Date: Fri, 21 Jan 94 17:11:04 -0600 From: adams@chuck.dallas.sgi.com (Chuck Adams) Message-Id: <9401212311.AA19695@chuck.dallas.sgi.com> To: qrp@Think.COM Subject: RadioKit Review January 21, 1994 Gang, RadioKit Notes by Chuck Adams, K5FO After helping one individual build one of these critters over the phone, I got another call this week from another ham in the mid-cities area of Dallas-Ft Worth about another one. Background: This ham had been QRT for 20+ years, so in June got active again by buying the MFJ-9020 at HamCom in this area. He is having a blast using this little setup with a simple dipole. He bought the RadioKit QRP-30 kit. Since he was having some difficulty, I drove 40 miles over to his place to take a look at it. I wound up spending about 3 hours working on odds and ends. His wiring job was pretty good and I found only one solder bridge. The spacing between some of the paths on the PC board were very tight. Here are just some of my notes on the kit: 1. Very small board compared to other K1BQT kits and NN1G rig. 2. Not silk screened, not solder masked. Just a single sided board. 3. I'd personally grade the parts quality as a B. 4. I'd grade the board as a D. We had to drill four of the holes that were missing. For a beginner with minimum tools, this would really cause problems. 5. About a half-dozen components were missing. 6. I don't like the relay for QSK. I think the solid state switching that everyone else uses is vastly superior. And cheaper too. 7. We didn't get to installation in case, as board was unfinished. 8. I think that everyone has to write a better set of instructions on the art of winding toroids or refer the builder to the ARRL handbook. 9. Schematic is wrong in some places, as I found out by tracing paths on board and components. I knew from the schematic, upon close examination, that there was no way this puppy was going to work if the board was really setup per the schematic. I have the K1BQT schematic, NN1G, and some of the others memorized. :-) 10. Air variable capacitor was shorted. I believe the plates are just too close and would think that more than 30% of them would be shorted from the factory. Also, this will tend to cause thermal drift problems for the vfo. More notes as he gets the missing parts and brings rig over for tuneup and alignment and the 'smoke test'. Overall, I'd spend my money on something else for a xcvr kit. And I have. :-) dit dit Chuck Adams K5FO CP-60 wpm adams@sgi.com ====================================================================== Date: Fri, 21 Jan 1994 19:10:24 -0600 (CST) From: "JEFF M. GOLD" Subject: Radio Kit review To: qrp@Think.COM all, I think overall Chuck was much too kind. I have spent a few weeks working on this. If the kit was done correctly I could have finished it on a Saturday and really enjoyed it. The board is the worst I have ever seen.. doesn't solder great, you get one chance..even with good desolder tools.. you have to drill out the board in many places.. like for an adjustable board mounted pots (think there are 3 or 4) holes are left out, holes aren't orderly or logical, parts stink.. many pages of mod to mod.. soldering modifications on the bottom of the board..the parts overlay doesn't look a whole bunch like the board and like Chuck says.. schematic isn't correct and I find it hard to follow. I use to like building before this one.. I don't think someone could make a worse kit..except by trying REAL hard. At first I thought the non-silk screened board was throwing me.. then looked over to the $29 Spider sitting on the bench waiting to get a VFO and it was fun to build.. no silk screen.. good board..bottom looks good and work fine the first time. The NN1G I built didn't have silk screen. AVOID unless you want a real challange.,. anyone want a partially completed one for $70 shipped US.. I am about ready.. have some projects I would like to start.. I have been forcing my self to put a few parts on every day.. sure don't look forward to it. 72 Jeff, AC4HF ====================================================================== Date: Sat, 22 Jan 94 15:53:29 EST From: epacyna@auratek.com (Edward Pacyna) Message-Id: <9401222053.AA06613@auratek.com> To: uunet!Think.COM!qrp@uunet.UU.NET Subject: Re: RadioKit Review Cc: ed@auratek.com I have built nearly a dozen of these transceivers, so here is my 2 cents. First of all, I can not comment on the parts quality, as I furnish my own parts (expect for the air variable which I do buy from RadioKit). I buy the boards direct from FAR Circuits, the same source that RadioKit gets them. There are two versions, one from HR January 89 called NE602 QRP CW 20M transceiver sells for $8, and the other from CQ September 90 called QRP 15 transceiver and sells for $12. The earlier version had two COHN XTAL filters (one on each side of the IF amp. to develop 8 poles of selectivity and reduce board band noise) and a doublely tuned band pass filter in front of the RX mixer. The later has only a single tuned cir- cuit in front of the mixer, a 4 pole Chebyshev XTAL filter and adds RIT. Both versions include 2 stages of active AF filtering after the product detector, AGC and can easily be put on any band by changing the VFO and XTAL filter frequencies. Regarding the PCB boards: As mentioned, they are not silk screened with the component shape and Actually, I prefer this as it makes for a nicer looking finished product. Most builders should have no problem just following the parts placement diagram. However, a paint by the numbers approach might make it easier for some. The boards are very well laid out with a lot nice ground plane (pads are lined up and logical), however the later version does come with a couple of pads missing drilled holes). None of the 50 or so boards I've goten from FAR were soder masked (including the NN1G board set). The usually good soder plate by FAR takes soder very nicely. It is always good pratice to lightly clean any PCB with Comet before beginning. The board is very compact 4"W x 3"D (25% smaller than the NN1G board set and contains more circuitry e.g. RIT, active