And you want your semiconductors to be in ceramic/lided packages with the bond wires flapping in free air. Bond wires embedded in epoxy like to break... don't ask how I found this out ;-) ... it brings back bad memories... and makes bad memories... Quantum chips have very elaborate/specialized bonding to survive liquid helium... even with that, thermal cycling still breaks them.

Many years ago, circa 1977, I was moved to try some crude tests on a few semiconductor devices at LN2 temperature (77K). These tests were very crude, involving dunking the parts into the LN2 bath, and many failed outright. Most of the devices tested were in plastic packages. Here are the results as I remember them, applicable only for the survivors: Silicon bipolar transistors: The DC beta fell to very low values. Junction forward voltages rose considerably. Silicon JFETs: Seemed to continue working reasonably well. Silicon MOSFETs: Same as JFETs Red LEDs: The junction forward voltages rose considerably, to about 5V as I recall. The light output per unit current rose truly spectacularly. My first experiences with seriously-cryogenic RF amplifiers were at the Arecibo Observatory beginning about 11 years ago. These were all either GaAs- or InP-based and we cooled them to ~15K, generally leading to input-referred amplifier noise temperatures of ~3K. Many of the devices needed continuous exposure to light to work properly when cold, and the metal block amplifier packages had holes in the lid directly over the active device chips. Small red LEDs in ordinary plastic packages were inserted in the holes and were driven at a few mA, generally in a series string. Since cool-down was fairly gradual over a span of at least a couple hours, there was little problem with thermal shock and almost all LEDs survived cooldown and warmup for the several cycles they experienced during my 10 years at the observatory. RF amplifier biasing was invariably done with opamp circuits to maintain set drain currents and drain voltages, with said bias control circuits outside the dewar at room ambient temperature. Failures were not too uncommon, largely attributed to connector misbehavior at low temperature. Formation of "ice" (really frozen air) inside the dewars was suspected because fine wires inside the dewar were often found to have fairly sharp bends at improbable locations upon warmup for diagnostic purposes (or due to cooling system failure). Cooling was done with a closed-cycle gaseous He system, using the Gifford-McMahon cycle. Note that He does not liquefy (at reasonable pressures) until around 4K. All dewars for this kind of work depend on high vacuum inside for thermal insulation, with black body radiation and direct conduction through wires and mounting structures being the principal remaining heat leaks. At these temperatures, maintenance of high vacuum inside the dewar was essentially automatic because all components of the inward-leaking air were known to freeze out. This could lead to a hazard because over time, months or years, enough air could freeze out to result in dangerously high internal pressures upon "thawing" when the dewar was warmed for any reason. For this reason, all dewars were equipped with blowout plugs to avoid high pressure's damaging the dewars themselves. Dana On Tue, Apr 3, 2018 at 12:26 AM, Mark Sims <holrum@hotmail.com> wrote: > And you want your semiconductors to be in ceramic/lided packages with the > bond wires flapping in free air. Bond wires embedded in epoxy like to > break... don't ask how I found this out ;-) ... it brings back bad > memories... and makes bad memories... Quantum chips have very > elaborate/specialized bonding to survive liquid helium... even with that, > thermal cycling still breaks them. > _______________________________________________ > time-nuts mailing list -- time-nuts@febo.com > To unsubscribe, go to https://www.febo.com/cgi-bin/ > mailman/listinfo/time-nuts > and follow the instructions there. >

Hi Mark: When Aetech started to make their own Tunnel Diodes there was a problem with the neck breaking. Note they were made by alloying a ball of metal onto a highly doped chip, bonding from the lip of the ceramic package to the ball then on to the opposite lip, then etching the chip away leaving something that in cross section looked like a mushroom.  The neck was a few microns wide and often broke.  The fix was to epoxy a glass rod on either side of the chip, between the metal bottom of the ceramic pill package and the bonding wire.  The glass was chosen to have a CTE that matched the die.  That solved the broken neck problem. http://prc68.com/I/Aertech.shtml#Prod -- Have Fun, Brooke Clarke http://www.PRC68.com http://www.end2partygovernment.com/2012Issues.html -------- Original Message -------- > And you want your semiconductors to be in ceramic/lided packages with the bond wires flapping in free air. Bond wires embedded in epoxy like to break... don't ask how I found this out ;-) ... it brings back bad memories... and makes bad memories... Quantum chips have very elaborate/specialized bonding to survive liquid helium... even with that, thermal cycling still breaks them. > _______________________________________________ > time-nuts mailing list -- time-nuts@febo.com > To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts > and follow the instructions there. >

