Simple SSB Rig: Part 6 - IF Amplifiers

[@johnwest7993]

Hi Charlie. As someone sitting here with a bag of 5000 (yes, thousand) surplus FH1100 Schottky diodes that I don't know what to do with, I find all of your rambling and experiments with mixers to be VERY interesting. I understand you are not a professor lecturing on the subject, but in 15 or 20 minutes of watching your videos I learn far more than I would have from staring at datasheets and reading technical papers all day long. Thank you.

[@CharlieMorrisZL2CTM]

Thanks for the feedback. I'm really pleased you found it useful. Again, just my view of the world.

[@peterwebb2224]

Hi Charlie, Great video, Can’t wait to see the new Analyser in operation............regards Peter GW4ZUA

[@CharlieMorrisZL2CTM]

Me too Peter. I haver not used one for 30 odd years. Baby steps!

[@bachkhoa1975]

For practicality, design and test with a standard impedance (such as 50 ohms) for all source/load and amplifier input/output impedance. I also recommend using LTSpice to run simulation before building circuits.

[@jerryfankhauser2959]

Thanks Charlie, I'm waiting on parts for my IF amp. I'm using 2 BF998's with the input to the first stage loaded down with a 50 ohm resistor to keep the mixer happy at all frequencies and the output of the second stage with a 1k ohm resistor to match the filter. It works in spice now I just need the parts to show up.

[@CharlieMorrisZL2CTM]

Sounds great Jerry.

[@cyrillehasdenteufel298]

Hi Charlie, many many thanks for all the great videos and for sharing. I like what you do not only because you share your schematic but also because you share with us your calculations and your approach. But I would like to ask you if you could give more information of how you made the transformers, in this case for the IF amplifier on input and on the ouput, what kind of materials, copper wires (section) and so on you are using .. maybe you can also share the part numbre .. it would be very nice .. thanks a lot ... Many greetings from FRANCE Cyrille

[@SpinStar1956]

Charlie, great video and you're doing pretty well to get the performance you do from a single transistor stage. If you use 2 discrete signal-cables I'd use a hybrid-combiner for sure. A few thoughts: When I was a youngin', all I could afford was the bulk un-marked transistor-pack from Radio-Shack. As anyone with this experience knows, this was one of the most insidious, devious and demoralizing crimes 'Fausted'-upon new and unsuspecting neophyte experimenters! As as result, I came to learn (other than how low Radio-Shack would stoop for profits) every defect that a transistor might present either immediately or subsequently when you installed one of these fine jewels in a customer's piece of equipment! After separating the total junk and rigorously (actually viciously!) testing for durability, I learned to segregate based upon certain characteristics as well as gain or beta. When I would build something, I would usually include a spare kit of repair parts to use in the future. So, I would color-code with paint the tops of similar/compatible transistors, signifying where they would go in the future. The point is, that you can now buy 100's of 2N390x transistors and cheaply grade them yourself; you'll find most are close to each-other with excellent performance especially from a reputable manufacturer. Furthermore, your beta-calculations will be much closer and truer because you're using 'your' 2-numbers from the actual lot you sorted. So, if you bought a few hundred of these and did this, you'd have enough to last your lifetime and could easily directly replace anything using that sorted lot. I just always had a problem of using the minimum-beta unless I was designing for production, and without control/inspection of the components used. Just and idea. 73...

[@CharlieMorrisZL2CTM]

Thanks for that. I did purchase 100-odd in the last lot. I'll take another look when I can.

[@Chris-ZL]

You make it all seem so clear

[@CharlieMorrisZL2CTM]

Not sure Chris :0) Lots of errors I'm sure, but it works for me.

[@curtstacy779]

Thank's Charlie, Nice looking new equipment. you are looking so modernized from your first video's. lol

[@CharlieMorrisZL2CTM]

Yes, I have been saving up my pocket money of the years and upgraded equipment. I'm really happy with the setup at the moment. Just debating if I want a nice SNA/VNA or stick with the new NanoVNA... We'll see.

[@user-su5sq5ib3i]

That 10ohm resistor you mentioned is a decoupling resistor. I thought it would be a current limiter. Can you explain how it decoupled? Also your math is solid. I am learning all of this right now in school and my instructor (engineer) is not explaining as well as you are. We are doing ac and dc analysis and trouble shooting. That is what brought me here on a search. Great channel and explanation

[@kd9kck376]

The rig i am working on, i am using 5v for the VCC on basically everything to allow for longer operation on my 12 LiFePO batterys and also maybe being able to run off a 9v or a cheap USB battery bank. For the PA i am going to use what ever the actual raw input from the power source is.

