Jerry, lets not forget about the oscillator for the Si5351a, be that 25MHz or 27Mhz, that is another potential source for birdies.
-Justin N2TOH?
|
Sadly the 8 output version of the Si5351 has no easy way to set a default frequency on boot up, as the 8 output version has better isolation between outputs, and could also have been used to drive the AVR chip.? along with enough additional outputs for a multitone signal generator to align the filters.?
-Justin N2TOH
|
WSJT-X 2.0.0: GA release
#ft8
Joe Taylor joe@... ?lists.sourceforge.net?
|
|
6:11 AM (3 hours ago)
|
The WSJT Development Group is pleased to announce the general availability (GA) release of WSJT-X Version 2.0.0. If you have been using version 1.9.1 or earlier, or one of the candidate releases v2.0.0-rc#, it's important to upgrade now.? The original protocols for FT8 and MSK144 are no longer supported.? With v1.9.1 and earlier you cannot communicate with WSJT-X 2.0 using these modes. We now recommend using WSJT-X 2.0 anywhere in the conventional FT8 and MSK144 sub-bands.? Everyone should upgrade to v2.0 by no later than January 1, 2019. NEW FEATURES IN WSJT-X 2.0 -------------------------- 1. Compound and nonstandard callsigns are automatically recognized and handled using new FT8 and MSK144 message formats. 2. The new FT8 protocol provides optimized message formats for North American VHF contests, European VHF contests, ARRL Field Day, and ARRL RTTY Roundup.? Similarly, the new MSK144 protocol provides optimized message formats for North American VHF and European VHF contests.? Full support is provided for "/R" and "/P" calls in the relevant contests. 3. The new protocols provide nearly equal (or better) sensitivity compared to the old ones, and lower false decode rates. 4. New logging features are provided for contesting and for "Fox" (DXpedition) mode.? Logging is optionally integrated with N1MM Logger+ and WriteLog. 5. Color highlighting of decoded messages provides worked-before status for callsigns, grid locators, DXCC entities, continents, CQ Zones, and ITU zones on a ¡°by band¡± and ¡°by mode¡± basis, and for stations that have uploaded their logs to Logbook of the World (LoTW) within a specified time interval. 6. The WSPR decoder now achieves decodes down to S/N = -31 dB.? For the particular benefit of LF/MF users, an option "No own call decodes" has been added. 7. The UDP messages sent to companion programs have been expanded and improved. A more detailed list of program changes can be found in the cumulative Release Notes: Upgrading from earlier versions of WSJT-X should be seamless.? There is no need to uninstall a previous version or move any files. Please do not continue to use any release candidate -- that is, any beta release with "-rc#" in the version name. Links to installation packages for Windows, Linux, and Macintosh are available here: You can also download the packages from our SourceForge site: It may take a short time for the SourceForge site to be updated. We hope you will enjoy using WSJT-X Version 2.0.0. ? ? ? -- 73, Joe, K1JT, for the WSJT Development Group ______________________________
|
A) Hi Clark, yes I recall there was an issue with the full swing vs low power with the ATmega644p/ATmega1284p the oscillator was interfering with the serial port in full power mode, it made for interesting times trying to use a serial bootloader.
that said changing the fuse bits in the NANO is very easy with the ATMEL ISP programer, step 1) make a backup copy of the firmware, eeprom, and fuse settings. 2) update the fuse settings from their current state to place the device in low power clock mode. 3) verify firmware, eeprom, and new fuse settings.?
B) Jerry, did you run the numbers for other common catalog crystal frequencies? for example, would running the AVR at 11.052MHz and leaving the IF at 12MHz be a viable option???
-Justin N2TOH
P.S. Ashhar?as an option could a person order the Randino with the NANO uninstalled? some people including myself need the real FTDI USB chip, the other types cause unresolvable usability problems.? the drivers for the chip provided in my kit rendered my Macintosh unusable, I had to use a 3rd party keyboard and mouse to recover the settings.??
|
Hello everyone. Here is the picture of the uBITX I got just today. So from what I am understanding is that I do not need to replace the relays. I ordered them from Arrow anyways and have them on hand. I had read this article before ordering.??. So I am a little bit lost as to what to do before assembly since i want to leave it together for a while after the build. Any suggestions what should be fixed before putting it all together? As far as spurs go. ?   ? -- KI4KJH
|
Hello Andy,
Yes sir, my uBITX is now FCC compliant (though just barely on 40 meters, CW mode) except for one spur on the 10 meter band.?
