Hacking Old 97, Part 2: Optics

In this story’s first installment, I described restoring a 12-year-old Fluke 97 Scopemeter to operational status by replacing the instrument’s proprietary NiCd battery pack with four rechargeable NiMH AA batteries converted into C cells with plastic sleeves. That at least got Old 97 running. The next step was to get it talking. The Model 97 Scopemeter was the first of the product line capable of talking to something else, namely a printer, either an Epson or an HP Thinkjet at either 1200 or 9600 bps. The printer could reproduce a hardcopy version of the Scopemeter’s screen.

Talking occurs through a Fluke-proprietary optical port in the side of the instrument. Originally, Fluke produced an optical-to-RS-232-C adapter called the PM 9080/001 Optically Isolated RS-232 Adapter/Cable for the Scopemeter line. Because many laptops no longer sport an RS-232 interface port (the serial interface’s usefulness only lasted 50 years or so), Fluke later introduced the OC4USB adapter that converts the Scopemeters’ optical port to USB. Both the PM 9080 and the OC4USB cost about $150. Even on eBay they go for $80 and up by auction’s end. The FlukeView software that works with these cables costs north of $200. (Fluke is on version 4 of FlukeView these days but version 3.04 was the last version to work with Old 97.) These prices are too steep for a hobby project, so I decided to invest some sweat equity to lower the price of optical communications.

Fluke 97 Scopemeter’s optical port

A little Web research reveals footprints in the sand. People have preceded me on this path. Jan Wagner in Finland has an excellent site that provides a wealth of information about building optical interfaces for the old Scopemeters as well as a pointer to ScopeGrab32, Wagner’s freeware Scopemeter screen grabber. The links to descriptions of DIY hardware for the Scopemeters include information about the critical element needed to build a no-sweat USB-to-RS-232 interface: the Silicon Labs CP2102. This terrific chip converts a USB port into a UART. Best of all, Silicon Labs has drivers that automatically cause Windows XP to recognize the chip as a serial COM port. No sweat.

Except that the CP2102 and its big brother, the CP2103, are surface-mount devices and my condo lacks a reflow-soldering oven. A little more Googling produced another big find: SparkFun Electronics in Boulder, Colorado sells assembled CP2102 and CP2103 breakout boards for $19.95 apiece. What a deal! Order one of those puppies up over the Web. (Coincidentally, SparkFun is on Central Avenue, just down the street from the building I worked in 25 years ago as a design engineer at pc-board EDA pioneer Cadnetix. What a small world it is.)

While waiting for the CP2103 breakout board to arrive, I focused on the optical end of my DIY cable. The holes in the Scopemeter 97 interface exactly fit T1-sized (3mm) opto components. A look in the Anchor Electronics catalog offers an obsolete Siemens phototransistor/IR LED SHF309/409 matched pair for $1. I hadn’t learned about Anchor Electronics until a few months ago when a friend at Tensilica, Frank Motta, told me about it. Anchor is a tiny, terrific little walk-in shop on Walsh in Santa Clara with an amazing number of electronic parts per square foot at great prices for such low quantities.

I bought the SHF309/409 matched pair at Anchor and a 10-pin, tenth-inch header to mount the devices because the hole spacing on the Fluke 97 just happens to be 0.2 inches. However, I didn’t want to leave the optics unprotected, which presented a problem because Fluke’s optical port design matches no connector backshell I’ve ever seen. Others who have traveled this path have fashioned makeshift shrouds of cardboard. Not too rugged. My first attempt at building a backshell for my Scopemeter cable’s optical head was made of Fun Foam, purchased at a local craft store called Michael’s. The experiment was a complete success in that it showed me something that absolutely would not work. The Fun Foam adhesive wasn’t up to the task. I considered sheet copper or brass from the hobby shop but the solution was simpler still: 28 cents of 12-gauge copper wire from Home Depot roughly (very roughly) formed into a wire bail and soldered to the extra pins on the 10-pin, tenth-inch header.

Optical head with Siemens SHF309/409 phototransistor/IR LED matched pair
(The purple device is the IR LED)

Before soldering the opto components to the header, I had to determine which hole in the Scopemeter 97 was the transmitter and which was the receiver. I couldn’t find the information on the Web and it wasn’t in the Scopemeter’s documentation. Bright idea: use the IR LED as a photodiode and stuff it in one of the holes. Connect the Scopemeter itself to the LED’s pins (it will generate a small voltage when illuminated with IR) and tell the Scopemeter to output a screen image. Brilliant, except that it turns out that the Scopemeter 97 can’t walk and chew gum at the same time. If it’s outputting a screen image, it’s not making measurements. Substitute a lowly DMM for the Scopemeter and get rewarded with about 0.4V ac out of one port, nothing out of the other. It’s a small signal, but enough to point the way.

With the proper hole identified, I soldered up the opto components and plugged the 10-pin header into a prewired DB25-to-header cable I bought for 99 cents at Halted (now known as HSC Electronic Supply, yet another Silicon Valley electronics store/landmark in Santa Clara).

All I needed now was that breakout board from SparkFun.