External feedback level calculations

User avatar
w-u-2-o
Posts: 5540
Joined: Fri Mar 10, 2017 1:47 pm

External feedback level calculations

Postby w-u-2-o » Sun Apr 23, 2017 1:37 pm

UPDATED 25 MAY 2018

There are three things to consider when configuring an external coupler and/or external attenuators to obtain a proper feedback level at the rear panel connector of the radio:

1. When using Puresignal, the Auto-Attenuate function drives the step attenuator to the setting necessary to obtain approx. a feedback signal level of +3dBm at the ADC input. This is -17dBm at the output of the step attenuator for all radios except the 7000 and 8000 due to the 20dB preamp used prior to the ADC. The 7000 and 8000 both use a 14dB preamp, so this is -11dBm at the output of the step attenuator in those radios, both of which use the newest Orion MkII SDR board.

2. On older radios, step attenuator levels of below 7dB or so generally result in poor performance because internal crosstalk dominates the external feedback signal. The newer the radio, the less of an issue this is. However, better to have it (a higher feedback signal level) and not need it then need it and not have it.

3. It generally pays to set levels such that the step attenuator obtains a rather high setting, say in the 25dB range when RF output power is at maximum. There are two reasons for this. First, this provides a good dynamic range situation such that, when operating at reduced power levels, Auto-Attenuate has room to reduce the step attenuator value when not QRO. Second, there is a better feedback signal to crosstalk ratio on the PA/RF board as discussed in (2) above.

With these three guidelines, one can easily calculate the coupling factor and/or any additional attenuation in the feedback path necessary to obtain the desired step attenuator value and proper feedback levels. For example, let's consider a typical legal limit scenario of an external amplifier at 1500W peak RF power and an Xtronic coupler with 44dB coupling factor, and a desired step attenuator setpoint of 25dB.

Step 1--calculate peak RF power in dBm. Power in dBm = 10log(power in Watts)+30. In this case 1500W = 61.76dBm.

Step 2--calculate output power level of the coupler at peak RF power input. 61.76dBm - 44dB = 17.76dBm.

Step 3--calculate required input power to the step attenuator to obtain -17dBm at the output of the step attenuator at your desired step attenuator setpoint. -17dBm + 25dB = 8dBm (assumes any radio except the 7000 or 8000, just change -17 to -11 for those radios).

Step 4--calculate the additional external attenuation required between the coupler output level determined in (2) above and the rear panel input to obtain the feedback level determined in (3) above. 17.76dBm - 8dBm = 9.76dB.

Step 5--beg, borrow, buy, make an attenuator of the value identified in (4) above and install it between the coupler output and the rear panel input (Bypass on newer radios using the Rev. 24 PA/RF board or the 7000, Ext1 on older radios using the Rev. 15 or 16 PA/RF board, PS Input on the 8000).

Of course you can simply use an attenuator that is close in value, it need not be exact. The step attenuator will make up the difference. However, do not use too low of a value, as a) you do not want to exceed +13dBm at any rear panel input otherwise you risk damaging the radio, and b) the attenuator will run out of dynamic range, 31dB is the maximum it can be set to.

One thing that you might consider is obtaining a rotary step attenuator with 1dB steps. They are usually available on eBay for not very much money compared to fixed attenuators. This way you can dial in just the right amount of additional attenuation you want or need.

Image
User avatar
W2PA
Posts: 166
Joined: Sun Apr 09, 2017 6:34 pm
Location: LaGrangeville, NY
Contact:

Re: External feedback level calculations

Postby W2PA » Sun Apr 23, 2017 2:36 pm

Scott,

This makes sense and reinforces my original calculation that I needed around 9dB of additional attenuation on the output of my XDC-1. It's working fine with my 8000DLE and Alpha 89 amp. But here's the quirk...

In OpenSDR-PowerSDR, the "S-ATT" window is reporting only 9dB. At one point it was reporting only 7dB.

So I conclude that maybe the 3.3.15 version has not been properly modified to display the correct value with the 8000. I'm guessing at this point, so I've got it on my list to investigate. I'll post another reply to your posting when I have a firmer conclusion about that.