Ripley The words are mine but this iPad does what it will with them.
toggle quoted message
Show quoted text
On Dec 9, 2018, at 3:43 PM, AndyH < ahecker@...> wrote: Hi Ripley, ? Thank you for the clarification!? My intent was to relate to Jerry that the bidirectional RF amps were still stock.? The amp immediately prior to the filter (that used to feed the mixer directly) has a trio of SMD 2N3904s marked Q20, Q21, and Q22 on the V3 schematic.? I didn't see that you'd replaced those with BFR106s yet, though might have missed plenty.? OM Farhan recommended changing R26 and R46 on the stock amps to help reduce distortion - something else for you to test.? hi hi?? /g/BITX20/message/61586? Is your signal in spec now except for 10 m SSB? ? ?Thanks again for? your work! ? ?Best, Andy On Sun, Dec 9, 2018 at 02:10 PM, Ripley wrote:
Hi Andy,
?
A slight clarification, actually I replaced Q92, Q93, Q96, Q97 with 2N2222A and the emitter resistors along with Q90 and the change of resistors per Allison¡¯s recommendation First. Then checked for effect in evening out the power across the bands. Then I did the 45MHz crystal filter second. Once again checking the before and after results followed by the relay replacement of KT1-KT3. As with the other mods checking the before and after to see if it improved the performance.
?
I haven¡¯t been able to spend much time on the uBITX lately but the next thing that I want to try is replacing L5 and L7 with surface mount inductors per the experiment that Raj tried but I am also going to keep 45MHz crystal filter in circuit. I want to see if those two mods together will reduce those pesky spurs on 10 meter SSB further.
?
Ripley
?
Sent from for Windows 10
?
From: AndyH
Sent: Sunday, December 9, 2018 1:42 PM
To: [email protected]
Subject: Re: [BITX20] Ashhar Farhan #ubitx
?
A number of people are working to meet different thresholds of 'take care'-edness (two more times in print and it's a real word).? A few (probably in a nod to the phase noise wars) are pushing for a standard just shy of unobtanium.? Others are working to meet regulatory requirements.? Ripley, KD8UYQ, found that relays and a 45 MHz crystal/transformer/cap was quite effective.? That's on a board using the stock 3904s for Q20/Q21/Q22 and before Farhan's recent emitter current increase for Q21/Q22.
Thanks for pointing out Farhan's recent post about current production boards.? That's a curious statement and one that reverses two previous posts that make clear that the relays had been updated and that fixes are added to production on the fly.? I guess I'll "have" to order another one after the holidays and see what's different.
Interesting on needing to match the 45 MHz filters.? After enjoying the process of matching crystals for my in-progress scratch build, it might be interesting to check the 45 MHz filters I purchased from AliExpress and Mini-Kits.? If I recall, the point of the 2nd filter is to remove the "stuff not 45 MHz".? That might mean that a wider variant of the filter would be useful/usable there and might not need to be a narrow filter closely matched to Y1.? TIme to fire-up the Sweeperino, maybe.? Since matched pairs are available, though...
73, Andy, KG5RKP
On Wed, Dec 5, 2018 at 10:34 PM, Jerry Gaffke wrote:
I don't see a post yet that clearly spells out how to take care of the spurs.
This should include pointers to specific parts, show how they get mounted to the board,
include fixes for both the 45mhz-DialFreq and the 12mhz spurs.
And not be spread across dozens of different posts over the past couple months.
|
Lee:
Did you get anywhere with this? ?I looked into parts but it would be dealing with 3-4 different vendors.?
toggle quoted message
Show quoted text
On Dec 4, 2018, at 1:21 PM, Lee < mr.olson@...> wrote: Any chance the the QRP Guys would like to put together a bag of parts in the BOM and sell them separately to be added on to the order of a JackAl board? Hint-Hint. -- Lee - N9LO? "I Void Warranties"
?
|
Hi Ripley, ? Thank you for the clarification!? My intent was to relate to Jerry that the bidirectional RF amps were still stock.? The amp immediately prior to the filter (that used to feed the mixer directly) has a trio of SMD 2N3904s marked Q20, Q21, and Q22 on the V3 schematic.? I didn't see that you'd replaced those with BFR106s yet, though might have missed plenty.? OM Farhan recommended changing R26 and R46 on the stock amps to help reduce distortion - something else for you to test.? hi hi?? /g/BITX20/message/61586? Is your signal in spec now except for 10 m SSB? ? ?Thanks again for? your work! ? ?Best, Andy
On Sun, Dec 9, 2018 at 02:10 PM, Ripley wrote:
Hi Andy,
?
