Interesting TIA design
in [Design]
|
From: Phil Hobbs on 24 Oct 2009 10:49 I just finished up an interesting new (to me) TIA design for a Far Eastern customer. (There's no NDA, so I can talk about it.) It made a pretty interesting study. It works up to about 30 uA photocurrent, but manages to be shot noise limited above 20 nA at 400 kHz and 100 nA at 1 MHz with 30 pF of photodiode capacitance, which is a factor of about 20 better than I originally expected. (A narrower current range device can be shot noise limited at 20 nA at 1 MHz.) It uses a pair of parallelled BF862s in a bootstrapped bootstrap, connected to another pair in a cascoded common-source configuration. It avoids the 300 kelvin resistor noise by having a main signal path with no resistors! It uses capacitive feedback on the common-source stage and a differentiator going into the second stage--really an odd design, but it works great. (Bootstrapping the drains of the bootstrap FETs was good for 3 dB of SNR, interestingly.) LTSPICE said I could get a bit better performance by running the bootstrap a bit above IDSS, but I wasn't that brave. It's for a relatively narrowband application, so in principle it could run at lower frequency, except that it has to work around fluorescent lights. Electronic ballast fluorescents produce not only EMI but also strongly modulated light. With a 40 kHz switching frequency the harmonics go up to above 1 MHz. The phosphor isn't fast enough to do that, of course, but there are mercury and argon emission lines all over the place, and the plasma responds pretty fast, especially near the ends of the tube, so you have to pick an operating wavelength that avoids the emission lines. 940 nm is pretty good, fortunately--there are lines at 870 and 1010 nm that are inconveniently close, but relatively weak. What I wound up with was a glorified AM radio--the above-mentioned TIA stage running into a 74VHC4053 analogue mux connected as a double balanced mixer, with a 1455 kHz LO and two 455 kHz ceramic IF filters, cascaded so as to increase the stopband attenuation. Besides the TIA, the two most challenging parts were: (1) Achieving a 1% linearity spec over the full range, including the gain error from a 1x-128x programmable gain amp. Doing a built-up PGA with good accuracy at 1 MHz is surprisingly hard with jellybean parts, I discover--the on-resistance specs of muxes are the pits. (2N7002s to the rescue.) (2) Getting rid of the DC and low frequency photocurrent without using resistors. I made a BJT current sink degenerated with 4 diodes in series with the emitter. Using N diodes reduces the transconductance by a factor of N+1, and increases the equivalent shot noise voltage by sqrt(N+1), so the resulting current sink runs 10*log(N+1) dB lower than full shot noise--7 dB for 4 diodes. So far, so vanilla. The problem was that the capacitance of the diodes dominated at low photocurrents, which essentially doubled the 1 MHz noise at 40 nA. I used BFT25A C-B junctions for the diodes, but that was the best I could get. I'd love to figure out how to make fast, wide range, sub-Poissonian current sources below 100 nA. So I learned something about a corner of the design space that I'd never worked in before, which I thought was pretty neat. Cheers Phil Hobbs -- Dr Philip C D Hobbs Principal ElectroOptical Innovations 55 Orchard Rd Briarcliff Manor NY 10510 845-480-2058 hobbs at electrooptical dot net http://electrooptical.net
From: John Larkin on 24 Oct 2009 12:15 On Sat, 24 Oct 2009 10:49:56 -0400, Phil Hobbs <pcdhSpamMeSenseless(a)electrooptical.net> wrote: >I just finished up an interesting new (to me) TIA design for a Far >Eastern customer. H? (There's no NDA, so I can talk about it.) It made a >pretty interesting study. > > It works up to about 30 uA photocurrent, but manages to be shot noise >limited above 20 nA at 400 kHz and 100 nA at 1 MHz with 30 pF of >photodiode capacitance, which is a factor of about 20 better than I >originally expected. (A narrower current range device can be shot noise >limited at 20 nA at 1 MHz.) > >It uses a pair of parallelled BF862s in a bootstrapped bootstrap, >connected to another pair in a cascoded common-source configuration. It >avoids the 300 kelvin resistor noise by having a main signal path with >no resistors! It uses capacitive feedback on the