 ohms, calibrated using a 5-1/2 > >>> digit, 10M DVM and a 10V HP precision power supply. A Russian 1 teraohm > >>> 10% resistor reads about 0.86 Tohms on the 610C and about 0.89 Tohms > >>> using the digital voltmeter trick. The fast setting pegs the meter in > >>> the resistance mode. (I didn't try re-zeroing it.) > >>> > >>> Voltage is within 1% where I looked. > >>> > >>> Stability is excellent--on the 10 fA FS range, it stays zeroed to within > >>> 1% of FS over several minutes, and that's after only about half an hour > >>> of warm-up. > >>> > >>> With a 100 pF polystyrene cap across the input (just so I know roughly > >>> what the capacitance is there), and measuring the 0-3V output on the > >>> back near full scale (2.986V), the voltage decay is less than 100 > >>> uV/min and is consistent with zero. > >>> > >>> That means the input current on the 10V range is less than > >>> 100uV * 10V/3V * 100 pF /60 s = 0.5 fA. Amazing. (Not bad on the > >>> polystyrene, either.) > >>> > >>> I measured a 2.2 uF 600V wound mylar, charged to 10V for a few seconds. > >>> On the 10-V range, it produced nearly full scale on the 0-3V DC output. > >>> This output dropped about 220uV out of 3V in 3 minutes, for a > >>> self-discharge time constant of about 2.4 million seconds (about a > >>> month). So with a 47 G resistor and an LM662, you could actually make a > >>> decent 10 microhertz sawtooth out of it. > >>> > >>> Not a bad gizmo for $125 plus $25 shipping! > >>> > >>> Cheers > >>> > >>> Phil Hobbs > >> > >> It's a beautiful instrument, as long as you don't look inside. It's > >> sort of a rat's nest. > >> > >> The change you are seeing on the mylar cap may be dielectric > >> absorption, maybe not leakage. They are awful that way. > >> > >> Someone here once posted a long-term measurement of film cap leakage, > >> and as I recall he saw essentially zero voltage change per year for > >> some. > >> > >> > >> > > > >Interestingly the major source of instability on the 10 fA range appears > >to be slow capacitive pickup from things moving around inside as it > >warms up. You can make it move > 10 fA in each direction by pressing > >gently on the front panel. Pushing on the range switch knob is about > >10x more sensitive. > > > >When you're down in the hundreds of electrons, these things matter! > > > >Cheers > > > >Phil Hobbs > > 1 fA is 6000 of them per second. > > Hey, back to the flame detector. What does a typical flame sound like? > On o/e converter and some headphones might be interesting. Burning > cotton must have features. > > Imagine listening to a fireworks display, or a sparkler, or a gunshot, > or a fountain. Lightning. > > When I was a kid, I used to walk around listening to magnetic fields > and electric fields and light. I was a weird kid.Electric fields and magnetic fields and light. I think that's how we all here make money? (Well except for the dang thermal thing. and of course mechanicals to hold it together...) Aside but 610 related The newark guy stopped by earlier in the week with another sample of the leaky mill max part. I hooked it up and it looks less leaky....(sigh) I didn't have time to do comparison testing. But I feel like the plastic mix is changing. I've got a circuit where I have to have a socket, Which means I have to test each batch... what if it ages over time? (I would have thought someone had already figured out the best plastic for sockets?) Bare single sockets are looking very appealing. They are easy to replace if you can solder! George H.> > > > > -- > > John Larkin Highland Technology, Inc > picosecond timing precision measurement > > jlarkin att highlandtechnology dott com > http://www.highlandtechnology.com
