That was a superb video Charlie. Thanks. My Siglent scope is perhaps 5 years older than yours and was about the same UK price - and wow! your FFT function is so much better than mine! Your photography is also to be commended. No shakes and rattles. Very clear. Thanks.
Hi Tony. Thanks for the feedback. That FFT function seemed to work well. I nice addition for the homebrewer. I must admit I do have my eye on a Siglent spectrum analyzer, which I'll probably get. Not that the scope is no good, just that I have the opportunity. While I am not out to build competition rigs, it is nice to see how the various modules perform. All good fun!
Hey Charlie, I would like to see more videos of you designing and homebrewing your own radio designs. I learn a lot from your videos. Please do more. I like seeing you show your work/math and explain it. I like the math (I saw in another video of yours) when a ham explains it and shows their work and brings it to a level where hams can understand. Many hams including me our math has gotten rusty since non-use. I like to hear more about how to design radios. Thanks.
At 11:35, when you increase the input level, the 3MHz leakthrough starts to increase sigificantly. My guess is that's due to imbalances "further up the curve" with higher current flow through the diodes. If so, that suggests that by careful selection of diodes, maybe plotting them and binning them at 1mA first, then re-runing each group at at 200uA, 1mA and say 10mA and sub-grouping them again, you could achieve better preformance. Of course, that's a heck of a lot of extra work and would probably need a much larger group of diodes to select from. I suppose it could be a thermal effect too (or instead), so by simply applying heat from a pinch of your fingers at your original DC measuring/selection stage, you could maybe pick the most stable ones within any similar bin group, assuming that there's any noticeable thermal effect. Conversely, it could be some differences or nonlinearity in the transformer cores too. If that's the case, larger cores that would show less saturation curve effect would be better. (I doubt there's much effect of this at these sort of drive levels though) It's already a pain selecting diodes/xtals etc (ask me how I know!) so I don't want to encourage you to do any extra work, just thought I'd put my hunch out on the internet in cae anyone wants to experiment.
getting ready to start building a new XMTR using some of your design elements, and Ive started watching all of your vids over again. I miss the updates and hope you are doing well! 73
Hi Charlie just completed the mixer using 1N4148 as I already had a pack of 100 !! being a newbie your video's are great the way you explain things are easy to understand hopefully by may I will have the next Level on my licence and can start transmitting on home-brew stuff just listening at the moment on the bit x that i made. great work all the best from terry UK
Another great video Charlie - thank you very much. I'm going to revisit homebrewing some DBMs now. Interesting about removing the 100 Ohm pot as I've previously put these into mine but never tried shorting them out. And what a difference the Schottky diodes made! Thanks again. Keep up the great work :)
Nicely explained Charlie i enjoyed it ,great seeing the display on what is going on with the freq's ,it helps alot well presented , and thanks for sharing with us. 73's.
Very nice. I would do another test where the LO, Signal and IF are not harmonically related as the results here also include the 3rd and 5th harmonic IMD.
Thank you Charlie. Your video answered all of my questions. You explained exactly the things, switched on and off, what are interested to me :-) These 20 minutes were worth it. Many thanks for your effort. 73 de Bernd
Effectively the 100R pot is a change in input impedance, so the effect you notice when shorted is due to mismatching? Interesting the 1N5711s work very well compared to 1N4148 in th e"old" design. As I noted there are two versions of SBL-1, one with split windings (pins 3 & 4) and one without SBL-1x as you have in your demo.
I'm looking to build my own mixers to improve this old HW-7 (w/ added preselector, etc) your videos are immensely valuable! Also please consider at some point coming over to Rumble (new replacement for RU-vid) .
Great Video! I've always wondered how much difference there was between the 1N4148's and the 1N5711's so really enjoyed your deep-dive on that one. Charlie, have you tried/thought-about comparing the effect of using a 15MHz diplexer to terminate the unwanted high-side product? Thanks Again! 73...
I did play around with a diplexer at one stage. My plan is to do it again once this simple SSB rig is out of the way and I revisit the more advanced one. Remind me if I forget.
