Erwin, There are no DSP filters in CW mode. We generate sine (I) and cosine (Q) wave signals in the DSP and send them without filtering. You still have to create buffers for the signals. As I recall the TX buffer size is actually ignored in CW mode but I would have to confirm with engineering. I can certainly confirm that there is no filtering needed or applied in that mode. There is however wave shaping.
73, Gerald Gerald Youngblood, K5SDR President and CEO FlexRadio Systems(TM) 13091 Pond Springs Road, #250 Austin, TX 78729 Phone: 512-535-4713 Ext. 202 Email: [email protected] Web: www.flexradio.com <http://www.flex-radio.com/> Tune In Excitement (TM) PowerSDR(TM) is a trademark of FlexRadio Systems On Wed, Nov 24, 2010 at 2:33 AM, Erwin van den Bosch (PA7N) <[email protected]>wrote: > Just a thought: > > When I install 2.0.16 all DSP buffer sizes are 2048. For SSB and Digi modes > I can understand that because the filtering is better with A large buffer > size instead of A small buffer size. > > But when transmitting CW I think the DSP transmit buffer (not the USB/audio > buffer) can be zero because you don't need filtering on the CW carrier. Just > generate a nice sinewave (via calculation or a simple lookup table). By > doing this without DSP buffers the latency may be a little less when > transmitting CW. > > Regards, > Erwin > > On 23-11-2010 18:56, Gerald Youngblood wrote: > >> David, >> >> CW is the primary focus of the current software work here since we >> released >> 2.0.16. We have done a "clean sheet of paper" rewrite of the entire CW >> system. That means from the key input to the tone output. It is a >> totally >> new architecture. The basic elements are functional and getting good >> feedback from some top level CW operators. This will be released when it >> is >> completed and tested by our internal beta team. >> >> Regarding your specific issue related to external keyers truncating the >> first dit, we are doing some investigation here to see if we can recreate >> your problem. Nearly all reports from customers using the internal keyer >> in >> 2.0.16 with the FLEX-1500 have been positive. >> >> Regards, >> Gerald >> >> >> Gerald Youngblood, K5SDR >> President and CEO >> FlexRadio Systems(TM) >> 13091 Pond Springs Road, #250 >> Austin, TX 78729 >> Phone: 512-535-4713 Ext. 202 >> Email: [email protected] >> Web: www.flexradio.com<http://www.flex-radio.com/> >> >> Tune In Excitement (TM) >> >> PowerSDR(TM) is a trademark of FlexRadio Systems >> >> >> On Sun, Nov 21, 2010 at 6:20 PM, David Feldman<[email protected]> wrote: >> >> >> >>> Thanks, Tim, for the quick reply: >>> >>> Here is the result of the recommended changes: >>> >>> >>> >>>> 1.) Look in the Setup->General->Options tab and make >>>> sure that the Enable TX Out Delay check box is not checked. >>>> >>>> >>> It was checked, and I unchecked it. >>> >>> >>> >>>> 2,) Use the smallest audio buffers that do not distort the >>>> audio (Setup->Audio->Primary). I would start >>>> with 512. >>>> >>>> >>> It seems to receive fine even at 128, so I used that value. >>> >>> >>> >>>> 3.) Use smaller CW RX/TX DSP buffers >>>> (Setup->DSP->Options) >>>> >>>> >>> I set to 512. >>> >>> >>> >>>> 4.) Change the CW Blanking Delay value >>>> (Setup->Transmit) >>>> >>>> >>> The minimum value appears to be 10; I tried increasing the value, doesn't >>> appear to address my specific type of issue (by the way, I did notice >>> some >>> noise on rx/tx transition as well as in the transmitted