For the moment, however, my 8000 seems to be working fine with PS-A turned on, 9dB of attenuation in the sense line, and displaying only 9 in S-ATT. PS works with the amp disabled, too, down to 40W. Below that it stops working.
73,
Chris, W2PA
User avatar
w-u-2-o
Posts: 5540
Joined: Fri Mar 10, 2017 1:47 pm

Re: External feedback level calculations

Postby w-u-2-o » Mon Apr 24, 2017 2:14 am

Yeah, that's very strange, Chris. If you are at 1500W, with a 44dB coupling factor and a 9dB attenuator, I would expect to see an S-ATT of 26dB, not 9. A difference of 17dB doesn't seem to correlate with any obvious scaling problems, unless it's a factor of three!

I'm running 3.3.15 with Angelia 5.8 and I'm getting within one dB of theoretical on the S-ATT.

I wonder if other 8000 owners would be willing to chime in with their peak power level, coupling factor, external attenuation and S-ATT results?

73,

Scott
User avatar
WA0VY
Posts: 83
Joined: Sun Apr 09, 2017 3:41 pm

Re: External feedback level calculations

Postby WA0VY » Mon Apr 24, 2017 7:57 pm

Based on my operational power level with an external amp, I ran the math as set out by Scott and found that I have about 6.6 db of phantom attenuation in my system with my 8000DLE. In other words, based on the math, where I should be seeing 26.6 db on the step attenuator, I'm seeing only 20. I generally use an additional 10 db of line attenuation. I then swapped in a 20 db attenuator and a 6 db attenuator. With those attenuators substituted, I saw 10 db on the step with the 20 and 24 db on the step with the 6. So, the line attenuators appear to be properly calibrated and the results exactly consistent with what would be expected. I used the same attenuators, XDC-1 coupler and cables with my 200D and had about the same amount of phantom attenuation. My conclusion is that the XDC is attenuating at about 50 db instead of the 44 db per spec. I don't believe there there is anything wrong with the software.

73 Brent WA0VY
73 Brent WA0VY
AB2EZ
Posts: 99
Joined: Sun Apr 09, 2017 2:29 pm
Location: Princeton, NJ

Re: External feedback level calculations

Postby AB2EZ » Tue Apr 25, 2017 4:14 pm

Brent

Make sure that the line attenuator that you are using is terminated in 50 ohms (i.e. the impedance looking into the feedback port of the 8000 or the 200D), and not 25 ohms (the impedance looking into the feedback port of the 8000 or the 200D ... in combination with an extraneous, parallel 50 ohm load that is across the output of the line attenuator).

If a 20dB line attenuator is terminated in 25 ohms, its output will be approximately 6dB lower than if it is terminated in 50 ohms.

Also, make sure that each of the output ports of the XDC-1 coupler is terminated in 50 ohms... with the ANAN's feedback input port serving as the 50 ohm load for the coupler output port that is feeding the ANAN.

Stu
User avatar
WA0VY
Posts: 83
Joined: Sun Apr 09, 2017 3:41 pm

Re: External feedback level calculations

Postby WA0VY » Tue Apr 25, 2017 7:24 pm

Stu, in the context of a run from the XDC-1 coupler to the PS sample input connector, I don't know what you mean by 50 ohm termination. The attenuator itself does not specify an impedance. The cable is 50 ohm coax. I assume the sample input on the 8000 is 50 ohm. The Ref. port on the XDC-1 is terminated with a 50 ohm terminator. What am I missing?
73 Brent WA0VY
AB2EZ
Posts: 99
Joined: Sun Apr 09, 2017 2:29 pm
Location: Princeton, NJ

Re: External feedback level calculations

Postby AB2EZ » Tue Apr 25, 2017 9:44 pm

Brent

You aren't missing anything. The XDC-1 is designed to operate as specified if both of its sample output ports are terminated with (i.e. looking into) 50 ohm loads.

As you described, the XDC-1 forward sample output port is looking into a 50 ohm coaxial cable that is connected to the input of the 50 ohm in-line attenuator. The output of the 50 ohm in-line attenuator is connected, via another 50 ohm coaxial cable to the 50 ohm input impedance of the ANAN's feedback input port... so all is as it should be.

I.e. The input side of the inline attenuator presents a 50 ohm load impedance to the XDC-1's forward sample output port if the output side of the inline attenuator is connected to the 50 ohm input impedance of the ANAN's feedback input port.

I just wanted to make sure that you hadn't added an additional 50 ohm termination in parallel with the ANAN's feedback input port, or in parallel with the input side of the in-line attenuator.

