Omni 6+ CW Keying Improvement Discussion

In 1998-1999, an extremely informative investigation took place on the Ten-Tec reflector concerning the quality of the Omni 6+ cw transmit signal and cw sidetone. Out of this investigation came a few suggested mods as well as the introduction of the Inrad 2.8Khz filter as a substitute for the stock 2.4Khz filter. W9AC was the principal contributor, and I've quoted some of his key postings below:

9/6/98
Anyway, I continue to enjoy my Omni 6+ for real CW work, particularly since I made my keyed waveform and chirp mods back in June, and eliminated the slight QSK "thump" altogether. The last mod makes a tremendous improvement and requires a non-invasive mod to only one wire on the I.F./A.F. Board. Everyone should at least try this since it can be restored simply with the push of a connector.

By now, I thought the W1AW 20-meter signal would have improved. While listening tonight, the leading edge chirp is clearly evident. I don't mean to start another raging discussion on this, but it's evident that some Omnis manifest this slight chirp. Hint: the source can be traced to the 9 MHz @ 2.4 kHz crystal filter circuit. This is the only filter that shares the TX and RX function. Those radios with chirp are the same radios that exhibit a truncated trailing CW waveform, particularly when adjusted to the most "soft" setting. The new Omni 6+ SMD Control Board no longer allows you to make this adjustment. Instead, a fixed resistor is now located where a pot used to reside and guess what? The fixed value corresponds to what used to be the most "hard" setting. Here's another hint for Ten-Tec: The source of the chirp is what also causes the USB receive audio to sound a bit thinner in low-end response and less "full" than in LSB. Has anyone else noticed this? The good news is that the chirp is correctable. Anyway, compare the 20-meter signal against the simultaneous 40-meter transmission.

9/7/98
The mod consists of unplugging connector 52 on the I.F./A.F. Board. That's it. Plug 52 should have only one wire attached to it. Those of you with more recent Omni Sixes should verify this as some of the circuit boards have been changed to SMD types. Once connector 52 is unplugged, use a nylon tie-wrap to keep it safely out of harm's way.

The theory behind the change: Plug 52 activates a 4066 CMOS quad bilateral switch which is configured as a sidetone wrap-around circuit. The audio click and thumps one hears is generated by the CMOS chip. These chips are used in the broadcast audio world and are known to cause the same effect when switching audio on complicated audio mixers and routers.

You will note after the change that you still have sidetone control and the ability to control its mix relative to incoming received audio through the front panel's "ST" button.

Use a set of headphones to listen to your QSK audio before the mod and make a mental note of what you're hearing. Then, using your headphones, critically listen to the QSK after the mod. There should be a marked difference.

9/20/98
Yes, the 20-meter W1AW signal does not do justice to the Omni Six/Six Plus, and I do not believe it is representative of most Omnis.

As you know the chirp is correctable, but a considerable amount of effort is required to get there. Most of my notes are back at the office, but here's a brief snapshot of what's involved with a late-generation Omni Six Plus. Older units like my Omni Six, require much more work. Ten-Tec took the Omni Six Plus in the right direction, they just didn't finish the job. BTW, do not confuse this mod with another chirp abnormality that has been discussed here recently. The issue there is noticed on receive and the two are not related.

1) A new crystal is required on the 9MHz BFO Board. The new frequency for the LSB crystal should be 9,000.500 kHz. I had mine made from JAN Crystals in Ft. Myers. The BFO board requires alignment after the mod. The crystal is now shifted slightly down in frequency instead of up. This type of VXO really hates to be pulled up. This accounts for, and addresses a partial manifestation of the chirp, step 2 below cures the rest.

2) The 9 MHz @ 2.4 kHz filter requires modification to widen its bandwidth. Several alternatives exist. In the stock Omni, the transmitter CW BFO frequency falls square on the filter skirt, exacerbating the chirp effect. (See additional notes below).

3) The waveform mod requires changing several R/C time constants on the 9 MHz Crystal Board. Working with SMD devices here is no fun, but the result is a beautiful, symmetrical CW waveform with absolutely no overshoot. Your amplifier will love your Omni.

4) The thump mod can be completed at any time. It requires the simple removal of one plug. That's it.

5) A 10K pot on the Control Board was once in standard production to allow the user to adjust waveform rise and decay time. In later Omnis, its gone and a fixed resistor is in its place. The good news is that the holes for the pot are still on the newer PC boards.

