Call Search
     

New to Ham Radio?
My Profile

Community
Articles
Forums
News
Reviews
Friends Remembered
Strays
Survey Question

Operating
Contesting
DX Cluster Spots
Propagation

Resources
Calendar
Classifieds
Ham Exams
Ham Links
List Archives
News Articles
Product Reviews
QSL Managers

Site Info
eHam Help (FAQ)
Support the site
The eHam Team
Advertising Info
Vision Statement
About eHam.net

donate to eham
   Home   Help Search  
Pages: Prev 1 2 3 4 5 [6] 7 Next   Go Down
  Print  
Author Topic: The Longest Journey Starts With A Single Step  (Read 52860 times)
AD5ZC
Member

Posts: 78




Ignore
« Reply #75 on: January 18, 2015, 10:04:46 PM »


Your tests confirm what Wayne said, that when running lower power the 100W K3 is very clean, so maybe the secret is to build a 300W amplifier and run it at 100W  Smiley

Peter DL8OV
I've been waiting for someone to suggest this...

All you'd be doing with this approach is ultimately putting just as dirty of a radio into the hands of amateurs except now they have a 300 Watt radio, or whatever they can get it up to with their golden screwdrivers.
You know it's true.  Sad
Logged
DL8OV
Member

Posts: 1057




Ignore
« Reply #76 on: January 19, 2015, 12:13:32 PM »

Currently I am experimenting along these lines. A dual 300W Motorola device run at 48V with heavy drain to gate feedback that reduces the power to 120W. Overdrive one of these devices and it will fail faster than a fuse.

Peter DL8OV
Logged
N6KR
Member

Posts: 21




Ignore
« Reply #77 on: January 21, 2015, 07:29:19 PM »

Hi all,

As previously mentioned, we purchased a new Icom M802 (from Amazon, $1813 with no options) last week. We wanted to test one of these transceivers in our own lab.

We very carefully measured transmit IMD at 100 watts on several channels covering the full range of the HF marine band. Here is an example plot from about 12.28 MHz:

    http://www.elecraft.com/Icom%20M802%20100-W%20IMD.jpeg

This plot shows the two 3rd-order products being down by about 27 and 30 dB, respectively. For reference, the plot for the Elecraft K3 posted earlier showed these same tones down 33 and 36 dB--about a 6-dB improvement over this particular Icom M802.

A bit on our test setup: We used a very clean 14.0 V DC power supply with short cables, a high-performance analog 2-tone generator, and a very hefty 50-ohm nonreactive dummy load. We set the top of the spectral plot at 100 W and equalized the amplitude of the tones at -6 dBc, consistent with the usual ARRL method. The tests were done by our senior RF engineer, and I'm sure he was not overdriving the spectrum analyzer Smiley

The sample plots someone posted for a different M802 showed a plot at 135 watts with about 10-15 dB better performance than our new M802 shows at 100 W. This was surprising.

73,
Wayne
N6KR
Logged
DL8OV
Member

Posts: 1057




Ignore
« Reply #78 on: January 21, 2015, 11:44:07 PM »

Good Morning Wayne

The results of your tests are surprising because I was given to understand that the output from marine transceivers were of higher purity than this. Looking at the output spectrum I am quite sure that you were not over driving the spectrum analyzer because a) they are expensive and b) the waveform of an overdriven instrument looks different to this.

I await comments by John, KA4WJA with interest.

However Wayne, although I did not contact Elecraft at the beginning of this thread would you be willing do devote some engineering time to this little quest and perhaps find some way to improve 12V amplifier output quality? 'Cleanest transmitter on the air' might be an interesting selling point.

Peter DL8OV

Edit: Looking at the earlier scans of the M802 your unit probably JUST fits the specification envelope as indicated by the dark line however the gross differences between the output of your unit and the earlier test device are remarkable and, currently, unexplained.
« Last Edit: January 21, 2015, 11:52:00 PM by DL8OV » Logged
K2GWK
Member

Posts: 707


WWW

Ignore
« Reply #79 on: January 22, 2015, 06:18:23 AM »

Waynes tests do not surprise me as there is nothing extraordinary about the PA circuit other than using NPN output transistors instead of MOSFets.
Logged

Guy
Lawn Guyland, New York

K2GWK Website
N6KR
Member

Posts: 21




Ignore
« Reply #80 on: January 22, 2015, 07:45:32 AM »

Good Morning Wayne

The results of your tests are surprising because I was given to understand that the output from marine transceivers were of higher purity than this. Looking at the output spectrum I am quite sure that you were not over driving the spectrum analyzer because a) they are expensive and b) the waveform of an overdriven instrument looks different to this.

I await comments by John, KA4WJA with interest.

However Wayne, although I did not contact Elecraft at the beginning of this thread would you be willing do devote some engineering time to this little quest and perhaps find some way to improve 12V amplifier output quality? 'Cleanest transmitter on the air' might be an interesting selling point.

