Pure Signal & external amplifier

DL5RBW
Posts: 41
Joined: Tue Sep 25, 2018 5:37 pm

Pure Signal & external amplifier

Postby DL5RBW » Fri Feb 15, 2019 11:45 am

2 weeks ago I built an rf-sampler to include my external amplifier in the predistortion loop. Without the amplifier in the loop but driving it with ANAN´s Pure Signal enabled I have seen 3rd order IMD down about 45 dB which is good enough any time however I wanted to push the envelope and find out what is possible. The rf sampler is designed for -60 dB and gives me a 0 dBm sample (for 1 kW) which is pretty much in the ballpark of what I measured at the coupler output on the ANAN (200 W output give me approx. + 4 dBm). The coupling factor of the rf sampler is flat from 1 to 58 MHz.

I didn´t see any improvement with the amplifier in loop or not so the question is whether there is any limit in phase and/or amplitude that once exceeded the software cannot correct anymore? With limit I mean the difference in (electrical) line length of the interconnecting cable (transmitter to amplifier) and the length of the cable from the amplifier output to the point where the rf sample is taken. Should they be equal if possible? In NR0V´s presentation it is pointed out that this, among other, is a critical factor which makes me believe that very different lengths of these two cables make it impossible for Pure Signal to work properly. Otherwise it should not matter -UNLESS there is a certain limit- as the software simply looks at the output sample and correct the input signal accordingly to make the ouptut signal maximum linear. I only ran a test on 20 m and didn´t try lower frequencies where at least the phasedifference between the cables could be within range again.

Another explanation could be that -45 dB is the maximum that can be achieved with the combination radio & amplifier in the loop. However as I get the -45 dB already with a clean driving signal and the amplifier out of the loop I think there should be some headroom for improvement left.

Any ideas or suggestions?
K9RX
Posts: 297
Joined: Fri Apr 14, 2017 3:47 pm

Re: Pure Signal & external amplifier

Postby K9RX » Fri Feb 15, 2019 2:06 pm

I have an 8000DLE, Expert 2K amp with a Xytronics coupler and routinely see -55 to -60.
Your findings seem low to me ...

Gary
K9RX
User avatar
w-u-2-o
Posts: 2273
Joined: Fri Mar 10, 2017 1:47 pm

Re: Pure Signal & external amplifier

Postby w-u-2-o » Fri Feb 15, 2019 3:08 pm

First, read this post on PureSignal external feedback levels:

viewtopic.php?f=20&t=2384

At 1KW max. I suggest you want a -50dB coupling factor for optimum operation, not -60dB.

Second, differential line lengths will make no difference (hah, a pun!) It will merely cause a constant phase offset at the output. One does not care what that constant is, just as long as the amp response is phase linear. If all phases are linear plus 10 degrees then they are all still linear. I apologize if that is as clear as mud, but I hope this short answer works ;)

73,

Scott
DL5RBW
Posts: 41
Joined: Tue Sep 25, 2018 5:37 pm

Re: Pure Signal & external amplifier

Postby DL5RBW » Sat Feb 16, 2019 10:58 am

I was aiming at a +10 dBm level at maximum output however wanted to start on the low side that´s why I included the additional 10 dB pad. The + 10 dBm are nothing magical but I found this level somewhere in wikidot/anan or the HPSDR website. Anyhow to make sure I measured the output level of the coupler on the ANAN-8000 before starting and found

+ 1,17 dBm for 100 W output
+ 4,17 dBm for 200 W output

As Pure Signal already works fine with 50 W output (equaling a level of -2,17 dBm) I was striving for 0 dBm @ 1 kW which is also what I measured at 1 kW output. The automatic attenuator adjusted itself to about 7 dB which I thought should be fine and verifying there is some headroom. Removing the 10 dB pad would raise the automatic attenuator level to 17 dB - even more headroom but maybe not necessary (?)

I assume that Apache Labs level calculation is correct (as is proven by the effect of Pure Signal on IMD) so with 0 dBm I should be right in the ballpark although 1 or 2 dB on the conservative side. Obvisouly the ANAN itself is doing fine with levels at least from approx. -2 dBm to + 4 dBm.

