Keyboard Shortcuts
ctrl + shift + ? :
Show all keyboard shortcuts
ctrl + g :
Navigate to a group
ctrl + shift + f :
Find
ctrl + / :
Quick actions
esc to dismiss
Likes
Search
P6131, T-coils and other probe stuff
|
John,
toggle quoted message
Show quoted text
I do have a manual that contains a schematic for the 6106 (most probably from tekwiki). I do see the two inductors, but they don't look like part of a T-coil in this case. They certainly do have some mutual coupling, though, given their proximity and parallel alignment, but I am inclined to think that the effect is just incidental. I'll have to give it some thought, though. Thanks to your previous post, I am now looking at a freshly disassembled comp box for a 6131. I'll see how much I can reverse engineer from staring at it. And thanks also for offering to mail me the manual for the 6131 -- the address below is indeed the right one to use. It'll be great to see how Tek's schematic might differ from mine.? I'll eventually get around to scanning it and uploading it to tekwiki, perhaps with a few notes. Finally, thanks for the homework -- studying the compensator is a heck of a lot more fun than trying to resurrect a Hameg 507 that a colleague killed by knocking over a Starbucks Venti something-or-other into its guts, and leaving it that way for months, hoping that somehow it would heal itself. -- Cheers, Tom -- Prof. Thomas H. Lee Allen Ctr., Rm. 205 350 Jane Stanford Way Stanford University Stanford, CA 94305-4070 On 1/17/2021 16:03, John Gord via groups.io wrote:
Tom, |
|
Tom,
toggle quoted message
Show quoted text
I'm happy to send the manual along. I think that the P6131 had some major changes over its production history, particularly a new comp box housing design. They show both box types in the 070-4210-01 manual, and the schematic shows a change in L1020 value between the box versions. (60nH in the old version, 50nH in the version, not the T-coil inductor.) --John Gord On Sun, Jan 17, 2021 at 04:44 PM, Tom Lee wrote:
|
|
Thank you very much, John -- I appreciate it greatly!
toggle quoted message
Show quoted text
I had fun this afternoon trying to figure out how the T-coil in the 6131 is wound and connected. It's a quadrifilar winding -- four colors, two of them similar unless the lighting is just so -- so not as easy to trace out as one would have thought. But I think I finally figured out what was connected to what, and have enough measurements on dimensions to figure out a crude model for the T-coil. I apparently disassembled a newer model, as my estimated inductance was 45nH. I'll call that 50nH, especially if the number is supposed to include the 7mm of interconnect from the inductor to the probe cable. That fact that you have that data makes me hopeful that the documentation also gives values for the trimcaps and one very small fixed cap. I'd rather not have to desolder them to make measurements. Thanks again! --Cheers, Tom -- Prof. Thomas H. Lee Allen Ctr., Rm. 205 350 Jane Stanford Way Stanford University Stanford, CA 94305-4070 On 1/17/2021 20:25, John Gord via groups.io wrote:
Tom, |
|
Dear Professor Lee,
Importantly, I would NEVER, in any way, try to discourage anyone! I thank you for what you called a "Rant"; I would, and still do, call it a very informative bit of information. It was very informative. I was not trying to discourage Mr. Cabot. I was simply trying to inform and I stand corrected in my misguided missive. No, I fully recognize that cutting off a minuscule portion of a probe cable would most likely effect the cable minusculely; I've taught The Calculus at the uni. When I wrote what I wrote I had not read further into what Mr. Cabot's efforts had accomplished. I simply wanted to add what turned out to be misguided information and then read everything in the series later that I might learn from others. I'm not sure the last bit of info has been imparted to all. I have sent to you a schematic of the P6131 from a Tek manual I have for same. I once posted this to I think this group asking for help in understanding it; I got no answers; I STILL DO NOT UNDERSTAND IT. I asked a couple of my professors at the uni and they were of no help; I think it a mater of time. I should ask again before the next semester starts. My professors told me it was a T-coil. Again, I thank you, larry |
|
On Sun, Jan 17, 2021 at 11:25 PM, John Gord wrote:
Hello Mr. Gord, My manual is 070-410-00; I think 'older' than yours. My manual shows L1020 to be 60nH for a 1.3 meter probe and 37nH for a 3 meter cable. Does the newer manual show different values for 1.3 and 3 meter cables? I have sent a photo copy of both lengths to Professor Lee. I would send you a copy, but I do not know your email. I once asked about this manual with reference to a T-coil. I received no answers. I STILL DO NOT UNDERSTAND THE CIRCUIT!!!! Thanks, larry |
