Talk:Baker clamp

2009

A schematic would be invaluable here instead of prolonged description.

And also some corrections because the text confuses the collector and the emitter. I don't have time right now, maybe I'll do it latter.89.137.246.65 (talk) 20:20, 2 November 2009 (UTC)Apass

Negative feedback

Girx, you keep taking out negative feedback; but that's how Baker describes it in the patent you cited. You should read it.

I believe the negative feedback statements add confusion without adding to the explanation. The simple view is the clamp is a comparator and a switch; when the collector voltage falls below a certain value, the switch turns on and diverts the excess base drive. Now there is a stable point where the diode is turning on just enough to divert the excess base current, but that description is subtle. The text was including modifiers such as "nonlinear" and "parallel". What's the loop gain? When the diode is reversed biased, the loop gain is zero. The diode is an odd summing device. When the transistor starts into saturation, its beta starts diminishing. If I put a voltage across a resistor and a diode, the circuit will reach a stable value. There's a feedback system working there, but I would be reluctant to explain it that way. Same issue with conventional diode clamps. Of course, you and C-D can overrule me. (And Motorola follows Baker and uses feedback.) Glrx (talk) 15:54, 12 April 2011 (UTC)

Your explanations above are temptingly attractive but they do not explain the very idea. The phrase "... the clamp is a comparator and a switch" gives only an idea how a diode can act as a switch (BTW I remember I saw in the past an "ideal diode" implemented by an op-amp acting as a comparator that drives a JFET acting as a switch). But the diode here is not simply a switch since the voltage drop across this switch would be zero. It acts here as a voltage-stable element maintaining some steady voltage across its terminals; this element is connected in series with the output voltage "source" (the collector-emitter junction). This "voltage string" is connected in parallel to the input base-emitter junction (or to the input "diode + base-emitter" string in the case of an additional base diode). Do you see the great "current steering" idea here? There are two voltage stable elements (the one - input, is steady and the other - output, is varying) connected in parallel; so, the most current flows through the element having lower threshold. The negative feedback is that controls this process and keeps up the collector voltage with V_F volts lower than the input voltage; thus it keeps the transistor in active mode. Circuit dreamer (talk, contribs, email) 17:36, 12 April 2011 (UTC)

It is interesting to see what happens if we reverse the feedback diode. Then the parallel negative feedback will keep up again steady collector voltage but now it will be above (with V_F) the base (input) voltage. If we connect the base and collector simply by a piece of wire, the collector voltage will be equal to the base voltage. These circuits are known as active diodes, active zener diodes, etc. (the input current-setting part of a current mirror is a good example).Circuit dreamer (talk, contribs, email) 17:53, 12 April 2011 (UTC)

I have trouble with your additions to the article. They focus on involved terminology and neat analogies. Although refs do refer to BC using feedback, do WP:RS refs say "nonlinear parallel negative feedback"? OK, the Baker clamp is a shunt regulator of the base current, but that is pretty opaque. The clamp diode could be an ideal diode (looking like a short when on). The clamp aspect is more important than the feedback aspect. There's no feedback until the catch diode turns on, but once the catch diode turns on the any additional base current is confronted with two paths: a high Z path to the base and a path to the collector that is (1/β) smaller. The collector looks like a good sink. The feedback is subordinate rather than prominent. Glrx (talk) 21:15, 12 April 2011 (UTC)

I don't know what the "parallel" means here, but it's certainly nonlinear. A loop gain and transfer function could be easily worked out for a model linearized about any particular operating point. It's probably important to treat the diode as an exponential, not a switch or constant voltage offset, to see what the feedback does. It's probably analyzed in Baker's original LL tech report, but we'll probably never find that; I sent a mesage to a friend there to see if they have a library of such. Dicklyon (talk) 22:29, 12 April 2011 (UTC)

Very interesting thoughts... They make me ponder over the roles of the clamp diode and the negative feedback phenomenon here... I have the feeling I need time to realize them... Circuit dreamer (talk, contribs, email) 16:41, 13 April 2011 (UTC)

"Parallel" = "shunt" as in "shunt regulator" or shunting a signal to ground. If you look at current rather than voltage (and its attendant exponential), you can solve by inspection. V_BE and V_F are essentially constant (25mV for factor of e); i_C = β i_B. Dynamic Z essentially h_ie. Glrx (talk) 16:55, 13 April 2011 (UTC)

What is called a Baker clamp?