Best regards
Stu
User avatar
w-u-2-o
Posts: 5540
Joined: Fri Mar 10, 2017 1:47 pm

Re: External feedback level calculations

Postby w-u-2-o » Wed Apr 26, 2017 10:44 am

I just wanted to make sure that you hadn't added an additional 50 ohm termination in parallel with the ANAN's feedback input port, or in parallel with the input side of the in-line attenuator.
For example, someone might have split the coupler output with just a simple "T" connector and sent one side to an oscilloscope for station monitoring. If the 'scope was set to 50Ohm input impedance instead of high input impedance it could cause this problem.

If you want to split the signal for a 'scope, Mini-Circuits makes a relatively inexpensive ($60) power divider: https://www.minicircuits.com/pdfs/ZFRSC-2050+.pdf. They are available in SMA, BNC and N connector formats.

Image
liam
Posts: 35
Joined: Mon Apr 10, 2017 7:26 am

Ferrite Toroid cores

Postby liam » Sun Apr 30, 2017 8:17 am

Hi I was looking at this site https://sites.google.com/site/vkonehw/h ... al-coupler

and when I click on the Ferrite cores the supplier reads out of stock nor do they post outside VK has anyone details of the ring core that can be purchased in the EU for this project


thanks
Liam
pd5gvp
Posts: 2
Joined: Sun Apr 09, 2017 2:31 pm

Re: External feedback level calculations

Postby pd5gvp » Sun Apr 30, 2017 8:42 am

You can build one yourself: like the dc6ny
http://www.hamsdr.com/data/GlobalFileUploads/9636__RF%20Sampler%20for%20Pre.pdf
Flat freq. response (<1dB), 50Ohm's, set the attenuation to your need etc.
Image
VK8BN
Posts: 1
Joined: Sun Apr 09, 2017 10:37 pm

Re: External feedback level calculations

Postby VK8BN » Sun Apr 30, 2017 9:05 am

https://www.rfparts.com/signal-samplers/4273-020.html
I have a couple of these .. not directional but about -40dB at hf should this work OK?
liam
Posts: 35
Joined: Mon Apr 10, 2017 7:26 am

Re: External feedback level calculations

Postby liam » Sun Apr 30, 2017 11:04 am

HI the DC6NY is ok but I would have to make a PCB and the RF Parts is 300 euro I like

https://sites.google.com/site/vkonehw/h ... rf-sampler

as its straighforward to make the only part i can not source are the moment
in the EU are ferrite toroid ring cores in either of these projects , and thanks to VK1HW
for sticking them on the net


Liam
User avatar
w-u-2-o
Posts: 5540
Joined: Fri Mar 10, 2017 1:47 pm

Re: External feedback level calculations

Postby w-u-2-o » Sun Apr 30, 2017 11:10 am

The Bird part should work ok. I don't know how flat it's frequency response is. As it's adjustable be careful to not wind up chasing your tail when doing so. With luck it's flat enough to set it and forget it.
User avatar
FM5GB
Posts: 130
Joined: Sun Apr 09, 2017 10:03 pm

Re: External feedback level calculations

Postby FM5GB » Tue May 09, 2017 9:51 pm

Hi Liam,

Have a try with the sampler proposed by Hauke VK1HW (to be found on his site
at : https://sites.google.com/view/vk1hw/hom ... rf-sampler . It needs no
PC board and has very fews components. I built one (using an Amidon FT50-43 toroïd) and
see the flat response it has 100 kHz to 80 MHz on my spectrum analyser :

73s Phil FM5GB
Attachments
Sampler 51 dB 100 kHz to 80 MHz.png
Sampler 51 dB 100 kHz to 80 MHz.png (55.62 KiB) Viewed 63103 times
Last edited by FM5GB on Wed May 13, 2020 2:27 pm, edited 1 time in total.
User avatar
w-u-2-o
Posts: 5540
Joined: Fri Mar 10, 2017 1:47 pm

Re: External feedback level calculations

Postby w-u-2-o » Wed May 10, 2017 2:06 am

Beautiful results there, Phil! I wish I had a Rigol spectrum analyzer, too :mrgreen:

73!

Scott
User avatar
FM5GB
Posts: 130
Joined: Sun Apr 09, 2017 10:03 pm

Re: External feedback level calculations

Postby FM5GB » Mon Jan 14, 2019 1:09 am

Hi all !