A slight clarification, actually I replaced Q92, Q93, Q96, Q97 with 2N2222A and the emitter resistors along with Q90 and the change of resistors per Allison¡¯s recommendation First. Then checked for effect in evening out the power across the bands. Then I did the 45MHz crystal filter second. Once again checking the before and after results followed by the relay replacement of KT1-KT3. As with the other mods checking the before and after to see if it improved the performance.
?
I haven¡¯t been able to spend much time on the uBITX lately but the next thing that I want to try is replacing L5 and L7 with surface mount inductors per the experiment that Raj tried but I am also going to keep 45MHz crystal filter in circuit. I want to see if those two mods together will reduce those pesky spurs on 10 meter SSB further.
?
Ripley
?
Sent from for Windows 10
?
From: AndyH
Sent: Sunday, December 9, 2018 1:42 PM
To: [email protected]
Subject: Re: [BITX20] Ashhar Farhan #ubitx
?
A number of people are working to meet different thresholds of 'take care'-edness (two more times in print and it's a real word).? A few (probably in a nod to the phase noise wars) are pushing for a standard just shy of unobtanium.? Others are working to meet regulatory requirements.? Ripley, KD8UYQ, found that relays and a 45 MHz crystal/transformer/cap was quite effective.? That's on a board using the stock 3904s for Q20/Q21/Q22 and before Farhan's recent emitter current increase for Q21/Q22.
Thanks for pointing out Farhan's recent post about current production boards.? That's a curious statement and one that reverses two previous posts that make clear that the relays had been updated and that fixes are added to production on the fly.? I guess I'll "have" to order another one after the holidays and see what's different.
Interesting on needing to match the 45 MHz filters.? After enjoying the process of matching crystals for my in-progress scratch build, it might be interesting to check the 45 MHz filters I purchased from AliExpress and Mini-Kits.? If I recall, the point of the 2nd filter is to remove the "stuff not 45 MHz".? That might mean that a wider variant of the filter would be useful/usable there and might not need to be a narrow filter closely matched to Y1.? TIme to fire-up the Sweeperino, maybe.? Since matched pairs are available, though...
73, Andy, KG5RKP
On Wed, Dec 5, 2018 at 10:34 PM, Jerry Gaffke wrote:
I don't see a post yet that clearly spells out how to take care of the spurs.
This should include pointers to specific parts, show how they get mounted to the board,
include fixes for both the 45mhz-DialFreq and the 12mhz spurs.
And not be spread across dozens of different posts over the past couple months.
|
Hi Andy, ? A slight clarification, actually I replaced Q92, Q93, Q96, Q97 with 2N2222A and the emitter resistors along with Q90 and the change of resistors per Allison¡¯s recommendation First. Then checked for effect in evening out the power across the bands. Then I did the 45MHz crystal filter second. Once again checking the before and after results followed by the relay replacement of KT1-KT3. As with the other mods checking the before and after to see if it improved the performance. ? I haven¡¯t been able to spend much time on the uBITX lately but the next thing that I want to try is replacing L5 and L7 with surface mount inductors per the experiment that Raj tried but I am also going to keep 45MHz crystal filter in circuit. I want to see if those two mods together will reduce those pesky spurs on 10 meter SSB further. ? Ripley ? Sent from for Windows 10 ?
toggle quoted message
Show quoted text
From: AndyHSent: Sunday, December 9, 2018 1:42 PM To: [email protected]Subject: Re: [BITX20] Ashhar Farhan #ubitx ? A number of people are working to meet different thresholds of 'take care'-edness (two more times in print and it's a real word).? A few (probably in a nod to the phase noise wars) are pushing for a standard just shy of unobtanium.? Others are working to meet regulatory requirements.? Ripley, KD8UYQ, found that relays and a 45 MHz crystal/transformer/cap was quite effective.? That's on a board using the stock 3904s for Q20/Q21/Q22 and before Farhan's recent emitter current increase for Q21/Q22.
Thanks for pointing out Farhan's recent post about current production boards.? That's a curious statement and one that reverses two previous posts that make clear that the relays had been updated and that fixes are added to production on the fly.? I guess I'll "have" to order another one after the holidays and see what's different.
Interesting on needing to match the 45 MHz filters.? After enjoying the process of matching crystals for my in-progress scratch build, it might be interesting to check the 45 MHz filters I purchased from AliExpress and Mini-Kits.? If I recall, the point of the 2nd filter is to remove the "stuff not 45 MHz".? That might mean that a wider variant of the filter would be useful/usable there and might not need to be a narrow filter closely matched to Y1.? TIme to fire-up the Sweeperino, maybe.? Since matched pairs are available, though...