Also, does the army book really call the Schottky diode a Baker clamp? Looking for evidence, I find evidence to the contrary, like here. Dicklyon (talk) 07:06, 12 April 2011 (UTC)

The army book does not use term Baker clamp and, being from 1959, is really about Ge devices and ignorant of Schottky. (The book does cover brute force clamps to avoid saturation: a stiff supply at couple volts and a catch diode supplies lots of collector current at a low voltage to prevent the saturation.) I believe the GE Transistor Manual (or possibly RCA TM) has a description of several "Baker" clamps with single and multiple diode configurations, but it's not handy.

One of the problems with early transistors is V_CE(sat) was often high, so even a Ge diode couldn't keep a Si transistor from saturating. DCTL needed low V_CE(sat) (Mot. High-speed Switching Transistor Handbook); switching transistors started being designed for that parameter; Si 2N2501 is < 0.2V at 10mA I_C β=10. The single Ge diode could then work with a Si transistor. We also get into other issues such as the turnoff time of the diode (Si 1N4148 were 4ns; some fast Si switching/signal diodes were 2ns). The single Schottky diode is a much better clamp because the diode is low forward drop and much faster than the transistor.

Motorola HS STH (1963) has the two diode Baker clamp on pg 32 (and refers to it as a feedback arrangement citing Baker, "Maximum Efficiency Switching Circuits", MIT Lincoln Lab Report TR-110, 1956.) Motorola, however, doesn't spend a lot of time on Baker clamps because if speed were important, it wanted current mode logic. Neither does Motorola identify them as Si or Ge; with low V_CE(sat) two Si diodes will work.

My take (and the switch mode and guitar refs suggest) that all diode circuits that divert base drive to the collector became "Baker" clamps. The logic circuit clamps didn't have the suck-out diode, but that diode became important for the power switching designers: they want to turn off the transistor fast. Somehow, Baker still gets credit. (And, at high voltages, a single Schottky diode clamp is not an option.)

Glrx (talk) 15:58, 12 April 2011 (UTC)

One more thing. The Kyttälä quote is odd, since the cited patent contains no hint that Baker is credited with inventing this thing. The claims are all to bistable circuits. He happened to use this clamp in his bistables, and that's why it got named for him, perhaps, but to say he is credited with the invention per this patent makes no sense. Do any RSs actually credit Baker with the invention? Dicklyon (talk) 07:09, 12 April 2011 (UTC)

The Motorola HS STH pg 32 states, "The use of diodes in a feedback arrangement devised by Baker..." (citing Lincoln Lab TR). Most texts just identify some circuits as Baker clamps; I don't know of the attribution. I went googling to find attrib and found the guitar ref that lead to Shea and another ref. Checking Shea and the TR would be appropriate. I doubt Baker devised all, but it seems that Baker understood the advantage of using low and high forward drop diodes. Glrx (talk) 16:09, 12 April 2011 (UTC)

Principles of Transistor Circuits. Richard F. Shea, Ed. Wiley, New York; Chapman & Hall, London, 1953, may be the book. Shea was at GE. Glrx (talk) 16:15, 12 April 2011 (UTC)

I've ordered the Army and Shea books and some others to see what they say. And I'll check my seven editions of the GE transistor manual. So far, I don't see any reliable sources referring to anything but the two-diode thing as a Baker clamp. Dicklyon (talk) 22:27, 12 April 2011 (UTC)

