Directional Coupler - SWR Issues - Calibration?

[George Smart (M1GEO)]

Some time ago my HL2 got itself very confused I had to restore from broken gateware. At the time, I was happy to have recovered the radio. I mostly use the radio for receiving, but lately have found myself wanting to have a few QSOs (lockdown fever?) and what better radio to experiment with.

However, I notice that the SWR reading on the radio is totally weird. At low power, it seems that things are behaving normally. For example, I can see discontinuities in the reverse power detection between drive value 214 and 215. Looking at the graph with the best fit curve, you see that the forward power always falls on the line, and the reverse power does not. It also seems to report an SWR of around 1.6:1, but my Keysight VNA shows the VSWR to be around 1.07:1 (certainly less than 1.1:1). I've swapped out all cables, connectors, etc., and I'm sure it isn't anything like that. I checked the power output with a 10W HF insert on a Bird Thruline and I can see around 5.7W forward power and no detectable reflected (not even a needle twitch).

Any advice for what to look at next would be appreciated. I guess it could be the SWR calibration is off - maybe overwritten when the gateware went weird some months ago. Perhaps a physical defect on the board? Poor joint? Faulty diode?

Thanks,

George, M1GEO.

-- 

Dr. George Smart.

https://www.m1geo.com

[Alan Hopper]

Hi George,

where are you reading the power levels from? Quisk and Spark use a calibration table to convert adc value to watts, in spark it is stored in the settings directory for each radio and is called FwdPwrCalibration, the same table is used for fwd and rev so If it were corrupted I'm not sure it would cause your problem but deleting the radio file (hpsdr_yourmac.json) will put the defaults in on restarting spark.  Not sure how other sw handles the calibration.

73 Alan M0NNB

[Matthew]

The power values are raw ADC bits that come via the protocol, software then turns this into a power reading. Are your reported power values from some software? If so, which software?

Are you able to measure the voltages on the board to cross check?

73 Matthew M5EVT.

[George Smart (M1GEO)]

They were from Alan's SparkSDR. I hadn't realised that the ADC value was converted on the PC, otherwise I would have plotted ADC data directly.

I deleted the JSON file, and re-setup the radio and no luck. I also observe the behaviour in Quisk, so I assume it's a hardware issue, as in the ADC isn't supplying the correct value.

The FwdPwrCalibration JSON structure looks like the following: 

"FwdPwrCalibration": [

    {

      "inp": 0.0,

      "outp": 0.0

    },

    {

      "inp": 25.865385055541992,

      "outp": 0.002550253877416253

    },

    {

      "inp": 101.0245361328125,

      "outp": 0.012752044945955276

    },

    {

      "inp": 265.2901306152344,

      "outp": 0.05060093104839325

    },

    {

      "inp": 647.9155883789062,

      "outp": 0.2164583057165146

    },

    {

      "inp": 1196.593505859375,

      "outp": 0.6654804944992065

    },

    {

      "inp": 1603.7032470703125,

      "outp": 1.155722975730896

    },

    {

      "inp": 2012.3271484375,

      "outp": 1.8118921518325806

    },

    {

      "inp": 2616.772705078125,

      "outp": 3.008584976196289

    },

    {

      "inp": 3173.818115234375,

      "outp": 4.39274263381958

    },

    {

      "inp": 3382.792236328125,

      "outp": 4.979132652282715

    },

    {

      "inp": 3721.071533203125,

      "outp": 6.024750709533691

    },

    {

      "inp": 4093.178466796875,

      "outp": 7.289948463439941

    },

    {

      "inp": 4502.49658203125,

      "outp": 8.82083797454834

    },

    {

      "inp": 4952.74609375,

      "outp": 10.673213958740234

    }

  ],

[Alan Hopper]

Hi Group,

this has reminded me that I had always meant to have separate tables for forward and reverse, if anyone has a need for this please shout and I'll move it up the list.

73 Alan M0NNB

[Steve Haynal]

Hi George,

The AD9866 TX programmable gain control is in dB, not Watts, ranging from 0 to -7.5 dB in 0.5 dB increments. If that is what SparkSDR's PA drive is controlling, then you'd expect the plot you show. I suspect it will be a straight line if plotted using a log scale.

