Sunday, October 22, 2017

Squeeze Keying - Iambic mode operation

IAMBIC modes of electronic keying

Iambic mode keying occurs when you squeeze dual lever paddles, and then release both paddles simultaneously.  If the keyer is in mode A, it will finish sending the current element (dit or dah), and then stop. If the keyer is in mode B, it will finish sending the current element, then send another element, and stop.
Electronic keyers provide Iambic modes to allow for fewer strokes while keying, thus improving keying efficiency
 Mode B requires fewer strokes than Mode A

You can stop reading now, if you'd like.  The remainder is just me prattling on.


Iambic modes were a mystery when I began learning the code.  My first transceiver defaulted to Mode A and I had Bencher paddles so I got used to squeeze keying a bit but only in Mode A.  When I got my TenTec Eagle, its internal keyer only supported Mode B and that totally screwed me up.  I didn't know what was going on and thought something was wrong with the radio.  I ended up using my old HamKeyer external keyer with the TenTec Eagle rather than figure out Iambic mode B.  But, as with most things in this hobby, I eventually decided I wanted to understand what the Iambic modes were about and how to use them.

Iambic refers to a pattern of syllables, typically in a line of poetry.  There are various "meters" that describe different syllabic patterns.  Iambic is a pattern that has a short sound followed by a long sound, as in deDUM or ditDAH. If you've ever used a CW dual paddle key with an electronic keyer and squeeze the paddles you'll be greeted with  ditDAH ditDAH ditDAH ditDAH ditDAH ...  Aha! Iambic meter.

Iambic keying saves strokes.  Comparing the number of strokes necessary to key all letters of the alphabet and all digits with different keys and keyers yields the following result, provided the operator is consistently squeezing with twin-lever keyers:
 straight key  132
 semi-automatic "bug" key  100
 single-lever keyer  73
 iambic keyer  65

Iambic Paddles?

Unless you like to practice sending Walt Whitman poems, there's not much direct relationship between ham radio and poetry.  But I've read CW operators referring to Iambic "paddles" or advertising an "Iambic paddle".... Well, I'll argue there's no such thing as an Iambic paddle because a paddle by itself whether it is a single lever paddle or dual lever cannot make an Iambic pattern on it's own.  A paddle is only half a CW key, it needs a keyer, and the code in the keyer is what makes Iambic possible!

Whew! Alright, I got that off my chest. I'm not sure what get's me so worked up about this sort of "yes it matters" things... I blame being locked in the basement as a child, but that's fodder for another blog post...

Iambic keying requires a dual-lever paddle and an electronic keyer

Ok, back on topic; a dual lever paddle can be used with an electronic electronic keyer, that supports Iambic mode(s).  A single lever paddle when used with an electronic keyer cannot make use of Iambic mode, so called, squeeze keying modes.  So when we talk about using a paddle with an electronic keyer for Iambic mode keying we are referring to a dual-lever paddle connected to an electronic keyer (or the electronic keyer circuit built into a transceiver).

A two lever paddle is required because the Iambic modes of the keyer are employed while both levers are pressed and released at the same time.  What happens when both levers are released simultaneously depends on the Mode that the keyer is using.

A dual lever paddle

The Modes only matter when the paddles are squeezed and released together

The difference in the two modes only matters when you release both paddles simultaneously.  So if you never release both paddles together, you won’t see a a difference.
Mode A -- The keyer will complete the element being sent (either a dit or a dah) when the paddles are released
Mode B -- The keyer sends an additional element when the paddles are released

You can avoid Iambic Modes altogether

You can avoid interaction of either Iambic mode if you release each paddle as it finishes its final element rather than releasing them together.  For example with the letter "C", DAHditDAHdit, release the DAH paddle during the final DAH, and release the dit paddle during the final dit.  For a letter ending in DAH like "A", ditDAH, release the dit paddle during the final dit, and then release the DAH paddle during the final DAH.  This will produce the correct code whether the keyer is in Iambic mode A or B.


The following video shows the different modes in action...

To Squeeze or Not to Squeeze? That is the question

Ok, I'll admit squeeze keying allows you to send kinda lazy, and it's cool when you demonstrate CW to someone how little motion you can use with paddles compared to a straight key.  However, there are numerous articles out there for and against squeeze keying, and I think they both make good points but some proponents of each treat it as a somewhat religious doctrine and I just can't get that serious about it.