73, Andy, KG5RKP
On Wed, Dec 5, 2018 at 10:34 PM, Jerry Gaffke wrote: I don't see a post yet that clearly spells out how to take care of the spurs.
This should include pointers to specific parts, show how they get mounted to the board,
include fixes for both the 45mhz-DialFreq and the 12mhz spurs.
And not be spread across dozens of different posts over the past couple months.
|
Re: 2nd 45MHz filter - Power loss
TNX for the suggestions... Values are with RV1 adjusted to near maximum power.? Reach max before end of pot travel. FWIW, I have substituted the factory 45 MHz xtal filter with one of the two Tellurian units I have.? The other Tellurian is the second filter addition.? They said matched, but..?? Without major surgery, couldn't get a 15 pf across just the filter.? Tried across the whole xformer/filter assembly.? Not a significant change... Have most all the other spur / harmonic mods in place.? Trying to decide to leave it a very QRP TX, or put the 47 resistor back and get my 10+ watts back..... Woody ------------------------------ On 12/9/2018 18:33, ajparent1/KB1GMX wrote: Woody,
Yes but....? Did you try resetting RV1 to regain some of the power?
Also testing I did showed that for some filters the center frequency was
not nearly as offset as others as Mine and two other were closer to 45mhz than
stated and implemented in the Raduino firmware.? The result is adding a second
filter can have you down the band pass skirt and less power.
Try a small amount of C say 10 to 15pf across the filter input and output to "tune it".
Allison
_._,_._,_
------------------------------------------------------------------------
--
|
A number of people are working to meet different thresholds of 'take care'-edness (two more times in print and it's a real word).? A few (probably in a nod to the phase noise wars) are pushing for a standard just shy of unobtanium.? Others are working to meet regulatory requirements.? Ripley, KD8UYQ, found that relays and a 45 MHz crystal/transformer/cap was quite effective.? That's on a board using the stock 3904s for Q20/Q21/Q22 and before Farhan's recent emitter current increase for Q21/Q22.
Thanks for pointing out Farhan's recent post about current production boards.? That's a curious statement and one that reverses two previous posts that make clear that the relays had been updated and that fixes are added to production on the fly.? I guess I'll "have" to order another one after the holidays and see what's different.
Interesting on needing to match the 45 MHz filters.? After enjoying the process of matching crystals for my in-progress scratch build, it might be interesting to check the 45 MHz filters I purchased from AliExpress and Mini-Kits.? If I recall, the point of the 2nd filter is to remove the "stuff not 45 MHz".? That might mean that a wider variant of the filter would be useful/usable there and might not need to be a narrow filter closely matched to Y1.? TIme to fire-up the Sweeperino, maybe.? Since matched pairs are available, though...
73, Andy, KG5RKP
toggle quoted message
Show quoted text
On Wed, Dec 5, 2018 at 10:34 PM, Jerry Gaffke wrote:
I don't see a post yet that clearly spells out how to take care of the spurs.
This should include pointers to specific parts, show how they get mounted to the board,
include fixes for both the 45mhz-DialFreq and the 12mhz spurs.
And not be spread across dozens of different posts over the past couple months.
|
Re: LSB signal looking more like AM
joshuaclough4@
Start with carrier balance pot and adjust with NO MODULATION for minimum power out.
There should be a noticeable null at about mid rotation.
Likely the carrier frequency setting is off as well and inside the filter band pass.??
Allison
|
Re: 2nd 45MHz filter - Power loss
Woody,
Yes but....? Did you try resetting RV1 to regain some of the power?
Also testing I did showed that for some filters the center frequency was?
not nearly as offset as others as Mine and two other were closer to 45mhz than
stated and implemented in the Raduino firmware.? The result is adding a second
filter can have you down the band pass skirt and less power.
Try a small amount of C say 10 to 15pf across the filter input and output to "tune it".
Allison
|
2nd 45MHz filter - Power loss
With stock R27, Power output close to
most reported.? Added 45MHz xtal filter, no transformer.? Power
dropped significantly.? Added cap and transformer.? Power up
slightly, not much.? Interesting, the power drop below 10 MHz...
For those who have done this mod, how does it compare to your
result regarding power output?
Here is my result:
Thanks, Woody? -? KZ4AK
|
We believe the audio tones on some rigs come from harmonics of the 16mhz processor clock
beating with harmonics of the 12mhz BFO.? Most rigs don't have this, since the 12mhz
ceramic resonator is not very accurate, can be off by 100khz or so.?
