I'm interested in doing a UBITX build. It costs about what I'd spend on a used transceiver, but I'd rather put one together.
I like the design and it can do more than CW. Has anyone here had any experience with it?
I'm not shilling for the people who manufacture it, but if anyone is interested in specs and would care to comment, they can be found here:
https://www.hfsignals.com
I've been watching the uBitX for a while and listening to Soldersmoke Podcast with Bill Meara and Pete Juliano, they're die hard uBitX fans, worth a listen for much more than just the uBitX though.
One thing though, I think the uBitX is a ready assembled board so it's not so much a build as an assembly isn't it?
I'm interested in doing a UBITX build. It costs about what I'd spend on a used transceiver, but I'd rather put one together.
I like the design and it can do more than CW. Has anyone here had any experience with it?
I'm not shilling for the people who manufacture it, but if anyone is interested in specs and would care to comment, they can be found here:
https://www.hfsignals.com
I have one from india if that's the same one. Not really much to put together you just have to put it in a box connect your own mic and speaker and build a power supply for it.
Works over short distances but the low watts gets frustrating when people keep saying "I thought I heard someone, go ahead? No one? Never mind." Fun to listen to or if you have a friend you can use it with.
Sounds like you need a license and to understand how to use it. I get 3000km+ on less watts than that puts out.
Built this one
CW kit a year ago. Replaced all caps and resistors with quality ones. Also replaced with real amidons. Worked best with the original transistors. Got 6.5W out of it. Tried with bd139
but got only 3W max (did my best with various T1/2 winding ratios).
Also used my own pic firmware and some small mods. Otherwise a standard ne612 stuff.
If you try an older 1980s Philips branded BD139 instead of a modern NXP or ST part it'll probably kick out the high end. The newer ones are crap. I've got a big bag of really old ones I bought off ebay
Rockmite does look like fun and you can't moan at the price
I got a bunch of BD139-16 (ST, code "JB633") with an intention to have spare finals while experimenting (that was the original goal of mine - to experiment and make mods with the kit). No fun with those new bd139s, however. Hardly got 3W. The original D882 worked great. The same with the other transistors in the TX chain. I replaced them with various similar types, especially the Q1 is the absolutely most critical one there. It only worked with the original 9018 fine. Big fun with the kit!
I have the BITX 40 and uBITX, they are both really fine QRP transceivers. In fact these are my only two pieces of gear right now, I sold my old Yeasu after I purchased the BITX 40 and haven't looked back. As others have stated, SSB voice at QRP power levels can be a challenge at this stage of the solar cycle. If you know CW or are into digital modes, then you will have no problem working the world. If SSB voice is you thing then you might want to consider buying or building an external PA, however I've been quite happy at QRP power levels and have had no lack of success with both voice and digital modes.
The design of the transceivers are really easy to understand and infinitely hackable/modable. One of the best features is the large and very active community of users/hackers/moders at
https://groups.io/g/BITX20Ralph
If you try an older 1980s Philips branded BD139 instead of a modern NXP or ST part it'll probably kick out the high end. The newer ones are crap. I've got a big bag of really old ones I bought off ebay
Rockmite does look like fun and you can't moan at the price
I don't really get the use of vintage parts in radios; it's not the 19th century and we should really be using LDMOS power transistors instead of something with not even a fT spec; the rqa0009 can output 5W all the way into VHF and costs less; for high powers I'd use a used gsm transistor, e.g. mrf9060m (search m9060m, m9060n, m59070n on aliexpress/taobao), ~$2 and up to 60W. I can provide more info about how to use these 900MHz transistors at HF and VHF if people are interested.
Fully aware of these. Three big problems though with such devices:
1. Difficult to mount compared to a TO39/TO220
2. Really not at all forgiving.
3. RF transistors come and go so often. By hitting the trailing curve we’re actually producing more repeatable designs. These designs tend to be shared and there’s nothing more annoying than finding out you can’t get some obscure RF LDMOS IC only 2 years down the line.
Not to mention they don’t actually change the outcome at all.
