Years ago, an old HP Santa Clara Engineer attributed HP's service and support policies to Dave Packard: "HP will provide outstanding service to our customers, and we will make money doing it." Even in the late 1980s, 50 years after HP was founded, the legacy of Packard's influence was still strong in the Test and Measurement half of HP. (The computer half, while bigger than T&M, wasn't the essence of historical HP like T&M.) HP was very proud of making the best test equipment in the world. Remember, HP stood for "High Performance", and "High Price". Keeping these products working well in customer hands was important. And making profits was important too. Packard made several famous speeches over the years where he made it clear that profits were the foundation of the company. (When delivered to a group of General Managers, he would also make it clear if they couldn't manage their divisions profitably, he would find someone else who could.) HP didn't exist just to make profits, but without profits we didn't exist. And profitability was the best measure of a valued contribution to the market.
Santa Clara Division, which had the Frequency and Time products, had a department devoted to service and support, which included the team that wrote and edited the Service and Maintenance manuals for Frequency and Time products. One of my jobs at the end of the 5061B development was to go through the 5061A manual, and update it to reflect the 5061B we were now selling. We had 15 years of 5061A production change orders to integrate, plus everything we fixed on the 5061B. Working with the tech writer assigned to the task, we went through the entire manual, line by line. This was a great lesson in being thorough in the details, as the Service Manual was often the only document, both inside and outside of HP, which covered the entire product. I've carried this lesson with me in the decades since. The Service Engineering team was managed by Chuck Little, a cheerful fun loving guy, with his own "time-nuts" history. He had been on some of the early 5060A "Flying Clock" around the word trips HP sponsored to demonstrate the theory of relativity, or to calibrate official time at different international locations. Someone in marketing got the idea to publish a newsletter to customers called "The HP Standard", which had articles about future products, applications, and service information. In this, they advertised "The Cesium Seminar," a one week class on tuning, servicing, troubleshooting and repairing the 5061A/B Cesium Beam Frequency Standards. As the Production Engineer for the 5061B, I was recruited to help. My job was to gather up a bunch of 5061s and assorted test equipment for use in the class, and loiter in the back of the room while Chuck taught the class. There would be about a dozen students, split into small teams each with a test patient 5061 and troubleshooting gear. The classes were great fun. The students were mostly senior technicians from high end standards labs. Many were X-military NCO "Tech Sergeant" type guys. Many worked for the hard-core defense industry companies or their contractors. They had all kinds of "interesting" stories to tell, especially to my young, unsophisticated, untraveled ears. It was enlightening to observe the difference between seasoned technicians, and "engineers." If you actually wanted something done, these were the guys to call. Their knowledge of electronics was a lot more intuitive than classically trained engineers, and many had an appropriately cynical view of engineering. Even HP had some engineers that were all theory, and frankly useless when it came time to actually do something useful. You all know the type. One of the biggest compliments I got from a few of them is that HP engineers (and me specifically) were not like the engineers at their companies. First off, they were surprised we were not unionized, which gave us a lot of freedom in what we did. And we actually showed the techs some respect, valuing their opinions and skills. In leading the class, Chuck did two brilliant things. First, he would go over the entire 5061A/B design, and explain from a high, system level perspective, how a Cesium Beam Frequency Standard worked. Attached is the overall block diagram if you are interested and want to follow along. He would spend a day, starting at the overall block diagram level, and then slowly zoom into each sub system, drawing simple but useful pictures and diagrams on a big white board. I have a photo of him standing at the white board in front of a crude drawing of a CBT, with a silly grin on his face while explaining the operation of this amazing machine, telling stories and cracking jokes along the way. His explanation would go something like this: (If you want to follow along, download the block diagram from Chapter 8 of the 5061B service manual from this location: http://www.ko4bb.com/getsimple/index.php?id=manuals&dir=HP_Agilent/HP_5061A_Cesium_Beam_Frequency_Standard ) It all starts with the oscillator. It is like the flywheel in an engine. Smooth and steady, but uncontrolled, it tends to drift over time. The whole point of the all the