2021 ~ A Shack Mystery Event
A WhoDunnit Shack Mystery...
May 10, 2021 ~ A Clue to Who!
Late Breaking News. I had a standard 40M LPF from Communications Concepts Inc. and swapped out the installed board with the CCI Board. That seems to have fixed the 40M issue. I was able to have the intermediate amplifier and SB-200 in line without any oscillations. All appears well! The bottom line -- A $1.30 SM Cap was the culprit.
As so often happens to me when a problem comes up, my brain goes into the fog zone and our analysis process can get confused.
So in looking back I remembered that with one of the QRP rigs that was on 20 Meters and connected to the intermediate amp -- there was not an issue. So I went back to testing other rigs connected to the intermediate amp on 20 Meters -- no feedback.
I should mention that I installed a "snoop loop" of wire through the last iron powdered core before the antenna on the LPF and monitor the signal on the scope. A clean signal is evident by the pattern displayed. If there is feedback into the rig -- the scope would show that.
So what would cause the feedback -- well one thing for sure would be a high SWR condition on the output of the amp being reflected back to the RUT (Rig Under Test). So this leads to questions if the amp works OK on some bands like 20M but not on 40M then the issue might be specific to the LPF in use.
My intermediate amp uses the CCI AN762 Amp board; but is band switched using a home fabricated band switch. I actually took two band switches and connected them with a common shaft that separates the inputs and outputs of the LPF's. This was done to separate the In's and Out's to eliminate any possibility of feedback. The photo below shows one of the filter boards and the home fabricated switch.
This is the actual board as the one above is from a prior time. It was at the same location, so just a pop out and install a new one.
The filter boards are combinations of filters bought from CCI and ones I fabricated. The ones built by N6QW follow the W3NQN template and for 40 Meters the front end and back end are 430 PF with the intermediary caps being 660 PF (two 330 PF's in parallel) the the shunt cap is 100 PF.
I proceeded to uninstall the board and noted that one of the 330PF caps looked slightly discolored. Could this be the issue? I am trying to noodle the two cases where the cap is Open where effectively this element is only 330 PF and not 660 PF. The discoloration might suggest it is shorted wherein the RF input from the amp itself is being shorted to ground. But then that would signal no output. from the amp. Possibly a 3rd case --it was shorted and then opened!
But certainly this opens up a new line of investigation. As it turns out the 430 PF are rated at 500 Volts and the 330 PFs rated only at 300 volts which puts them in a marginal state. I have ordered 4 new 330PF 500 Volt SM caps and will replace all four. That may not cure the problem but at least the filter will be more robust.
The amp can put out 150 watts so there is a lot of circulating current in the filter elements. Below is how I sandwiched the 4 filter elements into the underside of the chassis.
I have often complemented myself on having fabricated this band switch. There was a lot of cut and try and the mechanical coupling is critical using some robust coupler's to maintain the proper clocking! This may offer some ideas to homebrew switch builder's.
If this is the problem resolution -- this sure puts a damper on acquiring one of the new MRF300 Linear Amp boards. Can you imagine 5 watts in and 600 Watts out across 1.8 to 54 MHz? A complete 600 Watt MRF300 amp less the LPF's is about a dollar a watt.
May 7, 2021 ~ It has gotten worse --not better.
The mystery has deepened! But first I have received three inputs on the cause/resolution:
- The very 1st input was that because I was attempting to take ostensibly QRP radios and turn them into 600 Watt powerhouses, a group of QRP purists has made a clandestine raid on my shack and literally poured sugar into the gas tank!
- A second input was to remove all extraneous wires and switches and take the most direct path so that you are establishing a base line. That was indeed a great suggestion
- A third put was that there was indeed a "commando style" SWAT type raid on my shack and the works were "buggered".
My very first inclination was that somewhere in the tangle of wires and I had a coax problem such that on low power the problem was there but masked. However when you put juice to the wire -- big time issues. So I installed all new coax cable in various lengths ( 6 foot, 3 foot and 1.5 feet) and that did not fix the problem! Although I did note that some of the braid looked a bit "hinky" so the coax fix was a good idea anyway as it was only a matter of time.
Something I should have thought of at the outset but did not do was to test the setup using a 100 watt commercial transceiver as this approach had some critical factors that could be evaluated.
- Was it an issue with my intermediate amplifiers?
- Was it something in my homebrew transceivers?
- Interesting note --it is definitely a RF Feedback issue as the Arduino locks up and the display goes blank.
- If the system works with the commercial rig in line -- then that proves the integrity of the new coax and my SB-200 amp
Well putting the ICOM IC-735 in line and tuning up barefoot on 40M -- works perfectly (keep in mind new coax interconnects). Powering up the SB-200 being driven by the IC-735 -- perfect! No trace of RF Feedback and clean signal reports.
So indeed did the Commando Strike Force visit my shack? Noteworthy all of these homebrew radios worked previously with no issues! The mystery has just gotten more mysterious!