I’m actually playing with multiple MMBF170 based PA with 74hc driver at the moment. That pumps out 5W easy on 30m and doesn’t even get warm and appears not to be too fussy about mismatches. Also you can get 10 legit ON Semi ones for less than an aliexpress LDMOS transistor with free next day here in the UK
How many mmbf170s do you use to get 5W? each transistor can only dissipate 300mW (at 25*C) so I'm assuming you're doing class E/F? I'm curious about the design and would like some schematics
I'm still finding it hard to believe that a transistor not designed for RF can handle mismatch better, especially when operated in class E/F, whereas the RF LDMOS transistors are pretty much all spec'd for 1:10 vswr at all load angles. Unless you have a load-pull tester that covers 360 degrees of reflection most likely you just got lucky with the particular load angle.
I have a class F 100MHz amplifier for several watts using the rqa0009 and it can handle the output open circuited because I designed the matching network to give a sane impedance to the transistor for that case. I've tested several coax stub lengths at the output and some make the transistor heat up quite a bit, but I haven't managed to damage it yet.
Class E. Don’t have a schematic. Basically 4x MMBF170 in parallel driven off a 74hc14. Drain coil is a randomly wound FT37-43 followed by standard LC pi LPF. There’s a 52v Zener across the DS as well. Most significant harmonic is 40dB down. Also built one with the ubiquitous IRF640.
Note: this has some nasty key clicks. Still working on that.
Most of the mismatches I’m talking about is when some numpty (me) forgets to attach the load.
My experience is with VMOS sorry not LDMOS. Brian fart there.
I’m running mostly at 10 and 14Mhz. Things down that end are easier.
Sounds like you need a license and to understand how to use it. I get 3000km+ on less watts than that puts out.
I do. This area people use 100+ watts and don't try do qrp stuff unless you set it up with that intent. You get keyed over unless the people on the other end have their radios set to 5 watts. Most start at 5W: "can't hear you!" OK turn it up to 100W. Sure we have all done that before.
Not to derail the thread further
but I've been learning about class-E tx design and wondered if any of you have a suggested FET for good efficiency in the upper HF bands (20-10m) at ~ 1W output power. I read through Dan Tayloe's presentation (
http://www.norcalqrp.org/files/Class_E_Amplifiers.pdf) where he shows a design evolution using 2N7000 FETs but they seem a bit slow at 20m and above. When I play around with parametric search at Digikey, I see the new GaN parts like EPC2037 (
https://epc-co.com/epc/Portals/0/epc/documents/datasheets/EPC2037_datasheet.pdf) have by far the lowest gate charge, and they're cheap. Other than the difficulties handling bare-die packages, would these make good QRP class-E finals?
Certainly looks like it might work. Numbers work and you probably can get away without a driver IC or FET as well. If you glue it to a PCB upside down and solder to the balls it might be manageable as well.
There are some more conveniently packaged devices used for external switches in SOT23 and SOT89 id hit first myself because I’m a chicken
If you glue it to a PCB upside down and solder to the balls it might be manageable as well.
Ugh, I would definitely want to reflow it. I deal with leadless (QFN) parts all the time, so at first blush I don't see why this would be terribly difficult, with only four balls. OTOH I have never used a bare die before so I don't know what special challenges come with that. The thermal resistance is pretty high, 100 C/W junction-to-ambient when mounted in the recommended way, so something like 5W out might be pushing it, but my interest is mostly in low/slow digital modes and battery-powered operation.
I don’t have the kit to reflow solder. Nor the inclination if I’m honest. Bare die is probably pretty robust if you don’t dink the deposited side while mounting/soldering.
0.9mm x 0.9mm die size with 4 balls on top of it.. That would require a steady hand, indeed
One thing though, I think the uBitX is a ready assembled board so it's not so much a build as an assembly isn't it?
Interesting point. Yes, the board is already populated and ready to go. Yet there is no enclosure, no power supply, and most importantly, no antenna. I'm thinking of this project in terms of a "build" because there are a great number of permutations it could still undergo considering how tweakable it looks, not to mention the options for adding on. I've seen people do some impressively different things with this design already. Sure, I could buy a box and a battery pack and an antenna and use the software in its stock form... but where's the fun in that? I might as well buy an SK's whole Kenwood rig and spend an evening scraping his call off of all the surfaces. I'd rather try to make something my own, even if I'm not quite at a point where I can improve on the fundamental design.