|
Hi Larry,
toggle quoted message
Show quoted text
Please call me Tom! You may regret having asked this question, because the T-coil is a bit of an obsession of mine. Students in my RF classes are probably tired of hearing me drone on endlessly about the thing, and go into a deep coma when I start yet again. There have been deaths. I'll try to keep it short here, but if you have follow-up questions, as is likely, be forewarned that the rabbit hole is very deep! The T-coil is indeed a wonderful -- and mysterious -- bit of magic. The derivation of the full equations for their design was kept a trade secret by Tektronix until not that long ago. They shouldn't have worried -- the full derivation is still mysterious even when you have it in front of you. I published the first derivation in the open literature about 15 or so years ago, and it was the hardest derivation I'd ever done, even though it was only for the simplest case (lossless L and C, perfectly symmetrical network). Its chief virtue was that it ultimately led me to meet Bob Ross, who had done the original derivation at Tektronix for the complete case. So, after many years of staring at the damn thing, I'd finally gotten to hear the details from the man himself. A colloquial version of the history was written up by Paul Rako and can be found here: . Minor mistakes crept in (both of circuits and history), but it's close enough, and Paul tells the story in an engaging way (as he always does). Since the problem is one of driving a capacitance over a large bandwidth, it seems like a potentially good idea to embed the load in something that looks like a bit of transmission line. Why? Because a transmission line has a very large bandwidth (ideally infinite), even though it consists of inductances (which represent the wires) and capacitances (which represent the stuff between the wires). The bandwidth only approaches infinity in the continuum limit of an infinity of infinitesimal inductances and capacitances. If you build a discrete approximation with finite L and C, of course you get a low pass filter. But the more segments you use (while keeping the total L and C fixed), the greater the bandwidth. The first amplifier to exploit this notion of absorbing capacitances into a discrete T-line approximation was the distributed amplifier (invented in the 1930s, but revealed only in a patent). Tek mastered the art like no other, as the 585 showed. The irony is that Bill Hewlett had co-authored the first paper to explain the distributed amplifier. Tek, not HP, appreciated the magic's potential and ran with it. Treating an infinitesimal T-line segment as an LC network has strong intuitive appeal, but isn't the only possible choice. The bridged T-coil network is another. It is an all-pass network, and was originally developed as a phase equalizer in telephone networks, as far as I can tell. Being all-pass, it allows much faster convergence to T-line behavior than does an LC low-pass segment, as one takes the limit. That means that if we represent a T-line with a finite number of segments, making the segments out of bridged-T coils will give us a higher bandwidth than if using simple LC low pass ones. So the bandwidth-enhancing properties of the bridged-T coil comes from its ability to absorb a troublesome load capacitance into the network naturally. Using a bridged T-coil in, say, a textbook common-emitter amplifier, allows us to achieve nearly triple the bandwidth. That's spectacular, and it's what allowed Tek to go from the 50MHz 453 to the 150MHz 454 with relatively few modifications. There are some devils that are encountered when you start looking into the details. One is that the loads we drive are not always purely capacitive. There may be inductances and loss as well. These perturb the nice equations (and amplifier performance) in extremely ugly and baffling ways. Until Bob Ross came along, Tek's engineers were using the time-honored cut-and-try approach, guided by some intuition, and driven by desperation. It was clunky, and not always successful. Bob gave them the full math which, when understood, revealed that one could still get huge bandwidths boost, even with a lossy capacitance, if one used asymmetrical T-coils and twisted some other component values a bit. It is a marvelous result, and Tek hid it as well as it could. When driving bipolar transistors, the base impedance is a lossy capacitance. Bob's equations showed the optimal way to use T-coils to extract the maximum bandwidth even in that case, within the constraint of good pulse response. He paved the path for the 7904, 7104 and so on. Without the equations, those products likely wouldn't have happened until years later, if at all. The same idea