So OK, theere are no limp bloody bodies lying on the shack floor. The Safe Door is not ajar ( I don't even have a safe in the shack.) The phone has not been hacked --or at least I don't think so and there are no SWAT cars in the driveway.
But there is a mystery and it had (has) me scratching my head. So let me set the stage for you and see what you think.
In my quest to build a Spillsbury Tindal "Look Alike" SSB transceiver I began looking at some built rigs from the past to accomplish the following:
- To look for the 'best in class" features of construction aspects. You know "that was a really slick working Mixer stage".
- To look at past mistakes like -- man I should have done that differently.
- To evaluate if simply repackaging a built unit would fit nicely in my selected case
- Look at the physical layouts to achieve a compact layout.
There is another critical evaluation factor and that is performance. Some of my built rigs seemed to work better than others. That clearly is traced to the design topology. Do we use NE602's or Plessey Amps? Is it better to have bilateral stages or a single pass relay steered approach? Those choices do affect performance.
So I thought it best to develop some sort of crude evaluation criteria such as
- Internally generated noise
- Power Output
- Physical Size
- Signal Reports
- Transit time from transmit to receive
- Transmit artifacts
- Minimum circuits elements for the maximum performance
Thus with my crude list I started a process of placing the five or six rigs on the air. Typically my rigs are built with IRF510's in the output stage and so we see somewhere in the range of 5 watts output. I usually follow my rigs with a linear amp stage that is good for 100 watts.
This homebrew amp uses the AN762 RF board from CCI to which I added a band switch and 4 Low Pass Filters (some from CCI some Homebrew) which were modified to the W3NQN LPF template.
I have had this amp for about 15 years. There is also a second amp that uses the RF Brick from an Atlas 210X which has been fitted with a 40M LPF. I have had this amp for about 8 years.
So now the mystery part. I picked the first test rig and connected that to the antenna barefoot and placed it on the air. It loaded FB and the SWR was perfect. I could hear many stations and gave a quick CQ and worked one station at the QRP Power Level and of course the received signal on the other end was marginal so I flipped on the amp stage.
Boom the Power Meter jumped up with output; but I noted that the meter was acting up (sort of hanging on with voice peaks). The station at the other end reported RF Feedback on the signal. Stop the Press!
Into a dummy load, all sounded good and the power meter had more of a normal response. So then I tried a second rig firstly into a dummy load and then into the antenna. It appears that the same sort of event was occurring with the amp and with the antenna.
So with all 5 rigs tested they exhibit the same symptoms:
- OK at QRP into the antenna
- OK with the Amp into a Dummy Load
- Not OK with the Amp and Antenna connected.
- The Burning Question: Why apparently OK at QRP but not with juice to the wire? Or that it is indeed present but because the wimpy signal strength it is not appparent. I suspect that may be the case.
So then I connected the set up to the SB200 and OK into dummy load --but 600 watts of oscillation when connected to an antenna. Not good!
Since I had a second linear amp stage (Atlas 210X Brick) I repeated the tests with 5 rigs and it was the same result.
So time to collect data and test results as the mystery deepens.
- The Rigs work OK at QRP into the antenna. Maybe not really validated but only looking at the SWR bridge
- The Rigs work OK with either Amp into a Dummy load
- The rigs Do Not Work OK with either Amp into the Antenna
As an added test I tried three different power supplies with the rigs -- just to eliminate any possibility of RF Feedback through the power supplies. This had no effect!
This is a time for the Classic Statement...
SOMETHING HAS CHANGED! But WHAT?
My set up for the station is as follows.
- I have a two position switch with two three foot pieces of RG8X cable . On piece of cable has a standard PL259 on one end and the second cable has a RCA Plug that can be fitted with a RCA to BNC.
- From that switch a three foot chunk of RG8X to the CCI amp (or Atlas 210X) Stage.
- . The output goes to a SWR bridge. This arrangement lets me look at the SWR of just the rig connected or with the amp to the antenna.
- Following on to this is a chunk of RG8X to the SB200 Amp which when that amp is off is simply routed through the amp the a Drake W4 wattmeter.
- From the wattmeter the signal is routed to my homebrew 2.5 KW T Type Tuner.
- There is actually another switch in line following the W4.. That switch enables me to shunt the output into a 2 KW dummy load or to the antenna tuner.
So we actually have two SWR bridges in line and two switches with one to accommodate various rig output connectors and the second to select either a dummy load or the antenna.
So some initial conclusions -- the problem occurs with literally any rig and with either intermediate amp stage. Thus the mystery.
What could be the issue that has changed? Since with a Dummy Load we appear OK, Is there an antenna feedline problem? From my tuner the output is 75 feet of RG8 Coax to a 9:1 BALUN and then to about 40 feet of 450 Ohm line to the Delta Loop.
So now the mystery of why it is happening and what is the solution to whodunnit? I think I have the answer and proved that solution this morning. But what do you think is the reason?