Wow! super fantastic video. also, you proved the new diodes and got more gain. One thing that keeps coming to mind when I see the diode ring mixer sitting by your SDL1, I wonder if anyone has tried a shield with the ring mixer? I'm just assuming it's a grounded shield that's on the SDL1 of course.
@@CharlieMorrisZL2CTM I was doing a little research on the shielding. at this point, I don't think we need it. it could help later to keep better isolation and interference. I also ran into another mixer that could be fun later. the DDBM that's double double balanced mixer. It's supposed to have better isolation. so many things to play with, it's fun. If I'm making too many comments just let me know. I don't know anyone else I can bounce things off of around here. I don't know one Ham that gets into this near me. I know sometimes you feel like you are rambling on but I'll tell you that that's not how I see it. It's all about the thought and troubleshooting process.
Very interesting Charlie. Just wondering if the 1n914 silicon diode mixer needs a higher level VFO than 7 dbm? They have a higher forward voltage. Just wondering if that was the reason for the lower output signal. Cheers, Ian
@@CharlieMorrisZL2CTM I was thinking the same as Ian: a regular PN junction diode such as the 4148/914 would need a higher drive level than the Schottkys. So, with maybe 10dBm drive, the output might be closer to the other two mixers, and the effect on the VFO might be surprising. Anyway, you've already got very respectable results with the 5711s, and I'll bet your careful work on the transformers has kept the mixer nicely balanced as well, rendering the balancing pot an unnecessary drag on the output. I'm happy I now know that.
I’m curious what the effect and spectral output would look like if you would have set the IF waveform to be a square wave - similar to what an SI5351 would provide. Thanks for your channel!
I wondered the same thing! I think it doesn't really matter because these diode ring mixers are switching mixers (as opposed to using transistors or diodes as non-linear mixers). Switching mixers actually work better with a square wave carrier. There are harmonics (I think those are all those other small frequency spikes you see in the scope), but they will get filtered out by the IF filter. This link helped me think about it: wiki.analog.com/university/courses/electronics/electronics_lab_diode_ring_modulator
Ok Charlie. One of my questions in the newest video comments, you have answered with this one. 😊 Really great. Do you also have compared your homebrew mixer/ifamp with a sa612-kind? You have a fantastic lab! What are you doing there next to your radio hobby? 73
That's good Markus. No, I have not formally compared the homebrew mixers with the SA612 and like. For a start, one is passive and the other is active. I have used both, and both worked well in their respective rigs. As for the test equipment, i don't do anything other than amateur radio. I find what I have works well for me.
Hello Charlie - This is great. I have a question though - is the display output of the FFT linear? I only ask because I'm copying along at home, but displaying the output on a £50 TinySA (which is perfect for me). The output of the Tiny SA is displayed logarithmic and I see much more 'leakage' of the LO and RF signals in the output display of the TinySA than your FFT display. The 'leakage' is about 30 to 40 dB down but it is of course *very* visible on a conventional logarithmic display. So, I wondered if you get a chance to quickly repeat the experiment but using a spectrum analyser with a logarithmic display - just so I can compare my mixer output! All the best Kevin M0XYM
Nice video. However the FFT function of an oscilloscope is not the best instrument for measuring mixers. The dynamic range is only 10, maybe 15 dB (I'm guessing the numbers here from my experience with a signal analyzer). NanoVNA, although less convenient, would be a better choice for the job.
@@4youian I'm pretty sure it also depend at least on the number of samples and the noise floor of the instrument. The problem is that we are dealing with low signal levels here, around -14 dBm for IF signal and around -40 dBm for LO leakage (take a look on the typical spectrum of the DRM under test eax.me/diode-ring-mixer/ ). I have no doubt the oscilloscope has a good dynamic range when dealing with 200 Vpp signals, but this is not the case. The datasheet for the SDS1202X-E doesn't specify the noise floor, but there is a good reason why. You can't get a decent noise floor with a bandwidth of 100+MHz. But there is a simple test we can do to understand the issue - generate -14 dBm and -40 dBm signals and see how they look like on the oscilloscope. Now think about how many bits are actually used for these signals. Simply put there are a few factors beside the number of bits to consider.