73,

Steve

kf7o

[George Smart (M1GEO)]

Thanks, Steve.

I'm not sure exactly what Alan controls with the 0-255 drive level; 0 to -7.5dB is only 15 steps; I assume Alan is also controlling the waveform amplitude to get extra linearity. Given that a natural log best fit line fits perfectly with the forward power, and that the transmitted power matches that I see with Bird Thruline, I suspect this is not the issue.

The reverse power seems nonsense, so I suspect it's a hardware fault? Any other tests I could do before taking my screwdriver/scope to it ;)

Thanks,

George M1GEO

[George Smart (M1GEO)]

For the record, this ended up being a hardware issue. it was a poor connection on the jumper bridging the main board to the SDR. I rebuilt the jumper with a bit of veroboard, and it's all good :)

George M1GEO.

[Steve Haynal]

Hi George,

That is good to know. We can direct people to the jumper board if others have problems with reverse power measurement.

73,

Steve

kf7o

[George Smart (M1GEO)]

This turned out to be a hardware issue - a poor connection between the N2ADR filter board and the Hermes Lite 2 mainboard.

George M1GEO

[iw7dle]

Dear George and All,

nice to meet you, this is Mario IW7DLE.

I recently bought an HL2 that has exactly the same issue you described in this post. With slider at 0db, the real PWR out is 5.1W (measured with external meter and dummy load) but Thetis is only showing 3W and wrong SWR.

The only way I found to show real power in the software is to manually recalibrate the values (see following screenshot) but SWR reading is still wrong.. don't know if there's a way to recalibrate that value.

I checked all the connections between NADR filter and main board, they all seem to be fine. Any advice on that?

Thanks a lot in advance for your answer.

Mario IW7DLE

[Steve Haynal]

Hi Mario,

Are you using the version of Thetis specific for the Hermes-Lite 2? From you screenshot it looks like you are using PowerSDR. You can check with other software (Quisk, SparkSDR, SDR Console) to determine if this is a software or hardware issue.

73,

Steve

kf7o

[iw7dle]

Dear Steve,

thanks for your reply. The screenshot I attached refers to Thetis for HL2 by MI0BOT. I tried different releases, same behaviour.

I am attaching some other screenshots, as you can see external instrument's data are not matching with ADC/software data. Looks like i have exactly the same issue that Geroge Smart had with his HL2.

What do you think?

Thanks for your time.

73 Mario IW7DLE

[iw7dle]

[Greg VKthreeKV]

Accuracy of hobby metres is very questionable and unless you spend huge $ on calibrated industry meters will not be accurate across low to high power levels.

I really don't see an issue as as long as you set up the audio drive levels on computer first and the as recommended per YouTube Videos ON7OFF, W1AEX you will struggle to achieve the SSB and audio levels to drive good reports or an amp.

Have fun with your HL2 follow the excellent YouTube videos and instructions provided on the forum groups and you really experience the technology provided by so so many who are embracing and assisting the excellent Hermes Lite 2 projects

Cheers,

Greg VK3KV

--
You received this message because you are subscribed to the Google Groups "Hermes-Lite" group.
To unsubscribe from this group and stop receiving emails from it, send an email to hermes-lite...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/hermes-lite/5556a904-574c-41ad-91dd-0452692dd463n%40googlegroups.com.

[Ezio Insinna]

Hello Mario,

I solved the problem with a database reset (new data base).

73

Ezio F5MWA / K3EK

To view this discussion on the web visit https://groups.google.com/d/msgid/hermes-lite/CAL3GjzvS95Eb-iqsS70Ni8kzNbqQYtUhKAvYV5ny5-jOwp2mNw%40mail.gmail.com.

[iw7dle]

Dear Ezio,

thanks for the tip, I tried to reset the database many times but no luck, still get same values :-(

Thanks again,

73 Mario IW7DLE

[iw7dle]

Dear Group,

just to give you some updates about the test I have done recently, but unfortunately, with no luck:

- tried to  reflash gateware (it was already latest version) ;

- tried to perform hard reset.