As I have a switchable attenuator around I will remove the 10 dB pad and go for 4 or 5 dB additional attenuation which will make it 49 dB (the actual coupling attenuation of the rf tap) + 5 dB. This should be plenty for Pure Signal to work.

I too think that phase difference should not be a problem as with diversity the phase can be shifted any amount between 0 and 360 degrees and probably the same or a similar algorithm is used for the phase correction in Pure Signal. Actually what is important is to have a sample of the output signal, a reference how it should look like and correct the amplitude and phase accordingly.
User avatar
w-u-2-o
Posts: 2273
Joined: Fri Mar 10, 2017 1:47 pm

Re: Pure Signal & external amplifier

Postby w-u-2-o » Sat Feb 16, 2019 11:56 am

Again, you want +10 out at max. power. Read point #3 in the link I posted above.
K9RX
Posts: 297
Joined: Fri Apr 14, 2017 3:47 pm

Re: Pure Signal & external amplifier

Postby K9RX » Sat Feb 16, 2019 1:30 pm

just to be sure since this discussion got off to the coupler and not the IMD results ... the -55 to -60 I posted is my IMD results in db.

Gary
DL5RBW
Posts: 41
Joined: Tue Sep 25, 2018 5:37 pm

Re: Pure Signal & external amplifier

Postby DL5RBW » Sat Apr 06, 2019 5:48 am

Finally I found the time to roughly translate my measurement notes for the following summary:

The directional coupler in the ANAN has a coupling factor of about -49 dB (as measured with lab grade equipment). With 100 W output the automatic attenuator levels itself at -10 dB while transmitting which means the “perfect” or “tailored” coupling factor would be -59 dB (for 100 W). A -59 dB coupler would need 0 dB attenuation for 100 W however output powers <100 W would eventually not develop enough level for PureSignal to work correct or at all.The 100 W output power level therefore can be used as a reference and every other power level could be scaled to this level. In other words if you want to include your 1 kW amplifier in the predistortion loop (a 10 dB increase referred to 100 W) you have to add 10 dB of additional attenuation. There are 2 ways to accomplish this: external to keep the full dynamic range of the automatic attenuator or internal (= auto attenuator) with the fact that you limit your dynamic range should you decide to run as much output power (approx.. 12 kW!!) to exceed the maximum automatic attenuation level of 31 dB. I somewhere read the maximum permissible level for the predistortion input is +13 dBm – whether this is true or not I couldn´t verify. However accepting this information and taking a little bit of a safety margin I aimed for a maximum level of +10 dBm at the predistortion input port under any circumstances. The automatic attenuator level range of the ANAN is 31 dB which means a power ratio of 1258. In other words if you use an external -49 dB coupler you restore the situation as it would be by using the ANAN´s internal coupler. However the external coupler could (and should) be located at the output of an amplifier as this includes the amplifier in the predistortion loop and makes it really what it pretends to be: a LINEAR amplifier! Going back to the -49 dB coupler as explained above with 100 W the needed 10 dB of attenuation comes from the automatic attenuator of the ANAN. That means there are still 21 dB of attenuation range left (31 dB – 10 dB). 100 W + 21 dB means about 12 kW output maximum power. The minimum power that predistortion will work with is 10 W because we can sacrifice the 10 dB of attenuation that are in line with 100 W. So with a -49 dB coupler one is ready for employing predistortion from the 10 W to 12,6 kW level without any additional circuitry. The predistortion circuit won´t be able to differentiate whether a certain level comes from a -49 dB coupler + 10 dB attenuator or a -59 dB coupler. I personally use a -49 dB coupler and a switchable (1 dB & 10 dB steps) attenuator to keep the input level at between 0 dBm and +5 dBm for maximum power output. As I very seldom run 10 W or less there is always plenty of level for the predistortion to work properly without the risk of damaging anything should I forget to manually increase the attenuation with increasing power – the remaining 20 dB+ headroom of the automatic attenuator takes care of it! The result is that 3rd order intermodulation with the amplifier is down by 55-60 dB now supporting that the calculations are correct.
User avatar
w-u-2-o
Posts: 2273
Joined: Fri Mar 10, 2017 1:47 pm

Re: Pure Signal & external amplifier

Postby w-u-2-o » Sat Apr 06, 2019 1:17 pm

You results match up with those presented here:

https://apache-labs.com/community/viewtopic.php?f=20&t=2384

With three exceptions.