Personally I enjoy squeeze keying when I'm sending at 20 words per minute or less because it's relaxing and is a bit nerdy to let the keyer do a dit of extra work for me now and again (see what I did there).  But I start making more mistakes as the speed increases past 23 wpm and I absolutely can't squeeze-key above 27 wpm.  I have trouble using a dual paddle above 25wpm because I accidentally squeeze key when I shouldn't and Iambic stuff occurs accidentally.  I could likely improve with practice, but higher speeds are easier for me with my Bug or a single paddle key so I haven't really tried to speed up using dual paddles.  For now it's not much of an issue because I rarely QSO above 23wpm.  I only key above 25wpm when working a DX station and it's such brief bursts that I just use my bug or send my "5NN TU dit dit" using a keyer macro.

History of Keyers and modes

The following excerpt is used by permission of Karl Fischer, DJ5IL.  Karl has written an excellent article entitled "All about Squeeze-Keying" available in its entirety at: 

In 1951 an electronic single-lever key was described1 which used 5 tubes to send self-completing dot- and dash-elements with automatic spacing between letters and words. But its continuously running time-base resulted in an uncontrollable beast so that the author himself wrote he does "not feel that any but the most feverish electronic key enthusiasts will wish to build one of these infernal, maddening machines", but nevertheless he hoped that the idea might provide an inspiration for further development.

John Kaye, W6SRY, accepted that challenge and came up with a rather ingenious design which he published2 as the Ultimatickey in QST magazine in 1953 . The circuit is based on 3 tubes and 7 relays and sticks to the basic idea of a continuously running time-base, which is the only weak spot of his design: pulses from the time-base trigger the generation of dot and dash-elements, and so they do not start immediately with the closure of a key contact but only with the next pulse. By addition of dot/dash-memories he avoids dropping of leading elements and transforms the beast into a beauty: once a contact of the singlelever key has been closed, that closure is retained by a memory-relay contact parallel with the key contact and the associated dot- or dash-element is properly generated as soon as the trigger pulse arrives, even if that key contact is open again or the opposite key contact is closed by then. The dot/dash-memory relays are reset and their contacts opened by the closing contacts of the associated dot/dash-generator relays. The dot/dash-memories are independent of each other and because a dot and dash often are rapidly stored together before keying starts, a sequencing circuit retains the proper order in which the dot/dash-generators are initiated.

This combination of independent dot/dash memories which not only avoid dropping of leading elements but also provide tremendous timing leeway with sequencing allows the storage not only of a single dot or dash but of a whole dot + dash or dash + dot sequence. While this initial design still used a single key lever, his next version3 which appeared in 1955 was the first twin-lever electronic keyer and ancestor of the modern squeeze-keyers which we use today, and it is this key’s action that gave the "ultimatic" mode its name. The circuit is based on 11 tubes and only one relay and the time-base, memory and sequencer are functionally identical to the first version. But because contrary to a single-lever both contacts of a twin-lever key can be closed at the same time, a seizure circuitry was added: whenever a lever makes contact, it seizes control and the subsequent elements correspond to that lever until the other lever makes contact or the lever is released.
While one lever contact is closed the twinlever Ultimatic generates a string of dot- or dash elements,
exactly like any single-lever keyer does. However, when the levers are squeezed so that both contacts are closed, it generates a string of elements from whichever lever was pressed last. So any closure of a lever contact guarantees at least one element of that type, generated in correct relationship to the order of closure.  This key can be attacked as if it were a semi-automatic (bug) key or a single-lever electronic keyer or with any intermediate technique. And at that time it was considered the "ultimate" key because it sent perfect code without the need for the operator to send it perfectly, or in the words of W6SRY "a key that gives Klein output with Lake Erie input. It does everything for the operator
but spell and punctuate" (alluding to the characteristic "Lake Erie swing" of some bug operators).

