lose down this news group.
in [Basics]
|
From: George Herold on 27 Apr 2010 09:43 On Apr 27, 12:20 am, John Larkin <jjlar...(a)highNOTlandTHIStechnologyPART.com> wrote: > On Mon, 26 Apr 2010 20:19:25 -0700 (PDT), George Herold > > > > > > <ggher...(a)gmail.com> wrote: > >On Apr 26, 3:08 pm, terryS <tsanf...(a)nf.sympatico.ca> wrote: > >> Looks like this news group has gone the way of some others. > > >> It's inundated with advertising; for Gucci and Rolex cheap copies, > >> imitation or over-stock Nike shoes etc. doubtful services or site > >> references and so on. > > >> So it's now virtually useless for it's original purpose. i.e. > >> discussion of basic electronics and allied matters. > > >> Personally make a point of NEVER contacting or buying anything from > >> any one of these advertising interlopers to what once was a useful > >> discussion group. > > >> Anyone agree? Bye the bye ...................... goodbye! > > >OK, I still like this group. Here's today's problem. I was measuring > >the power supply noise that was feeding my input opamp. And when I > >did that the noise went down! opamp was opa134 with 1uF film cap to > >the power supply rail and 10k ohm to ground. The power supply came > >from a switcher ~100kHz, I have a ~30kHz LC filter in front of > >it(100uH torroids and ??? tantalum caps.), followed by a two pole > >capacitor multiplier. (2N4401/3..2N3904/6's oscillated too much.) 1 > >kohm, and 100uf Al electro. in parallel with 0.1 uF ceramic. > > Emitter followers love to oscillate when the base impedance is low. > The fix is a series base resistor, or a ferrite bead if you're afraid > of Johnson noise. > > This is an energetic, ballpark 100 MHz oscillator: Yeah with a 2N3904 the thing would sometimes 'sing' up at 2-300MHz.... > > +15 > | > 1K-----+ > | | > | c > +----b 2n3904 > | e > C | > C +-------To opamp supply pin. > | | > GND C > C > | > GND Ferrite beads on the base lead helped a bit... a 10ohm resistor worked better. But with a 2N4401 I never observed any 'singing'. So the beads and R are left in the base and I used the 2N4401's. > >The noise went down from ~8 nV/rtHZ when the 1uF cap was connected to > >ground to about 5% lower... 7.7 nV/ ... Maybe I should start hanging > >1uF metal films everywhere? > > >George H. > > The OPA is spec'd for 8 nv/rthz typ, so you're about there. Most > opamps have really ratty PSRR at high frequencies, so a clean supply > is mandatory. > > John- Hide quoted text - > > - Show quoted text - Yeah the noise is exactly what the spec sheet claims. I was going to have the students measure the noise on the power supply. But I don't know what to tell them when measuring it makes the amplifier noise smaller... very confusing. I sprinkled some film caps on the power supplies but that didn't change anything. When I put some load on the power supply the effect became smaller. (pulled about 100mA from it.) I'm going to try and measure the supply noise with an opamp powered from some other supply.... (Guessing it might be some weird coherence thing.) Oh I also tried a BJT opamp (opa228) Here the effect was much smaller! But still at perhaps the 1% level the noise was less when the opamp was looking at one of it's supply rails that when it was looking at ground. (ground either through a cap or a 1 ohm resistor.) George H.
