Option H99

On my line there’s an older series signal generator with a notorious optional circuit board we refer to as Special Handling option H99. Most of us just call it the H99 for short. What it does is push the available frequency range on this instrument a couple of GHz higher than a standard equipped box. For many customers, the H99 has proved to be a very desirable feature because of the performance specifications it delivers. Also it is much cheaper to upgrade a unit with option H99 rather than buy a new, more expensive Sig Gen to get similar or better frequency range.

H99 is a drop-in replacement RF output board. The secret ingredient that gives greater frequency range and performance is a special component, called an HFET. HFET stands for Heterostructure Field Effect Transistor. It’s a high performance hotrod of a transistor. I’m not going to delve into the theory behind how they work, if you’re interested you can read up on HFETs elsewhere. For now all you need to know is, the HFET in option H99 is critical to instrument functionality. Without it the signal generator is essentially dead.

During development of the H99, managers placed unrealistic schedule demands on the research and development team in the lab. They were under so much pressure to get this thing through prototype phases to production that the guys had little choice but to cut corners wherever possible. And that is exactly what they did. First, they purchased a competitor’s box that had better performance than one of our comparable products. Second, our R&D lab rats dismantled the rival’s instrument and reverse engineered portions of the circuitry. That’s when they discovered a new HFET they were previously unaware of. Third, wasting no time they contacted the component manufacturer responsible for making this part, and R&D quickly built a whole new RF output board around it.

Little time was given to rigorous testing of the new design output board for flaws or performance issues. Normally, months of environmental stress testing would take place in the lab along with thorough evaluation on the production line. Management’s business plan hinged on getting the H99 out the door to customers as soon as possible. As expected, when the option was officially released we sold a shitload of them. Less than a year after option H99 was commercially available though, a flood of those H99 boxes came back to the factory for warranty repair. Every last one of them had the exact same failure mode. The RF Output board was unlevelled, meaning there was no frequency output. Upon investigation of defective boxes at the factory we discovered those HFET components were blown on every single output PC board. What could possibly have happened to them?

In a crazed rush to develop the H99 someone missed a small detail about the HFET: it needed to be kept cool. In order to operate normally the HFET runs at an extremely high internal temperature. The more voltage that is applied to the part, the more it heats up. According to the part manufacturer, large heatsinks were supposed to be placed above and below the HFET on both sides of our output board. We completely failed to do that. No heatsinks were incorporated into the design for that component. There was no room inside the instrument to add heatsinks, anyway. Upon additional analysis of the problem our engineering team also discovered that the HFET was constructed using a kind of epoxy that thermally insulated it, further trapping heat in the component. So, under daily use a customer was lucky if their output board lasted six months before burning out. They were extremely fortunate if it lasted closer to a full year.

As customer complaints grew louder amid rapidly increasing H99 failures, R&D tried desperately to come up with a workable solution to the HFET problem. Engineering experimented with a number of alternatives like splitting the voltage across multiple HFETs in the circuitry, reducing voltage to the existing design in the hopes of running cooler, and using a different manufacturer HFET. Nothing they came up with fully resolved the issue. The best R&D could do was extend the service life of those HFETs from a year to maybe a year and a half. Ultimately those HFETs would all burn out, just not as quickly.

Fully aware we had a serious design flaw, Bill and Dave’s company continued selling option H99. Then someone in management thought it would be a great idea to reduce warranty coverage for this problem. Customers who bought our sig gens were basically given a big bony middle finger and told too bad, so sad. That kind of shoddy business attitude enraged big customers like Motorola, who had purchased large volumes of our product. Motorola was so pissed off that they threatened us with a potential loss of their future business because of this issue, among other things. While we screwed out smaller customers for repair service, Bill and Dave’s management bowed to Motorola’s demands. We agreed to continue fixing their defective H99 instruments at no-cost on the down low, but everyone else was left out in the cold.

~ by factorypeasant on June 8, 2007.

3 Responses to “Option H99”

  1. Investigative journalism. Kickass.

  2. why are bill and dave dead? oh ya a bunch of chukle headded new management schaems killed the company from the inside…sounds like a comunist plot.


    p.s. 1/2MachineLipMoves now signed.

  3. Wad- i’m pleased you dig it.

    devil-t cool man, guess Helios is a pretty good guy. seems like it anyway from everything i read about him. would have been hip to be there for the show with ya but just couldn’t swing it this week.

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s

%d bloggers like this: