Home brew T/R relay switch for Anan 10 pure signal

[W3RSW]

If I understand your last post, you have added the internal T/R switch as a precaution, in case K19 fails to short the K-19 end of the mini-cable on transmit. However, at this point, you believe that K19 is working as it should.

That's an external T/R switch I've added.    and yes K-19 is working,
and it'll take some time to digest all your remaining. Your right that dB differences, dBm, conversion to Watts, that and wondering how much same confusion, among other things are going on in the literature.  Pratt, et. al. somehow got a beautiful system working.

Hey, heard you on 40 a couple of days ago.  Wanna try 7210, open now?
I guess your at work, not a weekend come to think of it.

Later: Also you have 5 or 6dBm, (4 watts if CW) being dissipated across your 50 ohm resistor since the other end is essentially a short to ground with -20dB down left over in your hookup.  How can 3 inches of internal coax in my setup equal a 50 ohm resistor plus the add'l coax in yours?  Yeah, I know it's at RF and at freq.F  but the real component is 50 ohms.  If the j5 input is 50 ohms already then your 50 ohm resistor is a voltage divider?  And yes, I know your attenuation train takes this into account.  Still thinking. I'm on it.  Might try your hookup and see what transpires. Need a 50 ohm N.I.Res. and will put it in a BNC cylinder.

3:45pm  ok so I'm not so worried about your hookup but for mine on the Hermes board I have 5 or 6 dBm coming into Rx1 at what I assumed is already a 50 ohm input with a 1 megohm static drain and back to back protection diode to ground stack, then to the DAT 31-8P step attenuator, then to FL6 filter, then to always on LTC-6400 Preamp (enable grounded) then finally to the LTC-2208 ADC.  Granted that the 5 or 6dBm is normally less than 50% duty cycle, I'm looking at about 2 watts ave. +/- to be dissipated in the attenuator with worse already bypassed by the diodes.

I know this has been answered before about the ability of the dissipation resistors to handle pure signal, just have to find it.  You probably already touched on that in the many posts to the Yahoo site.
Here's one in the ballpark.
http://lists.openhpsdr.org/pipermail/hpsdr-openhpsdr.org/2015-May/071232.html

[W3RSW]

Disregard much of previous post, like you haven't already.    Only my mind Burned out from over thinking .  Nothing coming back to J5 on transmit is going harm anything since essentially all but a miniscule amount is shorted out at K19 as per pix you've used before.

And yes S-ATT has to really be pushed to show 2 db when I run the Alpha up to 900 plus watts and finally get S-ATT off zero.  To do that though, the Anan 10 nominal watt amp, 15 to 20 actual on 40 meters runs really ratty IM products.  So I've backed down to the 600 to 800 range and settled for restricted pure signal range even though it seems happy and in the green.

[AB2EZ]

Rick

Okay... I'm back from the opera at Lincoln Center with my wife and her two sisters

Next try:

a. remove the external 10dB series attenuator

b. click on 2-tone test

c. click on single calibrate

d. wait a second or two for the step attenuator to auto-adjust to the 10dB larger feedback signal

What the S-ATT read?

Stu

[W3RSW]

Well, very interesting.
1.) forgot to set S-ATT at 31 before so that pure signal could act on it.
2.) two tone set at 0.00 dB
3.) 10 dB att. removed from 43dB forward output.
So:
f=7180kHz, Po Anan around 11, Po LP-100a 12.1, Po Alpha 780, Pure sig. on, S-ATT @ 5, green level
                                        Lp100a at about 12.8       Alpha 820        S-ATT 5 to 9 , blue blinking, action erratic on panadapter.

Backed down Anan drive about 1/2 watt             Alpha 790      and got S-ATT 7, green steady.
    
Note the Alpha tune an load controls didn't change much. Peaked on scope w/o pure signal to start for sharpest 2 tone peaks.  

On pan , 2 tone peaks register at -12, first IM3's at -50 with pure signal off.
                                                 -12                     -65 with pure signal on.