Reply by ●December 11, 20152015-12-11
On Fri, 11 Dec 2015 19:18:37 -0500, Phil Hobbs <pcdhSpamMeSenseless@electrooptical.net> wrote:>On 12/11/2015 06:51 PM, John Larkin wrote: >> On Fri, 11 Dec 2015 18:18:05 -0500, Phil Hobbs >> <pcdhSpamMeSenseless@electrooptical.net> wrote: >> >>> On 12/11/2015 11:56 AM, John Larkin wrote: >>>> On Fri, 11 Dec 2015 11:50:08 -0500, Phil Hobbs >>>> <pcdhSpamMeSenseless@electrooptical.net> wrote: >>>> >>>>> On 12/10/2015 07:25 PM, Phil Hobbs wrote: >>>>>> On 12/10/2015 03:51 PM, John Larkin wrote: >>>>>>> On Thu, 10 Dec 2015 13:46:24 -0500, Phil Hobbs >>>>>>> <pcdhSpamMeSenseless@electrooptical.net> wrote: >>>>>>> >>>>>>>> On 12/10/2015 01:41 PM, John Larkin wrote: >>>>>>>>> On Thu, 10 Dec 2015 05:11:41 -0800 (PST), Phil Hobbs >>>>>>>>> <pcdhobbs@gmail.com> wrote: >>>>>>>>> >>>>>>>>>>> Teflon and some ceramics are the best capacitor types; 100 pF >>>>>>>>>>> gives you >>>>>>>>>> Iout = (1e-10) dV/dt >>>>>>>>>>> and I've made ramp generators that did about 10 mv/hour, and can keep >>>>>>>>>>> it up all day. that'd be Iout in the 3e-16 amps range... >>>>>>>>>> >>>>>>>>>> HP 3325As make nice ramps down to 1 uHz, and they're nearly as >>>>>>>>>> cheap as 610Cs. It's not a modern DDS/arb architecture, so I don't >>>>>>>>>> think its output is a staircase. >>>>>>>>>> >>>>>>>>>> >>>>>>>>>> Cheers >>>>>>>>>> >>>>>>>>>> Phil Hobbs >>>>>>>>> >>>>>>>>> You could charge a 10 uF cap through a 1G resistor, and couple that >>>>>>>>> into a picoammeter through a 10 pF cap. 10 volts makes 10 nA which >>>>>>>>> becomes 10 fA. >>>>>>>>> >>>>>>>>> >>>>>>>>> >>>>>>>>> >>>>>>>> >>>>>>>> Ceramic caps that big are pretty nonlinear to use for calibration, no? >>>>>>>> I have some 2-uf 200V dogbone mylars. When my 610C gets here this >>>>>>>> evening I'll test their insulation resistance. ;) >>>>>>> >>>>>>> A 10 pF NPO should be pretty good on the output side. I'd use a film >>>>>>> cap for the big one. >>>>>>> >>>>>>> >>>>>>>> >>>>>>>> An LM662 triangle wave generator with one of your patented >>>>>>>> bootstrapped-voltage-references could go pretty slowly. >>>>>>>> >>>>>>> >>>>>>> Yeah, if you want a linear ramp a bootstrap current source is a good >>>>>>> way to go, almost the only way to go if you want, say, 1 nA. >>>>>>> >>>>>>> Film cap, cmos opamp follower, bandgap level shifter, 1G resistor back >>>>>>> to the cap. Reset it with a paper clip. >>>>>> >>>>>> >>>>>> The $125 610C arrived just now, by UPS ground all the way from SoCal. At >>>>>> first blush, it looks like it works pretty well. It zeroes fine, which >>>>>> is one of the main indications of health. No trace of stickiness in the >>>>>> meter movement, which is another important indication. >>>>>> >>>>>> On the 10 fA FS range, with the input disconnected but with the cap off >>>>>> the connector, waving my hand a foot away from the front panel makes the >>>>>> meter move all over the place. >>>>>> >>>>>> Works on both the slow and fast settings. >>>>>> >>>>>> Very pretty! >>>>>> >>>>> >>>>> Resistance reads about 2% low up to 1E12 ohms, calibrated using a 5-1/2 >>>>> digit, 10M DVM and a 10V HP precision power