rest of the instrument is to detect the drifting of the oscillator, and tune it back on center. Well, it turns out that Cesium 133 is an amazing element. If you excite with an exact frequency, 9,192,631,772.5 Hz, the atoms undergo a change, where the magnetic spin will reverse. (Imaging Chuck holding his hand in the air, twisting his wrist one way and then the next.) The Cesium Beam Tube is there to excite and detect this change. If we take the 10MHz oscillator, and use it to generate the 9.192 GHz signal, we can detect when it is off, by noticing that the energy state change didn't happen. The 90MHz oscillator, and 12.631 MHz synthesizer, and the Harmonic generator are there to generate the 9.192GHz signal from the 10MHz. And then we do a clever thing, we modulate the 9.192GHz signal back and forth a few hertz at a 137Hz rate, and watch the signal out of the CBT go on and off either side of the peak, and exactly balance the signal out. (He would draw the Beam current response signal, looking a bit like a heartbeat, and show the 137 Hz modulation on either side.) Oh, why 137Hz? Because it isn't 50, 60, 100, 120, 150 or 180 Hz. No power line harmonics. (All the techs would grunt and nod - good thinking.) He would then explain how the cesium oven and oven controller got the cesium hot and ejected a "beam" of atoms through the u-shaped wave guide in the CBT so the 9.192GHz signals could interact with the atoms, get magnetically deflected if the frequency was just right, and then would then strike the heated tantalum wire ion detector, and then launch off electrons that would get multiplied through the electron multiplier (with a crude drawing of how it worked, and the reason we needed the -2500V power supply). From there he would walk through the components that made up the op-amp, etc on the feedback loop side, which then controlled the 10811 oscillator through the EFC, or electronic frequency controller. And then explain how the "logic board" module sensed if everything is working correctly go give you feedback on the health of the instrument. After a couple of hours, a white board would be full of drawings and notes, and we all would kind of know how a 5061A/B worked. All the while I was getting a great education in system thinking, and overall product architecture of very well engineered, fairly complex instrument. (When previewing this article, Tom Van Baak provided this link to a much better description of how a Cesium Standard works: http://www.leapsecond.com/museum/hp5062c/theory.htm ) The other thing he did was talk about troubleshooting, with practical trials. After reinforcing what every tech already knows, check the power supply first, he would talk about things that could go wrong with a 5061, and how the instrumented responded to various faults. And then the fun started. We'd have a team of techs step outside for a minute, and sabotage the instrument in some way. The team would come back in, and figure out what was wrong, and fix it. A great game of electronics hide and seek. As the week wore on, the tricks we'd deploy would get trickier and sneaker. We'd start with simply unplugging a connector or something like that. Our favorite at the end was to put a small scrap of paper between the switch contacts on the "logic reset" switch. Everything was actually just fine in the instrument, but you just couldn't get the green light to come on. Again, this was a great education for me, as I'd wander the room, and help the teams when they got stuck, learning along the way with them. The better teams would sometimes turn tables on us, and sabotage the 5061, and see if we could figure it out. If we were really sneaky, we'd mess with the test equipment, so it wasn't giving accurate readings. All sorts of real life stuff. (Later in my inkjet engineering career, I'd always start with checking out my scope probes, on the square wave test connection. I've wasted a lot time working on a "bad" circuit, only later to discover a bad scope probe. Those probes got cut in half, so no one would be tempted to "rescue" them, and thrown away.) I helped with the Cesium Seminars several times during my 5061B days, and they were always something to look forward to. As the time wore on, I was teaching and leading more, and Chuck would be able to do some of his real job during the week. Chuck was a great guy, great teacher, and a lot of fun to be around. Much of the valuable work I do today is teaching and training younger engineers. Until I wrote this, I hadn't realized how much I have modeled my own teaching style after his: Help the student see the bigger system, and then work your way down into the details. Draw lots of pictures. Tell stories. Keep it lively. A lot of my effectiveness now as a development engineer comes from what I've learned from this great service engineer. I probably should try to look Chuck up, and if he is still with us, thank him for what he taught me. Hugh Rice _______________________________________________ time-nuts mailing list -- [email protected] To unsubscribe, go to http://lists.febo.com/mailman/listinfo/time-nuts_lists.febo.com and follow the instructions there.