Peter DL8OV

Edit: Looking at the earlier scans of the M802 your unit probably JUST fits the specification envelope as indicated by the dark line however the gross differences between the output of your unit and the earlier test device are remarkable and, currently, unexplained.

Hi Peter,

Just in case the M802 has some magic elixir that applies to some examples of the radio and not others, we're going to open ours up and examine the PA stage, etc. If we find anything unusual we'll let you know.

We'll also do some K3 plots at lower power levels and post them. As someone alluded to earlier, all spurious products drop considerably when you don't run full power.

73,
Wayne
N6KR

Logged
N6KR
Member

Posts: 21




Ignore
« Reply #81 on: January 22, 2015, 09:03:10 AM »

FYI -- We just tested the M802's transmit-mode power efficiency. When the unit is set for "150 watts" in CW mode, output is actually 132 watts, and current drain is 27.6 amps with a 14.0 V supply. That's an overall efficiency of about 34%. By comparison, a K3 at 100 W typically draws 15 amps at 14 V, or close to 50% overall efficiency.

Lower efficiency may be of less importance in a marine application given the huge batteries, and you could perhaps justify the extra current drain if it resulted in a much cleaner output, per the current discussion. But the K3 is often run from a smaller 12-V battery at Field Day, etc., so we do need to pay attention to this. Part of the improvement comes from low power consumption by all of the rig's non-transmit circuitry (typically around 1 amp).

Wayne
N6KR
Logged
W3RSW
Member

Posts: 606




Ignore
« Reply #82 on: January 22, 2015, 11:01:52 AM »

Quote
We'll also do some K3 plots at lower power levels and post them. As someone alluded to earlier, all spurious products drop considerably when you don't run full power.
Wayne,

As previously posted on this thread.

http://amfone.net/Amforum/index.php?topic=37673.0
http://amfone.net/Amforum/index.php?topic=37674.0
http://amfone.net/Amforum/index.php?topic=37675.0
Last pix showing results from a lot of intervening cabling, switching and in-line linear on the way to a 50 ohm load.  Very messy.  All preceding done with direct 6 ft. of LMR400 to 50 ohm dummy load.  

K3 appears Very nice at 30 watts.
 Not so hot at 80 to 100.

These QS1R scans utilized the internal K3 two tone injection.

A K3 owner (5122) and pretty much satisfied with most of it,
Rick, W3RSW



« Last Edit: January 22, 2015, 11:17:32 AM by W3RSW » Logged

Rick, W3RSW
KA4WJA
Member

Posts: 1098




Ignore
« Reply #83 on: January 22, 2015, 03:07:38 PM »

Wayne,
Thank you for doing this testing and posting the scan...

I have no answer as to why it isn't as good as the scans of the lab that did the FCC Certification testing....
But since that was in 2002, is it possible that different vendors / sources for parts (particularly the PA components), over the years has had a serious impact on actual IMD performance?

[I know that since the March 2011 tsunami in Japan, many semi-conductors have either ceased to be manufactured, or have been 3rd party sourced / manufactured elsewhere (China?), and possible actual internal differences?....while I have no knowledge of whether this has happened for some parts in the Icom M-802, it was the first thing that came to my mind....]


As previously mentioned, we purchased a new Icom M802 (from Amazon, $1813 with no options) last week. We wanted to test one of these transceivers in our own lab.

We very carefully measured transmit IMD at 100 watts on several channels covering the full range of the HF marine band. Here is an example plot from about 12.28 MHz:

    http://www.elecraft.com/Icom%20M802%20100-W%20IMD.jpeg

This plot shows the two 3rd-order products being down by about 27 and 30 dB, respectively. For reference, the plot for the Elecraft K3 posted earlier showed these same tones down 33 and 36 dB--about a 6-dB improvement over this particular Icom M802.

The sample plots someone posted for a different M802 showed a plot at 135 watts with about 10-15 dB better performance than our new M802 shows at 100 W. This was surprising.

Again, I want everyone to understand that I'm not "Icom", and I don't work for Icom, nor have I ever worked for Icom (nor any marine electronics manufacturer / dealer / etc..)
I just a fellow ham, electronics professional, and ocean sailor....who owns and uses two Icom M-802 Marine HF transceivers (on both marine and ham bands) quite regularly (almost daily)....and...

And, in comparison to most modern ham transceivers, I was rather impressed by the transmit spectral purity / low transmit IMD of the marine HF transceivers that I've look at over the years (Icom, Furuno, SGC, etc.), by looking at the FCC certification tests (for maritime "Part 80", etc.), and since I own / use the Icom M-802's, and it is a current production 12vdc HF rig, I thought posting the spectral scans of its transmitter here, as an example of how good a 12vdc P can actually be, would be a good idea....

