7A11: Difference between revisions

The 7A11 input capacitance is 5.8 pF from 5 mV/Div to 50 mV/Div, goes down to 3.4 pF from 100 mV/Div to 1 V/Div, and down again to 2.0 pF from 2 V/Div to 20 V/Div.   

Because the AC coupling is attained with a plug-on capacitor (that adds 1.2 pF to the input C), a DC offset is supplied to reduce the need for AC coupling.

 

A nice thing about the 7A11 was that the probe cable length was adjustable in six steps to just over 7 feet in length.   

You just uncoil however much cable length you want.

 

The 7A11 was admittedly clumsy to use because of the probe size.  It had some trouble — DC drift, RF pickup at 5 mV, 100 mV and 2 V/Div.  

It was not very popular partially because it was expensive, $850 for single channel vs. dual channel [[7A12]] (105 MHz) at $700.   

But there was never another 1 MΩ 7000-series plugin as fast as the 7A11 ... and it was one of the originals!  

 

The [[7A16]] was then 225 MHz (for $625).  The 7A16 disappeared rapidly and became the [[7A16A]], still 225 MHz in 1973.  

 

The 485 used a faster IC process ("[[SH2]]", 3.5 GHz) but the 1 MΩ to 50 Ω converter used only discrete devices. (See Electronics June, 1972.)

 

By 1974, the 7A11 cost $950 and the new, popular [[7A26]] (dual trace) was $1,050, not counting probes.   

But that would still only get you to 50 mV/div with a (9.5 pF, 3 foot, × 10) probe, five years after the 7A11 introduction.

[[Tom Rousseau]] designed the 7A26 which used the faster IC process and a [[155-0078-00|vertical IC I designed for the 485]].  

The 7A26 sold so well that Tektronix presented Tom with an entirely gold plated 7A26!   

Obviously, he still has it.

 

Remember also that the 7A11 was introduced at the same show as the HP183A/1830A/1840A, faster (250 MHz), smaller, lighter, and less expensive with better triggering than 7000 series had.   

 

The star-crossed [[7A12]], which was supposed to be the dual trace flagship of the original plugins, was only 105 MHz, not the hoped-for 150 MHz.   

That was partially due to the fact that it used the existing Tek IC process (about 1 GHz).  The HP IC process was about 3 GHz.   

The 7A11 used discrete NPN and PNP transistors with 4 GHz f_t.

 

There is one noteworthy point I would like to make about the 7A11.  In those days, the different V/div settings were generally attained using fixed high impedance attenuators, usually stacked one after another and few with more than ×10 attenuation.   

I wanted the 7A11 to be able to handle the full useful range of sensitivites that other plugins attained when adding a ×10 probe ... that meant going to 20 V/Div.   

But you could certainly get to 5 mV/Div, which you could not attain with a 5 mV/Div plugin and a ×10 passive probe.   

 

But using that scheme, if the 7A11 went to 20 V/Div, it would have to stack three ×10 attenuators inside the probe, making it a behemoth, and even if you did, that would still leave the ×2 and ×5 attenuations up to gain switching in the amplifier.   

 

Problem with gain switching is, the bandwidth and transient response tend to change when you change amplifier gain.   

The greater the gain change, the greater the bandwidth/transient response change in the amplifier.   

The more gain settings you had, the worse it got due to longer leads and more parasitics.

 

To get a 1-2-5 sequence from 5 mV/Div to 20 V/Div, it had the worst of all possible worlds.   

The 7A11 needed two ×20 attenuators in the probe and that meant that it needed not just ×1, ×2 and ×5 gains in the amplifier, but ×1, ×2, ×2.5, ×4, ×5 and ×10!   

This had not been done before, but the 7A11 does it without any change in transient response!

 

The secret is three entirely passive, [[miniature relays | relay-switched]], O-pad attenuators:  ×2, ×2, and ×2.5 in a balanced transmission line environment.   

Stack them up (as you can do with matched pads) and you get all the combinations you need: ×1, ×2, ×2.5, ×4, ×5, and ×10.  From 5 mV/Div to 20 V/Div, twelve different V/div settings, more than any other high speed plug in!   

 

I got a patent on the variable attenuator, which was just a JFET shunting the O-pads.   

The patent had to do with making the gain vs. control rotation linear, which does not simply happen with linear gate-source voltage control.   

The JFET causes a small change in transient response, but not a bad one.

 

! 2023 value

|align=right| $6,400

|align=right| $6,400

|align=right| $7,900