Following a new screenshot that shows PWR, REV and SWR with full power, still output to dummy load and drive set at 0db:

73 Mario IW7DLE

[James Ahlstrom]

Hello Mario,

Please try this with Quisk. Remove your power meter and connect the dummy load directly to the HL2. Set the Spot slider to zero, press Spot, and then slowly vary the slider from 0 to 100%. The current should increase from 200 ma to 1200 to 1500 ma, and the power should increase from zero to 4 or 5 watts. The power and current will increase quickly at first and then slow down, but they should never stay the same or go backwards. Try with a different dummy load if you have one.

Check the filter bit settings to make sure the correct filter is used for the band. Make sure the test frequency is within the band and the Tx level is set to full. Try the test on several bands.

Make sure the "Power meter calibration" is set to HL2FilterE3 on the Hardware screen.

Manually inspect the power sensor on the filter board. Sensor L3 should have ten turns on the toroid and the wire through the center must not touch the toroid winding.

Report any odd behavior. If everything seems to be in order, the next step would be to measure DC voltage levels on the board.

Jim

N2ADR

[Mario Pagliari]

Dear James,

Thank you very much for the very detailed information. Will run all the test you suggested and get back to you with the results soon.

Thanks again.

Mario IW7DLE

You received this message because you are subscribed to a topic in the Google Groups "Hermes-Lite" group.
To unsubscribe from this topic, visit https://groups.google.com/d/topic/hermes-lite/bB7whTsvPVw/unsubscribe.
To unsubscribe from this group and all its topics, send an email to hermes-lite...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/hermes-lite/d529ae83-c53e-49bf-a42e-68e7f835ca42n%40googlegroups.com.

[iw7dle]

Dear James,

following the results of the performed test with HL2 directly connected to dummy load and "Power meter calibration" set to HL2FilterE3 :

-----------------------------------------------------------------------------

Spot Slider PA (mA)             PWR(w)     SWR
-----------------------------------------------------------------------------
0        250 -->   277       0                   ----
0.5        800 -->   855       1.2               1.33       
1      1500 --> 1590       2.8               1.4

-----------------------------------------------------------------------------

When I press the "Spot" button, current starts from the lowest current value and slowly increases to the highest.

I also inspected L3, everything seems to be fine, no shorts between toroid and front wire.

I also checked the voltage at pins 19 and 20 (AIN1 and AIN2), with full power out I measured respectively 1.95V and 0.21V

 Any other advice?

Thanks a lot for your time.

Mario IW7DLE

[James Ahlstrom]

Hello Mario,

I am running out of ideas. Since only a dummy load is attached, perhaps you could substitute a different dummy load. If you only have a low power load, perhaps you could test with only a few seconds of power. I admit I have never seen a faulty-but-not-dead dummy load. Since your external power meter works, this looks like a low percentage play.

After all this, it looks like a hardware problem with the filter board. If you can borrow another filter board, you could substitute it. My voltage at AIN1 and AIN2 is 0.09 and 2.78 volts.

Jim

N2ADR

[Steve Haynal]

Hi Mario,

With Quisk, do you see this problem with all bands or just 40M? Also, can you post a screenshot of your Quisk filter settings? If they are not correct you may see a problem like what you report.

https://www.youtube.com/watch?v=1pPbQplSBoo

See an example quisk filter setup here:

https://groups.google.com/g/hermes-lite/c/dNNyLkzODW4/m/GTks2dGQDAAJ

73,

Steve

kf7o

[Steve Haynal]

Hi Mario,

Can you please post pictures of the top and bottom of your N2ADR filter board, as well as of it connected to your HL2? I'd like to compare it with a working board, especially the components related to SWR. High quality phone photos usually work well.

73,

Steve

kf7o

[iw7dle]

Dear Steve,

I am attaching screenshots of Quisk, including the filter settings. I ran tests with the spot slider set to 0, 0.5, and 1 - RF output  connected to dummy load.