Apparently I'd gone over into the 'blue' before and didn't realize it was so critical.  Best I'm getting seems to be an S-ATT of about 7.     Have not tried on other bands yet.  I may try your 50 ohm direct feed system next, realizing that it's freq. dependent.                          

[AB2EZ]

Rick

There appears to be a point of ongoing confusion.

If you are using Auto Attenuate, the Pure Signal "Feedback Level" indicator should not be flashing blue, and... at the same time... "S-ATT" (in the main window) reading in the single digits. Something is probably wrong with the steps you are taking after you increase (or decrease) the PEP feedback power level.

Make sure "Auto-Attenuate" is checked, and never uncheck it.
You should never have to manually set "S-ATT" to -31dB or any other value.

When you increase the PEP feedback power level (by increasing the amplifier PEP output power, or by removing the 10dB series attenuator from the feedback path):

A. Make sure that the 2-tone generator is not active. Turn off "Pure Signal" (I.e. click on the "OFF" button in the "Pure Signal" window).
B. Click on "TUN" in the main window, to turn on the transmitter with full PEP output from the amplifier. Leave "TUN" active for a few seconds.
C. Click on "TUN" again to turn off the transmitter (make sure the ANAN-10 is no longer in transmit mode... by verifying that both the "TUN" and "MOX" indicators in the main window are not illuminated)
D. Click on "Two-tone Gen" in the Pure Signal window to turn on the 2-tone generator
E. Click on "Single Calibrate" in the Pure Signal window

The Pure Signal "Feedback Level" indicator may flash blue for a second or two... but it should turn green when the step attenuator adds in additional attenuation to compensate for the increased feedback PEP power level. The S-ATT reading should also increase (from what it was before you changed the feedback level) to reflect the additional attenuation of the step attenuator.

Stu

[AB2EZ]

Deleted duplicate post

[W3RSW]

I followed your steps exactly, no confusion.
Still kicks back to 5db ATT on xmit.

When Anan/Alpha run up to 900 watts, action same as before, didn't stop blue blinking and S-ATT xmit literally wandered from 5 to 9 with drop outs on screen pep readings but program format and GUI background stayed stable.

wonder now if it's RF feedback.

Looked better when 10 dB attn. was in line and auto pure signal run. fwiw.
2nd chart and notice it has phase zoom.

[AB2EZ]

Rick

One thing is clear to me:

It's not working as it is supposed to work.

What is wrong.... I don't know.

If it were my equipment:

I would start by removing the external T/R switch (in case it is a source of excessive crosstalk from the ANAN's output to the RX input, or some other problem). You can always add it back in later... to see what affect it has on the behavior of Pure Signal, if any.

I would verify that the 43dB coupler is installed properly.

I would not use the 10dB supplementary series attenuator

I would include the 50 ohm series resistor between the coupler output port and the ANAN RX input port (located close to the RX SMA connector). This is important to ensure that the coupler output port is feeding a 50 ohm resistive load, instead of an inductive load.

50 ohms in series with j5 is very close to 50 ohms. By placing this 50 ohm resistor close to the ANAN RX port, you can have a long coaxial cable between the coupler output port and this resistor (feeding the ANAN RX port)... and the coupler output port will still be looking into a 50 ohm load.


The absence of this 50 ohm series resistor may be the principal cause of the problems you are observing.

Stu

[AB2EZ]

Rick


addendum:

The absence of the 50 ohm series resistor, added near the RX input, may explain why things work somewhat better when you use the 10dB series attenuator. The 10dB series attenuator, feeding the RX input port, presents a load to the 43dB coupler that looks more like the 50 ohm resistive load that the coupler output port is designed to look into (via its coaxial cable connections).

Stu

[KD6VXI]

A thought:

A common mode choke at the I put of the Anan RX port?  May have common mode on the outside of coax.

I'm going through that now with my mod monitor and associated things in my shack.   Not sure how or if that small loop on the outside of the coax could or would interfere,  but a thought.

--Shane
KD6VXI

[W3RSW]

Yes I thought of using your resistor option in earlier posting too. If small enough or right form factor , stick it in a BNC barrel.  What are you using for a non inductive resistor, Mouser no? Or equivalent.