Incidentally, the older Omni V while using nearly the same CW generation scheme, does not exhibit the chirp problem. Here's why: The Omni V used a fixed BFO CW offset of 9,000.600 kHz. This frequency falls well within the 9 MHz @ 2.4 kHz filter passband, not on the edge of the skirt like the Omni Six/Six Plus. Here, the fixed offset is down at 9,000.400 kHz. It may not seem like 200 Hz is much of a difference, but when you get to the filter's edge, it makes all the difference in the world. (Ten-Tec take note of this). Ten-Tec HAD to use this fixed offset for the BFO. Here's why: The Omni's minimum software adjustable CW offset is 400 Hz. The Omni's software and synthesizer make up for the difference on a CW offset that's anything greater than 400 Hz. The OMNI V's CW offset was 600 Hz. No adjustment to the CW offset was capable. Period.

Another manifestation of the chirp can be found while observing the waveform on an oscilloscope. Those OMNIs that chirp on transmit also exhibit a slightly truncated trailing edge on the CW waveform. Interestingly, the new Omni Six Plus does not allow the user to make the waveform adjustment (at least on the new SMD Control Board I received from them back in June). Instead, a fixed resistor places the Omni at full "hard," where the visual effect of the chirp is not seen as on the adjustable Omnis…..only heard.

As you can see, it's a lot of work to make the rig sound gorgeous. For me, it was worth it. Others will undoubtedly disagree. I hope this helps.

10/7/98
Those of you who have an Omni Six/Six Plus that exhibit a slight transmitted CW chirp may be in for some good news. George, W2VJN at INRAD has agreed to manufacture a 2.8 kHz BW 9 MHz I.F. filter if we can demonstrate enough of a demand for it. This is a replacement filter for the stock 2.4 kHz BW 9 MHz filter. This is the only filter in the Omni that shares the receive and transmit function. This same filter also exacerbates the slight chirp caused by frequency-shift-keying the Omni's 9 MHz BFO circuit board on some Omnis.

How can you tell if your Omni is transmitting with chirp? Yes, of course you listen for it. But in many cases is can be subtle. It's when you compare the CW note against another Ten-Tec when you notice the difference the most. Here's a test you can use: key your transmitter in the CW mode, sending a string of characters, listening carefully to the CW note on another receiver. It's best to use a wide SSB bandwidth for monitoring. Next, depress the Omni's FSK mode button. Again. key the transmitter by sending a string of characters, listening for any change. In FSK, the CW transmit circuit is identical, with the only exception being that the sidetone monitor is disabled, but more importantly….the transmit BFO frequency is HIGHER than it is in CW, placing the CW well above the lower skirt of the 2.4 kHz filter. If CW sounds cleaner to you in FSK than it does in CW, you've got the problem I'm discussing here.

I want to make this point clear: the problem does NOT manifest on all Omnis. Nevertheless, on those that are not problematic of a chirp, the CW BFO transmit offset frequency of 9,000.400 is running dangerously close to the 2.4 kHz filter's lower passband skirt. When the CW transmit BFO frequency reaches the filter skirt, chirp results and the transmitted CW waveform can become truncated on the trailing edge.

Recall that the Omni's CW transmitted BFO frequency is 9,000.400 kHz. The center frequency for this filter is mid-way between the LSB and USB transmitted BFO frequencies at 9,001.500 kHz. Therefore, the margin of lower skirt freedom in the designed Omni Six/Six Plus is only 100 Hz; the filter cuts off at 9,000.300 kHz. If the filter is not near perfection in manufacturing, the CW BFO frequency can indeed fall square on the filter's edge as it does in two of my Omni Sixes.

Other's have repeatedly asked me why the Omni V sounds so consistently good on CW transmit in comparison with some Omni Sixes. With the Omni V's CW circuitry being a close 99% match to that of the Omni Six, the answer again lies with this filter. Only in the Omni V, the fixed CW offset is 9,000.600 kHz, well out of harms way of the lower filter skirt. Here, the Omni V has a margin of freedom of 300 Hz, not 100 Hz as in the Omni Six series.

A 2.8 kHz BW, 9 MHz I.F. filter will provide for the same 300 Hz of safety to the lower filter skirt. While this corrects the problem on CW, how will it affect SSB operation? If the USB and LSB BFO frequencies remain the same, the audio response goes from 300Hz - 2.7 kHz to 100 Hz- 2.9 kHz. As a practical matter, this should not present a problem. Yes, the occupied bandwidth is slightly greater, but not so much as to cause QRM to QSOs near your operating frequency. As a side note, Kenwood once sold a filter with this bandwidth and the Yaesu FT-1000MP has a menu choice for expanded transit SSB audio response. Also, since the USB and LSB frequencies are adjustable, the audio passband characteristics could be shifted slightly if desired.