From: John Larkin on 27 Apr 2010 10:05 On Tue, 27 Apr 2010 00:44:49 -0400, "Paul E. Schoen" <paul(a)pstech-inc.com> wrote: > >"George Herold" <ggherold(a)gmail.com> wrote in message >news:8c2d90ba-9e79-42c5-bf7c-db403e62eb85(a)v18g2000vbc.googlegroups.com... >On Apr 26, 3:08 pm, terryS <tsanf...(a)nf.sympatico.ca> wrote: >> Looks like this news group has gone the way of some others. >> >> It's inundated with advertising; for Gucci and Rolex cheap copies, >> imitation or over-stock Nike shoes etc. doubtful services or site >> references and so on. >> >> So it's now virtually useless for it's original purpose. i.e. >> discussion of basic electronics and allied matters. >> >> Personally make a point of NEVER contacting or buying anything from >> any one of these advertising interlopers to what once was a useful >> discussion group. >> >> Anyone agree? Bye the bye ...................... goodbye! > > >OK, I still like this group. Here's today's problem. I was measuring >the power supply noise that was feeding my input opamp. And when I >did that the noise went down! opamp was opa134 with 1uF film cap to >the power supply rail and 10k ohm to ground. The power supply came >from a switcher ~100kHz, I have a ~30kHz LC filter in front of >it(100uH torroids and ??? tantalum caps.), followed by a two pole >capacitor multiplier. (2N4401/3..2N3904/6's oscillated too much.) 1 >kohm, and 100uf Al electro. in parallel with 0.1 uF ceramic. > >The noise went down from ~8 nV/rtHZ when the 1uF cap was connected to >ground to about 5% lower... 7.7 nV/ ... Maybe I should start hanging >1uF metal films everywhere? > >====================================================================== > >Not big enough to filter spam. I put a couple of 30 uF 660 VAC oil filled >capacitors on my computer screen, and I no longer saw any spam! Wrap your monitor in tinfoil. I've heard that works. John
From: George Herold on 27 Apr 2010 11:10 On Apr 27, 9:43 am, George Herold <ggher...(a)gmail.com> wrote: > On Apr 27, 12:20 am, John Larkin > > > > > > <jjlar...(a)highNOTlandTHIStechnologyPART.com> wrote: > > On Mon, 26 Apr 2010 20:19:25 -0700 (PDT), George Herold > > > <ggher...(a)gmail.com> wrote: > > >On Apr 26, 3:08 pm, terryS <tsanf...(a)nf.sympatico.ca> wrote: > > >> Looks like this news group has gone the way of some others. > > > >> It's inundated with advertising; for Gucci and Rolex cheap copies, > > >> imitation or over-stock Nike shoes etc. doubtful services or site > > >> references and so on. > > > >> So it's now virtually useless for it's original purpose. i.e. > > >> discussion of basic electronics and allied matters. > > > >> Personally make a point of NEVER contacting or buying anything from > > >> any one of these advertising interlopers to what once was a useful > > >> discussion group. > > > >> Anyone agree? Bye the bye ...................... goodbye! > > > >OK, I still like this group. Here's today's problem. I was measuring > > >the power supply noise that was feeding my input opamp. And when I > > >did that the noise went down! opamp was opa134 with 1uF film cap to > > >the power supply rail and 10k ohm to ground. The power supply came > > >from a switcher ~100kHz, I have a ~30kHz LC filter in front of > > >it(100uH torroids and ??? tantalum caps.), followed by a two pole > > >capacitor multiplier. (2N4401/3..2N3904/6's oscillated too much.) 1 > > >kohm, and 100uf Al electro. in parallel with 0.1 uF ceramic. > > > Emitter followers love to oscillate when the base impedance is low. > > The fix is a series base resistor, or a ferrite bead if you're afraid > > of Johnson noise. > > > This is an energetic, ballpark 100 MHz oscillator: > > Yeah with a 2N3904 the thing would sometimes 'sing' up at > 2-300MHz.... > > > > > +15 > > | > > 1K-----+ > > | | > > | c > > +----b 2n3904 > > | e > > C | > > C +-------To opamp supply pin. > > | | > > GND C > > C > > | > > GND > > Ferrite beads on the base lead helped a bit... a 10ohm resistor worked > better. But with a 2N4401 I never observed any 'singing'. So the > beads and R are left in the base and I used the 2N4401's. > > > >The noise went down from ~8 nV/rtHZ when the 1uF cap was connected to > > >ground to about 5% lower... 7.7 nV/ ... Maybe I should start hanging > > >1uF metal films everywhere? > > > >George H. > > > The OPA is spec'd for 8 nv/rthz typ, so you're about there. Most > > opamps have really ratty PSRR at high frequencies, so a clean supply > > is mandatory. > > > John- Hide quoted text - > > > - Show quoted text - > > Yeah the noise is exactly what the spec sheet claims. I was going to > have the students measure the noise on the power supply. But I don't > know what to tell them when measuring it makes the amplifier noise > smaller... very confusing. I sprinkled some film caps on the power > supplies but that didn't change anything. When I put some load on the > power supply the effect became smaller. (pulled about 100mA from it.) > > I'm going to try and measure the supply noise with an opamp powered > from some other supply.... (Guessing it might be some weird coherence > thing.) > > Oh I also tried a BJT opamp (opa228) Here the effect was much > smaller! But still at perhaps the 1% level the noise was less when > the opamp was looking at one of it's supply rails that when it was > looking at ground. (ground either through a cap or a 1 ohm > resistor.) > > George H.- Hide quoted text - > > - Show quoted text - OK for anyone who cares, I was able to measure the ~ 1nV/rtHz of noise on my power rails when I power up the opamp from a different supply. My brain starts to hurt when I try to imagine how looking at the power rail could reduce the meausred noise. So I'm just ovserving it and moving on.... George H.