common-source stage >and a differentiator going into the second stage--really an odd design, >but it works great. (Bootstrapping the drains of the bootstrap FETs was >good for 3 dB of SNR, interestingly.) LTSPICE said I could get a bit >better performance by running the bootstrap a bit above IDSS, but I >wasn't that brave. There's no brick wall at Idss. We run phemts enhanced all the time. Of course, you can't enhance a jfet much before the gate starts conducting, but you can get mesfets up about 50% and some phemts to twice Idss. I seem to have noticed a hyper-enhancement in some phemts at gate conduction, as if they are going bipolar or something, but I haven't investigated. I hope they keep making BF862's. There's nothing else like them. Their transfer curve is pretty radical for a jfet, so they'd enhance pretty well; double Idss at roughly 1 nA gate current, something like that. > >It's for a relatively narrowband application, so in principle it could >run at lower frequency, except that it has to work around fluorescent >lights. > >Electronic ballast fluorescents produce not only EMI but also strongly >modulated light. With a 40 kHz switching frequency the harmonics go up >to above 1 MHz. The phosphor isn't fast enough to do that, of course, >but there are mercury and argon emission lines all over the place, and >the plasma responds pretty fast, especially near the ends of the tube, >so you have to pick an operating wavelength that avoids the emission >lines. 940 nm is pretty good, fortunately--there are lines at 870 and >1010 nm that are inconveniently close, but relatively weak. I noticed that in my spectroscopy project, huge amounts of 40 KHz when things were unshielded. That's smack in my range of interest. > >What I wound up with was a glorified AM radio--the above-mentioned TIA >stage running into a 74VHC4053 analogue mux connected as a double >balanced mixer, with a 1455 kHz LO and two 455 kHz ceramic IF filters, >cascaded so as to increase the stopband attenuation. > >Besides the TIA, the two most challenging parts were: >(1) Achieving a 1% linearity spec over the full range, including the >gain error from a 1x-128x programmable gain amp. Doing a built-up PGA >with good accuracy at 1 MHz is surprisingly hard with jellybean parts, I >discover--the on-resistance specs of muxes are the pits. (2N7002s to >the rescue.) I'm working on a VME digitizer module with ranges from +-10 mV to +-40 volts and > 10 volts of common-mode. I'm having to make my own pga out of parts, which is a huge nuisance. It's a classic 3-opamp diffamp with analog switches selecting the first-stage feedback from a balanced resistor string in 5 ranges... what a pain to calculate! The third "opamp" is an INA157. 16 channels, ADC per, and two trimpots per channel to fine-tune the cmrr. > >(2) Getting rid of the DC and low frequency photocurrent without using >resistors. I made a BJT current sink degenerated with 4 diodes in >series with the emitter. Using N diodes reduces the transconductance by >a factor of N+1, and increases the equivalent shot noise voltage by >sqrt(N+1), so the resulting current sink runs 10*log(N+1) dB lower than >full shot noise--7 dB for 4 diodes. So far, so vanilla. The problem >was that the capacitance of the diodes dominated at low photocurrents, >which essentially doubled the 1 MHz noise at 40 nA. I used BFT25A C-B >junctions for the diodes, but that was the best I could get. I'd love >to figure out how to make fast, wide range, sub-Poissonian current >sources below 100 nA. BFT25's make awesome diodes. Since metal film resistors have low shot noise, why can't you just use one of them with a fairly high voltage, or maybe cascode? > >So I learned something about a corner of the design space that I'd never >worked in before, which I thought was pretty neat. It's always fun to do something that's orders of magnitude better. John