supply. A Russian 1 teraohm >>>>> 10% resistor reads about 0.86 Tohms on the 610C and about 0.89 Tohms >>>>> using the digital voltmeter trick. The fast setting pegs the meter in >>>>> the resistance mode. (I didn't try re-zeroing it.) >>>>> >>>>> Voltage is within 1% where I looked. >>>>> >>>>> Stability is excellent--on the 10 fA FS range, it stays zeroed to within >>>>> 1% of FS over several minutes, and that's after only about half an hour >>>>> of warm-up. >>>>> >>>>> With a 100 pF polystyrene cap across the input (just so I know roughly >>>>> what the capacitance is there), and measuring the 0-3V output on the >>>>> back near full scale (2.986V), the voltage decay is less than 100 >>>>> uV/min and is consistent with zero. >>>>> >>>>> That means the input current on the 10V range is less than >>>>> 100uV * 10V/3V * 100 pF /60 s = 0.5 fA. Amazing. (Not bad on the >>>>> polystyrene, either.) >>>>> >>>>> I measured a 2.2 uF 600V wound mylar, charged to 10V for a few seconds. >>>>> On the 10-V range, it produced nearly full scale on the 0-3V DC output. >>>>> This output dropped about 220uV out of 3V in 3 minutes, for a >>>>> self-discharge time constant of about 2.4 million seconds (about a >>>>> month). So with a 47 G resistor and an LM662, you could actually make a >>>>> decent 10 microhertz sawtooth out of it. >>>>> >>>>> Not a bad gizmo for $125 plus $25 shipping! >>>>> >>>>> Cheers >>>>> >>>>> Phil Hobbs >>>> >>>> It's a beautiful instrument, as long as you don't look inside. It's >>>> sort of a rat's nest. >>>> >>>> The change you are seeing on the mylar cap may be dielectric >>>> absorption, maybe not leakage. They are awful that way. >>>> >>>> Someone here once posted a long-term measurement of film cap leakage, >>>> and as I recall he saw essentially zero voltage change per year for >>>> some. >>>> >>>> >>>> >>> >>> Interestingly the major source of instability on the 10 fA range appears >>> to be slow capacitive pickup from things moving around inside as it >>> warms up. You can make it move > 10 fA in each direction by pressing >>> gently on the front panel. Pushing on the range switch knob is about >>> 10x more sensitive. >>> >>> When you're down in the hundreds of electrons, these things matter! >>> >>> Cheers >>> >>> Phil Hobbs >> >> 1 fA is 6000 of them per second. > >Yup. The meter is a bit faster than that, of course, and the >fluctuations are smaller than 1/10 of full scale (fun). >> >> Hey, back to the flame detector. What does a typical flame sound like? >> On o/e converter and some headphones might be interesting. Burning >> cotton must have features. > >Diffusion flames such as candles and small sparks are pretty quiet, I >expect. Certainly quiet compared with all the vibrating duct work and >reciprocating machinery and cussing workers. ;) > >> >> Imagine listening to a fireworks display, or a sparkler, or a gunshot, >> or a fountain. Lightning. >> >> When I was a kid, I used to walk around listening to magnetic fields >> and electric fields and light. I was a weird kid. > >When did that end? > >Cheers > >Phil HobbsBeing a kid, or being weird? Neither. -- John Larkin Highland Technology, Inc picosecond timing precision measurement jlarkin att highlandtechnology dott com http://www.highlandtechnology.com
Reply by ●December 11, 20152015-12-11
