Improving T/R relay transition latency through an alternative implementation (e.g. solid state switch, different relays, T/R isolation & receiver muting, etc.) is really the only way to practically change the existing SDR for better QSK performance (for those that need it). The board stack has TX & RX duplexed statically (hardwire) at the TRX board at the "RF side" of the two quad analog switches and that common path extends from T1 onward through the entire board stack. The LNA for RX, the TX Driver and the 1W PA for TX, on the RFE, RFE and BPF respectively are also all duplexed, but through relays. Therefore it requires improvement on the relay chain RFE K1 & K13, BPF K13 and PA K13 & K14 in order to make improvement on TX/RX transition latency. The slowest relay of any of the 5 sets the minimum transition time for the system.
Modifying the SDR to separate TX & RX chains would require a wholesale rework of the entire rf path from TRX through to PA and duplication (at a minimum) of the BPF board in order to realize such a function. In the end, it would probably be easier to add an additional soundcard, make some relatively minor code changes and add a second SDR hardware set sans PA and maybe BPF to actually implement a segregated TX/RX function. I know that Gerald has moved from the original extensible board kit to a finished HW business (the experimentation is in SW), but would he consider selling 'RX only" SDR units sans tuning knob and all the other included goodies as an add-on receiver to existing systems? This would allow for a SW modification to implement separate TX & RX by unit (via the USB interface - no parallel) and bring all the benefits of separate TX & RX antenna, full QSK, real sidetone monitoring, etc. Lee Pedlow NG6B San Diego, CA -----Original Message----- From: [EMAIL PROTECTED] [mailto:[EMAIL PROTECTED] On Behalf Of Charles Greene Sent: Wednesday, June 21, 2006 4:29 AM To: Lee A Crocker; [email protected] Subject: Re: [Flexradio] Improved CW Performance SDR 1000 Lee and All, Well, yes, but to QSK the SDR-1000 alone which is the first step cheaper diodes can be used; 1N4007s work well at 100 watts. Check the K2 schematic. This type of T/R switch can be modularized and connected in place of a T/R relay, generically speaking, that is. A small, fast reed relay with or without some switching diodes can also be used if you time the keying sequence so you don't hot switch the relay. For all practical purposes, you don't have to QSK over 60 WPM. I fail to see what has been done since the 70s can not be done now as a back fit item. There undoubtedly will be some challenges in physical layout. 73, Chas, W1CG At 09:55 AM 6/20/2006, Lee A Crocker wrote: >Pin diodes would be an ideal but expensive and complex way to handle >this. The transmit signal chain would be transmitter/hi power linear >amp/pin switch/antenna and the RX chain would be antenna/pin >switch/dual diversity switch or phaser/RX. You would need some kind of >electronic bias on the linear that would allow you to cut off the amp >when the drive falls below some given level like a few hundred mW, as >the amp would be on all the time. > >73 W9OY > >--- Charles Greene <[EMAIL PROTECTED]> wrote: > > > Lee and All, > > > > The "GOOD OLD DAYS" of separate receivers. > > transmitters and antennas > > is a real step backwards in time, whether it be with a sound card or > > not. I got rid of my KWS1/R388 in 1972 and replaced them with a Ten > > Tec 940 Triton IV which still is one of the best QSK rigs around. > > Aside from reed relays, how about switching diodes? I have a 15 > > year old QSK 1500 with switching diodes which QSK's most any 1500 > > watt linear with most QSK transceivers. Or is the separate R/T just > > an easy way to use the existing SDR-1000 hardware? Seems it would > > not be that difficult to modify. Replace the existing relay(s) with > > a module; whether it has a fast relay or diodes inside is > > immaterial, and make some circuit mods. > > > > 73, Chas W1CG > > > > At 09:03 AM 6/19/2006, Lee A Crocker wrote: > > >Ken, > > > > > >Thanks for pointing out the future and yes a shiny future it > > >appears it will be. > > > > > >It's true that Janus will be full duplex, but the SDR-1000 is not. > > >However I think with a little thought it could be. The point is > > >that this radio > > was > > >conceived on the model of a transceiver instead of > > a > > >transmitter/receiver, which was the original configuration for QSK > > >back in the old old days. > > The > > >transceiver of today (like an Orion) is actually pseudo QSK. It > > >just happens the relays are faster than most people can send, so > > >you can recover some > > of > > >the information between elements, at least enough information to > > >realize someone is sending so you > > can > > >stop sending. The reason we went to the > > transceivers > > >is cost. It was too expensive to build separate boxes. Now > > >however the separate boxes are > > primarily > > >software, and are easily linked and locked etc. so maybe it's time > > >to reconsider the "flexibility" of that design instead of binding > > >ourselves to a less flexible transceiver design. > > > > > >Real QSK had a separate TX and RX and 2 antennas, > > or > > >the old Johnson electronic switch which was > > designed > > >around a separate TX/RX, and it was much faster > > than > > >anything we have today. I have been told all the relays in the > > >flex are 3ms relays except for the clunky PA relays. A typical > > >vacuum relay like they use in an Alpha amp is around 10ms, and the > > >latency > > in > > >the radio is now 10ms, very very close to the holy grail. If you > > >could move away from the transceiver idea with its ring buffers and > > >all that and go > > strait > > >to duplex or 2 half duplex circuits you may be able > > to > > >build the best QSK/digital radio on the market at > > very > > >little extra cost since most computers have a > > second > > >sound card already (or wait for the promise of > > Janus). > > > So my post is more about engaging a discussion > > about > > >the merits of re-ordering the design philosophy > > rather > > >than ignorance of Janus. > > > > > >The other thing that moving to the > > >transmitter/receiver format would provide is the automatic > > >inclusion of independent antenna pathways for TX and RX. This is > > >one thing that is lacking > > in > > >the present design. My guess is you would have to > > add > > >a second bandpass board, and some means to couple > > and > > >uncouple the antennas so you could either use the > > TX > > >antenna on the RX or a separate RX antenna, or both > > TX > > >and RX at the same time on receive for spatial diversity. > > > > > >Since we are now in the sub 10ms realm in the software, the first > > >thing to consider is can we replace the clunky PA relays with > > >something like > > the > > >Orion uses? > > > > > >73 W9OY > > > > > >--- Ken N9VV <[EMAIL PROTECTED]> wrote: > > > > > > > Hi Lee, I believe the new "Janus" board is FULL DUPLEX both ways > > > > and that is part of it's appeal as a total > > replacement > > > > for all the sound > > > > cards! the performance of the new Janus with > > premium > > > > ADC chips is > > > > better than the best Lynx card that currently > > sells > > > > for over $1,000. > > > > We are on the threshold of some amazing > > developments > > > > :-) > > > > de ken n9vv > > > > > > > > > > > > > > > > Lee A Crocker wrote: > > > > > Today I ran some tests on the latency of the > > radio > > > > to > > > > > a "receive" station compared the timing to my > > > > Orion. > > > > > The lag has been dramatically reduced, and I > > now > > > > feel > > > > > the actual limiting factor is the relay in the > > PA > > > > > stage (just as Jim Lux always said it would > > be). > > > > Is > > > > > there a reason a faster (and quieter) relay is > > not > > > > > employed in the PA? Has anyone looked into > > faster > > > > > relays? I don't have a scope, but has anybody measured the > > > > > latency between key closure and > > > > character > > > > > with the new software, possibly comparing the > > new > > > > and > > > > > the old software and say a separate rig like > > the > > > > Orion > > > > > or another good QSK radio? It would be nice > > to > > > > see > > > > > the visual progression of the improvement and > > for > > > > Flex > > > > > it would be a good thing to have on the > > website to > > > > > show the improvement to prospective customers. > > > > > > > > > > Has anybody considered actually running the > > rig as > > > > a > > > > > transmitter/receiver using a second sound card > > to > > > > > generate the CW, and plugging in > > > > attenuation/isolation > > > > > in the RX during transmit? If you ran > > separate > > > > > soundcards it seems to me you could actually > > do > > > > better > > > > > than most QSK radios on the market. For this > > all > > > > you > > > > > would need is a way to disconnect the antenna > > from > > > > the > > > > > receiver during transmit and plug in some > > > > attenuation > > > > > into the DSP so your agc wouldn't go nuts and basically leave > > > > > the transmitter and PA > > connected > > > > all > > > > > the time. In other words in terms of signal > > flow > > > > we > > > > > would be stepping back in time to the original > > > > days of > > > > > QSK when we were using separate TX and RX > > instead > > > > of a > > > > > transceiver. A sound card capable of > > producing >=== message truncated === > > >__________________________________________________ >Do You Yahoo!? >Tired of spam? Yahoo! Mail has the best spam protection around >http://mail.yahoo.com _______________________________________________ FlexRadio mailing list [email protected] http://mail.flex-radio.biz/mailman/listinfo/flexradio_flex-radio.biz Archive Link: http://www.mail-archive.com/flexradio%40flex-radio.biz/ FlexRadio Homepage: http://www.flex-radio.com _______________________________________________ FlexRadio mailing list [email protected] http://mail.flex-radio.biz/mailman/listinfo/flexradio_flex-radio.biz Archive Link: http://www.mail-archive.com/flexradio%40flex-radio.biz/ FlexRadio Homepage: http://www.flex-radio.com

