|
Re: New single band superheterodyne transceiver
What is your I.F. frequency?? I'm curious because I have a similar superhet receiver using the HYCAS board and an 8 MHz I.F.? I'm currently using a Si5351a for both L.O. and BFO. It sounds pretty good but I have a lot of birdies. Do you have issues with receive birdies? My Si5351a board (Adafruit clone) has a 25 MHz crystal. One step I might take is to go with a different reference oscillator frequency. Where did your 26 MHz Si5351a come from? Do you use a module, or start with just the IC? Moving my 8 MHz I.F. would be painful because I went to a lot of trouble designing and building CW and SSB bandwidth crystals for that frequency. 73/TU, Nick, WA5BDU On Sat, Apr 26, 2025 at 12:13?PM hwstar via <steve=[email protected]> wrote:
|
|
Re: f_T fixture and cheap transistors [long]
Chuck,
?
Thanks for lighting the fire under my rear. I have my Raspberry Pi Pico based jig working now, after a couple days of struggling with programming. Hardware design and soldering was the easy part, as I expected.
I finally got the idea that I don't need to measure the output precisely, if I can adjust the input in a linear way. Charles Wenzel adjusts the frequency.
I don't change the frequency, it is fixed at 10MHz. I change the input amplitude until the output is at 1.5V p-p (the maximum this kind of circuit can drive without much distortion).
Now I have all the data to calculate the current gain without calibrating the non-linear detector. Everything is automated, I just plug the device in and wait for a few seconds for the result.
Below are some teaser photos. I promise to put the actual documentation and software on my github page (eventually).
The Pico drives an R-2R DAC and generates sine wave at 10MHz with variable amplitude. The signal generator code is from this project:
The amplifier circuit is similar to Wenzel's, except that the base DC bias resistors are higher value and are connected straight to the base, not before the current source 1k resistor.? One is variable to adjust the collector current.
The detector is the voltage doubling circuit that you are also using. The diode is BAT54S.
I have one board that uses 12V for the collector supply, Vce=10V. The second one (pictured) uses 5V from USB, so the measurement is for about 3V Vce.
?
Lower fT transistors show the expected values. The one plugged in when I took the picture is a fake 2N3904 2-cent part ($0.99 for 100), probably the same source you got yours from. Measured fT=195MHz. Genuine new 2N3904 from On Semi is not much better. An old stock Motorola part is 265MHz.
?
Now the high fT transistors don't work as expected. Both 2N5109 and 2SC5551 that I tried show 450MHz fT. I suspect this is because 10MHz is below their "knee" frequency. Need to add another range which uses 25 or maybe even 50MHz , which will be pushing capabilities of the RP2040 an of my construction technique.
?
73, Mike AF7KR
?
|
|
Re: Digi-key
On Sat, Apr 26, 2025 at 08:37 AM, Dave AA7EE wrote:
QST was more interesting then. Endless pages of full-color advertisements for brand new commercial gear makes my eyes glaze over. That's why I subscribe to SPRAT. Alas, I yearn for the [late] 70s when I semi-accidentally subscribed to Wayne Green's 73 Magazine. From 1977 through the early 80s, each magazine was chock full of projects, tech articles, and general interest stories.
?
Apparently Wayne had set up a collection of articles and, rather than do expensive cover pictures, from 1979 through 1980 or so, he ran a billboard table of contents type cover like shown below. Around 1981, this reverted to the picture covers and the articles thinned out. But there were a lot of interesting articles even then.
?
Mine were the CB to 10 articles which caused a lot of false starts for me along with a buildup of those old HyGain CB boards offered by Fair Radio Sales in Ohio.
I remember having a stockpile of them -- a bunch purchased at $2 per board Still have a couple lying around!
?
But, like Dave says, "them? whar th' days"!
?
?
Personal opinion: what has happened to QST (and the ARRL) is dreadful! Makes me glad resources like QRP Tech and other groups exist!?
--
William, k6whp
-------------------- "Cheer up, things could get worse. So I cheered up and things got worse." |
|
New single band superheterodyne transceiver
Here's my latest rig (see pic below). It is a 10 Watt PEP single conversion transceiver using 6 printed circuit boards in a copper clad PCB enclosure. I've been having a lot of fun talking to my friends on 40 meters with this rig.
?
The boards are as follows:
?
1. VFO board. (purple, front) This uses a raspberry pi pico. The pico has a really inexpensive debug probe you can get for $15. Debuggers make it really easy to find problems with your code. The board has an Si5351 and a 26 MHz 0.5ppm TCXO reference oscillator.? I used Platformio IDE to develop the C++ code for this rig. I chose the Arduino framework in Platformio.
?