And the idea of a common mode choke can't hurt either.

In the meantime I have a clean signal even without pure signal, and somewhat better with.

[AB2EZ]

Rick

1kW => 50 milliwatts of RF power flowing out of the -43dB coupler output port (if the port is looking into a 50 ohm load, as it is designed to). On 40m, the power coupled into the ANAN-10 receiver's step attenuator (due to the 20dB  of voltage divider loss) will be 0.5mW (-3dBm). The rest of the 50mW will be dissipated in the added 50 ohm series resistor.

If you switch (at some future time) to a -30dB coupler, then the power leaving the coupler output port will be 1 watt. Of that, 10mW (+10dBm) will be coupled into the ANAN-10 receiver's step attenuator, and the remainder will be dissipated in the added 50 ohm series resistor.

Either way, a pair of 100 ohm, half watt carbon film, or metal film, or carbon resistors in parallel would suffice for the added 50 ohm series resistor.

I used a coaxial jumper cable with an SMA connector at one end. At the SMA connector end, I cut the jumper cable to splice in the 50 ohm series resistor. I used a piece of heat shrink tubing over the sliced-in resistor. I used a piece of insulated hookup wire to restore the continuity of the shield across the spliced-in resistor.

Remember, you must install the 50 ohm series resistor close to the SMA connector. (I.e. close to the ANAN-10 RX port).  

I agree that Shane's suggestion of adding the common mode choke is a good one. I plan to add one. However, you still need to add the 50 ohm series resistor.

Stu

P.S. I just checked. I already have a ferrite core around the coaxial cable leading to my ANAN-10E's RX SMA port. I usually follow the same practice that Shane recommends... and I did this a year ago, when I set up the Pure Signal feedback path.

[AB2EZ]

I did some thinking to refresh my memory with respect to how a 43dB coupler might behave if the sample output port is not terminated in the 50 ohm load it is designed to look into.

Depending upon the design of the coupler, the result could vary over a wide range.

For some designs, the coupler would behave as a current source (i.e. having a high source impedance compared to 50 ohms). An example would be a 10:1 current transformer, terminated in a 50 ohm load, and with one end of the 10-turn output winding also feeding a 680 ohm series resistor... with the output taken between the other end of the 680 ohm series resistor and ground. Note that 20log(50/730) = -23dB. If a current source, that has a 680 ohm source impedance feeds a j5 ohm load, or if it feeds a j5 ohm load in series with 50 ohms... the voltage produced across the j5 ohm load will be the same. If a current source, that has a 680 source impedance, feeds a coaxial cable that is terminated in j5 ohms, the voltage produced across the j5 ohm load could be lower (less current flowing through the j5 ohm load) or higher (more current flowing through the j5 ohm load), or about the same... depending upon the length of the coaxial cable and the wavelength of operation.

For some designs, the coupler would behave as a voltage source (i.e. having a low source impedance compared to 50 ohms).  An example would be a 10:1 current transformer, with the output winding terminated with 50 ohm load... and with the 50 ohm load also attached to the input winding of a 14:1 step down voltage transformer. In that case, the voltage produced across a j5 ohm load would be larger (as much as 10x larger) than the voltage produced across a j5 load in series with a 50 ohm resistor... depending upon the length of the coaxial cable connecting the output of the step down transformer to the j5 load.

The above... while probably more than most readers would want to know... illustrates the importance of terminating the sample output port of an RF coupler with the 50 ohm load it was designed to look into.

Stu

[W3RSW]

Well, for your edification and delight.

