On the stuff below (see “What’s All This Hydraulic Ram Stuff, Anyhow?”), I see a lot of reference to switching, oscillating, etc.
A few times over the past 40+ years, I have had to get involved with customers who have had their applications slowly fade into the sunset (gradually reduced dynamic performance) and eventually quit working. The period of time involved was typically several days to several months after operating and meeting full dynamic expectations within their applications. When I got back the parts involved, their HFEs, particularly on the left side of beta peak, always had HFE-vrs-IC slopes that were three to eight times steeper than they should be in that region.
Thus, if they were using the parts queued up at a significantly lower average than where the peak HFE occurs (common in some switching power-supply and in most oscillator and other class-C operations), the HFE had “degraded” to the point (with the slope change involved) where the transistor would not queue up at the originally intended operating point any longer. This often becomes a problem first noticeable in relatively cold operating environments because that “much poorer slope” on the left side of HFE peak also flops around in the breeze more than normal vrs Ta (and/or Tj).
The problem always came back to being caused by very short but repetitive duration (usually at the working frequency or, occasionally, at the third upper harmonic if RF was involved) barely exceeding the actual BVEBO threshold of the transistor involved during the transistor-off portion of the full-cycle operating time period. In these instances, when passive component value changes were made around the part to ensure these short duration peak voltages were kept below the device BVEBO, the functions never encountered the problem again.
I don’t know if this is what is causing the problems being outlined or not, but it wouldn’t surprise me if I’m understanding and interpreting the circumstances being described correctly.
Until recently, I ran into this problem most often in RF multiplier applications, but I am seeing it more often now in some of the bipolars used as drivers in higher-speed (PWM rate) MOSFET power-supply/converter applications. I am also encountering more recently in bipolars used as outputs, particularly in low voltage (<10- to 15-V dc inputs), in power supplies/converters as designers try to reduce system and component costs or reduce space by taking the transistors closer to “the edge” than in the past and milking them for everything they’re supposedly capable of. The problem is that people can look for this situation and/or limitation (violation?) as part of a design evaluation, but instrumentation used can load down the applicable feedback loop equations such that the loop intensity then momentarily decreases enough that everything looks okay across the emitter-base junction during the “off” portion of the device operation. And the higher the frequency goes, the more apt that the potential problem will be missed.
The BVEBO ranges of devices with similar FTs and intended to be used in similar function segments and/or families will also tend to have similar BVEBO ranges. I’ve run into a few exceptions but they are not common, so it probably wouldn’t much matter whose device was involved if this is what is going on.
Here’s a pretty good technical answer.
If you had asked me how to degrade beta, I’d have said, “Zenering the Vbes.” Well, my old friend Gus had already put two and two together. Here’s what he said: He never distinguished between NPNs and PNPs, so you ought to protect your PNPs, too. In my LM331 circuit, the leakages at room temp were negligible, but a PNP junction that leaks a few pA at room temp could leak many nA when hot. I put in a diode to clamp that current and not let it zener the NPN. It seems (33 years later) to have worked well. But when the datasheet says “Do not exceed absolute max ratings,” such as Vbe, I guess they really do mean it!
Dear Mr. Pease:
Your essay on the hydraulic ram brought back a memory. During the summer of 1950, between my freshman and sophomore years in college, I took a job at a steel mill, with the bricklayer gang on the open-hearth floor. Molten pig iron from the blast furnaces would come to the floor in huge ladles. It would be poured into the open-hearth furnaces to be converted into steel. I would watch the glowing particles of molten iron that fountained up, like fireworks, from whatever the iron was being poured into. The particles would shoot up higher than the ladle. I always figured it was some surface-tension effect, but never really investigated it. I hadn’t thought about the phenomenon in years until your article triggered the recollection.
(I expect that any moisture could cause what you saw. Obviously, those molds that the hot iron was poured into, even with a 0.1% amount of residual moisture, could make enough expanding steam to cause what you saw. If you had a mold for pig iron, and you baked it even at 240°C, that might cut down the fountain of sparks, because the moisture would be baked out. But most of the time, nobody cares. Even if it weren’t water, it could be any other substance that could expand when highly heated. If these were molds made of sand and they worked okay, nobody would ever worry about this. Do you have any idea how these molds were made? Who made them? What was the formula for the sand that went into them? /rap)
I always enjoy your essays in Electronic Design. They are one of the first things I turn to, before I dig into the technical articles.
Joseph P. Martino
Hello, Joseph M.,
Thanks for writing. Good question!
I really appreciated your article about hydraulic rams. As a kid, the first experience was magic, but after a few minutes it was clear how it worked. It was also clear that the noise and leaks were not efficient and the process could be better. I have probably been borderline obsessive about efficiency ever since. (I’m in favor of efficiency—if it doesn’t drive me nuts!! /rap)
This led to engineering school, small hydroelectric (2 MW) projects in the Sierras, and generators and power application engineering of many kinds. Now my son is also a power systems engineer. He was infected early on.
In case you missed this interesting article in Machine Design: http://beta.machinedesign.com/article/more-efficient-pressure-regulator-cuts-operating-costs-1117
(Are you sure this idea is going to work? I have got to say, I am skeptical. An electronic switcher doesn’t just have two terminals. It normally has three. If you want to put out 1 A into a 5-V output, you can put in an average of 0.5 A at 11 V, and the other half-amp comes from ground. This guy’s switch really doesn't look like he can do that. It will take me a while, but I think this guy Alfred Perz’s idea is bogus. I have got to study this some more. I'll be back. Hey, ask your son if this will work. /rap)
We deal with natural gas compressors that consume huge amounts of power. Often it is possible to greatly reduce the power consumption by changing process strategy. I have been looking for an application of this guy’s strategy for a year. Someday I will find one.
