Forums

fast ramp follies

Started by John Larkin August 14, 2012
On Monday, August 13, 2012 8:49:33 PM UTC-7, John Larkin wrote:
> On Mon, 13 Aug 2012 20:38:56 -0700, miso <miso@sushi.com> wrote: > > >> OK, I need to charge a capacitor with a stable constant current. The > > >> desired slope is about a volt per nanosecond.
> >I'd be more inclined to cascode the current source output rather than > > >add an inductor. > > > > Would that help? The beta error would increase, and cascodes can > > oscillate
If beta 'error' is a problem, either use a transistor that's sorted by beta (rather than the less-expensive 'general purpose' part number), or you can add a transistor base resistor and feed back to the op amp to completely remove base current (small base resistor, relatively large feedback resistors). There was a good treatment, I think, in Electronic Design, summer 2006 if my notes are accurate. If you know beta, the only part of the 'error' that need be considered is the temperature and aging variation. If beta = 50 is minimum and beta=100 is maximum, the current is 1 to 2 percent of collector current, so you call it 1.5% plus/minus 0.5%. That'd be good for most uses.
On Aug 15, 11:53=A0am, John Larkin
<jjlar...@highNOTlandTHIStechnologyPART.com> wrote:
> On Wed, 15 Aug 2012 04:56:28 -0700 (PDT), dagmargoodb...@yahoo.com > wrote: > > > > > > > > > > >On Aug 14, 11:03=A0pm, John Larkin > ><jjlar...@highNOTlandTHIStechnologyPART.com> wrote: > >> On Tue, 14 Aug 2012 19:00:49 -0700 (PDT), dagmargoodb...@yahoo.com > >> wrote: > > >> >On Aug 14, 8:10 pm, George Herold <gher...@teachspin.com> wrote: > >> >> On Aug 14, 5:57 pm, dagmargoodb...@yahoo.com wrote: > > >> >> > This ckt from Mr. Bloggs has pretty high cuteness: > > >> >> >. =A0 =A0 =A0 =A0 =A0 Vcc > >> >> >. =A0 =A0 =A0 =A0 =A0 | > >> >> >. =A0 =A0 =A0.----+----. > >> >> >. =A0 =A0 =A0| =A0 =A0 =A0 =A0 | > >> >> >. =A0 =A0 =A0| =A0 =A0 =A0 [120] > >> >> >. =A0 =A0 =A0|LM385adj | > >> >> >. =A0 .-----. =A0 =A0 =A0| > >> >> >. =A0 | =A0+ =A0|FB =A0 =A0| > >> >> >. =A0 | =A0 =A0 |------+ > >> >> >. =A0 | =A0- =A0| =A0 =A0 =A0| > >> >> >. =A0 '-----' =A0 =A0 =A0| > >> >> >. =A0 =A0 =A0| =A0 =A0 =A0 |< > >> >> >. =A0 =A0 =A0+-------| 3906 > >> >> >. =A0 =A0 =A0| =A0 =A0 =A0 |\ > >> >> >. =A0 =A0 =A0| =A0 =A0 =A0 =A0| > >> >> >. =A0 =A0[15K] =A0 =A0 =A0| > >> >> >. =A0 =A0 =A0| =A0 =A0 =A0 =A0V > >> >> >. =A0 =A0 =A0| > >> >> >. =A0 =A0 =3D=3D=3D > > >> >> > That gets you an accurate CCS that's slow, all it needs is speed. > > >> >> Maybe hang the fast bits down below the 3906? > > >> >> Or isolate the lm385 from the higher frequencies? > > >> >> George H. > > >> >Yes. =A0I like miso's cascode idea, a fast pnp. > > >> >John's original isolates the PNP with a collector bead/inductor. =A0T=
hat
> >> >offers constant-current compliance that's potentially very fast, but > >> >it's an LC tank. =A0Parasitic feedback collector-to-emitter makes it > >> >sing. I've done that many times at UHF on purpose--they oscillate > >> >pretty nicely. > > >> >So, I'd consider spoiling that LC's Q, in combo with Fred's ckt. If > >> >that's too slow, cascode it. > > >> >Total brute force: a resistor to +100v. =A0Fast, and 1% linear over t=
his
