8000DLE step attenuator values with auto-attenuate

User avatar
w-u-2-o
Posts: 5539
Joined: Fri Mar 10, 2017 1:47 pm

8000DLE step attenuator values with auto-attenuate

Postby w-u-2-o » Mon May 01, 2017 1:34 am

8000 owners:

I am wondering what sort of step attenuator values you are seeing. In talking with an 8000 owner (I don't have one, I have a 100D), it seems that the 8000 may have approx. 20dB of internal attenuation built into the external PureSignal feedback path prior to the step attenuator.

At 1500W, with an Xtronic coupler and 9dB of external attenuation, one would expect an step attenuator value of approx. 26dB. However he was getting 8dB. He dropped his external attenuation to 3dB and his step attenuator went up to 14dB, so that tracks well from that perspective.

73,

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

Re: 8000DLE step attenuator values with auto-attenuate

Postby W2PA » Sat May 06, 2017 11:52 am

Just to further corroborate, Scott, I'm now running the 8000 into an amplifier at 1500W with no extra attenuation in line with the output of my XDC-1, and seeing 17dB in the step attenuator (S-ATT) window.
73,
Chris, W2PA
User avatar
w-u-2-o
Posts: 5539
Joined: Fri Mar 10, 2017 1:47 pm

Re: 8000DLE step attenuator values with auto-attenuate

Postby w-u-2-o » Sat May 06, 2017 12:09 pm

Hi Chris,

That matches exactly with the other data point. Both data points are exactly 18dB off from what one would expect with a 100D or 200D. This implies there is an additional 18dB of loss inside the radio prior to the RX1 input to the Orion Mk 2 board.

Thanks for a confirmatory data point!

73!

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

Re: 8000DLE step attenuator values with auto-attenuate

Postby W2PA » Wed May 10, 2017 10:26 pm

I did a more complete survey of 8000 reaction to PS on several bands, at Scott's request. In all cases there was zero additional attenuation inserted between the XDC-1 coupler and the ANAN-8000DLE PS input. PS was activated by my own voice input ("helloooo radio"). I also tried driving the amplifier to 200W output

Code: Select all

        Power   
Band     100   200   1500
160       5     9   
80        5     8     17
40        4     8     17
20        4     7     17
17        3     6   
15        3     6   
12        4     7   
10        5     9


Looks consistent within the accuracy of the S-ATT window.

[I had a hard time getting the columns to line up, hence the "Code" construct. Any guidance on how to do this in the forums? ]
73,
Chris, W2PA
User avatar
w-u-2-o
Posts: 5539
Joined: Fri Mar 10, 2017 1:47 pm

Re: 8000DLE step attenuator values with auto-attenuate

Postby w-u-2-o » Wed May 10, 2017 11:18 pm

Fantastic data, Chris, thank you!

Can anyone else post similar data?

Thanks,

Scott

P.S. guidance = use the "code" bbcode ;)
User avatar
W2PA
Posts: 166
Joined: Sun Apr 09, 2017 6:34 pm
Location: LaGrangeville, NY
Contact:

Re: 8000DLE step attenuator values with auto-attenuate

Postby W2PA » Wed May 10, 2017 11:43 pm

That's what I used - Code. Thought there might be something more transparent, like a selection of a non-proportional font.
73,
Chris, W2PA
AB2EZ
Posts: 99
Joined: Sun Apr 09, 2017 2:29 pm
Location: Princeton, NJ

Re: 8000DLE step attenuator values with auto-attenuate

Postby AB2EZ » Thu May 11, 2017 3:51 pm

Chris
Scott

The LTC6400 low noise preamplifier... that is included in the front end amplification chain of each receiver... has an "enable" pin. Could it be that, in the 8000 transceiver, this 20dB gain block is being disabled on transmit (to protect the A/D converter)? That would explain the apparent discrepancy in the operation of Pure Signal auto-attenuate function.