From: John Larkin on 27 Apr 2010 11:30 On Tue, 27 Apr 2010 08:10:06 -0700 (PDT), George Herold <ggherold(a)gmail.com> wrote: >On Apr 27, 9:43�am, George Herold <ggher...(a)gmail.com> wrote: >> On Apr 27, 12:20�am, John Larkin >> >> >> >> >> >> <jjlar...(a)highNOTlandTHIStechnologyPART.com> wrote: >> > On Mon, 26 Apr 2010 20:19:25 -0700 (PDT), George Herold >> >> > <ggher...(a)gmail.com> wrote: >> > >On Apr 26, 3:08�pm, terryS <tsanf...(a)nf.sympatico.ca> wrote: >> > >> Looks like this news group has gone the way of some others. >> >> > >> It's inundated with advertising; for Gucci and Rolex cheap copies, >> > >> imitation or over-stock Nike shoes etc. doubtful services or site >> > >> references and so on. >> >> > >> So it's now virtually useless for it's original purpose. i.e. >> > >> discussion of basic electronics and allied matters. >> >> > >> Personally make a point of NEVER contacting or buying anything from >> > >> any one of these advertising interlopers to what once was a useful >> > >> discussion group. >> >> > >> Anyone agree? Bye the bye ...................... goodbye! >> >> > >OK, I still like this group. �Here's today's problem. �I was measuring >> > >the power supply noise that was feeding my input opamp. �And when I >> > >did that the noise went down! �opamp was opa134 with 1uF film cap to >> > >the power supply rail and 10k ohm to ground. �The power supply came >> > >from a switcher ~100kHz, I have a ~30kHz LC filter in front of >> > >it(100uH torroids and ??? tantalum caps.), followed by a two pole >> > >capacitor multiplier. (2N4401/3..2N3904/6's oscillated too much.) 1 >> > >kohm, and 100uf Al electro. in parallel with 0.1 uF ceramic. >> >> > Emitter followers love to oscillate when the base impedance is low. >> > The fix is a series base resistor, or a ferrite bead if you're afraid >> > of Johnson noise. >> >> > This is an energetic, ballpark 100 MHz oscillator: >> >> Yeah with a 2N3904 the thing would sometimes 'sing' up at >> 2-300MHz.... >> >> >> >> > � � � � � � � � � � � +15 >> > � � � � � � � � � � � �| >> > � � � � � � � � 1K-----+ >> > � � � � � � � � �| � ��| >> > � � � � � � � � �| � ��c >> > � � � � � � � � �+----b � � 2n3904 >> > � � � � � � � � �| �� �e >> > � � � � � � � � �C �� �| >> > � � � � � � � � �C �� �+-------To opamp supply pin. >> > � � � � � � � � �| � ��| >> > � � � � � � � � GND � �C >> > � � � � � � � � � � � �C >> > � � � � � � � � � � � �| >> > � � � � � � � � � � � GND >> >> Ferrite beads on the base lead helped a bit... a 10ohm resistor worked >> better. �But with a 2N4401 I never observed any 'singing'. �So the >> beads and R are left in the base and I used the 2N4401's. >> >> > >The noise went down from ~8 nV/rtHZ when the 1uF cap was connected to >> > >ground to about 5% lower... 7.7 nV/ ... Maybe I should start hanging >> > >1uF metal films everywhere? >> >> > >George H. >> >> > The OPA is spec'd for 8 nv/rthz typ, so you're about there. Most >> > opamps have really ratty PSRR at high frequencies, so a clean supply >> > is mandatory. >> >> > John- Hide quoted text - >> >> > - Show quoted text - >> >> Yeah the noise is exactly what the spec sheet claims. �I was going to >> have the students measure the noise on the power supply. �But I don't >> know what to tell them when measuring it makes the amplifier noise >> smaller... very confusing. �I sprinkled some film caps on the power >> supplies but that didn't change anything. �When I put some load on the >> power supply the effect became smaller. �(pulled about 100mA from it.) >> >> I'm going to try and measure the supply noise with an opamp powered >> from some other supply.... (Guessing it might be some weird coherence >> thing.) >> >> Oh I also tried a BJT opamp (opa228) �Here the effect was much >> smaller! �But still at perhaps the 1% level the noise was less when >> the opamp was looking at one of it's supply rails that when it was >> looking at ground. �(ground either through a cap or a 1 ohm >> resistor.) >> >> �George H.