From: Joerg on 24 Oct 2009 12:48 Phil Hobbs wrote: > I just finished up an interesting new (to me) TIA design for a Far > Eastern customer. (There's no NDA, so I can talk about it.) It made a > pretty interesting study. > > It works up to about 30 uA photocurrent, but manages to be shot noise > limited above 20 nA at 400 kHz and 100 nA at 1 MHz with 30 pF of > photodiode capacitance, which is a factor of about 20 better than I > originally expected. (A narrower current range device can be shot noise > limited at 20 nA at 1 MHz.) > > It uses a pair of parallelled BF862s in a bootstrapped bootstrap, > connected to another pair in a cascoded common-source configuration. It > avoids the 300 kelvin resistor noise by having a main signal path with > no resistors! It uses capacitive feedback on the common-source stage > and a differentiator going into the second stage--really an odd design, > but it works great. (Bootstrapping the drains of the bootstrap FETs was > good for 3 dB of SNR, interestingly.) LTSPICE said I could get a bit > better performance by running the bootstrap a bit above IDSS, but I > wasn't that brave. > As John said, be brazen there, go for it. Just like we used to run power amp tubs with plenty of control grid current in the old days. I've never made a grid glow but I did run plates well past cherry-red. > It's for a relatively narrowband application, so in principle it could > run at lower frequency, except that it has to work around fluorescent > lights. > > Electronic ballast fluorescents produce not only EMI but also strongly > modulated light. With a 40 kHz switching frequency the harmonics go up > to above 1 MHz. The phosphor isn't fast enough to do that, of course, > but there are mercury and argon emission lines all over the place, and > the plasma responds pretty fast, especially near the ends of the tube, > so you have to pick an operating wavelength that avoids the emission > lines. 940 nm is pretty good, fortunately--there are lines at 870 and > 1010 nm that are inconveniently close, but relatively weak. > > What I wound up with was a glorified AM radio--the above-mentioned TIA > stage running into a 74VHC4053 analogue mux connected as a double > balanced mixer, with a 1455 kHz LO and two 455 kHz ceramic IF filters, > cascaded so as to increase the stopband attenuation. > > Besides the TIA, the two most challenging parts were: > (1) Achieving a 1% linearity spec over the full range, including the > gain error from a 1x-128x programmable gain amp. Doing a built-up PGA > with good accuracy at 1 MHz is surprisingly hard with jellybean parts, I > discover--the on-resistance specs of muxes are the pits. (2N7002s to > the rescue.) > If you do a low-volume (meaning non-consumer) design check out canned PGAs. One of my favorites is the AD603. They raised the price a couple of bucks over the years but it is a great chip. Another option is to get a chip with at least two uncommitted N-channcels on there, such as the CD4007 (cheap). Unbuffered logic can work as well, even cheaper. Now servo them by running one of them in a current source with opamp feedback. So your DAC commands a voltage and this sets the exact Rdson because it fudges the gate until that condition is met. The other FET that is part of your RF divider will follow, as long as this is monolithic and not a multi-chip part. But even the Rohm multi-chips are pretty good because the devices usually come off the same wafer. No guarantees with multi-chip, of course. > (2) Getting rid of the DC and low frequency photocurrent without using > resistors. I made a BJT current sink degenerated with 4 diodes in > series with the emitter. Using N diodes reduces the transconductance by > a factor of N+1, and increases the equivalent shot noise voltage by > sqrt(N+1), so the resulting current sink runs 10*log(N+1) dB lower than > full shot noise--7 dB for 4 diodes. So far, so vanilla. The problem > was that the capacitance of the diodes dominated at low photocurrents, > which essentially doubled the 1 MHz noise at 40 nA. I used BFT25A C-B > junctions for the diodes, but that was the best I could get. I'd love > to figure out how to make fast, wide range, sub-Poissonian current > sources below 100 nA. > No chance to chop DC and LF the old-fashioned way, with inductors and capacitors? > So I learned something about a corner of the design space that I'd never > worked in before, which I thought was pretty neat. > Always nice to enter a new field. I just entered that of injectors and high pressures. It does command some respect, if something flies off in a 4500 psi situation things can get ugly. -- Regards, Joerg http://www.analogconsultants.com/ "gmail" domain blocked because of excessive spam. Use another domain or send PM.