On 12/11/2015 08:01 PM, John Larkin wrote:> On Fri, 11 Dec 2015 19:18:37 -0500, Phil Hobbs > <pcdhSpamMeSenseless@electrooptical.net> wrote: > >> On 12/11/2015 06:51 PM, John Larkin wrote: >>> On Fri, 11 Dec 2015 18:18:05 -0500, Phil Hobbs >>> <pcdhSpamMeSenseless@electrooptical.net> wrote: >>> >>>> On 12/11/2015 11:56 AM, John Larkin wrote: >>>>> On Fri, 11 Dec 2015 11:50:08 -0500, Phil Hobbs >>>>> <pcdhSpamMeSenseless@electrooptical.net> wrote: >>>>> >>>>>> On 12/10/2015 07:25 PM, Phil Hobbs wrote: >>>>>>> On 12/10/2015 03:51 PM, John Larkin wrote: >>>>>>>> On Thu, 10 Dec 2015 13:46:24 -0500, Phil Hobbs >>>>>>>> <pcdhSpamMeSenseless@electrooptical.net> wrote: >>>>>>>> >>>>>>>>> On 12/10/2015 01:41 PM, John Larkin wrote: >>>>>>>>>> On Thu, 10 Dec 2015 05:11:41 -0800 (PST), Phil Hobbs >>>>>>>>>> <pcdhobbs@gmail.com> wrote: >>>>>>>>>> >>>>>>>>>>>> Teflon and some ceramics are the best capacitor types; 100 pF >>>>>>>>>>>> gives you >>>>>>>>>>> Iout = (1e-10) dV/dt >>>>>>>>>>>> and I've made ramp generators that did about 10 mv/hour, and can keep >>>>>>>>>>>> it up all day. that'd be Iout in the 3e-16 amps range... >>>>>>>>>>> >>>>>>>>>>> HP 3325As make nice ramps down to 1 uHz, and they're nearly as >>>>>>>>>>> cheap as 610Cs. It's not a modern DDS/arb architecture, so I don't >>>>>>>>>>> think its output is a staircase. >>>>>>>>>>> >>>>>>>>>>> >>>>>>>>>>> Cheers >>>>>>>>>>> >>>>>>>>>>> Phil Hobbs >>>>>>>>>> >>>>>>>>>> You could charge a 10 uF cap through a 1G resistor, and couple that >>>>>>>>>> into a picoammeter through a 10 pF cap. 10 volts makes 10 nA which >>>>>>>>>> becomes 10 fA. >>>>>>>>>> >>>>>>>>>> >>>>>>>>>> >>>>>>>>>> >>>>>>>>> >>>>>>>>> Ceramic caps that big are pretty nonlinear to use for calibration, no? >>>>>>>>> I have some 2-uf 200V dogbone mylars. When my 610C gets here this >>>>>>>>> evening I'll test their insulation resistance. ;) >>>>>>>> >>>>>>>> A 10 pF NPO should be pretty good on the output side. I'd use a film >>>>>>>> cap for the big one. >>>>>>>> >>>>>>>> >>>>>>>>> >>>>>>>>> An LM662 triangle wave generator with one of your patented >>>>>>>>> bootstrapped-voltage-references could go pretty slowly. >>>>>>>>> >>>>>>>> >>>>>>>> Yeah, if you want a linear ramp a bootstrap current source is a good >>>>>>>> way to go, almost the only way to go if you want, say, 1 nA. >>>>>>>> >>>>>>>> Film cap, cmos opamp follower, bandgap level shifter, 1G resistor back >>>>>>>> to the cap. Reset it with a paper clip. >>>>>>> >>>>>>> >>>>>>> The $125 610C arrived just now, by UPS ground all the way from SoCal. At >>>>>>> first blush, it looks like it works pretty well. It zeroes fine, which >>>>>>> is one of the main indications of health. No trace of stickiness in the >>>>>>> meter movement, which is another important indication. >>>>>>> >>>>>>> On the 10 fA FS range, with the input disconnected but with the cap off >>>>>>> the connector, waving my hand a foot away from the front panel makes the >>>>>>> meter move all over the place. >>>>>>> >>>>>>> Works on both the slow and fast settings. >>>>>>> >>>>>>> Very pretty! >>>>>>> >>>>>> >>>>>> Resistance reads about 2% low up to 1E12 ohms, calibrated using a 5-1/2 >>>>>> digit, 10M DVM and a 10V HP precision power supply. A Russian 1 teraohm >>>>>> 10% resistor reads about 0.86 Tohms on the 610C and about 0.89 Tohms >>>>>> using the digital voltmeter trick. The fast setting pegs the meter in >>>>>> the resistance mode. (I didn't try re-zeroing it.) >>>>>> >>>>>> Voltage is within 1% where I looked. >>>>>> >>>>>> Stability