I look forward to your columns.
Hello, John Carroll,
I think it won’t work. If he built it and measured the input power, and the output, and he thinks it is efficient, I think his measurements are wrong!
If you put in 10 lb of air at 100 psi and draw out 10 lb of air at 40 psi, you still have to put in a lot of energy. He is a bit overwhelmed by the thought of dealing with millions of cubic feet per day, instead of the workstation rate of consumption he discusses. I don’t think this will work for small or large applications.
If you run a turbine off the 60-psi pressure difference and use that energy to pump more air, you can get 10 lb of air out at 40 psi, with perhaps 6 lb of air at 100 psi, because the pump can compress 4 lb of air from the atmosphere. That is how a switcher works, and not like Perz says. His will still waste power.
Dear Mr. Pease,
I have been reading your column for years and enjoy it immensely. My first paying job in electronics was at Thompson Radio & Television Service in San Antonio, Texas.
In 1968, my junior year in high school, Mr. Thompson put me to work after school and on weekends fixing car radios. They had about five to seven tubes and a vibrator for the plate voltage. Tuning was mechanical: a gear and worm moved ferrite rods in/out of coils for each channel assigned to a key. Then you pulled the key out, tuned the desired channel, and then pushed the key back in to set the frequency. A trimmer cap adjacent to the antenna connector was adjusted for best overall band response (matching). (That sounds about right, even though I never fooled around with them. /rap)
I have been privileged to work on a few of these lately for antique car buffs. When the vibrators quit, you can buy a solid-state replacement but that’s no fun. I like to fix the old ones. I cut notches around the bottom edge of the can with a pair of sharp side cutters, peel the edge back to remove the vibrator, and then polish the contacts with 600-grit wet/dry sandpaper and Caig De-Oxit. I then bend the edge back over and coat the seam with RTV. It’s more original that way.
Often there is no visible sign of pitting or burning on the contacts. They just develop a high resistance. My friend the chemist tells me that the rubber used to wrap the inside of the can out-gasses over time and leaves an invisible insulating film on the points. I’ll just have to take his word for it. I miss the “good old days” when a technician knew how to sweep an IF.
Your Obedient Servant In This and All Things
Hello Tom T.,
Since this radio takes so much work to open up and get inside, I think the solid-state version would sound good to me, if I just want a radio that runs. However, I do understand “old-car buffs” who like to have everything as original and authentic as possible, no matter what the trouble or expense. But, I am in favor of rehabilitating the old ones. In general, I do like to repair old things, not just junk them.
Thanks for explaining how you do it. I really do like to take old things apart to see how they worked, see how they failed, and see if they can be repaired. I’m sure if I had one of those old vibrators, I’d pull it apart. But I never got into this old radio stuff. I’m sure there are things that are not easy to repair.
I have an old Harman-Kardon Stereo Festival 5000 and I had to replace its caps when I bought it. I may have to do it again.
Hey, old Sidney Harman just died at ~ 94? And you can even “google” him up.
I enjoyed your column in the March 24, 2011 issue of Electronic Design on “What’s All This George A. Philbrick Stuff, Anyhow?.
A hundred years ago there was considerable discussion about what frequency to use for electrical power. Edison preferred dc because it made some things simple, although it required MG sets where we now use transformers. (Uh, yeah. /rap) Fortunately, most people saw the advantages of ac (Well, it took old Nick Tesla to invent good ac motors. /rap) and it was only a question of what frequency.
Higher frequencies meant smaller transformers (less iron), but greater problems due to line inductance. Lower frequencies meant not only bigger transformers but bigger motors and generators, for a given power, because of an inherent speed limit on machines since power is torque times speed. For example, at 25 Hz the maximum speed of a synchronous machine (or near synchronous induction motor) is 1500 rpm. At 60 Hz it’s 3600 rpm. (Yeah, but 25 Hz is still used for some trains... /rap)
At high frequencies, in addition to the problems you mentioned, I would think one of the biggest problems would be transmission stability. Even at 60 Hz, transmission stability runs into problems when the line length starts being a significant fraction of a wavelength. For a 1000-km line, the line end phase difference due to this effect alone is 72°, which is a problem right there unless it’s properly handled. At 60 kHz the same issues would apply at only 1 km, and at 5 MHz they’d be horrendous. (Not to mention skin effect. /rap)
I could see going to radio frequencies, at least in principle, for short-haul situations like downtown areas, but you’d need frequency converters (to 60 Hz, say, or even dc) when transmitting for anything like a kilometer or more. (I think George was just playing games. /rap)
Taking into account all the pros and cons, most of the world quickly settled on 50 to 60 Hz as the best compromise, although there are pockets such as Toronto, where I live, where 25 Hz and even dc persisted until around the 1950s. (In some parts of Boston, some old dc power continued into the 60’s. /rap)
Tony Griffin, P. Eng.
Have you heard about the fiasco in Japan? Most of the east coast is on a 50-Hz grid and the west is on a 60-Hz grid, and as F-converters are so expensive, there are very few of them. So now the east is power-starved and cannot borrow much from the west! My goodness!