tube
in [Design]
|
From: John Larkin on 12 Feb 2010 10:15 On Fri, 12 Feb 2010 00:16:00 -0600, "Tim Williams" <tmoranwms(a)charter.net> wrote: >So? Light it up and get to work! Should be roughly comparable to a 2N5179, >I'd guess. >http://www.mif.pg.gda.pl/homepages/frank/sheets/049/9/955.pdf >Plate resistance is slightly high and Gm noticably low, but capacitance is >quite small and the plate curves look nice (mu and Rp are fairly constant in >the operating range). Offhand, give it a ~20k plate resistor and you'll get >around 7k || 1.4pF = 16.2MHz -3dB point if driven hard (not counting miller >or probe C). Okay, I suppose a 2N5179 will switch faster than that, but in >terms of fundamental performance, once it cuts off at fT, it doesn't really >do anything anymore; with tubes, add some L to cancel the C and you'll get >narrow-band performance for another decade or two. Tubes get killed, eventually, by transit time problems. There were some lighthouse (planar) tubes with really tiny spacings that would work at 3 GHz or some such. It took other tricks, like bunching, to break that limit. > >Funny how tubes do that. BJTs don't, they kind of just stop working, fT >limited by recombination more than circuit parameters. Do FETs do that? I >know power MOSFETs are typically limited by gate spreading resistance (~a >few ohms, so IRF540 stops being practical at ~10MHz). Do JFETs? What's the >step response of a JFET? > >Good MOSFETs, too, but they're hard to find. You'd think a few more microns >of sputtered aluminum over the gate connection wouldn't be worth $20 more, >or whatever it is they do with 'em. MOSFETs seem almost infinitely fast if you drive their gates hard enough. Here's a 50 volt pulse being generated by two 5-cent 2N7002s in parallel: ftp://jjlarkin.lmi.net/HV_mosfet_pulse.jpg which is a lot faster than the datasheet suggests. People push mosfets to places like 500 volts in 2 ns. DEI invented some tricky packaging to get the packaging parasitics down, and that seems to help a lot. John
From: Tim Wescott on 12 Feb 2010 11:12 On Fri, 12 Feb 2010 00:16:00 -0600, Tim Williams wrote: > So? Light it up and get to work! Should be roughly comparable to a > 2N5179, I'd guess. The packaging details make it very hard to keep stable -- the inner element's gain goes up to UHF or higher, and those wires start looking awfully inductive. AFAIK "pencil" tubes and lighthouse triodes were designed in large part to be a better mechanical fit to coaxial cavities. > http://www.mif.pg.gda.pl/homepages/frank/sheets/049/9/955.pdf Plate > resistance is slightly high and Gm noticably low, but capacitance is > quite small and the plate curves look nice (mu and Rp are fairly > constant in the operating range). Offhand, give it a ~20k plate > resistor and you'll get around 7k || 1.4pF = 16.2MHz -3dB point if > driven hard (not counting miller or probe C). Okay, I suppose a 2N5179 > will switch faster than that, but in terms of fundamental performance, > once it cuts off at fT, it doesn't really do anything anymore; with > tubes, add some L to cancel the C and you'll get narrow-band performance > for another decade or two. > > Funny how tubes do that. BJTs don't, they kind of just stop working, fT > limited by recombination more than circuit parameters. Do FETs do that? > I know power MOSFETs are typically limited by gate spreading resistance > (~a few ohms, so IRF540 stops being practical at ~10MHz). Do JFETs? > What's the step response of a JFET? > > Good MOSFETs, too, but they're hard to find. You'd think a few more > microns of sputtered aluminum over the gate connection wouldn't be worth > $20 more, or whatever it is they do with 'em. > Which is why tubes are still king for really high power VHF and microwave stuff. Uneasy on the throne, though. -- www.wescottdesign.com
