32,070
edits
7A11: Difference between revisions
no edit summary
No edit summary |
No edit summary |
||
| Line 2: | Line 2: | ||
In includes a permanently attached FET probe. The bandwidth of a 7A11 in a [[7904]] | In includes a permanently attached FET probe. The bandwidth of a 7A11 in a [[7904]] | ||
is 250MHz. | is 250MHz. | ||
Regarding the 7A11, its designer John Addis says: | |||
<blockquote> | |||
The capacitance at the probe tip of the 7A11 is 6pF max, and | |||
goes down to 3.4pF at 100mV/div | |||
and down again to 2.0pF at 2V/div. | |||
The 7A11 was expensive: $850 for single | |||
channel vs dual channel [[7A12]] (105MHz) at $700. There was never another 1 megohm 7000 | |||
series plugin as fast as the 7A11...and it was one of the originals! The [[7A16]] was | |||
150MHz, but was a year later. Then the 7904 came out in late 1971. The 7A11 was | |||
250MHz (for $950). The 7A16 was then 225MHz (but for $625). The 7A16 disappeared | |||
rapidly and became the 7A16A, still 225MHz in 1973. In 1972, the [[485]] came out with a | |||
250MHz 1 megohm input. The scope used a faster IC process ("SH2", 3.5GHz) but the 1 megohm to | |||
50 ohm converter used only discrete devices. By 1974, the 7A11 was up to $950 and | |||
the popular [[7A26]] (dual trace) was $1,050, not counting a probe, and that would | |||
still only get you to 50mV/div with a X10 probe, five years after the 7A11 | |||
introduction. The 7A26 used the faster IC process and a vertical IC I designed for | |||
the 485. Remember also that the 7A11 was introduced at the same show as the | |||
HP183A/1830A/1840A faster (250MHz), smaller, lighter, less expensive with better | |||
triggering than 7000 series had. HP cleaned our clock until the [[7904]] came out. | |||
Their secrets were: Al DeVilbiss and a faster IC process. We had neither. The | |||
star-crossed 7A12, which was supposed to be the dual trace flagship of the original | |||
plugins was only 105MHz. That was partially due to the fact that it did use the | |||
existing Tek IC process (about 1GHz), but it was a slow process compared with HP's | |||
(about 3GHz). The 7A11 used discrete NPN and PNP transistors with 4GHz Ft. | |||
</blockquote> | |||
<blockquote> | |||
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 none more than X10 attenuation. I wanted the 7A11 to be able to | |||
handle the full useful range of sensitivites that other plugins attained when | |||
adding a X10 probe...that meant going to 20V/div. But why not go to 5mV/div, which | |||
you could not attain with the X10 probe? If you did, you could not stack three X10 | |||
attenuators inside the probe without making it a behemoth, and even if you did, | |||
that would leave the X2 and X5 attenuations up to gain switching in the amplifier. | |||
Problem with that is the bandwidth tends 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 with | |||
longer leads and more parasites. | |||
But if you did 5mV/div to 20V/div, you had the worst of all possible worlds, | |||
Turns out you needed two X20 attenuators in the probe and that meant that you | |||
needed not just X1, X2 and X5 gains in the amplifier, but X1, X2, X2.5, X4, X5 and | |||
X10! This had not been done before, but the 7A11 does it without any change in | |||
transient response! The secret is an entirely passive, relay-switched, balanced O | |||
pad attenuator: X1 (straight through), X2, X2, and X2.5 in a balanced transmission | |||
line environment. Stack them up (as you can do with matched pads) and you get all | |||
the combinations you need, X1, X2, X2.5, X4, X5, and X10. Twelve different V/div | |||
settings, more than any other high speed plug in! I got a patent on the variable | |||
attenuator, which was a JFET shunting the O pads. The patent had to do with making | |||
the gain vs control rotation linear. The JFET causes a small change in transient | |||
response, but not a bad one. | |||
Ron Peltola (of Peltola connector fame) designed the probe. | |||
</blockquote> | |||
* [http://bama.edebris.com/download/tek/7a11/tek-7a11.pdf Tektronix 7A11 Manual (PDF)] | * [http://bama.edebris.com/download/tek/7a11/tek-7a11.pdf Tektronix 7A11 Manual (PDF)] | ||