Stu
AB2EZ
User avatar
w-u-2-o
Posts: 5539
Joined: Fri Mar 10, 2017 1:47 pm

Re: 8000DLE step attenuator values with auto-attenuate

Postby w-u-2-o » Thu May 11, 2017 4:07 pm

Actually, I just found out this morning that the preamp on the Orion Mk II has only 14dB of gain, as opposed to the 20dB preamp used on all other radios. That would account for 6dB of the 18dB difference that Chris is seeing, leaving 12dB as a mystery.

Also this morning I was able to confirm with Brent, WA0VY, that he is seeing step attenuator values within 1dB of theoretical once the difference in preamp gain is taken into account.

I would really like to get a couple of other data points from other 8000 users in this thread. Right now Brent's results are spot on, Chris' off by 12dB. If we could get a couple of other people to chime in then it would tell us if Chris has a problem or not.

Thanks,

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

Re: 8000DLE step attenuator values with auto-attenuate

Postby AB2EZ » Thu May 11, 2017 4:41 pm

If the preamplifier gain is 14dB instead of 20dB* and the volunteers who develop the software don't know that... then the S-meter readings and the panadapter readings will also be 6dB low.

*This sounds like only one side (+ ground) of the balanced input to the preamplifier is being used.

The extra ~12dB of discrepancy that Chris is observing could be caused by a not-enabled or defective preamplifier

Stu
Last edited by AB2EZ on Thu May 11, 2017 8:13 pm, edited 1 time in total.
User avatar
w-u-2-o
Posts: 5539
Joined: Fri Mar 10, 2017 1:47 pm

Re: 8000DLE step attenuator values with auto-attenuate

Postby w-u-2-o » Thu May 11, 2017 6:18 pm

Stu--the part has changed to an LTC6400-14. I'm pretty sure Warren and Doug know that and accounted for it. However, your point that Chris' preamp could be broken somehow is interesting.
User avatar
W2PA
Posts: 166
Joined: Sun Apr 09, 2017 6:34 pm
Location: LaGrangeville, NY
Contact:

Re: 8000DLE step attenuator values with auto-attenuate

Postby W2PA » Fri May 12, 2017 12:11 am

The only reason I'm doubtful I have a defective preamp is that I know of at least one other amateur using the 8000 who is also running with no additional attenuation in line. I like Stu's suggestion of a deliberately disabled preamp. The unit operated fine with PS from day one with the factory installed jumper, too. At that time I wasn't paying attention to the S-ATT window.
73,
Chris, W2PA
User avatar
W2PA
Posts: 166
Joined: Sun Apr 09, 2017 6:34 pm
Location: LaGrangeville, NY
Contact:

Re: 8000DLE step attenuator values with auto-attenuate

Postby W2PA » Fri May 12, 2017 1:02 am

How about maybe someone from Apache Labs could weigh in on all this and end the speculation?
73,
Chris, W2PA
User avatar
w-u-2-o
Posts: 5539
Joined: Fri Mar 10, 2017 1:47 pm

Re: 8000DLE step attenuator values with auto-attenuate

Postby w-u-2-o » Fri May 12, 2017 5:36 pm

I just received another data point by email. That data has a 5dB discrepancy in it. So now I've got three data points: one spot on, one with a 5dB discrepancy and another with a 12dB discrepancy from theoretical.

I wish I could put a spectrum analyzer on the feedback signal in all three cases in order to detect any problems in couplers, cables, attenuators, etc.

73,

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

Re: 8000DLE step attenuator values with auto-attenuate

Postby AB2EZ » Fri May 12, 2017 5:54 pm

I have an inexpensive 7.040 MHz S9 (-73dBm) signal reference that I purchased from Elecraft a few years ago. They sell an upgraded (but more expensive) version now... that has three frequency selections. It would be very easy to make one of these (see the XG-1 schematic in the XG-1 manual on the Elecraft web site).

Anyway... in receive operation... using a known 50 ohm signal reference, it would be interesting to know what these 8000 owners are seeing on their S meters and panadapters. It would also be interesting to know what the value of the S-meter calibration setting is (to see if a significant correction has been dialed in during factory unit testing... to compensate for low gain in the receiver RF amplifier chain).

If there is a problem with the preamplifier, it would affect both receive and Pure Signal feedback by the same amount... unless the preamplifier is being disabled on transmit.