- Hide quoted text - >> >> - Show quoted text - > >OK for anyone who cares, I was able to measure the ~ 1nV/rtHz of noise >on my power rails when I power up the opamp from a different supply. > >My brain starts to hurt when I try to imagine how looking at the power >rail could reduce the meausred noise. So I'm just ovserving it and >moving on.... > >George H. Opamp front ends tend to rectify ambient RF noise, like from radio stations and cell phones. If they do, minor alterations of the environment, like scope probing or moving cables around or just touching stuff, can change things a lot. Ditto if some other part on the board is oscillating. The most fun is a stage that only oscillates at some operating points, like a class AB power stage. We recently did a complicated bootstrapped power amp that only oscillated when the output was in two narrow ranges in the vicinity of +4 and -4 volts, and the oscillation got into a front-end opamp. Yuk. Whenever we have an unexpected DC offset or noise level, I do the laying-on-of-hands thing. Fingers have amazing RF damping properties. A side benefit is that you can tell if something is running too hot. It's surprising how many techs and EEs won't touch things because they are afraid of being shocked, even when they know none of the supplies exceed 30 volts. I know otherwise sensible people who are afraid of 9 volt batteries. Lots of people are afraid of car batteries. John
From: John Fields on 27 Apr 2010 12:36
On Tue, 27 Apr 2010 07:05:10 -0700, John Larkin <jjlarkin(a)highNOTlandTHIStechnologyPART.com> wrote: >On Tue, 27 Apr 2010 00:44:49 -0400, "Paul E. Schoen" ><paul(a)pstech-inc.com> wrote: > >> >>"George Herold" <ggherold(a)gmail.com> wrote in message >>news:8c2d90ba-9e79-42c5-bf7c-db403e62eb85(a)v18g2000vbc.googlegroups.com... >>On Apr 26, 3:08 pm, terryS <tsanf...(a)nf.sympatico.ca> wrote: >>> Looks like this news group has gone the way of some others. >>> >>> It's inundated with advertising; for Gucci and Rolex cheap copies, >>> imitation or over-stock Nike shoes etc. doubtful services or site >>> references and so on. >>> >>> So it's now virtually useless for it's original purpose. i.e. >>> discussion of basic electronics and allied matters. >>> >>> Personally make a point of NEVER contacting or buying anything from >>> any one of these advertising interlopers to what once was a useful >>> discussion group. >>> >>> Anyone agree? Bye the bye ...................... goodbye! >> >> >>OK, I still like this group. Here's today's problem. I was measuring >>the power supply noise that was feeding my input opamp. And when I >>did that the noise went down! opamp was opa134 with 1uF film cap to >>the power supply rail and 10k ohm to ground. The power supply came >>from a switcher ~100kHz, I have a ~30kHz LC filter in front of >>it(100uH torroids and ??? tantalum caps.), followed by a two pole >>capacitor multiplier. (2N4401/3..2N3904/6's oscillated too much.) 1 >>kohm, and 100uf Al electro. in parallel with 0.1 uF ceramic. >> >>The noise went down from ~8 nV/rtHZ when the 1uF cap was connected to >>ground to about 5% lower... 7.7 nV/ ... Maybe I should start hanging >>1uF metal films everywhere? >> >>====================================================================== >> >>Not big enough to filter spam. I put a couple of 30 uF 660 VAC oil filled >>capacitors on my computer screen, and I no longer saw any spam! > >Wrap your monitor in tinfoil. I've heard that works. --- A little pricey, though... http://www.alfa.com/en/gp140w.pgm JF |