iceNINE Tech

- The P6046 Differential Probe-

There have been several discussions on the P6046 Differential Probe, over the years, on the Tekscopes group. I happen to have an interest in these units; I bought one set a while back on eBay for $100, and recently picked another up at auction for all of $35, for a spare. If you are unfamiliar with this probe, it is well-described in Tektronix catalogs from the 70's and 80's (and well into the 90's, but the catalogs aren't as interesting!), and has an entry on the web site. It was made for almost 30 years (68-97)!

There is a comparison chart available, listing all the changes made in various versions of the peobe, amp, accessories, and manual, made by one Dave Robbins. It is available for download from various websites, including at

The basic probe itself originally was designed to plug into the 1A5 amplifier, but most folks will want the Amplifier for P6046 box as well, which has a 50ohm output to connect to any scope that can do 10mV/div (or so). It's got a pretty impressive bandwidth of 100MHZ, but pay close attention to the rolloff in CMRR as frequency increases. It's good, but not perfect, and you will see some significant common-mode signal feedthrough at full bandwith. Up to 60KHz, the CMRR is >= 10K:1, but it is down well under 1000:1 at full bandwidth (depending on volts/div). Source impedance (ie, an imbalance) also has a big effect.

One handy thing you can do, also, is use this probe as a single-ended FET probe (with the attendant low Cin). You'll need one of the special grounding tip adapters to short the - input, though. These are sometimes hard to find, but it's well worth chasing down a full set of the little tip accessories if you're planning on doing any significant work with your probe.

Manuals - you should get one of these. There are all sorts of things you'll need to know, most of the characteristics of the probe vary over frequency, and need to be taken into account if you want to have an accurate reading. The P6046 (and amp) were made for almost 30 years, and varied some over that time. Two manuals were made during that time, with the cutover at B032108. I bought mine at Artek Media for a reasonable price. I had to get both, because my first probe set actually has the early probe and the later amp. So someone had some mix-and-match fun.

I'm really here to talk about calibration stuff here, though. There are plenty of pieces of Tek hardware that claim to require special fixtures to calibrate, where you can actually hack something into shape, or work around a given step. Not so, with the P6046. You'll need three less-than-common items, two of which are single-purpose and fairly rare: the dual-input tip adapter, and the shield. Few people seem to have these. You'll also need a 100MHz sine gen, levelled is really the way to go (a 191 or SG503). And a fast-rise, clean pulse gen is recommended. The old manual wanted a 106, the newer one does with a PG506. This is not an easy item to calibrate.

(click the picture for a bigger version)

What you're seeing here is the 067-0562-00 and the 067-0563-00. I can't say for certain that my shield is the real deal (it may just be a spare or salvaged regular old probe cover), but it works.

There's some misunderstanding about this shield. Some people think that it is needed because Tek used the inner surface of the case (which is painted in conductive material) as a signal or power ground from one part of the board to another, rather than use a wire or run a trace. Well, yes, they did do that, but you can use a jumper wire and get by that problem (and you do, in part of the cal procedure). But we're talking about some really sensitive, high-speed circuitry. Jim Williams claims that the HF sections of this probe are "extrordinarily hard to trim up for CMRR". If he thinks that it's tricky, you know you're really in hot water. This probe really needs the cover in place to effectively be calibrated, both for shielding and (I strongly suspect) for calibrating out stray capacitance. My shield is simply a cover in which correctly-placed access holes have been drilled.

The dual tip adapter is also critical, but slightly more fake-able. Use something like a real metal box, though, with good solid connections, to prevent radiated pickup and minimize stray resistance. It shorts the grounds together, and then the tips, inside, and runs them to a BNC connector, for the common-mode rejection trims.

Feel free to drop me a line if you feel that I've made a mistake, or have something I could add for the Tek community. I make these pages not for any personal gain, but rather to spread knowledge of these fine devices.


Last updated April 5, 2007
Original content copyright © Christian Weagle unless otherwise indicated.