Same problem on all bands, last test is on 40m with full power.

Later I will post high resolution photos of the boards.

Thanks again for your help.

Mario IW7DLE

[Steve Haynal]

Hi Mario,

Your settings look good. Just to make sure I understand, the power and SWR values reported by Quisk or other software is incorrect as confirmed by an external meter where you do see 5W and near 1:1 SWR (your september 19 post). And sometimes you hear relay chatter.

It would be very helpful to see pictures of your boards. I am wondering if the FWD/REV power detector is damaged (wire not going straight through) or built incorrectly.

73,

Steve

kf7o

[iw7dle]

Dear Steve,

Your description of the issue is correct. As agreed, I am attaching a Google Drive link to the high-resolution pictures of the boards:

https://drive.google.com/drive/folders/1x7NAVYuPX20ChluYDfyZ-asDSitDIxeS?usp=sharing 

Thanks for your feedback.

73,

Mario iw7dle

[Steve Haynal]

Hi Mario,

Thanks for the pictures. Everything looks okay. I'm trying to determine if the defect is on the N2ADR board or on the small ADC U13 on the HL2 board. The small ADC U13 is also used for temperature measurements. Is the reported temperature reasonable? Quisk reports the temperature at the bottom of the screen. If you power on the HL2 after it has cooled to room temperature, the temperature should be only 1 to 2 degrees warmer than your room temperature immediately after powering on. Is this what you see?

73,

Steve

kf7o

[iw7dle]

Hi Steve,

The reported temperature seems reliable; I never had the feeling that the values were incorrect. Either Quisk or Thetis report same values.

Thanks and 73,

Mario

[Steve Haynal]

Hi Mario,

I am leaning towards having your N2ADR board replaced, but request one more measurement. With the N2ADR board connected to your HL2, dummy load attached and a digital volt meter, please report the voltage you measure at AIN2 and AIN1 referenced to ground while your HL2 is receiving and while it is transmitting into the dummy load with correct filter settings. You can find the pins for AIN2 and AIN1 here:

https://github.com/softerhardware/Hermes-Lite2/wiki/IO

They are at the end of the HL2/N2ADR jumper and can and should be measured with the jumper installed.

I see ~1mV for both during receive, and 3V on AIN2 (fwd power) and ~1mv on AIN1 (rev power into dummy load) during full 5W transmit.

This should help us further distinguish if the problem is in the N2ADR board or HL2.

73,

Steve

kf7o

[iw7dle]

Hi Steve,

I finally had access to HL2 and took measurements. 

Here are the results:

RX MODE
AIN1 = 0.003V
AIN2 = 0V

TX MODE FULL PWR --> DUMMY LOAD
AIN1 = 0.325V
AIN2 = 2.103V

I also have a feeling that the issue might be with the N2ADR board because the values reported by the filter board to the ADC appear to be incorrect.

The final verdict is with you. :-)

Thanks and 73,
Mario IW7DLE

[Steve Haynal]

Hi Mario,

Thanks for all the testing. I will request Makerfabs to send you a replacement N2ADR board. You AIN2 value produced by the N2ADR board is lower than expected but the HL2 is reporting the appropriate lower power for that value.

73,

Steve

kf7o

[iw7dle]

Hi Steve,

Thanks a lot for the amazing support; it’s much appreciated. 

Once I receive the new filter board, I’ll let update you... fingers crossed.

73,

Mario 

iw7dle

[iw7dle]

Good morning,

I finally have great news! I received the NA2DR filter board, replaced it, and boom! Everything works perfectly now. I’m now getting accurate SWR and PWR measurements. I’m attaching a screenshot of the measurements I made with the HL2 connected to the antenna this time.

I want to sincerely thank the HL2 community, and in particular, Steve Haynal, Sales Vivi (Makerfabs), and James Ahlstrom for their invaluable and prompt support.

All the best and 73 from Italy.

See you on-air,
Mario IW7DLE

[Steve Haynal]

Hi Mario,

I am happy to hear your problems are resolved!

