I get random hiccups when using cw.

mike_w1bfn

The best I can come up with is that it seems like a timing problem in software Pauses occur between dots. Haven't been able to anticipate it. Unit goes back to proper operation until the next time. like what I have experienced with Interrupt problems when a key is being sampled what it sounds like is for example a V turns into e...e...e...t.

Ed, K0KC

Mike, I have experienced the same problem; it really negatively affects your cw keying. I am almost embarrassed to operate cw on my 6700 and that is a mode that I use often. I suspect an interrupt issue as well, but since I am not that familiar with the radio architecture and software, I am not qualified to speculate. I hope this gets fixed soon; I know Flex is working on CW performance (specifically QSK) prior to the release of v. 1.0.

W5XZ - dan

no problems here, using 'stomp to talk' foot switch...not even trying QSK or even 'semi-break in'...most of my other t/r relays are too slow, but i'm fixing them, one at a time...FWIW,,,W5XZ..

Ed, K0KC

My problem is not T/R switching as I typically run with a semi-break-in delay of 600 msec and these "interrupts" occur in the middle of a long character such as BT before the rig switches back to receive.

K2CM

I too seem to have a hiccup periodically on CW. It seems like I am missing dots. Occurs frequently. Case in point, when operating the EU contest and sending traffic, frequently I miss dots requiring me to start over.

Ned K1NJ

I'm using an external keyer with a 6500. I've spent some time with a second receiver and a dummy load to listen to keying latency and quality. I haven't seen that problem here. Latency is undetectable by ear and the keying and note are clean. I've only used the radio on-air a few times, but it seems to key just as well into an antenna.

Ed, K0KC

Ned, The problem is not really latency in the sense that we experience with PowerSDR when the audio buffers are too large. If it is latency, it is intra-character latency. For example, if I try to send dah dit dit dit dah dit dah (break symbol), the 6700 will "sometimes" transmit dah dit dit dit dah dit pause dah. Except for this issue, the keying does sound good.

Al_NN4ZZ

Ned, It would be interesting to know if the problem is present in one or all of these scenarios below. (if you can reproduce it) - hold the dots paddle firmly closed and listen for the pause - hold the dashes paddle firmly closed and listen for the pause - hold both paddles firmly closed and listen for the pause - only happens when you randomly use each paddle (i.e. sending code normally) Also, is it worse at higher speeds than lower speeds or the same? I've seen a problem before related to the last scenario but will wait to hear back. And test into a dummy load, I'm sure you would, but just saying so no one thinks I'm advocating QRM. Regards, Al / NN4ZZ

Ed, K0KC

Ok, Al, to put your mind at ease, I turned the output power down to zero before I tried your suggested tests! I definitely hear the pause when I send dots, especially at high speed. I could not detect a pause when I send dashes at any speed. My key is not set-up for "squeeze keying", so I could not do that test. I do hear it when I send random code and it seems to occur more often when I have a dash following a dot in one of the longer characters. The extra pause occurs about once a minute or so during random code. I am definitely not a high-speed CW operator...usually send about 18 wpm.

Al_NN4ZZ

Ed, Ok, good info. I don't think it is the same weird problem I found with one of my paddles but will mention it anyway in case you want to try swapping to a different one as a test. Here is a description of the problem.... The dot side resistance would vary over time. The actual contacts were fine but the ground connection was through a bearing and the resistance varied between 5-100 ohms. We suspect it was due to the low amount of current passing through the bearing on this particular radio. Other paddles on the same radio worked fine. And this problem paddle worked fine on other radios. The solution was to bypass the bearing with a short flexible braid and the problem went away. With a solid ground the resistance at the plug measured about 1 ohm. If no one else is having the problem it may be paddle related so it might be worth a try to make sure it's not a bad connection somewhere along the line. (contacts, ground, cable or plug). Good luck troubleshooting and let us know what you find. Regards, Al / NN4ZZ

Al_NN4ZZ

Hope this reply above formatted okay, I noticed when I used a "less than" symbol it seemed to scramble the text.....I changed it to "about 1 ohm" and it seems okay now.....

Reggie-k6xr

Using a Winkeyer USB here and cw appears to be pretty perfect. I don't do Qsk.