> >> >range. > > >> Most any transistor with a resistor in its emitter is a candidate for > >> oscillation. The collector circuit could be soft or stiff. The usual > >> fix is a base resistor to kill the Q of the base circuit. > > >The fix depends on the mode of oscillation, which I guess you could > >figure out probing the phase. The emitter-follower feeds back e-to-b. > >A resonant base ckt usually makes it oscillate easily at HF-to-VHF, so > >killing the base ckt Q cures that. > > I once theorized that the open-loop output impedance of an opamp would > be about right to kill the q of a base. That was all wrong. > > > > > > > > > > > > >> This is a great interview quiz: > > >> =A0 =A0 =A0 =A0 =A0 =A0 =A0+10 > >> =A0 =A0 =A0 =A0 =A0 =A0 =A0 | > >> =A0 =A0 =A0 =A0 =A0 =A0 =A0 | > >> =A0 =A0 =A0 =A0 =A0 =A0 =A0 | > >> =A0 =A0 =A0 =A0 =A0 =A0 =A0 | > >> =A0 =A0 =A0 =A0 =A0 =A0 =A0 c > >> +5-----------b > >> =A0 =A0 =A0 =A0 =A0 =A0 =A0 e > >> =A0 =A0 =A0 =A0 =A0 =A0 =A0 | > >> =A0 =A0 =A0 =A0 =A0 =A0 =A0 | > >> =A0 =A0 =A0 =A0 =A0 =A0 =A0 1K > >> =A0 =A0 =A0 =A0 =A0 =A0 =A0 | > >> =A0 =A0 =A0 =A0 =A0 =A0 =A0 | > >> =A0 =A0 =A0 =A0 =A0 =A0 =A0 | > >> =A0 =A0 =A0 =A0 =A0 =A0 =A0gnd > > >> If they can tell you the b/c/e voltages, hire them. Currents, offer > >> more. If they mention oscillation, offer them stock. > > >> 100 volts * 10 mA is, sadly, 1 watt. > > >Yes, of course. =A0Since you're kind of a devil-may-care guy, watt's a > >what or two, eh watt? > > The specfic gadget here is powered by USB. What with crappy laptops > and ultra-cheap cables (copperclad steel conductors? 28 gage? Copper > is expensive!) you're lucky to get 4 volts at half an amp. Gotta start > groveling for milliwatts again. > > I wish I had a fast current source that ran directly from V+, which > might be as low as, say, 4.5 volts. That leaves 2 volts at the top of > my ramp. > > > > >> 10 volts through a resistor > >> wouldn't be bad. A simple software hack could take out the curvature. > >> I probably should have done that. Resistors don't often oscillate. > > >> I don't see how a cascode helps much. Beta error doubles, and now you > >> have two transistors to oscillate. > > >You're right about the beta error of course. =A0My first impulse would > >be to try RF-stabilizing the original stage by damping the collector > >(or VHF-bypassing the emitter), with Fred's LM385-adj for d.c. > >stability. > > Bypassing the emitter is interesting. That would be way outside the > opamp loop territory. The LM7301 is 4 MHz, the oscillation is probably > hundreds of MHz, so there's room to work. I can't simulate this, so > I'd have to breadboard it. > > The base resistor works OK, but probably lowers the collector > impedance and degrades the ramp a little. > > > > >Cascoded transistors (somewhat) protect each other from oscillating by > >breaking (or at least suppressing) the c-to-e feedback path. Hard- > >bypassing the cascode base is standard practice in UHF--prevents > >feedback nasties at the base. =A0Cascodes don't oscillate, generally. > >If they did, we wouldn't love them as we do. > > >From your frequencies, Q1's oscillation feedback is likely c-to-e > >(Cce). =A0Collector-load L1 makes the stage high-gain, and L1 resonates > >with strays at *some* frequency. If the gain is high enough, it'll > >oscillate. Damping the resonance attenuates the feedback. =A0Suppress > >the feedback sufficiently, kill the oscillation. > > The ferrite isn't high-Q, and eliminating it doesn't kill oscillation.
It's a CCS, so the collector load's the ramp cap. I guess that dominates. It's oscillating, so there's feedback somewhere. From another guy I'd ask about magnetic coupling in the layout, but I doubt your layouts would do that. That's why I figure it's in the transistor itself, and why it changes with Q1.
> It does tend to improve ramp linearity by disconnecting the transistor > capacitance from the ramp cap. Some. > > So the only difference between the upper transistor and the lower > cascode transistor is that the cascode has a hard base bypass, and the > upper is driven by the opamp?