First, it is better to drive the feedback signal as high as possible within the bounds of safety not just for the sake of achieving a wider range of output power levels that can be corrected, but also to obtain the highest feedback-to-internal crosstalk ratio possible.

Second, and more important, using a 49dB coupler at power levels in excess of 1500W could be dangerous. The reason is that one cannot rely on the step attenuator to reliably protect the preamp. At 12KW, if the step attenuator is at 0dB for any length of time, for instance if PureSignal is turned off, the preamp will see over 21dBm. This could damage the preamp (although it is fine for the step attenuator).

Third, those numbers are only good for Orion MKII based hardware (7000 and 8000 models). The preamps in those units have 14dB of gain. In all other models the preamp has 20dB of gain, so the numbers will be different by 6dB.
Danie
Posts: 1
Joined: Sun Jun 30, 2019 12:08 am

Re: Pure Signal & external amplifier

Postby Danie » Mon Jul 01, 2019 3:08 am

Hello, I going to do my best with my writting because, I'm french from Montreal, I got my Anan MK II 7000 dle with no pc, so my question is, I'm using a coupler extern with an Amplifier Acom 2100, everything are correcting with puresignal everything is green, but I got nothing on ATT on Tx all time is 0 for correcting, so I check all pl-259 I check everything, so maybe I forgot something in setup HPSDR ?????

I brought my anan on another hamradio and it was giving an ATT correcting like 8-9-10 so I thing I missing something in setup in HPSDR I'm using v3.4.9.

Thank a lot to your help,

ve2da
Danie
73
Last edited by Danie on Tue Jul 02, 2019 3:47 am, edited 1 time in total.
User avatar
w-u-2-o
Posts: 2273
Joined: Fri Mar 10, 2017 1:47 pm

Re: Pure Signal & external amplifier

Postby w-u-2-o » Mon Jul 01, 2019 9:03 pm

Danie,

In Setup > General > Ant/Filters > Antenna you need to make sure that ATT on TX is ON, and RX BYPASS on TX is ON.

Finally, you need to have the coupled output from the coupler attached to the BYPASS input. Do not exceed +13dBm on that input.

73,

Scott
dieseldawg
Posts: 12
Joined: Tue Apr 25, 2017 7:15 pm

Pure Signal on 160

Postby dieseldawg » Mon Jan 27, 2020 3:29 pm

I have a 7000 with AC2IQ coupler and Pure Signal works fine except for 160 meters. I get major IMD, -20dB with the external coupler, although the feedback is green, and S-Att is good. Pure Signal works on 160 with the internal sampler. Any ideas? Power level doesn't make any difference.

Greg
N8MW
User avatar
w-u-2-o
Posts: 2273
Joined: Fri Mar 10, 2017 1:47 pm

Re: Pure Signal & external amplifier

Postby w-u-2-o » Mon Jan 27, 2020 5:36 pm

Greg--what kind of amp? Tuner? Antenna? Just wondering if you might have some RF coming back into the coupler on the coax shield on 160.
dieseldawg
Posts: 12
Joined: Tue Apr 25, 2017 7:15 pm

Re: Pure Signal & external amplifier

Postby dieseldawg » Tue Jan 28, 2020 6:29 am

Scott, I don't think it's rf because the Pure Signal doesn't work even when the amp is at low power or in standby with less than 50 watts. I can get the needed feedback at that power level by changing the external attenuator. I did consider that the sample line from the sampler to the radio could somehow be picking up rf, it's 25 feet long, but that doesn't seem likely since it doesn't happen on other bands and has several toroids, both at the sampler and the radio. Do you think there could be some sort of resonance on the sample line at that length on 160?