Hi If the objective is great phase noise far removed from carrier, there’s a gotcha. Let’s say you have a 10 dbm source at room and it’s broadband is at KTB of -174 + 1db. That gives you -183 dbc. You cool your oscillator to whatever and KTB goes down to -194. You do a bang up job at that temperature and get within a db there was well. You now have a 10 dbm source with -203 dbc. Run the super cooled signal through a coax out to the room environment. Pass it through a 50 ohm gizmo and …. KTB is back at -174. Your source is at -184 dbc. To *use* the signal, you likely need to cool whatever it’s driving as well. While one might say …. that pretty weird. Well, similar things do happen. Many an ultra low phase noise OCXO gets sold, only to find that the “next stage” isn’t as quiet as the system guys hand hoped. Hmmm …. errrr ….oops !! Bob > On Apr 3, 2018, at 5:29 AM, Dana Whitlow <k8yumdoober@gmail.com> wrote: > > Many years ago, circa 1977, I was moved to try some crude tests on a few > semiconductor devices at LN2 temperature (77K). > > These tests were very crude, involving dunking the parts into the LN2 bath, > and > many failed outright. Most of the devices tested were in plastic packages. > > Here are the results as I remember them, applicable only for the survivors: > > Silicon bipolar transistors: The DC beta fell to very low values. > Junction > forward voltages rose considerably. > > Silicon JFETs: Seemed to continue working reasonably well. > > Silicon MOSFETs: Same as JFETs > > Red LEDs: The junction forward voltages rose considerably, to about 5V as > I recall. The light output per unit current rose truly spectacularly. > > My first experiences with seriously-cryogenic RF amplifiers were at the > Arecibo Observatory beginning about 11 years ago. These were all either > GaAs- or InP-based and we cooled them to ~15K, generally leading to > input-referred amplifier noise temperatures of ~3K. Many of the devices > needed continuous exposure to light to work properly when cold, and the > metal block amplifier packages had holes in the lid directly over the active > device chips. Small red LEDs in ordinary plastic packages were inserted > in the holes and were driven at a few mA, generally in a series string. > Since cool-down was fairly gradual over a span of at least a couple hours, > there was little problem with thermal shock and almost all LEDs survived > cooldown and warmup for the several cycles they experienced during > my 10 years at the observatory. > > RF amplifier biasing was invariably done with opamp circuits to maintain > set drain currents and drain voltages, with said bias control circuits > outside > the dewar at room ambient temperature. Failures were not too uncommon, > largely attributed to connector misbehavior at low temperature. Formation > of "ice" (really frozen air) inside the dewars was suspected because fine > wires > inside the dewar were often found to have fairly sharp bends at improbable > locations upon warmup for diagnostic purposes (or due to cooling system > failure). > > Cooling was done with a closed-cycle gaseous He system, using the > Gifford-McMahon cycle. Note that He does not liquefy (at reasonable > pressures) until around 4K. All dewars for this kind of work depend on > high vacuum inside for thermal insulation, with black body radiation > and direct conduction through wires and mounting structures being > the principal remaining heat leaks. > > At these temperatures, maintenance of high vacuum inside the dewar was > essentially automatic because all components of the inward-leaking air > were known to freeze out. This could lead to a hazard because over time, > months or years, enough air could freeze out to result in dangerously high > internal pressures upon "thawing" when the dewar was warmed for any > reason. For this reason, all dewars were equipped with blowout plugs > to avoid high pressure's damaging the dewars themselves. > > Dana > > > On Tue, Apr 3, 2018 at 12:26 AM, Mark Sims <holrum@hotmail.com> wrote: > >> And you want your semiconductors to be in ceramic/lided packages with the >> bond wires flapping in free air. Bond wires embedded in epoxy like to >> break... don't ask how I found this out ;-) ... it brings back bad >> memories... and makes bad memories... Quantum chips have very >> elaborate/specialized bonding to survive liquid helium... even with that, >> thermal cycling still breaks them. >> _______________________________________________ >> time-nuts mailing list -- time-nuts@febo.com >> To unsubscribe, go to https://www.febo.com/cgi-bin/ >> mailman/listinfo/time-nuts >> and follow the instructions there. >> > _______________________________________________ > time-nuts mailing list -- time-nuts@febo.com > To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts > and follow the instructions there.

