Showing posts with label test of time. Show all posts
Showing posts with label test of time. Show all posts

Thursday, August 23, 2012

Early Power Transistor Evolution, Part 1, Germanium


We recently completed our “Test of Time” power supply contest. Contestants told us about how they were using their Harrison Labs/HP/Agilent DC power supplies and the older the power supply, the better. It was pretty fascinating to see the many innovative way these power supplies were being used. It was also fascinating to see so many “vintage” power supplies still functional and in regular use after many decades. Several of them even being vacuum tube based!

One key component found in most all power supplies from the mid 1950s on is, no surprise, power transistors. Shortly after manufacturers were able to make reliable and reasonably rugged transistors in the mid 1950s they also developed transistors that would handle higher currents and power. Along with higher power came the need to dissipate the power. This led to some interesting packaging; some familiar and others not as familiar. Hunting through my “archives” I managed to locate some early power transistors. In review of their characteristics it was quite enlightening to see how they evolved to become better, faster, and cheaper! I also found it is quite challenging to find good, detailed, and most especially, non-conflicting information on these early devices.

Germanium was the first semiconducting material widely adopted for transistors, power and otherwise. One early power transistor I came across was the 2N174, shown in Figure 1.



Figure 1: 2N174 Power Transistor

Following are some key maximum ratings on the 2N174 power transistor:

  •  VCEO = -55V
  • VCBO = -80V
  •  VEBO= -60V
  •  IC = 15A
  • PD = 150W
  • hfe= 25
  •  fT = 10 kHz
  •  Thermal resistance = 0.35 oC/W
  •  TJ= 100 oC
  • Package: TO-36
  • Polarity: PNP
  • Material/process: Germanium alloy junction

The alloy junction process provided a reliable means to mass produce transistors. Most of the earlier transistors are PNP with N type semiconductor “pellets” or “dots” of typically indium alloyed to a P type germanium wafer. This process favored PNP production as the indium had a lower melting point than the N-type germanium bases. Still, this was a relatively slow and expensive process as they were basically manufactured one at a time. These early alloy junction transistors were not passivated and therefore needed to be hermetically packaged to prevent contamination and degradation. Often referred to as a “door knob” package, the TO-36 stud mount package was quite a piece of work and was no doubt expensive to as a result. It had a pretty impressive junction-to-case thermal resistance but given the maximum temperature of just 100 oC, low thermal resistance was necessary in order to operate the transistor at a reasonable power level. The low maximum operating temperature of germanium was one of most limiting attributes, especially for power applications. The transition frequency, fT of just 10 kHz was also extremely low. This is the frequency where current gain, hfe, drops down to 1, ceasing to be an effective amplifier. The 2N174 appears to have originated in the later 1950’s.

Another early power transistor we used in our HP 855B bench power supplies is the 2N1532, as shown in Figure 2.



Figure 2: 2N1532 power transistors used in a Harrison Labs Model 855B power supply.

Following are some key maximum ratings on the 2N1532 power transistor:

  • VCEO = -50V 
  • VCBO = -100V
  • VEBO= -50V
  •  IC = 5A
  • PD = 94W
  • hfe= 20 to 40
  • fT = 200 kHz
  • Thermal resistance = 0.8 oC/W
  •  TJ= 100 oC
  • Package: TO-3
  • Polarity: PNP
  • Material/process: Germanium alloy junction
The 2N1532 is also a germanium PNP power transistor, similar to a number of other power transistors of the time. It is packaged in the widely recognizable TO-3 diamond-shaped hermetic package.  Being a much less complex case design it must have been considerably less costly than the TO-36 package in Figure 1, and has become one of the most ubiquitous hermetic power semiconductor packages of all times. To keep junction temperature rise down the Harrison Labs Model  855B power supply used three 2N1532 transistors in its series regulator to deliver just  18 volts and 1.5 amps output. It’s no wonder why these power supplies have stood the “Test of Time” as these transistors are running significantly de-rated, at just a fraction of their maximum power here.  It is also noteworthy to see the transition frequency of 200 kHz is 20 times that of the 2N174. This is one of the more questionable data I had found but if it is accurate then clearly design and process improvements contributed to this performance improvement.  While date codes on some of the capacitors in this model 855B power supply place its manufacture in 1962, early germanium PNP power transistors in TO-3 packages like these also typically originate back in the later 1950’s.

While germanium transistors have much greater conductivity, lower forward- and saturation voltage drops compared to silicon transistors, silicon ultimately won out in the end, especially for power transistor applications. Stay tuned for my second part in an upcoming posting. Discover how silicon evolved to rule the day for power transistors!

Saturday, June 30, 2012

Test of Time power supply contest winners announced


Earlier this week, the winners of Agilent’s Test of Time power supply contest were announced. Here is a link to the press release:

I found the Test of Time contest to be quite interesting and I was honored to be one of the judges for the contest. The contest invited engineers who were using vintage Agilent or Hewlett-Packard (is there a “vintage” Agilent supply?) or even the older Harrison Labs power supplies (HP power supplies started as Harrison Labs power supplies) to describe their application, writing about how the instrument has been used over the years and how they are using it today. We received quite a few entries and we had extensive discussions to choose what we considered the best entry based on the contest rules. Go to this link for the home page of the contest and select the Gallery tab to see all of the entries:
https://powercontest.tm.agilent.com/

Richard Factor, of Little Ferry, NJ (I did not know he was from NJ until after I submitted my choices as a judge) won one of the prizes: an N6705B DC Power Analyzer with three modules installed. Quite a nice prize, and well deserved based on Richard’s entry!

Richard used an old HP 6186B in an application that basically turned his Toyota Prius into a backup generator for his house during a power failure. Now that’s what I call a unique application! You can read more about it at these links:
http://www.priups.com/ (you must select the correct answer….I’m sure you’ll figure it out….)

Simon Jensen of Husum, Germany, also won an N6705B for his entry. Simon’s entry was chosen by readers of the stories who voted for their favorite. Simon used an Agilent 6632B power supply as an inexpensive load to sink current from a switching power supply he built. As Simon correctly points out, the nameplate on this power supply does not reveal all of its capabilities. It says 0 to 5 A, but it can also sink a programmable, regulated current like an electronic load. So the nameplate should say -5 A to +5 A, as pointed out by Simon!

Having worked for HP/Agilent on power products for more than 32 years (since 1980), it was not too surprising for me to see so many interesting applications for our products. I was also delighted, and not too surprised, to see how many of our older power supplies are still out there, providing power, decades after they were introduced! Now that’s what I call “vintage voltage”!!