A minimum of five brave souls are needed for George to make this filter. Cost of the filter to us is $110. Shipping is 2 months ARO.

If you have an interest in this, please e-mail me directly and I will forward your commitment to George. Thanks.

1/3/99
In a nutshell, more or less, Ten-Tec's stock 2.4 kHz filter in the 9 MHz IF slot will excerbate a slight chirp eminating from the 9 MHz BFO circuit board. From my experiments earlier this summer, the degree of note degradation is a function of two factors: 1) Some Omni Sixes have a BFO which exhibits a slight degree of chirp, most don't. This can be confirmed by tuning an external receiver to 9 MHz and lightly coupling the BFO to the external receiver. 2) If the 2.4K filter has a sharp knee at the Omni Six's CW transmit BFO frequency of 9,000.400 kHz, the filter will not only magnify the effct of the chirp, but passing this frequency through the filter's sharp lower skirt can lead to a distorted CW waveform whose trailing edge becomes slightly truncated. I confirmed that the 2.4K filter was the sole contributor to this effect by augmenting the CW transmit BFO frequency slightly upward by 100 Hz, then downward by 100 Hz. On my particular Omni, a BFO frequency adjusted ever-so-slightly out of spec (or from drift) will cause noticable visual distortion of the waveform and audibly, this is when the note becomes dirty and soft. It bears pointing out that chirp is not the ONLY manifestation of poor 2.4K filter quality control.

In an effort to resolve the problem, I took drastic measures: I purchased a used Omni V. I then compared every stage of circuitry and after several weeks of non-stop research, I came to the conclusion that although the Omni V and Omni VI's CW and critical transmit path circuitry is nearly identical, the only difference could be attributed to the fact that the Omni V uses a fixed CW transmited BFO offset that is designed exactly 200 Hz higher than the Omni VI. At first glance, 200 hz seems insignificant. However, consider that the 2.4K filter with its 9,001.500 kHz center has a designed lower cutoff of 9,000.300 kHz. This leaves only 100 Hz of QC variability. This is why some Omni Sixes sound good on CW and some sound poor. The worst case, is like that of my own: my BFO circuit once chirped until I redesigned it with a switching scheme, and my stock 2.4 khz filter was very asymmetrical, placing the lower cutoff at 9,000.400 KHz. There's no magic here…no rocket science…just a rational reason, justified mathematically and confirmed with empirical testing.

Anyone wishing to confirm and duplicate my tests can easily do so by adjusting the CW transmit BFO trimmer. By monitoring with a station monitor or oscilloscope and frequency counter, adjust the BFO trimmer to produce a value exactly at the designed frequency of 9,000.400 kHz. Next, slowly decrease the value to 9,000.300 kHz and observe what happens on the scope. Now listen to the note on an external receiver. Another useful test is to compare the CW note while alternating between the CW and FSK modes. In FSK, the CW circuitry is identical, only the sidetone is muted. You know you need a better filter if FSK produces a more pleasing CW note than in CW. Why? Because in FSK the transmitted BFO is shifted slightly UPWARD, well out of harms way of the lower filter skirt.

The INRAD 2.8 kHz filter for the 9 MHz IF position was designed to provide for the same degree of passband freedom as that designed into the Omni V. The choice of 2.8 kHz was not arrived at recklessly. I wanted to duplicate the excellent CW characteristics of the Omni V. Becuase of the Omni Six's fixed CW transmit BFO of 9,000.400 kHz, nothing less than 2.8 kHz will do unless the filter is manufactured with exceptional QC standards. A side benefit of the INRAD filter is that the USB/LSB passband is much more predictable and symmetrical. There's an easy test to confirm filter asymmetry: if USB receive audio sounds slightly "thinner" than than that heard in LSB, you've got a defective 2.4 kHz filter.

(K0CQ comment)

The phase delay of a crystal filter changes rapidly at the knee. A little frequency shift could easily be magnified by that phase shift to sound like a wider chirp. It would seem to me to be best to make sure the CW signal was inside the filter a ways to get away from that. Either by selection of CW carrier frequency or filter frequency. Maybe it would be better to introduce a totally separate CW crystal oscillator that didn't need to go through the filter. I don't have any idea what that
would do to the switching, but ought to make it sound perfect. Though its not easy to key a crystal oscillator without either clicks or chirps. The crystal doesn't like stopping and starting… Probably better to key a mixer to get a better on/off ratio that just keying a straight gain stage with the oscillator running continouously.

1/4/99

Perhaps ALL Omni-VI can be fixed that way. E.g. crank in more >shift until they are clean. TRY IT!