I have the BITX 40 and uBITX, they are both really fine QRP transceivers. In fact these are my only two pieces of gear right now, I sold my old Yeasu after I purchased the BITX 40 and haven't looked back. As others have stated, SSB voice at QRP power levels can be a challenge at this stage of the solar cycle. If you know CW or are into digital modes, then you will have no problem working the world. If SSB voice is you thing then you might want to consider buying or building an external PA, however I've been quite happy at QRP power levels and have had no lack of success with both voice and digital modes.
The design of the transceivers are really easy to understand and infinitely hackable/modable. One of the best features is the large and very active community of users/hackers/moders at https://groups.io/g/BITX20
I gave the BITX40 a look, but I decided to go with this one, (mainly because I'm interested in 80 meters and I like having the CW option out of the gate, although from what I've read, it sounds as though the keyer really needs some work).
Thank you all for weighing in. I've always been interested in QRP; it just seems like there's a bit more finesse involved from an operation standpoint, and I have much to learn.
Agree with your points about an SK’s kenwood. I bought a nice Yaesu rig and have hardly touched it. Made a few contacts and got bored with McDonald’s radio and went back to fiddling with things. It gets used as a signal generator more than anything.
One thing though, I think the uBitX is a ready assembled board so it's not so much a build as an assembly isn't it?
Interesting point. Yes, the board is already populated and ready to go. Yet there is no enclosure, no power supply, and most importantly, no antenna. I'm thinking of this project in terms of a "build" because there are a great number of permutations it could still undergo considering how tweakable it looks, not to mention the options for adding on. I've seen people do some impressively different things with this design already. Sure, I could buy a box and a battery pack and an antenna and use the software in its stock form... but where's the fun in that? I might as well buy an SK's whole Kenwood rig and spend an evening scraping his call off of all the surfaces. I'd rather try to make something my own, even if I'm not quite at a point where I can improve on the fundamental design.
[/quote]
Absolutely agree, the hobby is a lot of things but one of the main things is an opportunity for experimentation and learning, I was mostly just pointing out that there's less construction with a BITx than some other kits, there are some great builds with it and plenty of opportunity to learn from it and develop it.
I have a commercial HF rig, an Icom, and TBH, it leaves me cold, I play FT8 with it as I'm still not able to have a shack set up and it's quick n easy to get my 'fix' but I've always got far too many projects on the go and I'm hankering after a QCX as well but will probably lose all interest in when they're finished and move on to something else.
More power to your QRP elbow and don;t forget to let us know how you're getting on (and might I suggest you join the GQRP club?)
Yes, I also enjoy getting my hands dirty, although winding my own toroids isn't something I've tried yet... and I also have to take into consideration that a Yaesu all-band is not within my budget.
Here's what I've been considering as an antenna.
http://www.hamuniverse.com/k5uss80loop.html
Toroids are easy. Don’t let anyone tel you orherwise
... apart from one in the QCX which is a bitch.
I am liking Ft8. Means I don’t have to talk to anyone
Not to derail the thread further but I've been learning about class-E tx design and wondered if any of you have a suggested FET for good efficiency in the upper HF bands (20-10m) at ~ 1W output power. I read through Dan Tayloe's presentation (http://www.norcalqrp.org/files/Class_E_Amplifiers.pdf) where he shows a design evolution using 2N7000 FETs but they seem a bit slow at 20m and above. When I play around with parametric search at Digikey, I see the new GaN parts like EPC2037 (https://epc-co.com/epc/Portals/0/epc/documents/datasheets/EPC2037_datasheet.pdf) have by far the lowest gate charge, and they're cheap. Other than the difficulties handling bare-die packages, would these make good QRP class-E finals?
When you say fet do you mean mosfet or is a plain fet: something different like a jfet. Aren't jfets the silicon version of a vacuum tube? Where more current fed in equals less amplification? What would you use a jfet for? I bought some but never found a use for them. I would be a fun project to replicate tube circuits using solid state parts.