From: John Larkin on 24 Oct 2009 13:03 On Sat, 24 Oct 2009 09:48:32 -0700, Joerg <invalid(a)invalid.invalid> wrote: >Phil Hobbs wrote: >> I just finished up an interesting new (to me) TIA design for a Far >> Eastern customer. (There's no NDA, so I can talk about it.) It made a >> pretty interesting study. >> >> It works up to about 30 uA photocurrent, but manages to be shot noise >> limited above 20 nA at 400 kHz and 100 nA at 1 MHz with 30 pF of >> photodiode capacitance, which is a factor of about 20 better than I >> originally expected. (A narrower current range device can be shot noise >> limited at 20 nA at 1 MHz.) >> >> It uses a pair of parallelled BF862s in a bootstrapped bootstrap, >> connected to another pair in a cascoded common-source configuration. It >> avoids the 300 kelvin resistor noise by having a main signal path with >> no resistors! It uses capacitive feedback on the common-source stage >> and a differentiator going into the second stage--really an odd design, >> but it works great. (Bootstrapping the drains of the bootstrap FETs was >> good for 3 dB of SNR, interestingly.) LTSPICE said I could get a bit >> better performance by running the bootstrap a bit above IDSS, but I >> wasn't that brave. >> > >As John said, be brazen there, go for it. Just like we used to run power >amp tubs with plenty of control grid current in the old days. I've never >made a grid glow but I did run plates well past cherry-red. > > >> It's for a relatively narrowband application, so in principle it could >> run at lower frequency, except that it has to work around fluorescent >> lights. >> >> Electronic ballast fluorescents produce not only EMI but also strongly >> modulated light. With a 40 kHz switching frequency the harmonics go up >> to above 1 MHz. The phosphor isn't fast enough to do that, of course, >> but there are mercury and argon emission lines all over the place, and >> the plasma responds pretty fast, especially near the ends of the tube, >> so you have to pick an operating wavelength that avoids the emission >> lines. 940 nm is pretty good, fortunately--there are lines at 870 and >> 1010 nm that are inconveniently close, but relatively weak. >> >> What I wound up with was a glorified AM radio--the above-mentioned TIA >> stage running into a 74VHC4053 analogue mux connected as a double >> balanced mixer, with a 1455 kHz LO and two 455 kHz ceramic IF filters, >> cascaded so as to increase the stopband attenuation. >> >> Besides the TIA, the two most challenging parts were: >> (1) Achieving a 1% linearity spec over the full range, including the >> gain error from a 1x-128x programmable gain amp. Doing a built-up PGA >> with good accuracy at 1 MHz is surprisingly hard with jellybean parts, I >> discover--the on-resistance specs of muxes are the pits. (2N7002s to >> the rescue.) >> > >If you do a low-volume (meaning non-consumer) design check out canned >PGAs. One of my favorites is the AD603. They raised the price a couple >of bucks over the years but it is a great chip. > >Another option is to get a chip with at least two uncommitted >N-channcels on there, such as the CD4007 (cheap). Unbuffered logic can >work as well, even cheaper. Now servo them by running one of them in a >current source with opamp feedback. So your DAC commands a voltage and >this sets the exact Rdson because it fudges the gate until that >condition is met. The other FET that is part of your RF divider will >follow, as long as this is monolithic and not a multi-chip part. But >even the Rohm multi-chips are pretty good because the devices usually >come off the same wafer. No guarantees with multi-chip, of course. > > >> (2) Getting rid of the DC and low frequency photocurrent without using >> resistors. I made a BJT current sink degenerated with 4 diodes in >> series with the emitter. Using N diodes reduces the transconductance by >> a factor of N+1, and increases the equivalent shot noise voltage by >> sqrt(N+1), so the resulting current sink runs 10*log(N+1) dB lower than >> full