Yes, technically this works, but if the CW transmit offset is not fixed at the design frequency of 9,000.400 kHz, the frequency display will not track accurately. The Omni Six is capable of agile CW offsets in the range of 400 Hz to 990 Hz. Ten-Tec's choice of 9,000.400 coincides with the 400 Hz lower agile offset. Other offset choices are then determined through microprocessor control. This is where the Omni V and Omni Six part ways. The Omni V's offset is fixed at 600Hz. The Omni Six must be capable of shifting from 400 Hz to 990 Hz. The convenience of this feature is offset by the problem of having to choose a lower limit fixed offset of 400 Hz; square on the knee of the 2.4 kHz filter. Additionally, when an offset OTHER than 400 Hz is chosen, the chirp effect can be magnified even more. Want proof? Try this: listen to the output of the Omni Six while operating SPLIT. To start, try an exaggerated offset of 100 kHz or so between VFOs A and B. Listen on A, transmit on B. Now, listen to the CW note on an external receiver. To further magnify the effect, try an offset of 200 kHz. True, one does not generally use offsets so far apart except under unusual circumstances. This is probably a worst-case scenario. Again, the Omni V outperforms the Omni Six when using large offsets. Why? The Omni V introduces more T/R delay to allow the synthesizer to settle. The Omni V is capable of very large offsets before any chirp is heard that is caused solely by the synthesizer's settling effect. Operating the Omni V with large offsets takes some getting used to: it throws off my CW timing because of the additional T/R delay. The Omni Six is not affected by this, but the end result is synthesizer-generated chirp from inadequate synthesizer settling time. Bottom line: It's a trade-off between synthesizer-generated chirp and additional T/R delay.

To summarize:

-The Omni Six's best CW performance occurs with a 400 Hz CW offset in the Simplex mode. Synthesizer-generated chirp (not caused by the BFO chirp), becomes more pronounced at large offsets and splits.

- Changing the CW BFO frequency will move the BFO carrier into a safe portion of the 2.4 kHz filter passband, but dial accuracy suffers by the Delta of the resulting offset from 9,000.400 kHz. There's no other reasonable way to compensate for this without rewriting the control software.

- Newer generation Omni Sixes may be more susceptible to the BFO-generated chirp. Ten-Tec redesigned the 9 MHz BFO board somewhere in production. The concept is great, only the choice of the LSB crystal should be moved from 9,000.000 kHz to 9,000.500 kHz, allowing for a 100 Hz negative CW BFO shift rather than a 400 Hz positive shift. I made this change, and BFO-generated chirp is now undetectable.

2/14/99
Several weeks ago, someone on the list indicated that they could not rid their Omni Six of audible "clicks" while listening with headphones. I suggested removing plug 52 on the AF/IF Circuit board, but apparently, this did not help.

After making yet another mod to my Omni Six last weekend, I noticed the same problem for the first time: I was hearing audible clicks, to the point of being annoying regardless of any front panel setting, including power and AF loudness.

This weekend, I jumped in to find the cause. On my particular Omni Six, several areas were contributing to the clicks. Through a combination of regrounding, using shielded audio cable, and rerouting cables, I have my Omni back to the point where no or clicks thumps are heard whatsoever. The following steps may be of help to those who wish to minimize clicks in their Omnis:

1) I'm a firm believer in removing plug 52 from the IF/AF circuit board.
Pull it, and tie-wrap it out of harms way.

2) Ground the heatsink tab of U8, the audio power amplifier. By pushing U8 against the side chassis, a hole can be drilled for placement of a #4 screw which now grounds and heat-sinks U8 to the chassis.

3) Many audio cables use ribbon cable. I replaced several of these with single-pair shielded cable to the AF pot, FIXED AUDIO Plug 57, and cut two audio circuit traces on the AF/IF Board that were running too close to other witching circuitry.

4) Cut the nylon tie-wrap that holds the cable bundle along side the chassis near U8, separate and twist the wires, finding the right combination that minimizes the clicks.

3/21/99
In order to completely eliminate the "micro-chirp," you may need to modify the TX AUDIO BOARD (BFO Osc.). The INRAD filter helps to clean up the dirty-sounding CW note due to the CW TX BFO placement on the edge of the 2.4 kHz filter's passband.

The BFO mods may require replacement of the TX AUDIO BOARD. Omnis older than approximately Jan. 1997, require the new board in addition to a new CONTROL BOARD (this may have coincided with the introduction of the Omni Six Plus). Finally, Y1 should be changed from 9,000.000 kHz to 9,000.500 kHz. If you, or anyone else is interested, a full description of the changes appear in the Ten-Tec archives.