Usually, yes. Here I was thinking the upper (d.c.) transistor could be slow, the cascode transistor fast.
> This is the thing that JT missed. He assumed we had a classic > gain-phase loop dynamics oscillation (with an 800 KHz opamp and a 100 > MHz oscillation! He never explained that.) What's really happening is > that the transistor capacitances, the base wirebond
Wirebond stuff is usually well above the 70Mhz oscillation of the 1st bjt you tried.
> and leads, the pcb > trace,
Hopefully minimal, with SMD layout.
> and the equivalent stuff inside the opamp (leads, wirebonds, > esd diodes, finally transistors) are part of a resonant base circuit. > 50 or 100 ohms of base resistor has no significant effect on > closed-loop dynamics, but kills the Q of that resonant system.
-- Cheers, James Arthur
On Aug 15, 2:23=A0pm, Gerhard Hoffmann <dk...@arcor.de> wrote:
> Am 15.08.2012 18:47, schrieb Jim Thompson: > > >> I once theorized that the open-loop output impedance of an opamp would > >> be about right to kill the q of a base. That was all wrong. > > > That's because an OpAmp's open-loop output impedance is not resistive. > > Unfortunately most manufacturers incorrectly model it that way. > > It is not so much the model but the use. > A voltage source that drops 6 dB / octave looks inductive (unless one > has done additional esoteric buffering etc) and adding a capacitive load > yields a nice resonance. > > Also, someone above praised cascodes and that they would not oscillate. > Nothing could be more wrong.
That was me. They don't generally oscillate, meaning that's not a general, usual result. They certainly can--I made a cascode oscillator-frequency multiplier once. It worked amazingly well.
> At ReallyHighFrequency(tm) the circuit topology is not the same. > The load at the collector is way above its resonance -> low, > capacitive impedance. Base is grounded with the usual parasitic L,
Hopefully base inductance is minimal. If L(base bond wire) is a typical SOT-23-ish 1.5nH, that resonates with 10pF at 1.3GHz. I assumed that's not the feedback mechanism, since John's first transistor oscillated at 70MHz. If the op-amp output is some complex impedance, which it well could be, that falls apart.
> and the emitter looks into the high impedance of the driving stage. > That is like the circuit of the usual UHF VCO, aka loaded follower. > Base input impedance is capacitive with with say, -50 Ohm in series. > Oscillates with parasitic base L.
If that's the mode, a base resistor would kill it, agreed?
> Hi, JL, does the usual negative CML gate provide enough level change > to switch one of these SKY, NEC or Avago depletion FETs completely off? > I think, I'll try it for my pulse stretchers :-) > > regards, Gerhard.
-- Cheers, James Arthur
On Wed, 15 Aug 2012 20:23:56 +0200, Gerhard Hoffmann <dk4xp@arcor.de>
wrote:

>Am 15.08.2012 18:47, schrieb Jim Thompson: > >>> I once theorized that the open-loop output impedance of an opamp would >>> be about right to kill the q of a base. That was all wrong. > >> >> That's because an OpAmp's open-loop output impedance is not resistive. >> Unfortunately most manufacturers incorrectly model it that way. > >It is not so much the model but the use. >A voltage source that drops 6 dB / octave looks inductive (unless one >has done additional esoteric buffering etc) and adding a capacitive load >yields a nice resonance.
The -6 rolloff is usually done somewhere deep inside the chip. At 100 MHz, the opamp gain as such is very low and the rolloff is a lot steeper than -6. I think the transistor oscillation is more dependant on wirebonds and esd diode capacitance and stuff like that, nothing that an opamp Spice model is likely to sim usefully. Fast transistors used as classic emitter followers (stiff bypassed base and collector) tend to oscillate. They don't need additional inductance. A base resistor will kill the oscillation, but an opamp output impedance usually won't. Somebody with a VNA could measure the impedance of some opamp outputs in the hundreds of MHz... that would be interesting. I suspect the low frequency loop (follower, inverter, whatever) wouldn't change things much at high frequencies. Maybe I'll TDR some opamp outputs. I don't have a VNA but I can do that.