Very true Sent from my iPhone > On Apr 3, 2018, at 7:51 PM, Bob kb8tq <kb8tq@n1k.org> wrote: > > Hi > > If the objective is great phase noise far removed from carrier, there’s a gotcha. > > Let’s say you have a 10 dbm source at room and it’s broadband is at KTB of -174 + 1db. > That gives you -183 dbc. You cool your oscillator to whatever and KTB goes down > to -194. You do a bang up job at that temperature and get within a db there was well. > You now have a 10 dbm source with -203 dbc. > > Run the super cooled signal through a coax out to the room environment. Pass it through > a 50 ohm gizmo and …. KTB is back at -174. Your source is at -184 dbc. To *use* the > signal, you likely need to cool whatever it’s driving as well. > > While one might say …. that pretty weird. Well, similar things do happen. Many an ultra > low phase noise OCXO gets sold, only to find that the “next stage” isn’t as quiet as the > system guys hand hoped. Hmmm …. errrr ….oops !! > > Bob > >> On Apr 3, 2018, at 5:29 AM, Dana Whitlow <k8yumdoober@gmail.com> wrote: >> >> Many years ago, circa 1977, I was moved to try some crude tests on a few >> semiconductor devices at LN2 temperature (77K). >> >> These tests were very crude, involving dunking the parts into the LN2 bath, >> and >> many failed outright. Most of the devices tested were in plastic packages. >> >> Here are the results as I remember them, applicable only for the survivors: >> >> Silicon bipolar transistors: The DC beta fell to very low values. >> Junction >> forward voltages rose considerably. >> >> Silicon JFETs: Seemed to continue working reasonably well. >> >> Silicon MOSFETs: Same as JFETs >> >> Red LEDs: The junction forward voltages rose considerably, to about 5V as >> I recall. The light output per unit current rose truly spectacularly. >> >> My first experiences with seriously-cryogenic RF amplifiers were at the >> Arecibo Observatory beginning about 11 years ago. These were all either >> GaAs- or InP-based and we cooled them to ~15K, generally leading to >> input-referred amplifier noise temperatures of ~3K. Many of the devices >> needed continuous exposure to light to work properly when cold, and the >> metal block amplifier packages had holes in the lid directly over the active >> device chips. Small red LEDs in ordinary plastic packages were inserted >> in the holes and were driven at a few mA, generally in a series string. >> Since cool-down was fairly gradual over a span of at least a couple hours, >> there was little problem with thermal shock and almost all LEDs survived >> cooldown and warmup for the several cycles they experienced during >> my 10 years at the observatory. >> >> RF amplifier biasing was invariably done with opamp circuits to maintain >> set drain currents and drain voltages, with said bias control circuits >> outside >> the dewar at room ambient temperature. Failures were not too uncommon, >> largely attributed to connector misbehavior at low temperature. Formation >> of "ice" (really frozen air) inside the dewars was suspected because fine >> wires >> inside the dewar were often found to have fairly sharp bends at improbable >> locations upon warmup for diagnostic purposes (or due to cooling system >> failure). >> >> Cooling was done with a closed-cycle gaseous He system, using the >> Gifford-McMahon cycle. Note that He does not liquefy (at reasonable >> pressures) until around 4K. All dewars for this kind of work depend on >> high vacuum inside for thermal insulation, with black body radiation >> and direct conduction through wires and mounting structures being >> the principal remaining heat leaks. >> >> At these temperatures, maintenance of high vacuum inside the dewar was >> essentially automatic because all components of the inward-leaking air >> were known to freeze out. This could lead to a hazard because over time, >> months or years, enough air could freeze out to result in dangerously high >> internal pressures upon "thawing" when the dewar was warmed for any >> reason. For this reason, all dewars were equipped with blowout plugs >> to avoid high pressure's damaging the dewars themselves. >> >> Dana >> >> >>> On Tue, Apr 3, 2018 at 12:26 AM, Mark Sims <holrum@hotmail.com> wrote: >>> >>> And you want your semiconductors to be in ceramic/lided packages with the >>> bond wires flapping in free air. Bond wires embedded in epoxy like to >>> break... don't ask how I found this out ;-) ... it brings back bad >>> memories... and makes bad memories... Quantum chips have very >>> elaborate/specialized bonding to survive liquid helium... even with that, >>> thermal cycling still breaks them. >>> _______________________________________________ >>> time-nuts mailing list -- time-nuts@febo.com >>> To unsubscribe, go to https://www.febo.com/cgi-bin/ >>> mailman/listinfo/time-nuts >>> and follow the instructions there. >>> >> _______________________________________________ >> time-nuts mailing list -- time-nuts@febo.com >> To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts >> and follow the instructions there. > > _______________________________________________ > time-nuts mailing list -- time-nuts@febo.com > To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts > and follow the instructions there.