[@CharlieMorrisZL2CTM]

Very interesting. Keep us posted on how you get on.

[@__dm__]

Do you have a TinySA? I just bought one and it is amazing value for money. Super powerful and great for ham use.

[@CharlieMorrisZL2CTM]

No I don't sorry. All I have is the Siglent SSA 3032X. Now i need to learn how to drive it.

[@R2AUK]

Many thanks for the video, and congrats with the new spectrum analyzer. Looking forward for future videos!

[@CharlieMorrisZL2CTM]

Thanks for the feedback Alex. Much appreciated.

[@AllElectronicsChannel]

Hey! Is it missing a capacitor to couple the input transformer? The way drawed will short the DC bias!

[@gillywild]

That was my thought too, but I thought I must just be missing something.

[@sm4wwg]

I saw that too. A capacitor or having the secondary coil in series with the biasing resistors. I am sure it is just a simple mistake he made when he made the schematics 😁

[@simonmichaeluk]

just remember not all the stages have been shown yet and Charlie has done a similar thing before and had the capacitor on a previous stage. So don't worry too much at this stage ( after all we all know it belongs there from all the other radio builds, so the fact it isn't drawn on this stage is not really an issue ). just remember also the previous stage is a mixer and there is no DC bias, it's actually an AC ( RF ) stage ( this is a second reason why the capacitor is NOT an issue). Best regards and stay safe everyone Simon M6FJW

[@CharlieMorrisZL2CTM]

Yes, but the secondary of T1 is a DC short across R2. The 100 nF cap fixes that.

[@TheEmbeddedHobbyist]

I was just thinking while looking at the circuit, the the bias is wrong with the DC resistance of the transformer shorting out R2.

[@TheEmbeddedHobbyist]

Great video again, got me thinking that i've got to order a few ferite cores and have a go.

[@CharlieMorrisZL2CTM]

Go for it. I bought quite a large batch when I last purchased them. Not sure when I will get to the States next.

[@vladimir.3141]

Hi Charlie! You forgot aabout the capacitor between the secondary winding of T1 and base of the transistor 2N3904

[@CharlieMorrisZL2CTM]

Yes, sorry the schematic is wrong Vladimir. It's correct on the blog.

[@stevereed6946]

Hi Charlie, thanks for the explanation of the component values. I am trying to get my head around all of this and thanks to you, it is making sense. Probably a stupid question, but where does the re of 26 come from? Regards Steve

[@N_OR_]

It is from re = 26/I_e(in mA)

[@simonmichaeluk]

Hi Charlie, I have just caught up on your video's due to work commitments. Just a quick note about the ft37-43 toroid's and the transformers for your double balanced mixers. I am sure you know that the wideband transformer optimal range is 5 MHz to 400 MHz . Possibly a better choice is the ft37-75 these have a wideband transformer range of 1 MHz to 50 MHz . This maybe a better solution as you possibly wouldn't suffer greater losses below 5 MHz due to using the core outside it's optimal range. The second point is the number of turns may NOT be optimal for the frequencies of interest due to the inductance and the frequencies in use, This would change the impedance on the input/output of the mixers at the frequencies of use. This could give rise to more possible losses in both double balanced mixer stages. In my humble opinion your double balanced mixers seem to be assumed to have an impedance of 50 ohms rather than designed to be 50 ohms. This opinion is solely based on not seeing you calculate or discuss this aspect of the double balanced mixer design step. Looking at these points may yield a double balanced mixer that is a lot closer to commercially available mixers ( it also may not make any noticeable difference at all. It is merely offered here as food for thought when building mixers. Stay safe Simon M6FJW

[@simonmichaeluk]

@@tcarney57 yes i understand that but many others may not ( that was the point , that Charlie didn't say why or what sources were used, for anybody doing a parallel build). The other point that i was making is this... 9 MHz and 10 turns on a FT37-43 = 1972 ohms of reactance at the stated frequency ( just using a basic inductor calculator for the primary input winding). now 12.65 MHz ( the oscillator frequency that was partially discussed, 9 MHz IF and 3.65 MHz signal assumed and a high side oscillator) and 10 turns = 2.7 k ohms of reactance ( again using a basic inductor calculator). These things may make no difference in real terms, but are worth discussing. Which leads to the question " will a different number of turns be better?". Better is a relative term and merely points to better for this build. Yes the sources are tried and tested, but doesn't mean they are optimal for any particular radio build, Or that the design shouldn't be altered in any way. These radio builds are about discovery and learning. Sometimes discovery and learning requires you make mistakes while experimenting ( that how humans learn) to fully understand why things should be done in a particular way. Just following somebody else's design blindly doesn't offer any real learning, it just teaches how to copy. best regards Simon M6FJW

[@nyworker]

Optimization makes sense but isn't the point of using ferrite cores to suppress high frequency noise which is more of an optimization factor?