2. TX amp and filter board (left wall of enclosure). This board takes the output of the product detector (though a high pass filter) and amplifies it by 40dB. It has 2 stages.: a 2N3904 followed by a 2SC3357.? It then filters the output with a triple-tuned bandpass filter for 40 meters. The purpose of the triple tuned bandpass filter is to filter out any spurs from the mixer before they get to the predriver and final amplifier. I thought about relay switching this band pass filter with one used in the front end board, but that would have over-complicated things and maybe introduced some paths to cause oscillation. I wanted to avoid bi-directional amplifiers as I wanted the rig to have AGC.
?
3. Predriver and final amplifier board (red, rear). This board takes the -9dBm output from the TX amp and filter board and amplifies it to +40dBm (10 Watts PEP).? It has 2 stages. A pair of 2SC5706's followed by a pair of RD06HHF1's. If you push it you can get 12 watts PEP out of it, but I don't like running things at their limits. ?There is only one relay in this rig and it is on this board. The nice thing about push-pull amplifiers is that they attenuate the second harmonic significantly. The 5th order low pass filter on this board? is good enough to attenuate all harmonics below the -43dBm limit.
?
4. Front end and crystal filter.? (Green board closest to the rear and close to the pre-driver and final amplifier board).? This board has a triple tuned bandpass filter before the mixer and also injects transmit audio into the mixer (which serves as the balanced modulator) during transmit. The output of the mixer is amplified using a 2SC3357? and then put through a 6dB pad before being presented to the 6 pole crystal filter. The crystal filter uses 12.288 MHz crystals. One thing to note about the 2SC3357: You can run it with 40mA of quiescent current, but you need to keep the the Vce below 12 volt absolute maximum in the datasheet. I power all the 2SC3357's in this rig with a switched 9 volt regulated supply.
?
5. Hycas IF amplifier board with AGC. (Green board closest to the right side of the enclosure). This is a variant of the Hycas IF amplifier featured in QST: ?? with some switching modifications to allow adjustable gain on transmit.
?
6. Product Detector/TX mixer/Audio Amplifier board. (Green board closest to the left front mounted on the bottom of the enclosure). This board contains a mixer, an NE5534 op-amp, and an LM380 audio amplifier.
?
?
?
|
|
Re: LT Spice Alternatives
I went down the alternate road a while back and found QucsStudio (uSimmics). ?
The neat part is the tunable parameters section, change your L&C's on the fly with the sliders and watch the data plot(s) change in real time.
I believe there is a converter available to move LTSpice files into Qucs but never tried it.
72 John kc9on
|
|
Re: Earliest example of Manhattan construction?
I thought of this thread when I found something in my notes headed "Birth of Manhattan Style?". It was in reference to a QST article of May? 1972 called "Some 2-meter solid-state RF power amplifier circuits" by K7QWR.? He stated that rather than etching the board for all connections, he attached some pieces of board material to the top of the board for transistor connections.
?
?
73
?
Nick, WA5BDU
|
|
Re: Digi-key
Wow, great issue.? I was going to try to track down that legendary keyer but you saved me the trouble. I'd recently been researching early keyer designs while I was engaged in my Accu-Keyer project, but I missed this one. And wow, I had no?idea that QRP was such a hot topic in 1968. In this issue, there's both a 40 meter, 1-watt transceiver, and a "milli-gallon" (1 watt) 15 meter amplifier by Wes Hayward. And as a bonus, an RTTY audio filter by Ed Weatherhold W3NQN with his 88 mH telephone toroids. Also of interest is that the electronic projects are all solid-state. I noted that the keyer used the 2N4888 high voltage PNP for grid block keying. Same thing the Accu-Keyer used a half dozen years later. So these things propagate. 73 Nick On Fri, Apr 25, 2025 at 4:01?PM Gerald Wolczanski via <jerrywlinux=[email protected]> wrote:
|
|
Re: LT Spice Alternatives
Karl,
?
Thanks for sharing you video. I looked at Spice many years ago and found it too confusing. Your video inspired me to take another look, so I downloaded the LT to my Mac.?
When I started and found a blank screen with no toolbar, I said uh-oh and thought maybe not.?
analog Devices sent an email a moment ago with some helpful information. One thing is a .pdf with all the shortcuts for both Windows and Apple users. Just what I needed! I will now proceed to battle this beast lol.?
Thanks again.?
Newt
N4EWT
?
oh I should add the url from AD:
?
?
?
? |
|
Re: LT Spice Alternatives
Nick,
?
The problems I was having were, as I mentioned, getting LT Spice to "properly" display/plotting of a filter. Ironically,? the video that Karl made directly addressed the LPF I was working on. Then, by extension, his explanation pointed out that it addressed the means of how he did it. More to the point, it was how to set up the analysis to yield the results.
?
I am now reviewing the video as it is loaded with juicy tidbits. I am most grateful to both of you for assistance!