That, and us old guys have lost the ability to hear anything much over 3khz.
Could drive the processor clock from the BFO, near 12mhz.
Should not create tones when it beats against itself.
The si5351 can glitchlessly transition that BFO by 3khz if we wish to switch from
low side to high side BFO (to switch between LSB and USB, keep clk1 on the high side.)
The 16mhz crystal resonator on the Nano is extremely small, will be tough for some to remove.
Perhaps leave it in place, overpower one end of it with a wire from the 12mhz BFO.
So this fix might be a matter of just adding one wire, and some new firmware on the Nano.
The ATMega328P has an internal processor clock generator that can be selected,
though it is even less stable than the ceramic resonator, and thus the birdies unpredictable.
Might be worth trying, as no hack is required except to the firmware.
Will be required to use the internal clock at startup until the si5351 is programmed
if the BFO is to be used as the processor clock.
Moving the crystal filter frequency to 11.052mhz should work, maybe that will show up on v5.
But matching 8 crystals for frequency, replacing the existing 12mhz crystals, evaluating
the filter passband given whatever crystal parameters you happen to have, and adjusting
caps and impedances for the desired bandwidth and filter ripple may be more than
most here want to take on.
If you have a good inspection microscope and no coffee, might be possible to swap
out the 16mhz ceramic resonator.? Any funny number for a frequency is likely to work
better than what we have.? We think we have trouble with 3*16 = 48mhz = 4*12.
If the processor were moved to 13mhz,? the first beat note I see is at 13*16= 208mhz = 16*13.
Those high harmonics will be very weak, won't get much better than that.
Remember to also avoid beats with clk1, at 45-12=33mhz and 45+12=57mhz.
Jerry, KE7ER
toggle quoted message
Show quoted text
On Sat, Dec 8, 2018 at 04:57 PM, <freefuel@...> wrote:
the Atmel AVR chip can be clocked up to 20MHz, with that said what frequency would be best if we were to replace the crystal on the Arduino board to shift that birdiey to another less offensive location.?
-Justin N2TOH
|
Echoing the congrats, but ... Timothy and others: please note how Farhan spells the other part of his name (given or family? his locals evidently call him Ash). That part is pronounced as he writes it: two syllables, Ash and har. I'm not sure if there's an accent or dictionary-style diacritical mark or two in there - the Indian Sanskrit-based script might describe the pronunciation more fully. English gives a general idea but is sometimes half-hearted.
...just a mark of courtesy and respect ...
73 de ZL2DEX
|
I¡¯ve seen some gear where the CPU clock frequency can be changed dynamically so any birdies from the CPU can by avoided. ?The AVR328 internal clock can be divided down from the crystal frequency under software control in powers of two from 1 to 256. ?Obviously if it¡¯s the crystal clock circuit itself that is causing birdies this won¡¯t help. ?The code would have to dynamically reset any timing based parameters (baud rate, timers, I2C clock and SPI clock) as well as synchronize the change (not make the change will data is gobbling transferred via any of the serial interfaces).
However there are settings in the clock oscillator which let you change from ¡°full swing¡± to ¡°low power¡±. ?If the fuses are currently set to use the full swing mode, the low power mode may mitigate the birdies.
toggle quoted message
Show quoted text
the Atmel AVR chip can be clocked up to 20MHz, with that said what frequency would be best if we were to replace the crystal on the Arduino board to shift that birdiey to another less offensive location.?
|
it would be easier to just buy 11.052 mhz crystals and shift the IF out.? - f
the Atmel AVR chip can be clocked up to 20MHz, with that said what frequency would be best if we were to replace the crystal on the Arduino board to shift that birdiey to another less offensive location.?
-Justin N2TOH
|
Read "Ashar"? Farhan - I can't edit the post and I am too lazy to repost it to correct the proper noun. TEF
|
This URL may be of interest.? ? The software is a bit flaky but it is usable.??
https://in-the-sky.org/satmap_worldmap.php
ExseedSat after a quick perusal does not appear but? the data is updated from NORAD object data from time to time .. no doubt a? GIF of Ashan's face will? circumnavigate the globe on this site in the future.?
You can select? Amateur radio satellites under select satellites? - the default is mainly GPS satellites.? it is fascinating that as I watch Dec 9th Just after lunch there is only one GPS Sat over the continent? of Oz yet obviously GPS has not dropped out.? The other ones over the Indian ocean and the near Pacific? must be close enough to give the constellation of four? required for a GPS fix. The maps are of course live.
|
Tony Ray KI4KJH
Ripley