Greg
User avatar
w9ac
Posts: 72
Joined: Sun Apr 09, 2017 4:01 pm

Re: Pure Signal on 160

Postby w9ac » Thu Feb 27, 2020 3:07 pm

dieseldawg wrote:I have a 7000 with AC2IQ coupler and Pure Signal works fine except for 160 meters. I get major IMD, -20dB with the external coupler, although the feedback is green, and S-Att is good. Pure Signal works on 160 with the internal sampler. Any ideas? Power level doesn't make any difference. Greg N8MW


Greg's issue may be a combination of factors, but an external sampler with less than 56 dB of coupling is inadequate when operating at 1500W on 160m. See my QRZ page for a view of 160m IMD screen captures. Scroll to the heading: "TOROID RF SAMPLER"

https://www.qrz.com/db/w9ac

Sampling level into the Orion MKII is critical for good linearization, especially on 160m because a Mini-Circuits DAT-31A-SP+ GaAs-switched digital attenuator is more IM sensitive at lower RF frequencies. With 1500W and a coupling coefficient of 56 dB, input power to the MCL attenuator is 5.8 dBm (1500log10*10+30 = 61.76 dBm. 61.76 dBm - 56 dB = 5.8 dBm). Levels exceeding this amount on 160m generate high internal IMD. If this occurs, the PureSignal linearization algorithm cannot distinguish between amplifier non-linearity and the non-linear overdrive function of the DAT-31A-SP+ attenuator. I have verified these results using three different external samplers and an external spectrum analyzer monitored at the input to a 50 ohm dummy load.

Legal-limit amps with a -50 dBc sample port require an additional 6 dB pad ahead of Orion MKII. A coupling factor of 56 dB still allows good linearization at full amplifier output power AND when the amplifier is in bypass with exciter power less than 50W.

Paul, W9AC
User avatar
w-u-2-o
Posts: 2273
Joined: Fri Mar 10, 2017 1:47 pm

Re: Pure Signal on 160

Postby w-u-2-o » Thu Feb 27, 2020 6:24 pm

w9ac wrote:Legal-limit amps with a -50 dBc sample port require an additional 6 dB pad ahead of Orion MKII. A coupling factor of 56 dB still allows good linearization at full amplifier output power AND when the amplifier is in bypass with exciter power less than 50W.

Paul, W9AC
My apologies, Paul, but I'm forced to disagree with this assessment. There is no evidence in terms of an electrical analysis of the RF path on receive, and most people observe no problems whatsoever, including myself, in accordance with the design.

I have re-posted my analysis on this, which you and I discussed on the old Yahoo Group, in a topic here: viewtopic.php?f=20&t=3331

Note that this in no way diminishes the fact that your station seems to perform differently than others, and that you do obtain better performance with the additional attenuation. Perhaps Greg is has a similar problem. Whatever that problem is has not been identified, but it does not have anything to do with driving the DAT-31-SP+ into non-linear operation. This assumes, of course, that your DAT-31-SP+ is conforming to the Mini-Circuits data sheet.

73,

Scott
User avatar
w9ac
Posts: 72
Joined: Sun Apr 09, 2017 4:01 pm

Re: Pure Signal & external amplifier

Postby w9ac » Thu Feb 27, 2020 7:43 pm

Scott,

Thanks for the feedback. I have made actual measurements, repeated those measurements, validated my test procedure with others, and I stand by the result.

>"There is no evidence in terms of an electrical analysis of the RF path on receive, and most people observe no problems whatsoever, including myself, in accordance with the design."

The IMD products seen on a stand-alone spectrum analyzer will not be seen on the GUI display when the IMD threshold is reached. I don't expect ops without spectrum analyzers to see the result. You can't take a measurement below the threshold and claim that the GUi spectrum display tracks with the external analyzer. Below that threshold, mine tracks too. Please increase level with your external step attenuator and let us know the point where you see IMD abruptly rise. If it nears the maximum power rating of the MCL attenuator, stop. That's fine, but attempt a test. You may be at the ragged edge and you don't know it because you haven't tested for the threshold.

I would like to see a comparable spectral analysis from anyone else who has conducted a similar test by varying the sampled coupling coefficient from an amplifier, while observing spectra at the input to a test load. If that testing is complete, I would like to see the measured result and also compare test conditions.

Paul, W9AC

Return to “PureSignal Operations (PowerSDR & Thetis)”