73,

Steve

kf7o

[iw7dle]

Dear Steve and Group,

I apologize for reopening this topic but the issue with the SWR has resurfaced in less than two months. It seems the mainboard is somehow damaging the N2ADR board and particularly the Directional Coupler section.

I have repeated all the measurements at full power using a dummy load and have attached screenshots for clarity. My setup is straightforward: PSU -> RTX -> HEXBEAM ANTENNA. I don’t have a PA or any other components on the RF line that could cause feedback, I’m unsure what might be causing the problem but it is very frustrating. Additionally, I have installed toroids on the power line and the RF pigtail to minimize potential parasitic current. 

It is very frustrating, don't know what else to do...

What do you think? Any other advice?

Thanks and 73,
IW7DLE Mario

  

[Steve Haynal]

Hi Mario,

If I remember correctly, the AD8692 (U3) on your first N2ADR board was likely bad and the board was replaced by Makerfabs. After it was replaced, things worked for some time but now you are seeing the same problem? If so, can you check that the N2ADR connector you are using provides proper ground and power to U2 on the AD8692. See the schematic here:

https://github.com/softerhardware/Hermes-Lite2/blob/master/hardware/companions/n2adr/n2adr.pdf

Maybe Jim N2ADR can comment if anything on the HL2 side might cause this part to have problems.

73,

Steve

kf7o

[iw7dle]

Dear Steve,

thanks much for your reply...

The first N2ADR board was replaced because I was receiving incorrect SWR reports. The real SWR, FWD, and REV power readings did not match what Thetis and Quisk were reporting.

As you suggested, I checked the AD8692 chip on the filter board, and it appears to be properly connected and powered. The voltage at Vdd (pin 20) is 3.3V, and Vss (pin 9) is correctly connected to ground.

I haven't seen any comments from Jim N2ADR on this matter. Do you have any other suggestions before I give up :-( ?

Thank you.

73,
Mario
IW7DLE

[James Ahlstrom]

Hello Mario and Steve,

I studied the schematic again and I don't see how the HL2 can damage the SWR circuit on the filter board. The lines AIN1 and AIN2 connect from the ADC U13 to the output of opamp U3 (AD8692) through a 330 ohm resistor.. The inputs of U13 are not supposed to provide power, and even if they did, the 330 ohm series resistor provides some protection for U3.

If there are high voltage spikes on the antenna, that can cause problems. It would have to be a spike with a high current so as to generate a pulse at the other side of the SWR bridge transformer. That spike could damage diodes D1 and D2, or the input to the opamp U3. U3 has very high input impedance. But U3 is a rugged part rated for automotive applications, and the others are just diodes.

The only active components of the SWR bridge are U3, D1 and D2. It would be useful to know which component failed. The best way is to replace them one at a time. Some voltage measurements might be useful, for example, at the input and output of the opamps. If you want to ship the defective board to me, I will test it. Or maybe you could find a local ham who could test the board and replace parts.

Jim

N2ADR

[iw7dle]

Dear Jim,

Thanks for your message.

As per your advice, I am focusing on U3, D1, and D2. I have already desoldered D1 and D2 and tested them with an ohmmeter since the tester I have here doesn’t have a diode test mode. The resistance between the anode and cathode is very high, about 1 megohm, and when reversed, it is infinite. I have a feeling that both might be faulty, but to double-check, I will test them with another tester that has diode test mode.

If the diodes are indeed faulty, I hope they are easy to find.

I’ll keep you posted.

Thanks again!

73

[James Ahlstrom]

Please check the data sheet for the BAS40-07 diodes. There are two diodes in each package. A diode tester should not be necessary. An ohmmeter is OK for a first pass. But the diodes must be matched, so replacing them is the only sure test to see if they are good.

Jim

N2ADR

[Mario Pagliari]

Good news from this side! The diodes were good—I checked the voltage drop on each junction, and it was around 0.3V. I resoldered them back, then checked all the other components and realized that the R2 value was incorrect: 8.02 kΩ instead of 2.2 kΩ. I replaced it, and boom—everything works as expected!

At full power, with the RF output connected to a dummy load, I measure around 3V on ADC2 (FWD) and around 0V on ADC1 (REV). The software is also reporting reliable values.