[@simonmichaeluk]

@@nyworker using a core outside it's optimal frequency range could lead to losses. so picking a core that covers all the frequencies of interest ( LO, RF and IF frequencies) will have less losses than using the wrong core type. If one or more of the signals is outside the optimal frequency range of that core then it wont function as well, this may have the effect the altering the lowest receivable signal level ( if this makes a difference at all, you may live in a high noise area ). best regards Simon M6FJW

[@CharlieMorrisZL2CTM]

Thanks Simon. Yes, I am aware they are wideband. I tend you use them as they are easy to obtain here and have a wide range of uses.

[@diddleize]

And now as far as xtal filters are concerned the filter needs a specific output load impedance - in your case 250R. But it also needs a specific SOURCE impedance of 250R. It does not present this impedance to the driving amplifier, it is the amplifier that must present this impedance to the filter. So you need an amplifier with a specific output impedance (of 250R), not a random choice of a load of 200R or 1K and a transformer... I do not yet know how to set the output impedance of an amplifier, but an emitter follower with a series resistor of 250R may work... or you could put a 200R resistor across the primary of the output transformer to define the impedance? Or perhaps use an IC such as the MC1350 and a series resistor?

[@CharlieMorrisZL2CTM]

I've always thought that too. Even though this arrangement has worked well for me in the past, i might place a 200 ohm resistor across the primary of T2. I've seen that a couple of times before.

[@rfdave3980]

Oh I saw a new toy...sweet.

[@CharlieMorrisZL2CTM]

Yes indeed. Now to learn how to drive it, and NOT blow it up!

[@R2AUK]

@@CharlieMorrisZL2CTM :) Most importantly don't exceed the maximum input signal level. You are going to need a high power (10W...100W) 30 dB attenuator for testing power amplifiers for IMD, harmonic distortions and things like this. Such kind of attenuator are easy to build. The principle is the same as with 50 ohm dummy loads. Otherwise these things are difficult to damage. If you are not going to exceed 10W in the nearest future there are inexpensive 10W dummy loads and 20 dB attenuators with BNC connectors on eBay.

[@jweinacs]

So when is your book coming out? HIHI! I would purchase definitely. My goal someday is to build a Manhattan style rig like AA7EE Dave’s

[@CharlieMorrisZL2CTM]

We'll see! As for Dave AA7EE, he certainly builds some beautiful Manhattan style rigs.

[@zoltannagy1502]

Hi Charlie, Thank you for the inspiring videos. Your videos made me learn more electronics. I have a comment/question. I simulated your IF amplifier in SPICE and it seem to me that the impedance that is looking into the transistor is about Zin = (19 - i*76.55)Ohm. After transforming with T1 it becomes Zin = (11.3-i*45.56)Ohm (|Zin| = 46.948Ohm). This leads to a reflection coefficient that is Gamma = -0.05-i*0.78 (|Gamma| = 0.782). That is rather big. Since this stage connected to the mixer (with 50Ohm output impedance) I think it is not healthy if the amplifier stage reflects back the signal to the mixer. The IF frequency is fixed to 9MHz, would it be better to use a L-network to match the complex impedance? Cheers, Zoltan

[@CharlieMorrisZL2CTM]

I'll look into it. This approach has worked well for me in the past. Will LTSpice calculate input Z?

[@zoltannagy1502]

@@CharlieMorrisZL2CTM Not directly. I did the ac analysis with an ideal voltage source. I put a Rs = 0.001ohm resistor in series with the input. In the AC analysis I plotted the -V(source)/I(Rs) in Cartesian representation. Then you can read out the input impedance. It was important to terminate the output with 250 resistor (that is your filter input impedance), otherwise you get negative resistance at the input. That makes sense, with floating output the impedance at the collector is infinite and C_\mu of the transistor become so important, I think.

[@diddleize]

Perhaps I am faulty, but with the bias at Vb = 2V the collector swing can be 10Vp-p, with 10mA and 200R this canbe only 2V swing. So 200R is too low as a load to produce much output voltage? Surely you need to have a load of 1000R to allow a swing of 10V, so the optimal load to the Transistor should be 1k, not 200R, so the 10mA can swing the output 10V.... Or am I missing something? Then of course the 1000R output load has to be transformed to 250R for the xtal filter? Have you tried this? Does it affect the 37dB gain?