> > > > >Also, someone above praised cascodes and that they would not oscillate. >Nothing could be more wrong. > >At ReallyHighFrequency(tm) the circuit topology is not the same. >The load at the collector is way above its resonance -> low, >capacitive impedance. Base is grounded with the usual parasitic L, >and the emitter looks into the high impedance of the driving stage. >That is like the circuit of the usual UHF VCO, aka loaded follower. >Base input impedance is capacitive with with say, -50 Ohm in series. >Oscillates with parasitic base L.
Yeah, sounds like a cascode has about the same hazards as my current source transistor, except that the cascode transistor isn't connected to the opamp output, with its own unknown complexities.
> > > > >Hi, JL, does the usual negative CML gate provide enough level change >to switch one of these SKY, NEC or Avago depletion FETs completely off? >I think, I'll try it for my pulse stretchers :-)
If you terminate them as you're supposed to, you only get 0.4 volts single-ended swing, not enough. The NE3508/9 can be driven nicely from a full-swing ECL (eclips lite, eclips plus) gate, with margin to enhance a bit (+0.2 to -0.6 maybe), or full-swing ECL can drive the neat Avago enhancement parts. MC10EL/EP89 swings about 2 volts, which can drive meatier depletion fets that want more swing. Beyond that, it's hard to get fast edges with volts of swing. I took a bunch of data on the NE3508/9 if you're interested. Bummer that the RF boys usually furnish pitiful DC/time domain specs. -- John Larkin Highland Technology, Inc jlarkin at highlandtechnology dot com http://www.highlandtechnology.com Precision electronic instrumentation Picosecond-resolution Digital Delay and Pulse generators Custom laser drivers and controllers Photonics and fiberoptic TTL data links VME thermocouple, LVDT, synchro acquisition and simulation
On Aug 15, 2:58=A0pm, whit3rd <whit...@gmail.com> wrote:
> On Monday, August 13, 2012 8:49:33 PM UTC-7, John Larkin wrote: > > On Mon, 13 Aug 2012 20:38:56 -0700, miso <m...@sushi.com> wrote: > > > >> OK, I need to charge a capacitor with a stable constant current. The > > > >> desired slope is about a volt per nanosecond. > > >I'd be more inclined to cascode the current source output rather than > > > >add an inductor. > > > Would that help? The beta error would increase, and cascodes can > > > oscillate > > If beta 'error' is a problem, either use a transistor that's sorted by be=
ta
> (rather than the =A0less-expensive 'general purpose' part number), or > you can add a transistor base resistor and feed back to the op amp > to completely remove base current (small base resistor, relatively > large feedback resistors). =A0There was a good treatment, I think, > in Electronic Design, summer 2006 if my notes are accurate.
Is this the one you're thinking of? http://groups.google.com/group/sci.electronics.design/browse_thread/thread/= 5bcae9656debc0f5/35371afc2d377036?hl=3Den&lnk=3Dgst&q=3Dresistor+compenate+= arthur#35371afc2d377036 -- Cheers, James Arthur
> If you know beta, the only part of the 'error' that need be considered is > the temperature and aging variation. =A0If beta =3D 50 is minimum > and beta=3D100 is maximum, the current is 1 to 2 percent of collector > current, so you call it 1.5% plus/minus 0.5%. =A0 That'd be =A0good for m=
ost uses.