--
William, k6whp -------------------- "Cheer up, things could get worse. So I cheered up and things got worse." |
|
f_T fixture and cheap transistors [long]
Gang,
1.? First of all you are going to need something like the following RF probe to measure the output of the f_T tester we have been discussing in previous posts. There are a lot of RF probes that use the one diode, a 4.7M resistor and ....? For this project I wanted more sensitivity. RF probe? <> The reason for this is that I find that with the Rigol scope there is some jitter on the max values of the output from the CAL and OUTPUT.? I am going to use the probe and an analog VTVM, B+K Model 177 VTVM, for a more steady reading.??? This meter I found at a swapmeet for $20 USD and I refurbished to like new.? Big meter. 2.? You will need the f_T fixture found at f_T fixture <> Copyright by Charles Wenzel.? No call sign that I see. This will now allow you to check all your transistor stock, old and new (see below), and document your builds. Call it f_T#1 by CW for your measurements using the original unmodified version. Now make your measurements according to the instructions in the schematic.? V_in=1.0V_pk or 0.707V_RMS.? Whatever voltage you get a CAL, and it may not be 100mV, make that the voltage out, V_OUT, to increase f_input to match.? Then multiply the f_in by 20 to get f_T.? Easy as pie.? They just need to match. The value is not critical. 3.? [Advanced]? Now, Steve, AA7U, and I have mods to the fixture. Simple.? Change the 150ohm resistor from V_cc to the 1K and 82ohm node to a variable resistor.? I used a 2K and Steve uses a 500 ohm variable.? The 500 will work better, but I didn't have one. Use this variable to adjust the bias to get 10mA through the 100ohm emitter resistor for the transistor you are testing.? Since Steve and I are doing measurements for the V_ce in the spec sheet, we have a variable lab supply for the 5V source and we vary it to get the V_ce in the datasheet and adjust the variable R to get 10mA. You measure V_c, the collector voltage to ground, subtract 1.0V to get V_ce.? This is going to be like adjusting a regen receiver. Also, use a SPDT switch and just switch V_CAL and V_OUT to your voltage measurement scope or meter so you won't be moving leads back and forth.? Idea from Steve, AA7U.? Speeds things up. 4.? Cheap transistors. <> I have the 1,000 2N3904 NPN transistors received yesterday.? Using the f_T#1 configuration, so that you can compare your 2N3904s to these, here is what I get.? The cost was $14.64 USD from the site.? He has about 50,000,000 left, so rush and get yours before they run out.? There are some other transistors there and 1,000x2N7000 JFETs for $23.89+$7.94S&H. Now I did the following measurements as a first pass test to determine how critical the V_ce voltage is to the resulting f_T.? It does matter. Also, I'm using the Rigol scope, so mileage may vary until I can get round to it using the diode probe and eliminate the small jitter. V_cc replaces the 5V supply in the schematic. V_c is voltage from the collector of the DUT to ground. V_ce is 1.0V subtracted from V_c, since there is a 1.0V drop across the 100 ohm emitter resistor due to 10.0mA current through it. V_cc?? ? ? I_C??? ? ?? V_c???? V_ce?????? f_T 5.25V? 10.0mA? 1.58V ? 0.58V? ? ----? couldn't get a match V_OUT too low 9.00V? 10.0mA?? 5.42V? 4.42V?? 104MHz 12.0V? 10.0mA?? 8.44V? 7.44V?? 142MHz 23.6V? 10.0mA? 21.00V 20.0V?? 176MHz h_FE=331 as measured on the Chinese Critter Checker h_FE between 100 min and 300 max on Philips Semiconductor datasheet at I_C=10.0mA. Don't know what the emitter current is in the tester, not sure it's a valid comparison. f_T is shown to be 300MHz minimum in the datasheet. Because the h_FE=331 value, I'd expect the f_T value to be lower, but by how much is a guess.? At f_T=176MHz, the transistors should be good enough for up to 0.1*f_T or so, but that is to be determined.? For smaller quantities and better quality, you may want to shop around.? I'm just pointing out how cheap these things were and I have a large number of projects to experiment using them in.? It would be a waste to not use them up. They came in 12 days from the order date. These are the transistors I plan on using for the Manhattan projects. Attached is photo of bag open containing the 2N4401 critters. FYI -- chuck adams, aa7fo QRP WAS 17m started Apr 2, 2025 WAS-17 QRP --> TX,FL,SC,GA,OK,AR,KS,AL,AZ,MN PA,CA,CO,WI,NE,ID,IL,NC,NY,AL MA,OH DXCC W,VE,JA,PY,ZL |
Nick Kennedy
Lex PH2LB
k6whp
hwstar
Dan Romanchik KB6NU
Newt N4EWT