It would also be interesting to know whether the S-meter readings are the same for RX1 and RX2 ... using the same signal source applied first to ANT1 and then to ANT2.

Stu
Last edited by AB2EZ on Sat May 13, 2017 6:34 pm, edited 2 times in total.
User avatar
w-u-2-o
Posts: 5539
Joined: Fri Mar 10, 2017 1:47 pm

Re: 8000DLE step attenuator values with auto-attenuate

Postby w-u-2-o » Sat May 13, 2017 11:55 am

Another data set, generously provided by George:

Code: Select all

Power   Coupler      Atten      Calculated   Actual      Delta
W         dB         dB         dB            dB         dB

 986      44         20          6.9          5         -1.94
 933      44         10         16.7         16          -0.7
 997      44          6         21.0         19         -1.99
1500      44         20          8.8          7         -1.76
1488      44         10         18.7         18         -0.73
1493      44          6         22.7         21         -1.74

"Calculated" and "Actual" refer to the step attenuator value. Less than 2dB of error from theoretical, And all of the errors in the same direction, and nearly the same magnitude, which implies a systematic error. Considering that the power measurements, coupler and everything else is probably +/- 0.5dB, that's very reasonable.

I used to say that the auto-attenuate algorithm is looking for an level of -17dBm out of the step attenuator. Since the 8000 uses a 14dB gain pre-amp and the other radios a 20dB gain pre-amp that is no longer a good way to describe the situation. It would now be better to say that the auto-attenuate algorithm is looking for a level at the output of the pre-amp/input to the ADC of +3dBm.

Obviously more data points will be necessary to solidify this assertion, however this is effectively identical to what Brent was reporting. A few more data points like this and it will be reasonable to say that anyone not getting close to theoretical results either has a measurement problem or a hardware problem of some type (coupler, attenuator, cable, radio, etc.).

Interestingly, using 14dB pre-amp gain limits the available dynamic range in comparison to the older radios. This is because I still would not recommend putting more than +13dBm into the feedback port under any conditions.

73!

Scott

P.S. edited to add Stu's corrections--thanks, Stu!
User avatar
W2PA
Posts: 166
Joined: Sun Apr 09, 2017 6:34 pm
Location: LaGrangeville, NY
Contact:

Re: 8000DLE step attenuator values with auto-attenuate

Postby W2PA » Sat May 13, 2017 11:16 pm

Breaking news on the mysterious missing dBs...
I decided to probe my setup further, since PS seems to not mind at all. It's a fairly simple arrangement:

XDC-1,
4-foot (or so) RG-58/U BNC cable connected to the forward port,
other end goes to a 0.6-foot quite small diameter cable (50-ohm RG316) with BNC on one end and SMA on the other
SMA end connects to the PS Input on 8000.

There's a 50-ohm terminator on both the Reflected port of the XDC-1 and the unoccupied Feedback port of the 8000 (on the end of another one of those 0.6-foot converter cables).

What can go wrong? Something, evidently. Because I took it all off, re-attached the jumper between the PS Input and Feedback ports on the 8000, and voila! S-ATT now reads 16dB at 100W.

I don't know why I didn't try this rather obvious thing last week (or so).

So I must conclude that it's less likely that the XDC-1 is defective, and more likely that my two-cable hookup is flawed. I'm going to scrounge for some replacements and see what develops. Personally, my bet's on that funky RG-213 length.

Stay tuned.
73,
Chris, W2PA
User avatar
W2PA
Posts: 166
Joined: Sun Apr 09, 2017 6:34 pm
Location: LaGrangeville, NY
Contact:

Re: 8000DLE step attenuator values with auto-attenuate

Postby W2PA » Sun May 14, 2017 12:27 am

I replaced both lengths of coax with different, new, lengths. Same result as before - back to 5dB in the S-ATT window with 100W.

So I must conclude that either my XDC-1 has an extra 12dB of loss beyond its stated 44 dB, or these coax assemblies, which I bought together, are all defective. I really could use a calibrated scope.