On Wed, 15 Aug 2012 12:50:15 -0700 (PDT), dagmargoodboat@yahoo.com
wrote:

>On Aug 15, 11:53&#2013266080;am, John Larkin ><jjlar...@highNOTlandTHIStechnologyPART.com> wrote: >> On Wed, 15 Aug 2012 04:56:28 -0700 (PDT), dagmargoodb...@yahoo.com >> wrote: >> >> >> >> >> >> >> >> >> >> >On Aug 14, 11:03&#2013266080;pm, John Larkin >> ><jjlar...@highNOTlandTHIStechnologyPART.com> wrote: >> >> On Tue, 14 Aug 2012 19:00:49 -0700 (PDT), dagmargoodb...@yahoo.com >> >> wrote: >> >> >> >On Aug 14, 8:10 pm, George Herold <gher...@teachspin.com> wrote: >> >> >> On Aug 14, 5:57 pm, dagmargoodb...@yahoo.com wrote: >> >> >> >> > This ckt from Mr. Bloggs has pretty high cuteness: >> >> >> >> >. &#2013266080; &#2013266080; &#2013266080; &#2013266080; &#2013266080; Vcc >> >> >> >. &#2013266080; &#2013266080; &#2013266080; &#2013266080; &#2013266080; | >> >> >> >. &#2013266080; &#2013266080; &#2013266080;.----+----. >> >> >> >. &#2013266080; &#2013266080; &#2013266080;| &#2013266080; &#2013266080; &#2013266080; &#2013266080; | >> >> >> >. &#2013266080; &#2013266080; &#2013266080;| &#2013266080; &#2013266080; &#2013266080; [120] >> >> >> >. &#2013266080; &#2013266080; &#2013266080;|LM385adj | >> >> >> >. &#2013266080; .-----. &#2013266080; &#2013266080; &#2013266080;| >> >> >> >. &#2013266080; | &#2013266080;+ &#2013266080;|FB &#2013266080; &#2013266080;| >> >> >> >. &#2013266080; | &#2013266080; &#2013266080; |------+ >> >> >> >. &#2013266080; | &#2013266080;- &#2013266080;| &#2013266080; &#2013266080; &#2013266080;| >> >> >> >. &#2013266080; '-----' &#2013266080; &#2013266080; &#2013266080;| >> >> >> >. &#2013266080; &#2013266080; &#2013266080;| &#2013266080; &#2013266080; &#2013266080; |< >> >> >> >. &#2013266080; &#2013266080; &#2013266080;+-------| 3906 >> >> >> >. &#2013266080; &#2013266080; &#2013266080;| &#2013266080; &#2013266080; &#2013266080; |\ >> >> >> >. &#2013266080; &#2013266080; &#2013266080;| &#2013266080; &#2013266080; &#2013266080; &#2013266080;| >> >> >> >. &#2013266080; &#2013266080;[15K] &#2013266080; &#2013266080; &#2013266080;| >> >> >> >. &#2013266080; &#2013266080; &#2013266080;| &#2013266080; &#2013266080; &#2013266080; &#2013266080;V >> >> >> >. &#2013266080; &#2013266080; &#2013266080;| >> >> >> >. &#2013266080; &#2013266080; === >> >> >> >> > That gets you an accurate CCS that's slow, all it needs is speed. >> >> >> >> Maybe hang the fast bits down below the 3906? >> >> >> >> Or isolate the lm385 from the higher frequencies? >> >> >> >> George H. >> >> >> >Yes. &#2013266080;I like miso's cascode idea, a fast pnp. >> >> >> >John's original isolates the PNP with a collector bead/inductor. &#2013266080;That >> >> >offers constant-current compliance that's potentially very fast, but >> >> >it's an LC tank. &#2013266080;Parasitic feedback collector-to-emitter makes it >> >> >sing. I've done that many times at UHF on purpose--they oscillate >> >> >pretty nicely. >> >> >> >So, I'd consider spoiling that LC's Q, in combo with Fred's ckt. If >> >> >that's too slow, cascode it. >> >> >> >Total brute force: a resistor to +100v. &#2013266080;Fast, and 1% linear over this >> >> >range. >> >> >> Most any transistor with a resistor in its emitter is a candidate for >> >> oscillation. The collector circuit could be soft or stiff. The usual >> >> fix is a base resistor to kill the Q of the base circuit. >> >> >The fix depends on the mode of oscillation, which I guess you could >> >figure out probing the phase. The emitter-follower feeds back e-to-b. >> >A resonant base ckt usually makes it oscillate easily at HF-to-VHF, so >> >killing the base ckt Q cures that. >> >> I once theorized that the open-loop output impedance of an opamp would >> be about right to kill the q of a base. That was all wrong. >> >> >> >> >> >> >> >> >> >> >> >> >> This is a great interview quiz: >> >> >> &#2013266080; &#2013266080; &#2013266080; &#2013266080; &#2013266080; &#2013266080; &#2013266080;+10 >> >> &#2013266080; &#2013266080; &#2013266080; &#2013266080; &#2013266080; &#2013266080; &#2013266080; | >> >> &#2013266080; &#2013266080; &#2013266080; &#2013266080; &#2013266080; &#2013266080; &#2013266080; | >> >> &#2013266080; &#2013266080; &#2013266080; &#2013266080; &#2013266080; &#2013266080; &#2013266080; | >> >> &#2013266080; &#2013266080; &#2013266080; &#2013266080; &#2013266080; &#2013266080; &#2013266080; | >> >> &#2013266080; &#2013266080; &#2013266080; &#2013266080; &#2013266080; &#2013266080; &#2013266080; c >> >> +5-----------b >> >> &#2013266080; &#2013266080; &#2013266080; &#2013266080; &#2013266080; &#2013266080; &#2013266080; e >> >> &#2013266080; &#2013266080; &#2013266080; &#2013266080; &#2013266080; &#2013266080; &#2013266080; | >> >> &#2013266080; &#2013266080; &#2013266080; &#2013266080; &#2013266080; &#2013266080; &#2013266080; | >> >> &#2013266080; &#2013266080; &#2013266080; &#2013266080; &#2013266080; &#2013266080; &#2013266080; 1K >> >> &#2013266080; &#2013266080; &#2013266080; &#2013266080; &#2013266080; &#2013266080; &#2013266080; | >> >> &#2013266080; &#2013266080; &#2013266080; &#2013266080; &#2013266080; &#2013266080; &#2013266080; | >> >> &#2013266080; &#2013266080; &#2013266080; &#2013266080; &#2013266080; &#2013266080; &#2013266080; | >> >> &#2013266080; &#2013266080; &#2013266080; &#2013266080; &#2013266080; &#2013266080; &#2013266080;gnd >> >> >> If they can tell you the b/c/e voltages, hire them. Currents, offer >> >> more. If they mention oscillation, offer them stock. >> >> >> 100 volts * 10 mA is, sadly, 1 watt. >> >> >Yes, of course. &#2013266080;Since you're kind of a devil-may-care guy, watt's a >> >what or two, eh watt? >> >> The specfic gadget here is powered by USB. What with crappy laptops >> and ultra-cheap cables (copperclad steel conductors? 28 gage? Copper >> is expensive!) you're lucky to get 4 volts at half an amp. Gotta start >> groveling for milliwatts again. >> >> I wish I had a fast current source that ran directly from V+, which >> might be as low as, say, 4.5 volts. That leaves 2 volts at the top of >> my ramp. >> >> >> >> >> 10 volts through a resistor >> >> wouldn't be bad. A simple software hack could take out the curvature. >> >> I probably should have done that. Resistors don't often oscillate. >> >> >> I don't see how a cascode helps much. Beta error doubles, and now you >> >> have two transistors to oscillate. >> >> >You're right about the beta error of course. &#2013266080;My first impulse would >> >be to try RF-stabilizing the original stage by damping the collector >> >(or VHF-bypassing the emitter), with Fred's LM385-adj for d.c. >> >stability. >> >> Bypassing the emitter is interesting. That would be way outside the >> opamp loop territory. The LM7301 is 4 MHz, the oscillation is probably >> hundreds of MHz, so there's room to work. I can't simulate this, so >> I'd have to breadboard it. >> >> The base resistor works OK, but probably lowers the collector >> impedance and degrades the ramp a little. >> >> >> >> >Cascoded transistors (somewhat) protect each other from oscillating by >> >breaking (or at least suppressing) the c-to-e feedback path. Hard- >> >bypassing the cascode base is standard practice in UHF--prevents >> >feedback nasties at the base. &#2013266080;Cascodes don't oscillate, generally. >> >If they did, we wouldn't love them as we do. >> >> >From your frequencies, Q1's oscillation feedback is likely c-to-e >> >(Cce). &#2013266080;Collector-load L1 makes the stage high-gain, and L1 resonates >> >with strays at *some* frequency. If the gain is high enough, it'll >> >oscillate. Damping the resonance attenuates the feedback. &#2013266080;Suppress >> >the feedback sufficiently, kill the oscillation. >> >> The ferrite isn't high-Q, and eliminating it doesn't kill oscillation. > >It's a CCS, so the collector load's the ramp cap. I guess that >dominates. It's oscillating, so there's feedback somewhere. From >another guy I'd ask about magnetic coupling in the layout, but I doubt >your layouts would do that. That's why I figure it's in the >transistor itself, and why it changes with Q1. > >> It does tend to improve ramp linearity by disconnecting the transistor >> capacitance from the ramp cap. Some. >> >> So the only difference between the upper transistor and the lower >> cascode transistor is that the cascode has a hard base bypass, and the >> upper is driven by the opamp? > >Usually, yes. Here I was thinking the upper (d.c.) transistor could >be slow, the cascode transistor fast.