shot noise--7 dB for 4 diodes. So far, so vanilla. The problem >> was that the capacitance of the diodes dominated at low photocurrents, >> which essentially doubled the 1 MHz noise at 40 nA. I used BFT25A C-B >> junctions for the diodes, but that was the best I could get. I'd love >> to figure out how to make fast, wide range, sub-Poissonian current >> sources below 100 nA. >> > >No chance to chop DC and LF the old-fashioned way, with inductors and >capacitors? > > >> So I learned something about a corner of the design space that I'd never >> worked in before, which I thought was pretty neat. >> > >Always nice to enter a new field. I just entered that of injectors and >high pressures. It does command some respect, if something flies off in >a 4500 psi situation things can get ugly. Pinhole leaks, like from bad welds or defective hoses, can make invisible fluid jets that cut like a sword, or inject hydraulic fluid like a hypodermic. Bad stuff. Guys in ships, looking for steam leaks, would probe around carefully with a broomstick. A leak would slice the end off. John
From: Phil Hobbs on 24 Oct 2009 13:32
John Larkin wrote: > On Sat, 24 Oct 2009 10:49:56 -0400, Phil Hobbs > <pcdhSpamMeSenseless(a)electrooptical.net> wrote: > >> I just finished up an interesting new (to me) TIA design for a Far >> Eastern customer. > > H? Nope, further west. > > (There's no NDA, so I can talk about it.) It made a >> pretty interesting study. >> >> It works up to about 30 uA photocurrent, but manages to be shot noise >> limited above 20 nA at 400 kHz and 100 nA at 1 MHz with 30 pF of >> photodiode capacitance, which is a factor of about 20 better than I >> originally expected. (A narrower current range device can be shot noise >> limited at 20 nA at 1 MHz.) >> >> It uses a pair of parallelled BF862s in a bootstrapped bootstrap, >> connected to another pair in a cascoded common-source configuration. It >> avoids the 300 kelvin resistor noise by having a main signal path with >> no resistors! It uses capacitive feedback on the common-source stage >> and a differentiator going into the second stage--really an odd design, >> but it works great. (Bootstrapping the drains of the bootstrap FETs was >> good for 3 dB of SNR, interestingly.) LTSPICE said I could get a bit >> better performance by running the bootstrap a bit above IDSS, but I >> wasn't that brave. > > There's no brick wall at Idss. We run phemts enhanced all the time. Of > course, you can't enhance a jfet much before the gate starts > conducting, but you can get mesfets up about 50% and some phemts to > twice Idss. I seem to have noticed a hyper-enhancement in some phemts > at gate conduction, as if they are going bipolar or something, but I > haven't investigated. > > I hope they keep making BF862's. There's nothing else like them. Their > transfer curve is pretty radical for a jfet, so they'd enhance pretty > well; double Idss at roughly 1 nA gate current, something like that. Might be worth a try, then. This gizmo is going to be prototyped by someone else 7000 miles away, so I'm being reasonably conservative. > >> It's for a relatively narrowband application, so in principle it could >> run at lower frequency, except that it has to work around fluorescent >> lights. >> >> Electronic ballast fluorescents produce not only EMI but also strongly >> modulated light. With a 40 kHz switching frequency the harmonics go up >> to above 1 MHz. The phosphor isn't fast enough to do that, of course, >> but there are mercury and argon emission lines all over the place, and >> the plasma responds pretty fast, especially near the ends of the tube, >> so you have to pick an operating wavelength that avoids the emission >> lines. 