filter, AGC). With a higher part density, sodering does take a little more skill. Also, use only a low wattage iron with a point tip and a good quality thin soder. Those over 40 like myself, might be more comfortable with the magnifying eye glasses as sold in drug stores. I have been able to unsoder and remove parts over and over with a little care (I make lots of modifications along the way). Regarding the schematic: I recall that both the schematic and parts placement diagrams being correct. The problem is that they are sightly different. The difference is fairly minor and as long as you didn't get your license by memorizing the answers, should not be to difficult to resolve. I recall having to install one part on the soder side of the board. Pin 4 on the NE602 product detector needs to have a bypass capacitor on it (since it is unused and has RF floating on it). This is shown on both diagrams. Regarding the tuning capacitor: This is a very nice capacitor and I sometimes buy them from RadioKit. It has a silky smooth built in 7:1 reduction drive and is 5 - 55pF. The frequency vs shaft tuning is exceptionally linear. The plate spaceing is perhaps a bit wider than the broadcast band variables of yesterday. The stability is excell- ent. If it got shipped with the plates unmeshed or dropped on the floor, a rotor plate could bend. Possible modifications: The circuit works well as is. The RIT version is nice because the station your working made be using the opposite sideband or drift. The T/R switching is semi-breakin and as MFJ states in their ads "silky smooth". By the way, this circuit is the same as used in the MFJ 90XX. Full QSK is nice if its used (but I find few that do). I usually run my radios from a 13.8V regulated power source (and believe most others do). The on board 12V regulator is uneccessary and reduces overall RX gain. It would also be a good idea to insall some reverse polarity protection to protect your labor of love. A power diode in series with the power supply feed is simple (reduces voltage by about .7V) or install a series fuse in the line with a shunt power diode on the radio side of the fuse (cathode to fuse, The AGC is a mixed blessing. It is audio derived and pops on strong signals. I don't like any AGC on CW. First of all, if a strong signal is nearby, it reduces the RX gain making it more difficult to receive the weak signal and the usually pumping action is unpleasant. In a simular fashion, it reduces the opposite sideband surpression that your XTAL filter provides). On the other hand it reduces the signals dumped into the product detector. This is a weakness in the NN1G design. The NE602A is a wonderful device provided it's used correctly. Due to it's gain, the input intercept is -15dBM or so. If you give it more signal than it's meant to handle, you get a lot of IMD product. By the way, one nice feature of this design is that the VFO is buffered and has an adjustable pot allowing you to control the TX mixer drive and thereby reduce IMD products. On the receive side, I control the IF gain with a front panel pot on RX (use AGC for TX monitoring only). I do not install the volume control (it's kind of hockey installed on the AF power amp. anyway). You will have a much beter receiver with this approach. I always use IC sockets when I build. Use only low profile machined sockets (the ones that have round holes for the IC pins). There has recently been some discussion of NE602's oscillating (e.g. NN1G Marc II) when sockets are used. Although NE602's do have alot of gain up into VHF, oscillations are usually due to poor circuit layout or improper decoupling. As shown in the data sheet, the NE602 supply voltage sould be decoupled with a 5uH choke and bypassed at all frequencies with multiple bypass capacitors (e.g. .1uF, .001uF etc.). I usually use a 100 ohm resistor with a ferite bead for the choke. Also since maximum supply voltage is 8 volts and the best noise figure is at 6 volts, I use 6 volts (78L06 regulators). The advantage of sockets is that it makes the radio very easy to debug and or repair. If you have a problem, remove the IC so it doesn't load the circuit and using a multimeter make measurements to easily find soder bridges, components installed incorrectly etc.. I always build my radios backwards. For example in a RX start at the output and build toward the input (reverse in a TX). Just build the audio output stage and test it before you go on. You'll never have to deal with more than a few parts or 1 error at a time. Its easy to test stage by stage without a lot of fancy equipment too. After you build the audio amp., put your finger on the input. You should hear a buzz. Next build the product detector. You will hear a hiss if all is well (or you actually now have a DC receiver and so connect an antenna and your might hear some signals as you tune the BFO cap.). Next build the IF amp. and the signals you heard before should be louder etc.. The other two changes I usually make to this radio are to change the VFO circuit to a series tuned config. (for lower noise) and stagger tune the AF active filter. With a little thought, the series circuit will fit on the same PCB pads provided. Most active filters simply repeat the same stages N times (e.g. same Q, frequency etc.). What you get is a very sharp needle nose response that rings and is not very useful. With stagger tuning, the Q, center frequency and gain is changed to get a bandpass response instead. Different bandpass shapes can be synthesized (e.g. Butterworth, Chebyshev etc.) with superior skirts. The circuit is exactly the same, only the R,C values change. I also build the filters with a little more gain so the sig- nal into the product detector can be kept low. Some times instead of the in out switch, I use a simple audio fader circuit like in the TenTec Century to simulate variable bandwidth. Sorry for getting so carried away here guys. My main point is that these are very nice transceivers, and should not present difficultly to many homebrewers. However, they are not suitable for novice level. 73 Ed WE1AAZ