Since PS is working fine, however, I'll continue to operate this way until I can make a measurement. At least I've confirmed the 8000 is working as expected.
73,
Chris, W2PA
User avatar
w-u-2-o
Posts: 5539
Joined: Fri Mar 10, 2017 1:47 pm

Re: 8000DLE step attenuator values with auto-attenuate

Postby w-u-2-o » Sun May 14, 2017 12:49 am

You need a calibrated signal source!

73,

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

Re: 8000DLE step attenuator values with auto-attenuate

Postby W2PA » Sun May 14, 2017 1:15 am

Well, Scott, with one more test I've eliminated the cables as the source. I inserted the exact same cable assembly directly between the two ports on the 8000 and am getting the expected 17 in the window. So the cables are fine.

So either there is an unaccounted-for 12dB coming in via the XDC-1, or else the built-in sense (i.e. feedback) signal source in the 8000 has 12dB more signal at a given power level than we think (along with the input line expecting that level and compensating).

At 100W output,
Loop on back --> 17dB (with either the short jumper or the longer coax line)
Loop around thru XDC-1 sampler --> 5dB
73,
Chris, W2PA
AB2EZ
Posts: 99
Joined: Sun Apr 09, 2017 2:29 pm
Location: Princeton, NJ

Re: 8000DLE step attenuator values with auto-attenuate

Postby AB2EZ » Sun May 14, 2017 1:54 am

Chris

Without knowing the coupling loss of the internal coupler (I.e. how much loss there is between the 100W output of the 8000 and the Pure Signal feedback output port of the internal coupler)... all you are observing is that the coupling loss of the internal coupler is 12dB less than the coupling loss of your external coupler.

Stu
Last edited by AB2EZ on Sun May 14, 2017 2:14 pm, edited 1 time in total.
User avatar
W2PA
Posts: 166
Joined: Sun Apr 09, 2017 6:34 pm
Location: LaGrangeville, NY
Contact:

Re: 8000DLE step attenuator values with auto-attenuate

Postby W2PA » Sun May 14, 2017 2:50 am

Stu - yes exactly.

I've heard tonight of another unit like this via email. Thus it appears that some units, like mine, shipped with this 12 dB higher level of internal coupling (and the added 12dB of attenuation in the input line), and some units, like the others cited here in the forum, did not.
73,
Chris, W2PA
AB2EZ
Posts: 99
Joined: Sun Apr 09, 2017 2:29 pm
Location: Princeton, NJ

Re: 8000DLE step attenuator values with auto-attenuate

Postby AB2EZ » Sun May 14, 2017 2:36 pm

Chris

Attached, is the URL of the KC9XG-SDR Yahoo Groups web site (files section), that contains (February 26 listings) block diagrams of the 8000.

Note that, if the coupling loss of the internal coupler were 14dB too low (compared to what was desired to prevent overload of the step attenuator... e.g. 30dB v. 44dB), and if a compensating, additional 14dB attenuator were inserted, in series, in the portion of the feedback path that is shared by the internal and external paths (i.e. after the back panel jumper)... then that extra 14dB of attenuation would, indeed, be present when using the external coupler.

https://groups.yahoo.com/neo/groups/KC9XG-SDR/files

Stu
User avatar
w-u-2-o
Posts: 5539
Joined: Fri Mar 10, 2017 1:47 pm

Re: 8000DLE step attenuator values with auto-attenuate

Postby w-u-2-o » Sun May 14, 2017 4:16 pm

I was told there is no internal attenuation in the path in question.

73,

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

Re: 8000DLE step attenuator values with auto-attenuate

Postby AB2EZ » Sun May 14, 2017 5:11 pm

Based on Chris's observation that: using the internal coupler, and with 100W output, Pure Signal sets the step attenuator to 17dB......