What might work is a cascode with a capacitor to ground on the middle node; that would look weird. Transistors are less likely to oscillate if their emitters are at RF ground. The base could be bypassed, too... no soggy Miller effect. That one is worth thinking about. All it needs now is a cute way to fold the cascode base current error back into the loop. -- John Larkin Highland Technology, Inc jlarkin at highlandtechnology dot com http://www.highlandtechnology.com Precision electronic instrumentation Picosecond-resolution Digital Delay and Pulse generators Custom laser drivers and controllers Photonics and fiberoptic TTL data links VME thermocouple, LVDT, synchro acquisition and simulation
whit3rd wrote:
> > On Monday, August 13, 2012 8:49:33 PM UTC-7, John Larkin wrote: > > On Mon, 13 Aug 2012 20:38:56 -0700, miso <miso@sushi.com> wrote: > > > > >> OK, I need to charge a capacitor with a stable constant current. The > > > > >> desired slope is about a volt per nanosecond. > > > >I'd be more inclined to cascode the current source output rather than > > > > >add an inductor. > > > > > > > > Would that help? The beta error would increase, and cascodes can > > > > oscillate > > If beta 'error' is a problem, either use a transistor that's sorted by beta > (rather than the less-expensive 'general purpose' part number), or > you can add a transistor base resistor and feed back to the op amp > to completely remove base current (small base resistor, relatively > large feedback resistors). There was a good treatment, I think, > in Electronic Design, summer 2006 if my notes are accurate. > > If you know beta, the only part of the 'error' that need be considered is > the temperature and aging variation. If beta = 50 is minimum > and beta=100 is maximum, the current is 1 to 2 percent of collector > current, so you call it 1.5% plus/minus 0.5%. That'd be good for most uses.
The beta feedback trick is cute, but it relies on the op amp to compensate, and so only works at low frequency. It also doesn't fix Early effect at high frequency, whereas a properly designed cascode does. Cascodes do double the base current error, and at low supply voltages, you have to reduce the voltage across the emitter resistor of the current source by a V_BE drop, which hurts temperature stability as well as initial accuracy. (RF transistor data sheets rarely mention V_BE, and IME never specify it, so you have to assume that it might vary by 50 mV or so.) PNP RF transistors have fairly horrible betas--the BFT92 is 20 minimum, 50 typical, but no maximum specified--and don't come in beta bins. With matched transistors, e.g. the Intersil HFA series, you can do very well using a Wilson current mirror. It compensates for beta exactly to leading order, and that works well up to frequencies of order f_T/5 or something like that, where you have to worry about phase shifts and second-order effects. With a beta of 20, its error is only about 0.2%. Unfortunately those arrays cost $5 or so. The NPN/PNP series-shunt pair (the one that folks with nothing better to do are are slagging each other off about at the moment) runs the driver transistor at constant I_C, and lets you use a much faster NPN to stiffen the PNP at high frequency. Try simulating it--it's really good medicine. (I spent most of July up to my ears in high performance laser noise cancellers, which rely on a lot of those sorts of tricks, so I'm pretty well up on all this at the moment.) 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 845-480-2058 hobbs at electrooptical dot net http://electrooptical.net
On Aug 15, 10:34=A0pm, Phil Hobbs
<pcdhSpamMeSensel...@electrooptical.net> wrote:
> whit3rd wrote: > > > On Monday, August 13, 2012 8:49:33 PM UTC-7, John Larkin wrote: > > > On Mon, 13 Aug 2012 20:38:56 -0700, miso <m...@sushi.com> wrote: > > > > >> OK, I need to charge a capacitor with a stable constant current. T=
he
> > > > >> desired slope is about a volt per nanosecond. > > > > >I'd be more inclined to cascode the current source output rather tha=
n
> > > > >add an inductor. > > > > Would that help? The beta error would increase, and cascodes can > > > > oscillate > > > If beta 'error' is a problem, either use a transistor that's sorted by =
beta
> > (rather than the =A0less-expensive 'general purpose' part number), or > > you can add a transistor base resistor and feed back to the op amp > > to completely remove base current (small base resistor, relatively > > large feedback resistors). =A0There was a good treatment, I think, > > in Electronic Design, summer 2006 if my notes are accurate. > > > If you know beta, the only part of the 'error' that need be considered =
is
> > the temperature and aging variation. =A0If beta =3D 50 is minimum > > and beta=3D100 is maximum, the current is 1 to 2 percent of collector > > current, so you call it 1.5% plus/minus 0.5%. =A0 That'd be =A0good for=
most uses.