is excellent--on the 10 fA FS range, it stays zeroed to within >>>>>> 1% of FS over several minutes, and that's after only about half an hour >>>>>> of warm-up. >>>>>> >>>>>> With a 100 pF polystyrene cap across the input (just so I know roughly >>>>>> what the capacitance is there), and measuring the 0-3V output on the >>>>>> back near full scale (2.986V), the voltage decay is less than 100 >>>>>> uV/min and is consistent with zero. >>>>>> >>>>>> That means the input current on the 10V range is less than >>>>>> 100uV * 10V/3V * 100 pF /60 s = 0.5 fA. Amazing. (Not bad on the >>>>>> polystyrene, either.) >>>>>> >>>>>> I measured a 2.2 uF 600V wound mylar, charged to 10V for a few seconds. >>>>>> On the 10-V range, it produced nearly full scale on the 0-3V DC output. >>>>>> This output dropped about 220uV out of 3V in 3 minutes, for a >>>>>> self-discharge time constant of about 2.4 million seconds (about a >>>>>> month). So with a 47 G resistor and an LM662, you could actually make a >>>>>> decent 10 microhertz sawtooth out of it. >>>>>> >>>>>> Not a bad gizmo for $125 plus $25 shipping! >>>>>> >>>>>> Cheers >>>>>> >>>>>> Phil Hobbs >>>>> >>>>> It's a beautiful instrument, as long as you don't look inside. It's >>>>> sort of a rat's nest. >>>>> >>>>> The change you are seeing on the mylar cap may be dielectric >>>>> absorption, maybe not leakage. They are awful that way. >>>>> >>>>> Someone here once posted a long-term measurement of film cap leakage, >>>>> and as I recall he saw essentially zero voltage change per year for >>>>> some. >>>>> >>>>> >>>>> >>>> >>>> Interestingly the major source of instability on the 10 fA range appears >>>> to be slow capacitive pickup from things moving around inside as it >>>> warms up. You can make it move > 10 fA in each direction by pressing >>>> gently on the front panel. Pushing on the range switch knob is about >>>> 10x more sensitive. >>>> >>>> When you're down in the hundreds of electrons, these things matter! >>>> >>>> Cheers >>>> >>>> Phil Hobbs >>> >>> 1 fA is 6000 of them per second. >> >> Yup. The meter is a bit faster than that, of course, and the >> fluctuations are smaller than 1/10 of full scale (fun). >>> >>> Hey, back to the flame detector. What does a typical flame sound like? >>> On o/e converter and some headphones might be interesting. Burning >>> cotton must have features. >> >> Diffusion flames such as candles and small sparks are pretty quiet, I >> expect. Certainly quiet compared with all the vibrating duct work and >> reciprocating machinery and cussing workers. ;) >> >>> >>> Imagine listening to a fireworks display, or a sparkler, or a gunshot, >>> or a fountain. Lightning. >>> >>> When I was a kid, I used to walk around listening to magnetic fields >>> and electric fields and light. I was a weird kid. >> >> When did that end? >> >> Cheers >> >> Phil Hobbs > > Being a kid, or being weird? Neither. > >I'm relieved to hear that. Thought I might have missed something ultra-important. Cheers Phil Hobbs -- Dr Philip C D Hobbs Principal Consultant ElectroOptical Innovations LLC Optics, Electro-optics, Photonics, Analog Electronics 160 North State Road #203 Briarcliff Manor NY 10510 hobbs at electrooptical dot net http://electrooptical.net
Reply by ●December 11, 20152015-12-11