OK, I checked the transistor manuals. The GE third through fifth editions have a section on diode clamps to prevent saturation, and do show the Baker configuration with two diodes, Ge and Si, but don't mention the name Baker, so don't help clarify what the name applies to. The first and second editions are too thin to get into such things. The seventh is big and all revised, and I don't find it in there. I can't find my sixth, but if memory serves, it's a lot like the fifth. The RCA manuals SC-10 and SC-11 from 1962 and 1964 don't seem to have any such things. Dicklyon (talk) 05:15, 14 April 2011 (UTC)

Scratch that. I found my sixth, and rechecked my seventh, and it's the same in all of them (since the third), but the seventh is reorganized enough that I had to do some more looking. Dicklyon (talk) 22:57, 19 April 2011 (UTC)

Army book did not mention Baker, but it wasn't good on citing in general, so that doesn't say much.

I thought GE would mention Baker because that's the logical place for me to have learned about it. (BTW, my guess is Shea is the editor of the GE TM.)

If it's not in the GE or RCA transistor manuals, then it's even more of a mystery. How did so many people start calling it a Baker clamp?

Looks like patents are not citing Baker either.

x {{citation}}: Unknown parameter |country-code= ignored (help); Unknown parameter |inventor-last= ignored (help); Unknown parameter |patent-number= ignored (help)

x {{citation}}: Unknown parameter |country-code= ignored (help); Unknown parameter |inventor-last= ignored (help); Unknown parameter |patent-number= ignored (help)

Earlier, my take had been Baker's idea was diverting excess base drive to the collector, and that circuit variations on that theme were all termed Baker clamps. Now I have many doubts. Glrx (talk) 15:24, 14 April 2011 (UTC)

Single Schottky diode clamp called "'Baker clamp'". Paul Horowitz, Winfield Hill, The Art of Electronics, 2nd ed, Cambridge University Press, 1989, isbn 0-521-37095-7, page 908.

Linear Tech AN98f, 1 diode Baker clamp. http://cds.linear.com/docs/Application%20Note/an98f.pdf, page 3.

On Semi, AN1577D, and 4 diode Baker clamp. http://www.datasheetcatalog.org/datasheet2/4/09qljkxgr6i3wai7opqscsfx4afy.pdf, page 4

x {{citation}}: Unknown parameter |country-code= ignored (help); Unknown parameter |inventor-last= ignored (help); Unknown parameter |patent-number= ignored (help) and citing to Baker and Heald et al, "Design of Schottky-Barrier Diode Clamped Transistor Layouts", IEEE JSSC, vol. SC-8, No. 4, pp. 269-275, Aug. 1973

x {{citation}}: Unknown parameter |country-code= ignored (help); Unknown parameter |inventor-last= ignored (help); Unknown parameter |patent-number= ignored (help)

Then a surprise - another R H Baker patent

http://www.fleadh.co.uk/downloads/MultilevelAsymmetric_PEMD2002.pdf citing Baker and Nabae

x {{citation}}: Unknown parameter |country-code= ignored (help); Unknown parameter |inventor-last= ignored (help); Unknown parameter |patent-number= ignored (help)

Nabae A., Takahashi J. and Akagi H., 1980, “A new neutral-point-clamped PWM inverter”, IEEE-IAS ’80 Record, 761–766.

Glrx (talk) 00:31, 15 April 2011 (UTC)

Personally, I had never heard of a Baker clamp until recently, so I'm not surprised to not find it in my old books; I haven't encountered "so many people" using the term at all. I just received the copy of Shea 1953 that I ordered, and it does not contain anything resembling a base-collector diode clamp, not even any mention of the concept of saturation or stored charge or recovery time, as far as I can find. It's interesting that Horowitz and Hill call the Schottky thing a Baker clamp; maybe they started that. TI didn't, as you point out in their patents. I think you hallucinated Baker and Heald; Heald's co-author on the JSSC article was Dave Hodges. I had looked at a number of other Baker patents, but didn't find them very relevant. Dicklyon (talk) 03:08, 15 April 2011 (UTC)