> > The beta feedback trick is cute, but it relies on the op amp to > compensate, and so only works at low frequency. =A0It also doesn't fix > Early effect at high frequency, whereas a properly designed cascode > does. > > Cascodes do double the base current error, and at low supply voltages, > you have to reduce the voltage across the emitter resistor of the > current source by a V_BE drop, which hurts temperature stability as well > as initial accuracy. =A0(RF transistor data sheets rarely mention V_BE, > and IME never specify it, so you have to assume that it might vary by 50 > mV or so.) > > PNP RF transistors have fairly horrible betas--the BFT92 is 20 minimum, > 50 typical, but no maximum specified--and don't come in beta bins. > > With matched transistors, e.g. the Intersil HFA series, you can do very > well using a Wilson current mirror. =A0It compensates for beta exactly to > leading order, and that works well up to frequencies of order f_T/5 or > something like that, where you have to worry about phase shifts and > second-order effects. =A0With a beta of 20, its error is only about 0.2%. > Unfortunately those arrays cost $5 or so. > > The NPN/PNP series-shunt pair (the one that folks with nothing better to > do are are slagging each other off about at the moment)
We aren't slagging one another off about the complementary Darlington as such (or whatever you've chosen to call it) but about the right name to use for it. If you give Google "complementary Darlington", "Sziklai Pair" appears at the top of the list. "NPN/PNP series-shunt pair" doesn't do as well. "Sziklai pair" is probably the most direct reference, but it doesn't mean much if you haven't run into the patent. <snipped sensible stuff that doesn't happen to be relevant to this - very narrow - point> -- Bill Sloman, Nijmegen
On Wednesday, August 15, 2012 1:34:08 PM UTC-7, Phil Hobbs wrote:
> whit3rd wrote: > > > If beta 'error' is a problem, either use a transistor that's sorted by beta > > > (rather than the less-expensive 'general purpose' part number), or > > > you can add a transistor base resistor and feed back to the op amp > > > to completely remove base current
> The beta feedback trick is cute, but it relies on the op amp to > > compensate, and so only works at low frequency. It also doesn't fix > > Early effect at high frequency, whereas a properly designed cascode > > does.
Just as voltage regulators rely on bypass capacitors at high frequency, I'd think of a series inductor to keep the current source happy at high frequencies. I wonder, though, how far one could get with B-E capacitance (with a HF transistor that has relatively small B-C capacitance)? Either way, oscillation is possible and layout will matter. Early effect should only add repeatable curvature, I'd be concerned only because it isn't usually a well-tested parameter.
On Wed, 15 Aug 2012 14:24:16 -0700 (PDT), whit3rd <whit3rd@gmail.com>
wrote:

>On Wednesday, August 15, 2012 1:34:08 PM UTC-7, Phil Hobbs wrote: >> whit3rd wrote: >> >> > If beta 'error' is a problem, either use a transistor that's sorted by beta >> >> > (rather than the less-expensive 'general purpose' part number), or >> >> > you can add a transistor base resistor and feed back to the op amp >> >> > to completely remove base current > >> The beta feedback trick is cute, but it relies on the op amp to >> >> compensate, and so only works at low frequency. It also doesn't fix >> >> Early effect at high frequency, whereas a properly designed cascode >> >> does. > >Just as voltage regulators rely on bypass capacitors at high frequency, >I'd think of a series inductor to keep the current source happy at high >frequencies. I wonder, though, how far one could get with B-E capacitance >(with a HF transistor that has relatively small B-C capacitance)? >Either way, oscillation is possible and layout will matter.
A b-e cap on a fast RF PNP, inside my slow opamp loop, is interesting. I could get rid of some other parts, maybe. The cap could be pretty big, because it doesn't do any harm to the opamp loop dynamics. How's it gonna oscillate with the base shorted to the emitter? [1] [1] wishful thinking -- John Larkin Highland Technology, Inc jlarkin at highlandtechnology dot com http://www.highlandtechnology.com Precision electronic instrumentation Picosecond-resolution Digital Delay and Pulse generators Custom laser drivers and controllers Photonics and fiberoptic TTL data links VME thermocouple, LVDT, synchro acquisition and simulation