940 nm is pretty good, fortunately--there are lines at 870 and >> 1010 nm that are inconveniently close, but relatively weak. > > I noticed that in my spectroscopy project, huge amounts of 40 KHz when > things were unshielded. That's smack in my range of interest. > They're really truly horrible, and all to save one capacitor. See for example http://wireless.stanford.edu/papers/Ravi/iee1996ir.pdf . >> What I wound up with was a glorified AM radio--the above-mentioned TIA >> stage running into a 74VHC4053 analogue mux connected as a double >> balanced mixer, with a 1455 kHz LO and two 455 kHz ceramic IF filters, >> cascaded so as to increase the stopband attenuation. >> >> Besides the TIA, the two most challenging parts were: >> (1) Achieving a 1% linearity spec over the full range, including the >> gain error from a 1x-128x programmable gain amp. Doing a built-up PGA >> with good accuracy at 1 MHz is surprisingly hard with jellybean parts, I >> discover--the on-resistance specs of muxes are the pits. (2N7002s to >> the rescue.) > > I'm working on a VME digitizer module with ranges from +-10 mV to +-40 > volts and > 10 volts of common-mode. I'm having to make my own pga out > of parts, which is a huge nuisance. It's a classic 3-opamp diffamp > with analog switches selecting the first-stage feedback from a > balanced resistor string in 5 ranges... what a pain to calculate! The > third "opamp" is an INA157. 16 channels, ADC per, and two trimpots per > channel to fine-tune the cmrr. Isn't it a nuisance? All those nice analogue switch parts that turn out to be 450 ohm Rs and 5 pF Cout. Irritating. There are a few 10 MHz MDACs, but the feedthrough at low codes is still pretty bad, though nowhere near as bad as the 1 MHz variety. Of course that's a bit less of a worry when they're used as the bottom resistor of the voltage divider. >> (2) Getting rid of the DC and low frequency photocurrent without using >> resistors. I made a BJT current sink degenerated with 4 diodes in >> series with the emitter. Using N diodes reduces the transconductance by >> a factor of N+1, and increases the equivalent shot noise voltage by >> sqrt(N+1), so the resulting current sink runs 10*log(N+1) dB lower than >> full shot noise--7 dB for 4 diodes. So far, so vanilla. The problem >> was that the capacitance of the diodes dominated at low photocurrents, >> which essentially doubled the 1 MHz noise at 40 nA. I used BFT25A C-B >> junctions for the diodes, but that was the best I could get. I'd love >> to figure out how to make fast, wide range, sub-Poissonian current >> sources below 100 nA. > > BFT25's make awesome diodes. > > Since metal film resistors have low shot noise, why can't you just use > one of them with a fairly high voltage, or maybe cascode? The problem is the Johnson noise of the resistor--you have to drop at least 50 mV for the noise to be dominated by the photocurrent shot noise (250 mV is better). If the output is off the peg at 10 uA (R < 1 M, say), it's way over the shot noise at 10 nA, unless it's range switched, and the range switch would have more capacitance than the BFT25As. The basic problem is that the impedance at the emitter is 2.5 M at 10 nA, so getting 1 MHz bandwidth requires the total capacitance there to be something like 0.06 pF at most--very hard to do in a built-up circuit. I thought about using a current splitter at the sink output, so that the actual sink transistor could run at higher current, but doing the splitting introduces essentially full shot noise by itself, which just pushes the problem off one level. [It's an interesting fact that a BJT diff pair whose emitter current has full shot noise produces exactly full shot noise from each collector, regardless of DeltaV_BE (in the infinite beta limit, anyway)]. You can't run photodiodes in series usually, but I wonder if there isn't some cute feedback trick that would allow that. Hmm. Next hill to climb. Cheers Phil Hobbs > > >> So I learned something about a corner of the design space that I'd never >> worked in before, which I thought was pretty neat. > > It's always fun to do something that's orders of magnitude better. > > John > -- Dr Philip C D Hobbs Principal ElectroOptical Innovations 55 Orchard Rd Briarcliff Manor NY 10510 845-480-2058 hobbs at electrooptical dot net http://electrooptical.net |