I opened up my Xtronic XDC-1 Directional coupler , a sort of Amateur gold standard with these specs.:
1 to 60 Mhz, -44Db +/- 0.1 dB  (we've been saying 43 dB all along so there's another dB   )
insertion loss < 0.04 dB
Return loss < 30 dB
Serial no. 234

And guess what-    -- a derivative of your next to last paragraph.
Both the forward line and return line have 49.9 ohm  1 % resistors, (yes, five banded babies) soldered from the last multiple toroid links directly to the BNC output jacks.  Yes in series, nothing else coupling to the BNC's except those resistors and I suppose the internal cavity dimensions if it were a microwave device. There are three toroids inside, including the usual one around the coaxial cylinder running direct from Alpha to Ant. So-238's.  Amazing compensation I suppose.  Last one is very carefully built with an inserted brass tube, one turn link, either side of which goes to the 49.9 ohm resistors, thence to the forward and reverse BNC ports.  Hey even the cab. screws are stainless into captive nuts. -very well built.

Furthermore, I've installed the parallel, non-inductive, 100 ohm 1/2 watt resistors which I happened to have before as a BNC QRP load in a cylinder with BNC (back to the coupler) and SMA (direct to Rx , J5) with the resistors C/L and grounding shell no more that 1 1/2 inches back from Rx, duplicating your and others' direct to Rx lash up.  My outboard Ry and Add'l 10 dB attn. are gone.

Regardless of input and output now, ranging from 600 to 850 watts output, I'm still getting S-ATT of 3 dB. , steady, no wandering  - But the IM suppression is still beautiful. Yesterday's hash in some of the later runs may have been from loosening sampling line connection since that was all re-installed today.

I've seen G7CNF's youtube demo of running pure signal back through an Anan 100 with a clearly shown 40 dB, 500 watt coupler (with his reverse connection left unterminated no less).  So here I am running a 44 dB sampler back from an 800 watt amp, 10 times his 80 watts.  Mine appears to work just as well,  Comment ca?  
https://www.youtube.com/watch?v=0adHZOTqTlQ

Today's typical example; Anan 10 set at "60"  Tune yields 13 watts Anan, 12.5 Lp-100A
2 tone w/o pure signal yields 680 pep Alpha out , a tone at -16.4 dB on pan, IM3 at -45
2 tone with pure signal yields -16.4 dB, IM3 at -70.  From an already clean amplifier. As already mentioned S-ATT still at 3 dB.

Thought. Since the 50 ohm resistors are already internal in the sampler, I could mount the Anan right next to the forward coupler output and omit the add'l 50 ohm that I just built.  Probably too close to the Alpha for a variety of reasons though.

  

[AB2EZ]

Rick

It appears that things are working as they should (more or less... see a few additional comments below)

With 680 watts PEP at the output of the coupler (+58dBm), the signal level at the ANAN-10 RX port should be:

58dBm - 44dB (the coupling loss of the coupler, if it is looking into the 50 ohm load it is designed to look into) - 20dB (the loss of the voltage divider formed by the 50 ohm series resistor, located near the ANAN RX input port, and the j5 ohm input impedance to ground looking into the ANAN-10 RX input port, when transmitting) = 58dBm - 64dB = -6dBm.

As per my original post to this thread, this feedback level is too low to ensure that Pure Signal is really reducing the IMD at the output of the amplifier by as much as the panadaptor indicates... and not just reducing the IMD of the signal seen by the ANAN-10's receiver... which  consists of the sum of the feedback from the output coupler and the crosstalk from signals that precede the final amplifier's input.

The Pure Signal auto attenuate process will try to adjust the output of the step attenuator (when transmitting) to around -13dBm. At least, that's how it is working in my ANAN-10E. Some believe that the target for the output of the attenuator is -17dBm... but that's not what I am seeing with my ANAN-10E.

Therefore, the step attenuator should be reading 7dB.

You are seeing a reading of 3dB on the step attenuator... which is in the ballpark of what it should be... but lower than I would expect.

Additional comments:

You mentioned that the 50 ohm resistor you added was originally part of a QRP (dummy?) load. Since a dummy load would normally consist of a 50 ohm resistor from center conductor to shield/ground... I wonder if you are adding a 50 ohm resistor (located near the ANAN-10 RX port) from center conductor to ground. If so, that's not what is needed. The 50 ohm resistor needs to be added in series with the center conductor of the coaxial cable.