On Fri, 11 Dec 2015 19:35:13 -0500, Phil Hobbs <pcdhSpamMeSenseless@electrooptical.net> wrote:>On 12/11/2015 07:24 PM, George Herold wrote: >> On Friday, December 11, 2015 at 6:18:15 PM UTC-5, Phil Hobbs wrote: >>> On 12/11/2015 11:56 AM, John Larkin wrote: >>>> On Fri, 11 Dec 2015 11:50:08 -0500, Phil Hobbs >>>> <pcdhSpamMeSenseless@electrooptical.net> wrote: >>>> >>>>> On 12/10/2015 07:25 PM, Phil Hobbs wrote: >>>>>> On 12/10/2015 03:51 PM, John Larkin wrote: >>>>>>> On Thu, 10 Dec 2015 13:46:24 -0500, Phil Hobbs >>>>>>> <pcdhSpamMeSenseless@electrooptical.net> wrote: >>>>>>> >>>>>>>> On 12/10/2015 01:41 PM, John Larkin wrote: >>>>>>>>> On Thu, 10 Dec 2015 05:11:41 -0800 (PST), Phil Hobbs >>>>>>>>> <pcdhobbs@gmail.com> wrote: >>>>>>>>> >>>>>>>>>>> Teflon and some ceramics are the best capacitor types; 100 pF >>>>>>>>>>> gives you >>>>>>>>>> Iout = (1e-10) dV/dt >>>>>>>>>>> and I've made ramp generators that did about 10 mv/hour, and can keep >>>>>>>>>>> it up all day. that'd be Iout in the 3e-16 amps range... >>>>>>>>>> >>>>>>>>>> HP 3325As make nice ramps down to 1 uHz, and they're nearly as >>>>>>>>>> cheap as 610Cs. It's not a modern DDS/arb architecture, so I don't >>>>>>>>>> think its output is a staircase. >>>>>>>>>> >>>>>>>>>> >>>>>>>>>> Cheers >>>>>>>>>> >>>>>>>>>> Phil Hobbs >>>>>>>>> >>>>>>>>> You could charge a 10 uF cap through a 1G resistor, and couple that >>>>>>>>> into a picoammeter through a 10 pF cap. 10 volts makes 10 nA which >>>>>>>>> becomes 10 fA. >>>>>>>>> >>>>>>>>> >>>>>>>>> >>>>>>>>> >>>>>>>> >>>>>>>> Ceramic caps that big are pretty nonlinear to use for calibration, no? >>>>>>>> I have some 2-uf 200V dogbone mylars. When my 610C gets here this >>>>>>>> evening I'll test their insulation resistance. ;) >>>>>>> >>>>>>> A 10 pF NPO should be pretty good on the output side. I'd use a film >>>>>>> cap for the big one. >>>>>>> >>>>>>> >>>>>>>> >>>>>>>> An LM662 triangle wave generator with one of your patented >>>>>>>> bootstrapped-voltage-references could go pretty slowly. >>>>>>>> >>>>>>> >>>>>>> Yeah, if you want a linear ramp a bootstrap current source is a good >>>>>>> way to go, almost the only way to go if you want, say, 1 nA. >>>>>>> >>>>>>> Film cap, cmos opamp follower, bandgap level shifter, 1G resistor back >>>>>>> to the cap. Reset it with a paper clip. >>>>>> >>>>>> >>>>>> The $125 610C arrived just now, by UPS ground all the way from SoCal. At >>>>>> first blush, it looks like it works pretty well. It zeroes fine, which >>>>>> is one of the main indications of health. No trace of stickiness in the >>>>>> meter movement, which is another important indication. >>>>>> >>>>>> On the 10 fA FS range, with the input disconnected but with the cap off >>>>>> the connector, waving my hand a foot away from the front panel makes the >>>>>> meter move all over the place. >>>>>> >>>>>> Works on both the slow and fast settings. >>>>>> >>>>>> Very pretty! >>>>>> >>>>> >>>>> Resistance reads about 2% low up to 1E12 ohms, calibrated using a 5-1/2 >>>>> digit, 10M DVM and a 10V HP precision power supply. A Russian 1 teraohm >>>>> 10% resistor reads about 0.86 Tohms on the 610C and about 0.89 Tohms >>>>> using the digital voltmeter trick. The fast setting pegs the meter in >>>>> the resistance mode. (I didn't try re-zeroing it.) >>>>> >>>>> Voltage is within 1% where I looked. >>>>> >>>>> Stability is excellent--on the 10 fA FS range, it stays zeroed to within >>>>> 1% of FS over several minutes, and that's after only about half an hour >>>>> of warm-up. >>>>> >>>>> With