Ned K1NJ

Conditions: 6500 Delay 360 Iambic OFF 1 watt into dummy load External Keyer Mono plug into rig 25 wpm Tested sequences: All dots OK All dashes OK Alternating OK BT (many) OK Tested for several minutes repeating sequences. *Perhaps this is an internal keyer problem.* When I first got the rig, I was looking for latency and observing the keying was incidental. Not having found problems, I made no comment. Still looks good today. Ned

Al_NN4ZZ

Ned, good test info. Based on the other feedback it didn't sound like a paddle/cabling problem. The good news is that the next release has some CW fixes so maybe the problem is solved. That would be good, being a CW only OP, a good solid internal keyer is important but as you noted there is a work around. Regards, Al

Tim - W4TME

I am investigating this behavior report. Thanks for all of the good feedback.

Tim - W4TME

I have a question. I presume you are hearing the the interrupt issues with CW using the internal keyer via the side tone. Has anyone verified that theexact same anomaly heard as transmitted RF? I need this info to continue debugging....

Ed, K0KC

I have not tried, but I will do so this afternoon.

Ed, K0KC

Tim, Ok, I listened to the RF with my IC-7700 and could not detect the "interrupt" problem which suggests an issue with the internally-generated side tone. This delay does impact one's CW sending even though it apparently is not going out on the air.

Tim - W4TME

OK. This indicates the issue is with the codec. We have made a lot of improvements in the codec code that will be available with the next version of SmartSDR. I think it will resolve the issues you are hearing.

mike_w1bfn

Tim:as hapens a lot, once I have put a problem in writing,it becomes easier to think about: I can make the problem happen any time I want now:: I have a single paddle keyer, so there is a finite time between when one contact goes off and before the other makes contact. All I have to do is get a little bit sloppy(tired, etc) and let the Time when both contacts are off get just a Little longer, and the keyer thinks it is at end of word. So it inserts a space. Mike_w1bfn

Ed, K0KC

That "might" be what is happening with me, although I take great care in trying to send good CW. I have not been doing much CW over the past couple of years as I became enamored with the JT65 and JT9 modes. Once I knew that the 6700 would be coming soon (and that it would not do DAX right away), I started making several CW contacts each day. It seemed like l had removed all of the "rust", but who knows (maybe it is partly old age)! On the other hand, as I mentioned in a previous post, I did notice an occasional "glitch" when I sent a long string of dots a couple of days ago.

Eric-KE5DTO

This may also relate to the muting code which will need to be scrubbed for QSK.

Ed, K0KC

CW sending performance seems to be much improved with 0.13.10, although I have not had the opportunity to make very many contacts. CW seems to have a slightly different "feel" to it...not bad, just a bit different.

Mickey N4MB

CW (paddle key down to xmit) latency seems higher with the new code (0.13.10) Previous version was similar to KX3 and Icom transceivers, this version is considerably different. I can set it up on a scope and send you traces if you'd like to see it. Keyer set at 22WPM, usually. I suspect now that it is more on the transition between a dah and dit that is a bit too long. If you hold the key to the dit side to send a long string of dits, it skips one after about 30. I know, fix one thing, break something else... Whack-a-mole. 73, Mickey

Al_NN4ZZ

Mickey, It would be interesting to see a trace like the example below showing the latency. Regards, Al / NN4ZZ

Tim - W4TME

I can not reproduce it. Are you using a FLEX-6500?

Mickey N4MB

Yes. Drops bits periodically, like it is a blocked process or something. I've been streaming bits with the Keyer watching the output on a scope and it drops a complete bit, so something in the Keyer code? It gave an unhandled exception today during a long dit stream (15s?) Loaned my external Keyer out, so I can't produce Al's suggestion to look at anything but "first bit" latency... Yet.

W5XZ - dan

I could see this behavior if the radio was connected to the local network, but not when it's connected directly to the shack pc.. FWIW...73

Al_NN4ZZ

Tim, Anything new on this issue? I received my 6700 this past week and in the past few days I've seen the "dropped dit" sporadically. I'm running 0.13.10 and it occurs every few minutes. I've been trying to watch the display to see if the dit is being transmitted but haven't been able to catch it. (Ed noted earlier that it's only the sidetone that is missing the dit). Also, has anyone had a chance to scope the latency? (not sure if it's related but would be good information) Are other CW users also still seeing this issue in the V0.13.10 ? Regards, Al / NN4ZZ .

Al_NN4ZZ

Dan, That may be a clue. It feels like an interrupt problem. In my case both the radio and 6700 are connected to a desktop switch. There could be some other network traffic affecting the timing. I wonder if the ethernet process is running on the same CPU as the sidetone process. Steve or one of the developers could comment on whether this is useful or not. But it might be helpful if others could report on whether they see a difference in the ethernet connection (direct vs a network switch). Regards, Al

W5XZ - dan

Al, have you tried cabling the rig directly to the pc yet? I could be 'all wet'...wouldn't be the first time...