| Electronics > Projects, Designs, and Technical Stuff |
| A "Poor Man's Adjustable Delay Line" |
| << < (4/4) |
| dnessett:
--- Quote from: tomato on August 09, 2018, 09:10:38 pm ---Really, not at all curious? The two graphs are essentially identical, i.e. every little feature in Figure 7 appears in Figure 8. That just screams out something is wrong with one or both of the measurements. --- End quote --- Not curious enough to tear the delay device apart. The figure in this post shows the results of a 9 KHz - 20 MHz sweep with all delay coaxes switched in and with the ferrites in place, which supersedes Figure 8. The following image shows the results of a 9 KHz - 20 MHz sweep with all delay coaxes switched out and with the ferrites in place, which supersedes Figure 7. Notice that the two spectra look pretty much the same, except the 10 MHz marker indicates -2.3 dBm when all delay coaxes are switched in and -.4 dBm when all coaxes are switched out. Even if there was a measurement error when Figures 7 and 8 were generated, it is no longer germane. If you build the delay device described, use snap-on ferrites to suppress cross-talk on the coiled coax braided sheaths. |
| tomato:
--- Quote from: dnessett on August 09, 2018, 11:08:52 pm --- The figure in this post shows the results of a 9 KHz - 20 MHz sweep with all delay coaxes switched in and with the ferrites in place, which supersedes Figure 8. The following image shows the results of a 9 KHz - 20 MHz sweep with all delay coaxes switched out and with the ferrites in place, which supersedes Figure 7. --- End quote --- I'm going to assume you meant to state this graph is with the coax switched out, in which case you now have two good graphs and all is good. |
| Gerhard_dk4xp:
This |O board software cancels the entire post just if a picture is 100 KB too large. Therefore, much shorter: The digital delay lines are a bad idea for phase noise measurement because they are really long strings of capacitively loaded inverters with a tap now and then. I'm also making such a delay line for 5 MHz to some 100 MHz. Since I did not want a lot of intermittent relay contacts in series, I use 3 pairs of 1:6 coax relays. 1 pair for coarse, medium and fine. Still working on mechanics. These relays are cheap on ham flea markets, everybody needs only 1:2 relays. Some relays have TTL drivers; these are usually defunct since they do not like the 28V on their inputs. That paved their way to the flea market. We don't need the drivers, we can power the coils directly. I must admit that the open relays don't look that precious anymore. regards, Gerhard |
| profdc9:
A suggestion for this approach... You may want to consider using signal relays like the Axicom d2n that are intended for high frequency switching / RF. This might give you better results. I have a similar project here that uses variable lengths of coax to adjust the delay https://github.com/profdc9/ModularTuner/tree/master/Tuner/DelayLine for a phase shifter to be used in a phased array antenna I am working on that uses such relays. A rendering of the PCB is below. If you want really high frequency use, you probably need to use genuine coaxial relays like https://www.rfparts.com/relays/relays-tohtsu.html but these are very expensive and likely overkill for what you want to do. Dan |
| Navigation |
| Message Index |
| Previous page |