That Shea doesn't discuss clamping raises doubt about the claim it was known before Baker. The idea may be Baker's. We also have the problem that some WP:RS term variations as Baker clamps and others are silent. Not single cite to a Baker and Heald but a cite to Baker pat and a cite to Heald et al. My surprise was I thought Baker was getting credit for power switch applications, but he actually was working with power switches -- and implying he might have done the suck out diode. Glrx (talk) 01:10, 16 April 2011 (UTC)

I checked Shea 1953 and confirm that it does not address base-collector clamping. Glrx (talk) 20:04, 22 April 2011 (UTC)

And I got the Shea 1957 book; it's in there, with reference to Baker's TR. It doesn't call it a Baker clamp, but has the footnote for its development; I'll put a quote later. And it has the other resistor/diode circuit like the Army book, implying that's earlier and not as great. Dicklyon (talk) 21:51, 22 April 2011 (UTC)

Here's a 1971 patent that refers to Baker clamp and uses the two-diode form. Here's a somewhat different clamp called a Baker clamp. Here is another conventional two-diode one. And this is a minor mod.

And here is a 1969 paper that very clearly contrasts the Baker clamp with the Schottk-barrier-diode (SBD) clamp. I can send you a copy if you email me. Dicklyon (talk) 03:59, 15 April 2011 (UTC)

Pressman's 2009 book has a big section on Baker clamps, and references his 1977 book (this one). It's all about the two-diode and three-diode version and elaborations; not the one-diode Schottky-style circuit. Dicklyon (talk) 21:54, 15 April 2011 (UTC)

Yes, I'd looked through Pressman a few days ago. I don't see power-switch guys as interested in the one-diode Schottky clamp; the breakdown voltages on Schottky were small for a long time. Low BV for a logic circuit is not a problem. Glrx (talk) 01:10, 16 April 2011 (UTC)

I received the Alley & Atwood first (1962) edition today, and no Baker clamp in there. It's in the 1966 edition; I'm looking for a cheap one. Dicklyon (talk) 00:58, 16 April 2011 (UTC)

Shea's 1957 edition does talk about clamps using diodes, and references the 1956 Baker paper, but doesn't otherwise mention Baker by name, looks like. That's about all I can tell, so I ordered a copy of that edition, too. Dicklyon (talk) 03:20, 16 April 2011 (UTC)

I've ordered a copy of Baker's TR. I'm also wondering if Long or similar legend is the source of Kyttälä's statement. Kyttälä's book is self-published. Glrx (talk) 01:54, 20 April 2011 (UTC)

Did you find an LL librarian to order from? I asked my bud there and he said he'd send my query to the librarian, but since then I haven't heard anything. Dicklyon (talk) 04:02, 20 April 2011 (UTC)

I ordered it from the NTIS bitbucket. Glrx (talk) 14:40, 20 April 2011 (UTC)

After one month, the report finally arrived. The max efficiency aspect was about power consumption: he used complementary logic (essential CMOS logic with bipolar transistors). He cites 8 references, but none are for clamp circuits. Baker's TR terms the technique "back clampling" and calls the collector-base diode a "back clampling diode". (Figure 20.) He gives examples of the single diode connected to tapped R_b resistors (figure 19a), Ge transistor with Si base diode and Ge back diode (Figure 19b [with base bias resistor] and 20.), and Si transistor with single Ge back diode (figure 26; Schottky clamp configuration). Baker used 0.2V at 1mA for Ge diode drop and 0.7V at 1mA for Si diode drop. He knew importance of overdrive and used speed up capacitors.

"The circuity techniques described in Figs. 7 through 18 may be extended to nonsaturating circuity. The primary gain in designing the circuits to operating in the nonsaturating mode is in the maximum operating speed.