can apply to scope probes, where the input impedance of the scope is now the lossy capacitance to be driven (e.g., 1 megohm shunted by 15pF, or what have you). I note that, in the schematics you kindly sent me, Tek does not actually show a T-coil -- very sneaky! They are trying to pretend that the compensation box only contains a textbook equalizer. I opened up a 6131 compensation box yesterday for the first time ever to see what's inside, thanks to John Gord's telling me that there was a T-coil in there. If Tek had shown the T-coil in the manual, you would have seen that it is highly asymmetrical, with the center tap (which drives the scope), not at the center at all -- there's a 3:1 turns ratio, rather than 1:1. The desirability of that asymmetry is what Bob Ross figured out. Very smart guy (and a very kind one)! That's enough for now, I hope. If your uni professors would like to see a derivation of the T-coil equations, you can send them to mine in Planar Microwave Engineering. It does not use anything like Bob Ross's approach, but he never published his, so I can't direct you to another reference (although there are some slides online where he gives summaries and hints). But mine uses standard techniques that most EEs would find accessible, though perhaps not entirely familiar. Unfortunately, my approach is also pretty much a one-off dead end -- it will not extend gracefully to accommodate the lossy-load case that Bob's method can handle. If there is a network theorist left on your faculty (very rare these days), you can utter the words "Wang Algebra" or "Duffin-Boff Algebra" and see if they light up. If they do, they'll have identified themselves as the right ones to talk to about the T-coil -- Cheers, Tom -- Prof. Thomas H. Lee Allen Ctr., Rm. 205 350 Jane Stanford Way Stanford University Stanford, CA 94305-4070 On 1/18/2021 07:36, Lawrance A. Schneider wrote:
Dear Professor Lee, |
|
Tom Lee wrote:
But are the number statistically significant? In any sample of a thousand people you will have some number die of heart attacks, strokes, aneurisms, defenestration, or whatever. It could hardly be the fault of your fine lectures if a few students in any year suffer these tragic, but entirely unavoidable mishaps. I'm still reading through the T-coil papers from the last time this came up. It's a bit brain bending, but I also started to notice pairs of little inductors in my 475 and 475A. -- Jeff Dutky |
|
:)
toggle quoted message
Show quoted text
It's remarkable that adding a little inductance in the right place can produce such a huge boost in bandwidth. Tek kept HP at bay by combining T-coils with the ft-doubler invented by Carl Battjes and, of course, CRT technology that was second to none -- a trifecta of magical technogeek goodness. --Tom -- Prof. Thomas H. Lee Allen Ctr., Rm. 205 350 Jane Stanford Way Stanford University Stanford, CA 94305-4070 On 1/18/2021 13:36, Jeff Dutky wrote:
Tom Lee wrote:There have been deathsBut are the number statistically significant? In any sample of a thousand people you will have some number die of heart attacks, strokes, aneurisms, defenestration, or whatever. It could hardly be the fault of your fine lectures if a few students in any year suffer these tragic, but entirely unavoidable mishaps. |
|
Tom,
toggle quoted message
Show quoted text
I put schematics of the P6131 1.3 meter and 3 meter probes in the photos section, album name "P6131". The full manual is following by snail mail. /g/TekScopes/album?id=259558 --John Gord On Mon, Jan 18, 2021 at 01:43 PM, Tom Lee wrote:
|
|
Thank you very much, John!
toggle quoted message
Show quoted text
-- Cheers, Tom -- Prof. Thomas H. Lee Allen Ctr., Rm. 205 350 Jane Stanford Way Stanford University Stanford, CA 94305-4070 On 1/18/2021 15:00, John Gord via groups.io wrote:
Tom, |
|
Very interesting, John -- the 1.3m schematic does show a T-coil, and confirms the 3:1 ratio that I counted (a relief to know that I can still count to three). The schematic for the 3m compensator you sent matches Larry's -- it doesn't have a T-coil. Hmm...
toggle quoted message
Show quoted text
-- Cheers, Tom -- Prof. Thomas H. Lee Allen Ctr., Rm. 205 350 Jane Stanford Way Stanford University Stanford, CA 94305-4070 On 1/18/2021 15:00, John Gord via groups.io wrote:
Tom, |
|
Tom,
toggle quoted message
Show quoted text
I think the 3 meter version is only rated for 150MHz vs 300MHz for the 1.3 meter version. --John On Mon, Jan 18, 2021 at 03:48 PM, Tom Lee wrote:
|
|
The part that surprises me is that the T-coil was deemed unusable for the case that could most benefit from it. I'm just wondering which of several possibilities was the showstopper for the 3m cable. That was the "hmm" part. :) I have some speculations, but nothing sufficiently baked to say anything yet.