The vast majority of today's electronic twinlever Morse code keyers operates in iambic mode, derived from the iambus which is a metrical foot in poetry with alternating short and long syllables like "dah-di-dah-di-dah". In 1967 Harry Gensler Jr., K8OCO, described5 his Iambimatic keying concept together with an adapter for the Hallicrafters HA-1 single-lever keyer in QST magazine. Pressing one
lever generates a string of dot- or dash-elements only, exactly as in ultimatic mode - but contrary to that, squeezing both levers generates a string of alternating dot- and dash-elements with the  commencing element corresponding to the lever which was hit first. So basic iambic keying with self-completing dots and dashes can be generated by executing this simple set of instructions: poll both levers alternately, if the lever is pressed generate the corresponding element and continue polling. The iambic mode is most effective for characters with alternating elements. All characters of the alphabet, except for the "P" and "X", and all digits can be generated with less than three strokes. However, only the "C" needs less strokes than in ultimatic mode.

the Curtis-keyer (came to be called Mode A)

The first iambic keyer EK-38 by John Curtis, K6KU, which appeared on the market in 1969, already
extended that basic iambic logic by a dot-memory. As we already know, this feature was originally developed by W6SRY, but his very ambitious Ultimatic key did not gain too much popularity. Then it was reinvented by Dave Muir, W2YVO, who recognized the problem of dropping single embedded or final dots e.g. in letters like "K" or "G" because the operator is too quick (continuously running time-bases were outdated and hence dropping of leading elements was no more a problem) and who filled the gap between simple circuits and the Ultimatic with his Penultimatic single-lever electronic keyer described6 1962 in QST. And because it is more likely to press and release the short dot too early during the long dash than the long dash during the short dot, the first Curtis-keyer had a dot-memory only but no dash-memory exactly like the single-lever keyer by W2YVO. In 1973 John Curtis brought out the 8043 CMOS chip, the first integrated-circuit iambic keyer with dot-memory.

the Accu-keyer (came to be called Mode B)

The Accu-keyer by James Garrett, WB4VVF, featuring dot- and dash-memory as well as automatic
character spacing, was published7 in QST magazine shortly after John Curtis' 8043 chip appeared. The behaviour of the Accu-keyer can be described by the basic iambic set of instructions together with the following dot/dash-memory rule: if anytime during generation of an element the alternate lever was pressed, generate an extra alternate element. So neglecting the fact that the Accu-keyer has both dot and dash-memory, the only procedural difference is that it just remembers a state "pressed" of both levers while the Curtis-keyer remembers a change of state or a transient "from unpressed to pressed" of the dotlever during the generation of an opposite element, and if that happened both keyers generate an extra alternate element.

iambic type "A" and "B"

In 1975 the 8044 chip was introduced by John Curtis, an improved version of the 8043 with dot- and
dash-memory. At that time most telegraphy operators already used iambic keyers - but scarcely anybody in basic iambic mode without dot/dash-memory, because neither the Curtis-keyer nor the Accu-keyer allowed to disable that feature. So over the years two schools of iambic keying developed, differing only in the dot/dash-memory logic which the operators initially learned but rarely scrutinized or even changed: Curtis keyer and Accu-keyer. In light of that fact, Curtis named his own logic iambic type "A" and that of the Accu-keyer iambic type "B" and in 1986 he introduced
the 8044ABM chip which offered selectable "A" or "B" type of iambic keying.


1. Jack W. Herbstreit, W4JNX: "Automatic Spacing of Letters and Words for the Electronic Key", QST, April 1951, p. 46
2. John Kaye, W6SRY: "The 'Ultimatic' - The Key with a Memory", QST, February 1953, p. 11
3. John Kaye, W6SRY: "The All-Electronic 'Ultimatic' Keyer", QST, April 1955, p. 11
4. Alvin F. Kanada, K0MHU: "The 'Ultimatic' - Transistorized", QST, September 1920, p. 27
5. Harry Gensler Jr., K8OCO: "The 'Iambimatic' Concept", QST, January 1967, p. 18
6. Dave Muir, W2VYO: "The Penultimate Electronic Key", QST, March 1962, p. 15

Please leave a comment regarding you experience with squeeze-keying.

That's all for now.

So lower your power and raise your expectations...

Richard, AA4OO


  1. Good write up...shared with the QRP group

  2. It must have been about 1967 when I built an Iambimatic keyer based on a design in QST magazine. It used a couple of small relays intended for R/C aircraft and some general purpose TO5 can GE_ transistors which were relatively expensive at roughly $1 each if I remember correctly. (Type designations were GE4, GE5; something like that.......I do not remember the digits.) I hand wired this on phenolic terminal strips on a small aluminum chassis. I bought a proper set of dual paddles. This all worked properly, except when I attempted to actually key the transmitter. It seems that too much RF got back into the keyer circuit and caused it to "freeze". I would be curious to find that original QST article which I followed. I quit amateur radio (and my subscription to QST magazine) in the 1970's when I went to college and became an EE.