From: John Larkin on 12 Feb 2010 11:43 On Fri, 12 Feb 2010 10:12:33 -0600, Tim Wescott <tim(a)seemywebsite.com> wrote: >On Fri, 12 Feb 2010 00:16:00 -0600, Tim Williams wrote: > >> So? Light it up and get to work! Should be roughly comparable to a >> 2N5179, I'd guess. > >The packaging details make it very hard to keep stable -- the inner >element's gain goes up to UHF or higher, and those wires start looking >awfully inductive. AFAIK "pencil" tubes and lighthouse triodes were >designed in large part to be a better mechanical fit to coaxial cavities. > >> http://www.mif.pg.gda.pl/homepages/frank/sheets/049/9/955.pdf Plate >> resistance is slightly high and Gm noticably low, but capacitance is >> quite small and the plate curves look nice (mu and Rp are fairly >> constant in the operating range). Offhand, give it a ~20k plate >> resistor and you'll get around 7k || 1.4pF = 16.2MHz -3dB point if >> driven hard (not counting miller or probe C). Okay, I suppose a 2N5179 >> will switch faster than that, but in terms of fundamental performance, >> once it cuts off at fT, it doesn't really do anything anymore; with >> tubes, add some L to cancel the C and you'll get narrow-band performance >> for another decade or two. >> >> Funny how tubes do that. BJTs don't, they kind of just stop working, fT >> limited by recombination more than circuit parameters. Do FETs do that? >> I know power MOSFETs are typically limited by gate spreading resistance >> (~a few ohms, so IRF540 stops being practical at ~10MHz). Do JFETs? >> What's the step response of a JFET? >> >> Good MOSFETs, too, but they're hard to find. You'd think a few more >> microns of sputtered aluminum over the gate connection wouldn't be worth >> $20 more, or whatever it is they do with 'em. >> >Which is why tubes are still king for really high power VHF and microwave >stuff. Uneasy on the throne, though. GaN fets are getting up into tube territory. You can't get an SO8 tube with Ft in the GHz and a Gm of 5. The terawatt radar-as-a-weapon things are apparently solid state, some sort of impulse generators based on (maybe?) dumping capacitors into antennas through laser-triggered (diamond?) films. John
From: Tim Williams on 12 Feb 2010 11:48 "Tim Wescott" <tim(a)seemywebsite.com> wrote in message news:0_CdnWT3Qt5s4OjWnZ2dnUVZ_jdi4p2d(a)web-ster.com... > The packaging details make it very hard to keep stable -- the inner > element's gain goes up to UHF or higher, and those wires start looking > awfully inductive. AFAIK "pencil" tubes and lighthouse triodes were > designed in large part to be a better mechanical fit to coaxial cavities. Don't forget nuvistors. :) And then there's these guys; http://webpages.charter.net/dawill/Images/Raytheon%20RK-707B.jpg although now we are actually talking electron drift and bunching. I've always wondered if I could operate this thing as a planar tetrode though. S'pose I should try some time. I don't have the equipment to detect microwave oscillations if it misbehaves though. > Which is why tubes are still king for really high power VHF and microwave > stuff. Uneasy on the throne, though. Y'think magnetrons will be around forever? GaN and etc. would have to get pretty damn cheap to offset them. Tim -- Deep Friar: a very philosophical monk. Website: http://webpages.charter.net/dawill/tmoranwms
From: Spehro Pefhany on 12 Feb 2010 11:58 On Fri, 12 Feb 2010 10:48:46 -0600, "Tim Williams" <tmoranwms(a)charter.net> wrote: >"Tim Wescott" <tim(a)seemywebsite.com> wrote in message >news:0_CdnWT3Qt5s4OjWnZ2dnUVZ_jdi4p2d(a)web-ster.com... >> The packaging details make it very hard to keep stable -- the inner >> element's gain goes up to UHF or higher, and those wires start looking >> awfully inductive. AFAIK "pencil" tubes and lighthouse triodes were >> designed in large part to be a better mechanical fit to coaxial cavities. > >Don't forget nuvistors. :) > >And then there's these guys; >http://webpages.charter.net/dawill/Images/Raytheon%20RK-707B.jpg >although now we are actually talking electron drift and bunching. I've >always wondered if I could operate this thing as a planar tetrode though. >S'pose I should try some time. I don't have the equipment to detect >microwave oscillations if it misbehaves though. > >> Which is why tubes are still king for really high power VHF and microwave >> stuff. Uneasy on the throne, though. > >Y'think magnetrons will be around forever? > >GaN and etc. would have to get pretty damn cheap to offset them. > >Tim Probably disappear about the same time as hard disk drives disappear.
First
|
Prev
|
Next
|
Last
Pages: 1 2 3 4 Prev: White LED appropriate for illumination? Next: Anyone imported/export WinQCad <-> OrCad? |