toggle quoted message
Show quoted text
--Cheers, Tom -- Prof. Thomas H. Lee Allen Ctr., Rm. 205 350 Jane Stanford Way Stanford University Stanford, CA 94305-4070 On 1/18/2021 17:19, John Gord via groups.io wrote:
Tom, |
|
On Mon, Jan 18, 2021 at 01:36 PM, Jeff Dutky wrote:
Given that a T-coil is sometimes construed to be a kind of peaking coil, you might read Oliver Heavyside's paper. As a related aside... Heavyside is an absolute hero of mine... his story is both fascinating and sad: "important" people endeavoured to silence his voice, and the debate, about his mathematical, and scientific, discoveries.... why? ... simply because they were "better" than him. |
|
Hi Tom,
<SNIP> A colloquial version of the history was written up by Paul Rako and can be found here: I have three comments: 1) This was a wonderful article with a very clear, easy to follow explanation. The Spice models will allow us to experiment with the underlying math without worrying about trying to understand the mathematical equations that always put me to sleep). 2) The hat was not as good as the article :) 3) Next time you are in Seattle I'll take you shopping and make sure you leave with a proper hat. Dennis -- Dennis Tillman W7pF TekScopes Moderator |
|
Hi Dennis,
toggle quoted message
Show quoted text
Generous offer accepted. :) See you in Seattle one of these days, after the Age of Covid passes! -- Cheers, Tom -- Prof. Thomas H. Lee Allen Ctr., Rm. 205 350 Jane Stanford Way Stanford University Stanford, CA 94305-4070 On 1/18/2021 20:45, Dennis Tillman W7pF wrote:
Hi Tom, |
|
When it is safe come and stay with us as our guest. We have a large house with plenty of bedrooms and lots of Tek toys to play with. Summertime we have a small house by the beach on an island on the south (Puget) Sound and there is a little guest house for you there as well.
toggle quoted message
Show quoted text
Dennis -----Original Message-----
From: [email protected] [mailto:[email protected]] On Behalf Of Tom Lee Sent: Monday, January 18, 2021 9:44 PM To: [email protected] Subject: Re: [TekScopes] P6131, T-coils and other probe stuff Hi Dennis, Generous offer accepted. :) See you in Seattle one of these days, after the Age of Covid passes! -- Cheers, Tom -- Prof. Thomas H. Lee Allen Ctr., Rm. 205 350 Jane Stanford Way Stanford University Stanford, CA 94305-4070 On 1/18/2021 20:45, Dennis Tillman W7pF wrote: Hi Tom, -- Dennis Tillman W7pF TekScopes Moderator |
|
I can't resist commenting about:
"If there is a network theorist left on your faculty (very rare these days)," In fact, it appears that circuit theory classes have nearly vanished. One of the few I know of is taught by none other than Gabor Temes: I find that some of the old classiques, such as Van Valkenburg or Kuo are truly edifying, for example, conversion from one 2-port form to another. Great stuff, highly recommended, , etc. Once you've seen transfer matrices (ABCD), you'll never look at circuit analysis the same way again. I'll stop before I wax even more ecstatically. On Tue, Jan 19, 2021 at 5:19 PM Dennis Tillman W7pF <dennis@...> wrote: When it is safe come and stay with us as our guest. We have a large house |
|
Hi Mark,
toggle quoted message
Show quoted text
Very nice to see you here! Your review article on sampling scopes is a true classic. I blame and thank you for the happy hours I've spent trying to chase down and then read, all the references you cite. :) I can't imagine the hours you spent amassing all of that information. Simply amazing. Gabor is a dear friend, one of the last remaining links to that earlier generation, a rarity who knows that "Bartlett bisection" isn't a medical procedure. Cheers Tom Sent from an iThing, so please forgive the typos and brevity On Jan 31, 2021, at 16:29, "Mark Kahrs" <mark.kahrs@...> wrote:
I can't resist commenting about: |
to navigate to use esc to dismiss