[@CharlieMorrisZL2CTM]

Not sure sorry Anthony. I measured the 700mVrms across the 240 ohm load.

[@lawrencebrown3582]

Great explanation - very clear!

[@CharlieMorrisZL2CTM]

Glad it was helpful.

[@diddleize]

Following my comments about the amplifier I have looked up a few things. My thoughts are on my blog GanymedeHam.blogspot.com.

[@kitcartel9390]

good piece new equipment

[@CharlieMorrisZL2CTM]

Yes! A late Christmas present. Hopefully, I can put it to good use.

[@SixWildKids]

Charlie, where does the value of 26 come from in the re part of the Rin calculation? The reason I ask is I built up the circuit and am trying to use a NanoVNA to measure the input impedance and I am having trouble measuring the input impedance. If I put a 40 dB input attenuator on the input I end up just measuring the "perfect" 1:1 VSWR of the attenuator so that doesn't work. If I drive the Amplifier directly without an input attenuator and put the 40 dB attenuator on the output the Amplifier goes into compression (saturation?) so I don't trust the input VSWR to be accurate. Maybe it's still accurate but I don't feel like I can trust it. The NanoVNA is measuring the VSWR with a "U" shape dipping to 1.5:1 at 15 MHz and rising a bit to 1.6:1 at 9 MHz. According to the NanoVNA that is an input impedance of 81-j10 at 9 MHz. I figured I'd remove a couple of windings on the input transformer to drop the impedance to closer to 50 ohms for a closer match but am not sure that with the compression I can trust the NanoVNA to measure the S11 (or VSWR as a surrogate). I understand all the other numbers in the input impedance calculation, it's just the 26 value in the maths that I don't follow. The rest of the numbers are much higher than the Beta AC times 26 `divided by 10 value so they don't much matter in the input impedance calculation. Any help would be greatly appreciated. I've watched all of your videos many times and they are really good! You've reignited a dormant passion to build radios from pieces in my life. I like to characterize the parts as best I can and this one has me stumped.

[@SixWildKids]

As a side note, I think things are generally working since I see an S21 gain of 24.3 dB at 9 MHz and a power supply current of 12 mA.

[@SixWildKids]

I dragged out my copy of SSDRA and found the answer on page 20. That's where the 26 (they have 25 but close enough) comes from.

[@SixWildKids]

I set the NanoVNA to CW output at 9 MHz and added a 50 ohm load to the output. It surprised me that the output wasn't clipped so I feel like the S11 return loss is probably accurate-ish. The Amplifier seems to clip at frequencies lower than 7 Mhz due to higher higher outputs from the NanoVNA at lower frequencies. I also messed around with the input transformer a bit and got some small improvements to the gain and bandwidth. The VSWR after these changes is 1.47 which is 67+j15 impedance or a return loss of -14.4 dB. Closer to 50 ohms, but still a bit off. The only other term in the input impedance is the Beta AC - maybe it's a bit hotter in the junk box 2N3904 I am using resulting in a higher input impedance than the maths? Probably makes no real difference, but I'm one of those folks who likes to understand the "whys" Sorry for the rambling comments. Trying to work my way through this.

[@SixWildKids]

Found a W2AEW video that shows how to use NanoVNA with IF amps. kzbin.info/www/bejne/bYXXfHhpbrmKpqs

[@4DRC_]

Could you make fractional turns depending on the core geometry used? (Such as around a ferrite rod)

[@CharlieMorrisZL2CTM]

I don't believe so. For toroid's it integers only as it each time the wire passes through the core. As for a rod, I'd assume it's for each full turn. The only way, that I know, to fine tune the inductance is to vary the spacing.

[@scharkalvin]

Yeah you must have left out a capacitor between the base and the transformer, or should have put the junction of r1/r2 at the bottom of the transformer with the top going wto the base. Then would have needed a bypass cap across R2.

[@CharlieMorrisZL2CTM]

Grrrr, so annoying. Yes, the circuit has a 100nF cap from the transformer to the base. I've updated the schematic on the bog, and updated the video description. Thanks!

[@madalinbetea9871]

What about 10k toko transformer ???

[@CharlieMorrisZL2CTM]

Another good idea Madalin. I'll add it to the list.

[@AmateursRadioYi1hxh]

Buongiorno a te

[@user-su5sq5ib3i]

I love you channel!! 73, ve1qet

[@CharlieMorrisZL2CTM]

I'm pleased you find it useful. I'm having fun!