a 100 pF polystyrene cap across the input (just so I know roughly >>>>> what the capacitance is there), and measuring the 0-3V output on the >>>>> back near full scale (2.986V), the voltage decay is less than 100 >>>>> uV/min and is consistent with zero. >>>>> >>>>> That means the input current on the 10V range is less than >>>>> 100uV * 10V/3V * 100 pF /60 s = 0.5 fA. Amazing. (Not bad on the >>>>> polystyrene, either.) >>>>> >>>>> I measured a 2.2 uF 600V wound mylar, charged to 10V for a few seconds. >>>>> On the 10-V range, it produced nearly full scale on the 0-3V DC output. >>>>> This output dropped about 220uV out of 3V in 3 minutes, for a >>>>> self-discharge time constant of about 2.4 million seconds (about a >>>>> month). So with a 47 G resistor and an LM662, you could actually make a >>>>> decent 10 microhertz sawtooth out of it. >>>>> >>>>> Not a bad gizmo for $125 plus $25 shipping! >>>>> >>>>> Cheers >>>>> >>>>> Phil Hobbs >>>> >>>> It's a beautiful instrument, as long as you don't look inside. It's >>>> sort of a rat's nest. >>>> >>>> The change you are seeing on the mylar cap may be dielectric >>>> absorption, maybe not leakage. They are awful that way. >>>> >>>> Someone here once posted a long-term measurement of film cap leakage, >>>> and as I recall he saw essentially zero voltage change per year for >>>> some. >>>> >>>> >>>> >>> >>> Interestingly the major source of instability on the 10 fA range appears >>> to be slow capacitive pickup from things moving around inside as it >>> warms up. You can make it move > 10 fA in each direction by pressing >>> gently on the front panel. Pushing on the range switch knob is about >>> 10x more sensitive. >>> >>> When you're down in the hundreds of electrons, these things matter! >>> >>> Cheers >>> >>> Phil Hobbs >> >> I assume you're keeping yourself ground during these proddings, >> (I grab the front panel ground... what did you grab?) >> Isn't there some bootstrap/feedback thing to get rid of the >> input C? That will depend the panel shape... >> and input knobs/switches are sensitive to pickup. >> >> (I just shipped this one/ two/ few of circuit in a plastic >> box. I had to put a ground strapped lug onto a toggle switch. >> Else when you touched it, you got 10's mV into 100 k ohm. >> well that's the pickup in my lab.) >> (Almost all my circuits suffer the cardinal sin of >> running the signal through control panel pots and >> switches... I tuck all the active elements right next >> to the switches/ pots, and that's worked so far. :^) >> >> George H. >>> > >This was done with the metal blank cap across the input jack (a Keithley >special teflon SO-239). With the cap in place, it makes no difference >if I ground myself to the chassis or not. > >Interestingly, with the cap off, if I grab hold of the connector body >(ground) I can still make the meter move a whole lot by flexing my >fingers. > >Cheers > >Phil HobbsMine's pretty steady at 1 pF full scale. The most sensitive current range, 1e-14 amps full scale, looks useless. It drifts like crazy and tiny mechanical forces on the case move it several fA. My homebrew meter had a similar background drift, 5 fA ballpark. It would take some serious work to measure a few fA with any sort of accuracy. -- John Larkin Highland Technology, Inc picosecond timing precision measurement jlarkin att highlandtechnology dott com http://www.highlandtechnology.com