Even though the coupler has a 50 ohm resistor feeding each of the coupler's BNC output ports, each of the the coupler's output ports  still has to look into a 50 ohm load... if the coupler is going to work as designed. If each output port is not looking into a 50 ohm load, the directionality will suffer significantly (it is a directional coupler)... and the forward sample port output power will not be 44dB below the forward input power. The series combination of j5 ohms and 50 ohms... from center conductor to ground... provides a load across the output of the coupler that is very close to 50 ohms. This also allows the coupler sample port to be connected to the ANAN with a length of coaxial cable... without messing up the impedance that the coupler's sample output port is looking into.

Note that an ANAN-100 (or an ANAN-100D or an ANAN-200D) does not have a j5 ohm input impedance to ground, looking into the feedback input port on transmit,  as is present at the RX port of the ANAN-10 or ANAN-10E. The impedance looking into the feedback input port, with those transceivers,  is 50 ohms. There is no 20dB voltage divider loss in those transceivers.

Stu

[W3RSW]

50 ohm R's not grounded. In series with sample line and parallel with each other. -they and the BNC connector just borrowed from the qRP load and re-purposed, slight rewiring and added an SMA.

And thanks for the difference explanation of the 100 vs. 10 / 10e RX on Xmit attenuations.  Makes sense and explains 20 dB.

[W3RSW]

I have now hooked back up the TAPR derived T/R outboard and disconnected the Anan 10 internal coax jumper between Hermes j29 and the Apollo PA j3.

In my attached pix, Hermes J29 coax plug is just left of spkr. out jack in rear of Anan 10.

With no 10dB additional attenuator on pure signal feed back output of 44dB sampler hooked to Alpha 89 output I'm looking at the following stats:

(Note that quite a lot varies with the load and tune settings of the Alpha. Pure signal can be adjusted with the PwrSDR pan on the fly for best products with the Alpha load, tune while watching grid current.
The longer I play with  testing power the more the thermal  rise in the Apollo transistors affect the correction as advertised.  I suspect SSB will affect pure signal less than AM from that thought.)

7290 KHz, LSB
drive    S attn                    Anan Pout  Lp100A Pout    Alpha Pout(pep)  

70       10         two tone        14         12.6               Off, Anan only
70       30       single tone       14         12.7               790                    
70       26         two tone      varies      12.6               709                    

Sometimes pure sig. max's out at 32 dB SAttn and pan signal shows slight oscillation.   May put 10 dB atten. back in sample line when running Alpha.
RF feedback experienced before so will add clamp on choke that I had on Stu's 50 ohm resistor input or track it down since I'm feeding a dummy load.

This is what the Anan 10 is now supposed to be with pure signal.
 44dB @ constant freq. sampler feeding back to now only the Anan's declining amp./freq. factor as the main variable to fight.  

(Data edited to clean up formatting , omitted Alpha tune and load settings.)

[w5tts]

So in a nutshell what was the best solution to use pure signal with the anan 10-e? I have the ac2iq coupler, various inline sma attenuators, spare sma female connectors if I need to add one etc. I use a homebrew amplifier (2xblf188xr) that does around 1200-1500 pep.

[W3RSW]

Over a year later; sorry but another asked about the same question lately.
Stu, AB2EX has the best solution.

I reversed my last noted Anan 10 internal jumper modification and put it back to stock condition,
and now have no external relay lashup that was built for thread's beginning.

So in a stock Anan 10 or 10e:

Simply put a 50 ohm, shielded resistor, (2 ea. 100's 1/2 watt in staggard parallel, for small ID) close to the Rx port on the 10e and then place attenuator in line as needed (if any) along with your ac2iq sampler to get the needed 7dB or so action required from the auto step attenuator in pure signal.

Amp. sampler coax line >SMA---50--->very short length to BNC Anan 10 Rx port.
 
Also clip on a small coaxial barrel ferrite on the 1 or 2 inch coax stub between the 50 ohm resistor to the BNC Rx port if you need it.  (My lash up is a SMA to the resistors to a BNC  pigtail with tape around the resistors and clamped over with the clamshell ferrite. total about 3 in. long.)