envelope, along >>> with >>> the still unidentified off-frequency low-level carrier (a few hundred Hz >>> away) in the transmit signal, but those are not my primary focus, as they >>> don't fundamentally prevent usability in CW; the "hesitation" as others >>> have >>> called it, is a fundamental usability problem - a show-stopper. >>> >>> In a nutshell, my test is to send a single "dit" from my external keyer >>> (set to about 15 WPM now, pretty slow) and see if that "dit" survives on >>> the >>> air. >>> >>> From what I can tell, there is enough delay from when the "dit" begins >>> at >>> the Flex-1500 key input jack until RF appears on the transmit path that >>> there is no way the first CW symbol could emerge unscathed. The first >>> symbol >>> is incomplete, and it sounds like it may also be damaged in terms of >>> waveform, but I've not tried using an oscilloscope to see. I haven't >>> tried >>> to measure it, but I'm guessing 50-150 mSec, but rather than delaying the >>> whole transmit stream, the whole setup isn't getting into transmit state >>> quickly, and instead the first symbol is dropped or grossly >>> foreshortened. >>> By the way, I've done this test at a variety of break-in delays >>> (generally >>> I'm using about 100 mSec, which should be a decent "semi-break-in" value >>> given that I've pretty much foresaken the prospect of getting QSK.) >>> >>> So, what to do - wait some more? Basically is the flex1500 I ordered >>> within >>> a few weeks of when flexradio first accepted orders, delivered a few >>> weeks >>> after flexradio began shipment, and sitting on the shelf awaiting the >>> integrated USB driver which ostensibly targets among other things CW >>> operation, just destined to be a hangar queen? I can accept that it's not >>> destined to be a QSK rig (toooooo bad, I really got used to that about >>> 1972 >>> when I got my ten-tec argonaut 505), but I can't accept that it's not >>> destined to be a CW rig at all, using either internal iambic or external >>> iambic keyer (I'm not going to limit myself to sending on a keyboard - if >>> that's the case this rig is headed for writeoff-land). Based on >>> off-reflector messages received so far since my post yesterday, my >>> experience is not unique. Certainly flex beta testing would include a few >>> host machines of different capability levels so that a 1.6 GHz Atom with >>> 2 >>> GB RAM running only XP and >>> Powersdr 2.0.16 could be quickly identified as insufficient, if that's >>> the >>> issue? >>> >>> So, back on the shelf awaiting update, unless there's any other >>> configuration change I can try. >>> >>> Thanks for the attempted assist, >>> >>> Dave >>> >>> --- On Sun, 11/21/10, Tim Ellison<[email protected]> wrote: >>> >>> >>> >>>> From: Tim Ellison<[email protected]> >>>> Subject: RE: [FlexEdge] First CW transmit/QSO attempt with> Flex 1500 >>>> To: "David Feldman"<[email protected]>, "[email protected]"< >>>> >>>> >>> [email protected]> >>> >>> >>>> Date: Sunday, November 21, 2010, 3:13 PM >>>> It does help. Junk from a >>>> previous DB could have been getting in the way, but that >>>> does not seem to be the issue. >>>> >>>> Here are a few things to check or fiddle with. >>>> >>>> 1.) Look in the Setup->General->Options tab and make >>>> sure that the Enable TX Out Delay check box is not checked. >>>> 2,) Use the smallest audio buffers that do not distort the >>>> audio (Setup->Audio->Primary). I would start >>>> with 512. >>>> 3.) Use smaller CW RX/TX DSP buffers >>>> (Setup->DSP->Options) >>>> 4.) Change the CW Blanking Delay value >>>> (Setup->Transmit) >>>> >>>> I am not a CW op so I do not want to venture a guess at >>>> what settings to try, but these are the things that you can >>>> change which can improve latency. >>>> >>>> -Tim >>>> >>>> >>>> -----Original Message----- >>>> From: [email protected] >>>> [mailto:[email protected]] >>>> On Behalf Of David Feldman >>>> Sent: Sunday, November 21, 2010 2:58 PM >>>> To: [email protected] >>>> Subject: Re: [FlexEdge] First CW transmit/QSO attempt with >>>> >>>> >>>>> Flex 1500 >>>>> >>>>> >>>> When I first started 2.0.16, I am pretty sure it created a >>>> new database, because I had uninstalled 2.0.5 prior to >>>> installing 2.0.16. In any event, I also tried resetting the >>>> database and restarting the program, and got the same >>>> results (the reset/restart left an XML file on the desktop >>>> which I didn't look at.) >>>> >>>> Hope this is helpful info, >>>> >>>> Dave >>>> >>>> >>>> >>>>> Message: 7 >>>>> Date: Sat, 20 Nov 2010 21:28:08 -0500 >>>>> From: "Tim (W4TME)"<[email protected]> >>>>> To: [email protected] >>>>> Subject: Re: [FlexEdge] First CW transmit/QSO attempt >>>>> >>>>> >>>> with Flex 1500 >>>> >>>> >>>>> and 2.0.16 beta on 1.6 GHz Atom XP box >>>>> Message-ID:<[email protected]> >>>>> Content-Type: text/plain; charset=windows-1252; >>>>> >>>>> >>>> format=flowed >>>> >>>> >>>>> Did you import the database from 2.0.5 into 2.0.16? >>>>> >>>>> -Tim >>>>> --- >>>>> W4TME >>>>> FlexRadio Systems Internet Systems Admin. >>>>> Product Verification Team >>>>> Tune In Excitement? >>>>> >>>>> >>>>> On 11/20/2010 6:54 PM, David Feldman wrote: >>>>> >>>>> >>>>>> This afternoon I had my first QSO with the Flex >>>>>> >>>>>> >>>>> 1500 >>>> >>>> >>>>> and PowerSDR 2.0.16 using an external keyer. I'm an >>>>> >>>>> >>>> absolute beginner >>>> >>>> >>>>> with any flexradio product - I bought the rig long >>>>> >>>>> >>>> ago, received it >>>> >>>> >>>>> when they began shipping, and put it on the shelf (as >>>>> >>>>> >>>> I'm exclusively >>>> >>>> >>>>> CW op) until 2.0.16 beta was released. So, today is my >>>>> >>>>> >>>> first on-air >>>> >>>> >>>>> experience with the rig. I was able to upgrade from >>>>> >>>>> >>>> 2.0.5 >>>> >>>> >>>>> (out-of-the-box CDROM) to 2.0.16 beta without any >>>>> >>>>> >>>> difficulty. >>>> >>>> >>>>> Where I'm struggling: I notice a pronounced >>>>>> >>>>>> >>>>> "hesitation" effect with the first transmit symbol, >>>>> >>>>> >>>> and also hear a >>>> >>>> >>>>> slightly off-frequency carrier (on another receiver in >>>>> >>>>> >>>> the shack) >>>> >>>> >>>>> during transmit - a kind of "backwave" but not on the >>>>> >>>>> >>>> exact TX >>>> >>>> >>>>> frequency. I don't think either of these is news to >>>>> >>>>> >>>> the developers or >>>> >>>> >>>>> reflector, but I just wanted to see if I should be >>>>> >>>>> >>>> making any >>>> >>>> >>>>> configuration/operational changes before putting rig >>>>> >>>>> >>>> back on shelf >>>> >>>> >>>>> awaiting later software version. I did try adjusting >>>>> >>>>> >>>> CW TX buffer size >>>> >>>> >>>>> between 256 and 2048 and didn't see much difference. I >>>>> >>>>> >>>> did find it >>>> >>>> >>>>> necessary to run keyer's own sidetone (with "Disable >>>>> >>>>> >>>> Monitor" selected >>>> >>>> >>>>> on powersdr) as powersdr sidetone is delayed enough vs >>>>> >>>>> >>>> actual sending, >>>> >>>> >>>>> and powersdr sidetone audio level is dependent on >>>>> >>>>> >>>> transmit drive level >>>> >>>> >>>>> (which means no powersdr sidetone when tx power turned >>>>> >>>>> >>>> down to 0 for >>>> >>>> >>>>> testing.) >>>>> >>>>> >>>>>> To work around the first-symbol hesitation, I was >>>>>> >>>>>> >>>>> able >>>> >>>> >>>>> to complete a QSO by setting the "break in delay" >>>>> >>>>> >>>> parameter to ~750 >>>> >>>> >>>>> mSec (so basically I could send a whole transmission >>>>> >>>>> >>>> without the rig >>>> >>>> >>>>> dropping back into receive.) I tried a simple test of >>>>> >>>>> >>>> sending a string >>>> >>>> >>>>> of dits at modest speed, and adjusted the Delay >>>>> >>>>> >>>> parameter down to 0 mS >>>> >>>> >>>>> (QSK hoped for) and discovered the shack receiver was >>>>> >>>>> >>>> not hearing any >>>> >>>> >>>>> useful signal (certainly not a string of dits). Until >>>>> >>>>> >>>> the delay was >>>> >>>> >>>>> set long enough to keep the rig in transmit between >>>>> >>>>> >>>> dits, I could not >>>> >>>> >>>>> obtain a workable transmit signal. >>>>> >>>>> >>>>>> The machine is an ATOM 1.6 GHz box with XP and >>>>>> >>>>>> >>>>> is >>>> >>>> >>>>> doing nothing but running powersdr. CPU utilization >>>>> >>>>> >>>> seems to be >>>> >>>> >>>>> 50-60%. >>>>> >>>>> >>>>>> I looked at the flexedge archives and 2.0.16 >>>>>> >>>>>> >>>>> release >>>> >>>> >>>>> notes and didn't see any advice that seems directed at >>>>> >>>>> >>>> the >>>> >>>> >>>>> first-symbol hesitation which is my main concern. >>>>> >>>>> >>>>>> Thanks for any advice, >>>>>> >>>>>> 73 Dave WB0GAZ >>>>>> >>>>>> >>>>> >>>>> >>>> >>>> >>>> >>>> _______________________________________________ >>>> Flexedge mailing list >>>> [email protected] >>>> http://mail.flex-radio.biz/mailman/listinfo/flexedge_flex-radio.biz >>>> This is the FlexRadio Systems e-mail Reflector called >>>> FlexEdge. It is used for posting topics related to SDR >>>> software development and experimentalist who are using beta >>>> versions of the software. >>>> >>>> >>>> >>> >>> >>> >>> _______________________________________________ >>> Flexedge mailing list >>> [email protected] >>> http://mail.flex-radio.biz/mailman/listinfo/flexedge_flex-radio.biz >>> This is the FlexRadio Systems e-mail Reflector called FlexEdge. It is >>> used >>> for posting topics related to SDR software development and >>> experimentalist >>> who are using beta versions of the software. >>> >>> >>> >> _______________________________________________ >> Flexedge mailing list >> [email protected] >> http://mail.flex-radio.biz/mailman/listinfo/flexedge_flex-radio.biz >> This is the FlexRadio Systems e-mail Reflector called FlexEdge. It is >> used for posting topics related to SDR software development and >> experimentalist who are using beta versions of the software. >> >> > > > _______________________________________________ > Flexedge mailing list > [email protected] > http://mail.flex-radio.biz/mailman/listinfo/flexedge_flex-radio.biz > This is the FlexRadio Systems e-mail Reflector called FlexEdge. It is used > for posting topics related to SDR software development and experimentalist > who are using beta versions of the software. > _______________________________________________ Flexedge mailing list [email protected] http://mail.flex-radio.biz/mailman/listinfo/flexedge_flex-radio.biz This is the FlexRadio Systems e-mail Reflector called FlexEdge. It is used for posting topics related to SDR software development and experimentalist who are using beta versions of the software.