"The circuit configurations to accomplish this are shown in Fig. 19."

"The reason that the back-clamped configuration of Figs. 19 and 20 does not allow the transistors to be saturated may be seen by comparing the diode characteristics shown in Fig. 24.

"If silicon transistors are used, the nonsaturating circuitry does not require the four silicon diodes. This may be seen from the silicon collector family shown in Fig. 25. The required basic configuration using silicon transistors is shown in Fig. 26."

Glrx (talk) 00:07, 22 May 2011 (UTC)

Awesome. Can you perhaps email me a scan of it? Dicklyon (talk) 04:01, 4 June 2011 (UTC)

Discharging diode

I wonder how the "second base diode connected antiparallel to the base diode" having V_F = 0.7 V will discharge a transistor base charged to the same V_BE0 = 0.7 V? Circuit dreamer (talk, contribs, email) 19:15, 20 April 2011 (UTC)

The antiparallel (suck out) diode is used in power switching applications that have bipolar drive (e.g., from a transformer). See, for example, [1] page 176, that has about ±6 V drive. (That ref also points out that a slow base diode is an advantage.) Glrx (talk) 20:22, 20 April 2011 (UTC)

Unsourced tangents

I took out the unsourced tangents on ECL, TTL, and DTL. If we find a source that connects these to Baker clamps in some way, we should say something about that connection, but lacking that, these asides are not very suitable, and putting in a notes sections specially for them is not justified either. Dicklyon (talk) 00:58, 23 April 2011 (UTC)

Dicklyon, please understand me! Glrx, please second me! These notes make connections (associations) between the present circuit and other circuits to help understanding of so interesting, important and powerful techniques as: "decreasing the gain just before the saturation point by turning on a nonlinear (in ECL - series) negative feedback" and "current steering (connecting a pn-junction in parallel to two or more series connected pn-junctions with the purpose to divert the current". We will certainly not find sources making these connections but this does not mean that we should not use them to explain to visitors odd circuit phenomena and odd circuits implementing them! These associations serve as "bridges" between apparently different circuit solutions. If it is not so clear, I can explain the written in more details! Circuit dreamer (talk, contribs, email) 04:32, 23 April 2011 (UTC)

I understand your desire to explain things and make connections. If you do that in a paper and get it published, we can use that as a source. Otherwise, it's outside the scope of wikipedia to present unverifiable connections, per WP:V. Dicklyon (talk) 04:39, 23 April 2011 (UTC)

I agree with Dicklyon about the removed ECL, TTL, and DTL statements being tangents and being unsourced. I also have technical problems with the statements (including the reinserted TTL comment).

Commercially, I'm not sure that the Baker clamp was ever significant to DTL. It wasn't used in the basic IBM 1401 gate (and speed up caps were used on level translators). I don't know if diode clamps were used in the SE 110 series or the μL 930 series; IIRC, the (possibly simplified) schematics did not include clamps. Historically, TI has a patent using a single Schottky diode on a DTL gate, but I don't know if that made it into production. There's no clear indication that anti-saturation clamping was needed for basic DTL gates; the R_LC_load time constant is a significant penalty.

So I'd revert the reinsertion about TTL.

I would keep the DTL see-also because DTL links to this article.

Glrx (talk) 17:05, 23 April 2011 (UTC)

I have mentioned TTL as another example of the powerful current steering idea. Ideas are important, not the specific circuit implementations. These circuits have gone down in history but the ideas behind them are living and can be implemented by new components. Ideas are immortal and eternal and we have to reveal and show them to Wikipedia readers. Circuit dreamer (talk, contribs, email) 17:36, 23 April 2011 (UTC)

The connection between Baker clamp and TTL gate

In both the circuits there are the same configurations - one p-n junction is connected in parallel to two connected in series p-n junctions. As a result, there is the same current steering effect - the common current is diverted to the single p-n junction.