I've forgotten to post pictures of my sampler. I finally decided to built the sampler inside the SPE Expert 1K-FA amplifier.
There is not much room inside and I wanted to insert the sampler between the amp final stage and the built in ATU.
So I built it direcly on a piece of military grade coax (M17/60 RG 142 wich stands 2.5 kW @ 30MHz). The toroid and the two resistors
are wrapped in teflon tape and shielded with copper foil. Sample is 50 dB down from the outpout (that's +10 dBm @ 1 kW and -10 dBm @
10 W output) To run safe I added a 10 dB SMA attenuator in the sample path to the orion board.
Correction is green between 10 W & 1 kW; S-ATT values range -19 to 0 dB. Correction yields 50 dB suppression of 3rd order products on all bands.
Only draw back : the amp has to be switched ON (operating or standby) to have the outpout pass thru the ATU. If the amp is switched off input
is diretly fed thru not going thru the sampler.

73s Phil FM5GB
Attachments
P1010509.JPG
P1010509.JPG (541.19 KiB) Viewed 57310 times
P1010510.JPG
P1010510.JPG (467.7 KiB) Viewed 57310 times
P1010511.JPG
P1010511.JPG (370.38 KiB) Viewed 57310 times
NS8K
Posts: 7
Joined: Mon Dec 19, 2022 4:49 pm

Re: External feedback level calculations

Postby NS8K » Mon Aug 14, 2023 9:16 pm

Scott,

I understand the calculations at the beginning of this thread. I'm in the process of buying a wattmeter and have the option of a coupler with an additional -50dB port for PS. My radio is a 7000DLE MKIII. Here's my dilemma.

I assume my radio has a 14dB pre-amp before the ADC. Working backwards through the math, if I desire the step attenuator to nominally be at -25dB, that would require an input level of +14dBm. I haven't been able to find what the limit is for this signal other than your warning to not exceed +13dBm. Based on the warning, I would feel comfortable at +10dBm. This would back the step attenuator down to -21dB. That might work but I'm giving up a lot of dynamic range. I'm not excited about doing that. In your example, the level was +8dBm. Maybe +10 is even too high.

Now, considering the coupler option is -50 dB, its output power would be 7dBm at 500 watts. That would bring the step attenuator down 3dB more to -18dB. Even more dynamic range lost.

Going through the calculations with 20dB of pre-amp gain shows 30 of the available 31 dB of step attenuation are useable without exceeding the 13 dBm input limit. Nicely balanced. At 14 dB gain, max attenuation is 24dB at 13dBm input. Not so well balanced. If I leave the amp off and drop down to 100 watts, there would go another 7dB with the step attenuator at a nominal -11dB. I understand I could change the radio settings and go back to the internal coupler in this situation.

Does my math make sense? Have I overlooked something? What are your thoughts on the -50dB coupler? A -47dB coupler at 500 watts would result in a 10dBm signal and a nominal step attenuation of -21. I don't know if such a thing is even available although I understand I could use more coupling and an external manual step attenuator too.

Thanks for reading. I look forward to your advice.

Tom - NS8K
User avatar
w-u-2-o
Posts: 5540
Joined: Fri Mar 10, 2017 1:47 pm

Re: External feedback level calculations

Postby w-u-2-o » Tue Aug 15, 2023 2:33 pm

Hi Tom,

You math is 100% spot on.

At 500W and 50dB coupling factor the resulting 18dB of attenuation is plenty fine and PureSignal performance will be good.

At 50W (10dB less) and 50dB coupling factor the resulting 8dB of attenuation will still be OK and PureSignal performance will not be appreciably degraded. And, as you say, if there is any degradation you can switch back to the internal coupler, something you'll want to do at less than 50W anyway.

With the older models that used the 20dB preamp internal crosstalk was much worse than it is in the 8000, 7000 and newer models. In those older models it was worth pushing the feedback level as high as it was safely possible to get the highest feedback-to-crosstalk ratio. Giving up a few dB for the convenience of built-in couplers with the newer models is not the problem it once was.

73,

Scott
NS8K
Posts: 7
Joined: Mon Dec 19, 2022 4:49 pm

Re: External feedback level calculations

Postby NS8K » Tue Aug 15, 2023 7:36 pm

Scott,

Thanks for the reply and advice. Pure Signal with just the radio performs admirably. I'll go ahead and order the wattmeter with the special coupler that provides the additional -50dB output. In case anyone ever reads this, the wattmeter is an LP-100A that has an optional coupler incorporating a Pure Signal tap. Larry is well aware of Pure Signal and doing double duty with the coupler. It will be a while, but I hope to post an update on the final results.

Thanks again,

Tom - NS8K

Return to “PureSignal Operations (PowerSDR & Thetis)”