Next goal is to repair a second filter board, actually the one received with the first order, that has the same issue. In this case, the resistors are all good, so I’ll focus on the Schottky diodes and U3. I’ll keep you and the group posted—hope this information can be useful.

Thanks again for your help. Have a good weekend!

Mario IW7DLE

--
You received this message because you are subscribed to a topic in the Google Groups "Hermes-Lite" group.
To unsubscribe from this topic, visit https://groups.google.com/d/topic/hermes-lite/bB7whTsvPVw/unsubscribe.
To unsubscribe from this group and all its topics, send an email to hermes-lite...@googlegroups.com.
To view this discussion visit https://groups.google.com/d/msgid/hermes-lite/1fa02203-48d4-4b23-9824-f39d8249d8ffn%40googlegroups.com.

[Steve Haynal]

Hi Mario,

I am glad you found and fixed your issue! I have added Makerfabs to this thread so they are aware of this potential issue: R2 on N2ADR board was incorrect 8.02 K instead of 2.2 K. Thanks to Jim, N2ADR, for his help also.

73,

Steve

kf7o

[KK Ze]

Hi Mario，

This is zeqing from Makerfabs.  we are sorry that your encountered hardware problems in the usage, and glad  you have caught it and solved.  but i checked our production files and checked 10 pcs sample, all R2 2.2K... do you still have a board that with the wrong R2 and if yes please take a photo of the board..to help us to track the problem and potentially avoid the alike mistakes . 

[iw7dle]

Dear All,

To clarify regarding R2: The resistor's value was initially correct, and the filter board worked properly for about two months. When the issue started, I inspected the board and discovered that the value of R2 had suddenly increased to around 8K. This could possibly be due to a defective component or a static discharge from the antenna. After replacing R2, the directional coupler started reporting correct values again.

I encountered a similar issue with another filter board (first order from Makerfabs) that was also reporitng incorrect FWD and REV power values  to the Hermes ADC. Upon inspection, I found the problem was related to D2 (BAS40-07). Replacing diodes component resolved the issue.

I hope this information is helpful.

Thanks again, and 73.

Best regards from Italy,
Mario IW7DLE

[Clifford Heath]

Perhaps it’s time to use a proper RF limiter diode to prevent front-end damage to the receiver?

A TVS is ok for ESD discharge, especially if it’s a really low-voltage TVS (unlike the 15V one that GSG fitted to the standard HackRF). You can get TVS diodes down to about 3.7V, which *may* be enough to present some level of protection from excessive RF.

To do the job of RF protection properly requires a purpose-built PIN limiter diode. I used the Skyworks CLA4611-085LF (and its earlier counterpart) on my version of the HackRF and have not had one single failure report from thousands of units sold over five years. Whereas failure was rife with every other HackRF.

The Skyworks limiter diodes are only a dollar, and require no extra components. It’s really a no-brainer, every receiver should have one.

Clifford Heath.

You received this message because you are subscribed to the Google Groups "Hermes-Lite" group.

To unsubscribe from this group and stop receiving emails from it, send an email to hermes-lite...@googlegroups.com.
To view this discussion visit https://groups.google.com/d/msgid/hermes-lite/0f3e2cf2-4302-494b-b91d-1e4eeec79606n%40googlegroups.com.

[Ed Grafton]

I use 3 megohm resistors across all of my antennas for continuous static discharge. It also allows a basic test of feedlines with an ohmmeter.

Ed

[Clifford Heath]

On Tue, 11 Feb 2025, 05:10 Ed Grafton, <yankn...@gmail.com> wrote:

I use 3 megohm resistors across all of my antennas for continuous static discharge. It also allows a basic test of feedlines with an ohmmeter.

It's an excellent thing to do but doesn't clamp excessive RF. I use 47k but still... 

13dBm at the receive input turns these limiter diodes into a 1 ohm short for a microsecond or so. They can dissipate 1W, but the mismatch reflects 100W while passing only 20mW into the receiver. Magical stuff.

Clifford Heath