Baker clamp: the collector-emitter junction of the present transistor TO connects the diode D1 in parallel to the two connected in series diode D2 and TO base-emitter junction. As a result, the input current I2 (limited by Rb) is diverted through TO collector-emitter junction to ground.

TTL gate: the collector-emitter junction of the previous output transistor T2' connects T1 emitter-base junction in parallel to the two connected in series T1 base-collector junction and T2 base-emitter junction. As a result, T1 base current (limited by Rb1) is diverted through T2' collector-emitter junction to ground.

Circuit dreamer (talk, contribs, email) 09:49, 25 April 2011 (UTC)

Sources. We need WP:V and WP:RS if you want to expound on such a connection. Dicklyon (talk) 15:45, 25 April 2011 (UTC)

Sources for "one diode connected to other two"? Circuit dreamer (talk, contribs, email) 16:56, 25 April 2011 (UTC)

Sources for a connection of Baker clamp to TTL. Dicklyon (talk) 04:26, 29 April 2011 (UTC)

Well, BC is a "one diode connected to other two"... and TTL is a "one diode connected to other two" as well... so there is a connection between them... Circuit dreamer (talk, contribs, email) 04:51, 29 April 2011 (UTC)

The TTL input with multiple emitters and a Baker clamp are fundamentally different circuits. The TTL input is the same as having multiple common base transistors that have the collectors tied together and the bases tied together, with the emitters used independently. It is also very similar to a current mirror. Each emitter can independently supply current to other TTL outputs that are low. The current supplied does not change with the number of low outputs. The Baker clamp is not a current mirror circuit. When the base-emitter junction is forward biased and the transistor is in the active region, D1 is reverse biased. As the collector current increases enough current goes through D2 D1 to keep the transistor from going into saturation. Not what I would call negative feedback, but more important not how a current mirror works.Zen-in (talk) 04:06, 29 April 2011 (UTC)

Common-base? Current mirror?!? "As the collector current increases enough current goes through D2 to keep the transistor from going into saturation" ?!?!? Circuit dreamer (talk, contribs, email) 04:44, 29 April 2011 (UTC)

Dead link

The "[2]" link to R. H. Baker (1956), "Maximum Efficiency Switching Circuits", MIT Lincoln Laboratory Report TR-110, has gone dead. Does anyone have a copy? It is mentioned here: http://www.dtic.mil/docs/citations/AD0096497 (with a slightly different title), but there is no link to the actual document. NCdave (talk) 02:49, 21 January 2015 (UTC)

You are correct, it should be Maximum Efficiency Transistor Switching Circuits. DTIC never had an online copy -- it was always a link. Glrx (talk) 03:33, 21 January 2015 (UTC)

External links modified

Hello fellow Wikipedians,

I have just modified one external link on Baker clamp. Please take a moment to review my edit. If you have any questions, or need the bot to ignore the links, or the page altogether, please visit this simple FaQ for additional information. I made the following changes:

• Added archive https://web.archive.org/web/20091222074842/http://www.diodes.com:80/_files/products_appnote_pdfs/zetex/an22.pdf to http://www.diodes.com/_files/products_appnote_pdfs/zetex/an22.pdf

When you have finished reviewing my changes, please set the checked parameter below to true or failed to let others know (documentation at {{Sourcecheck}}).

This message was posted before February 2018. After February 2018, "External links modified" talk page sections are no longer generated or monitored by InternetArchiveBot. No special action is required regarding these talk page notices, other than regular verification using the archive tool instructions below. Editors have permission to delete these "External links modified" talk page sections if they want to de-clutter talk pages, but see the RfC before doing mass systematic removals. This message is updated dynamically through the template {{source check}} (last update: 18 January 2022).

• If you have discovered URLs which were erroneously considered dead by the bot, you can report them with this tool.
• If you found an error with any archives or the URLs themselves, you can fix them with this tool.

Cheers.—InternetArchiveBot (Report bug) 10:44, 24 October 2016 (UTC)