The transceiver is set to 100W output = +50dBm
The input power to the A/D converter (that is set by Pure Signal... using the "auto attenuate" algorithm in combination with the step attenuator) = +3dBm
The step attenuator setting (that is set by Pure Signal... using the "auto attenuate" algorithm) = 17dB
The preamplifier gain = 14dB

Therefore... the input power to the step attenuator is +6dBm.
I.e. +6dBm (input to the step attenuator) - 17dB (step attenuation) + 14dB (preamplifier gain = +3dBm (input to the A/D converter)

Therefore, the total attenuation associated with the use of the internal coupler is: +50dBm - (+6dBm) = 44dB

If Chris were to remove the Pure Signal external coaxial jumper, terminate the Pure Signal output connector from the internal coupler (i.e. the output port that feeds the jumper when internal coupling is used) with a 50 ohm scope (or a 50 ohm termination + a properly grounded high impedance scope probe), set the transceiver to tune... with 100W of output into a dummy load, and measure the voltage across the 50 ohm-terminated Pure Signal output port (i.e. across the 50 ohm termination) of the internal coupler...

Then, if there is no additional series attenuation between the Pure Signal output port of the internal coupler and the input to the step attenuator... that voltage should correspond to +6dBm being dissipated in the 50 ohm termination.

+6dBm corresponds to 4 milliwatts. 4 milliwatts of dissipation corresponds to a peak (i.e. not RMS and not peak-to-peak) voltage across the 50 ohm termination of: 1.414 x the square root of (.004W x 50 ohms) = 0.63V. If the peak voltage is significantly larger that 0.63V, then there must be additional attenuation in the post-jumper path between the coupler and the input to the step attenuator (at least, in Chris's unit).

Stu
Last edited by AB2EZ on Mon May 15, 2017 12:47 am, edited 4 times in total.
User avatar
W2PA
Posts: 166
Joined: Sun Apr 09, 2017 6:34 pm
Location: LaGrangeville, NY
Contact:

Re: 8000DLE step attenuator values with auto-attenuate

Postby W2PA » Mon May 15, 2017 12:11 am

Stu - thanks for the link to the diagrams.
_ . . . _

I'm going to pick up a digital scope (because I need one around here) and will then do Stu's experiment properly.

Meanwhile, I received this note via email from Joe, K2GX: "The 8000 has about 12db more attenuation then the 100-200D at the PS input. Using internal coupler(jumper in) supplies a 12db higher output then the 100 – 200D. hope the helps."

I replied, thanked him, and asked if maybe only some units shipped that way. He replied, "I am experiencing the same levels as you on my 8000. I have two of Dave’s production couplers, and two beta versions of his couplers, they all give me the same levels. With my 200D’s I needed about 10db additional attenuation. I am not sure if Abhi made changes in production radio’s."

So there is at least a sample of two units behaving this way (i.e. requiring no additional attenuation with the XDC-1) and probably others.

While I'm waiting for a scope, I'm going to open the unit to do the display firmware update. At that time I'll also examine the PS board and see what can be discerned. Work week starts in the early morning so this'll take me a couple of days to get to. PSE QRX.

(Joe's email comments copied with permission.)
73,
Chris, W2PA
AB2EZ
Posts: 99
Joined: Sun Apr 09, 2017 2:29 pm
Location: Princeton, NJ

Re: 8000DLE step attenuator values with auto-attenuate

Postby AB2EZ » Mon May 15, 2017 12:46 am

If an oscilloscope is not available, one can do the following:

Adjust the drive so that, in "tune", the transmitter produces 25W (not 100W) of output. Turn the transmitter off. Remove the Pure Signal coaxial jumper. Connect the Pure Signal output port of the internal coupler directly to the RX2 input (no additional terminations). Set the RX2 step attenuator to 30dB (to prevent overloading of the RX2 preamplifier and/or the RX2 ADC, and for additional protection of the RX2 components that follow the step attenuator). De-select the "ground RX2 on transmit" option. Now, with the transmitter output power, in "tune", set to 25W (not 100W)... place the transmitter in "tune" mode. Observe the RX2 panadapter reading (with RX2 tuned to the transmit frequency). If the reading is +6dBm - 6dB = 0dBm* on the RX2 Panadapter, then there is no additional post-jumper attenuation in the Pure Signal feedback path between the internal coupler and the RX1 step attenuator input.

*I.e. 25W is 6dB less than 100W.

Stu

Return to “PureSignal Operations (PowerSDR & Thetis)”