Remember, I did not design / build the radio, I just own/use it....and I have not done IMD testing of it myself (as I don't sail around with a spec analyzer, Bird load, and two-tone test generator...but I DO have them at home / office.....and while I do have 3 mics for my two M-802's, I do NOT have a spare mic connector (a non-standard connector), so I'd need to cannibalize one of my mics to inject the two-tones into the mic amp, and since I sail offshore / across the Atlantic ocean, etc., with these radios being  my only long-distance communications devices on-board, I did not wish to take apart the mic and possibly cause me future reliability issues....), I simply posted the scans from the FCC Certification test report for FCC Part 80, Part 87, and Part 90....

So again, other than my speculation regarding different actual performance of the discreet components of the radio's PA (tsunami, etc.), I have no way of knowing for sure why Wayne's tests are so significantly worse that the FCC Cert tests...  
But, this IS surprising to me!!!

Here is the FCC Certification test scan of the Icom M-802, operated on 12.230mhz at 135-140 watts PEP....directly from the FCC....



This IMD testing was done by Rhein Tech Laboratories, Inc. in Herndon, VA, in June 2002....

And, here again directly from the FCC....here is their test set-up...
 
Quote
5 OCCUPIED BANDWIDTH - §2.1049

5.1 OCCUPIED BANDWIDTH - §80.211 TEST PROCEDURE

The antenna output terminal of the EUT was connected to the input of a 50W spectrum analyzer through a matched 40dB attenuator. The radio transmitter was operating at maximum output power with and  without internal data modulation. 100% of the in-band modulation was below the specified mask per §80.211.

Specified Limits:
A. On any frequency removed from the assigned frequency by more than 50 percent up to and
including 150 percent of the authorized bandwidth: At least 28 dB;
B. On any frequency removed from the assigned frequency by more than 150 percent up to
and including 250 percent of the authorized bandwidth: At least 35 dB; and
C. On any frequency removed from the assigned frequency by more than 250 percent of the
authorized bandwidth: At least 43 plus 10log10 (mean power in watts) dB



5.2 OCCUPIED BANDWIDTH TEST EQUIPMENT

TABLE 5-1: OCCUPIED BANDWIDTH TEST EQUIPMENT

 Hewlett Packard 8564E Portable Spectrum Analyzer (9kHz – 40 GHz)    Serial Number 3943A01719



Oh, and BTW....as I wrote earlier, the M-802 serv manual states that final PA drive levels / alignment be adjusted for 140 watts PEP into 50 ohm resistive load....and I did have one of my Bird 43's and 1000-wat Bird load on board a few years ago, and did confirm mine was at ~ 142 watts on my Bird (w/ a 250-watt slug), but I have never "cal'd" my slugs....so it could be +/- a few watts....
So yes, going by the current draw figures that I posted, the radio's transmitter is not very efficient, but according to the FCC cert tests it is very clean...


Perhaps someone else has a definitive answer as to why these two scans are so different???
And, while I completely understand that Wayne/ Elecraft have no interest in improving Icom's radio!!  If I had just bought the radio, and found that it didn't meet the specs that it is supposed to meet (under US / FCC rules), I'd be getting in direct contact with Icom Japan, and Icom USA, and I would not stop until I got an Icom engineer on the phone who can explain this....


Maybe we will all learn something here....maybe not.... Smiley


73,
John,  KA4WJA

« Last Edit: January 22, 2015, 03:09:39 PM by KA4WJA » Logged
N6KR
Member

Posts: 21




Ignore
« Reply #84 on: January 22, 2015, 06:27:46 PM »

Hi John,

We can't account for the difference between the FCC results and our results, either. However, we studied the PC boards and schematics today, and found that the M802 uses virtually the same amplifier design as the K3. Just a few component variations that allow the K3 to put out 100 W up through 6 meters.

You're absolutely right that Icom would not cherry-pick a radio for FCC testing. Maybe some of them are as good as that FCC plot shows. Short of doing a valid statistical sample, we have no way knowing. Maybe the FCC tech bumped a knob? Or as my 10-year old son might say, they could have had a "gravity surge" at just the right moment that dragged the IMD products lower.

Yes, the BJTs (PA devices) used in the M802 are now obsolete. I can't say whether there may have been other component changes since 2002 that would affect their results.

One other data point: From our design review, we can't account for the very low power efficiency (34% in TX mode). Relative to the M802, the K3 typically draws about half as much current in TX mode, has TX efficiency of around 50%, and shows better IMD numbers at 100 W. Go figure. RF design is a black art....

73,
Wayne
N6KR


Logged
DL8OV
Member

Posts: 1057




Ignore
« Reply #85 on: January 22, 2015, 11:29:06 PM »

I wonder if the M802 has higher bias on the transmit output stages than an amateur rig? This would account for the extra current draw.

Peter DL8OV
Logged
N6KR
Member

Posts: 21




Ignore
« Reply #86 on: January 23, 2015, 08:07:59 AM »

No, the bias is about the same. The lower efficiency may be due in part to higher current drain contributed by non-transmit circuitry. The M802 manual says max in RX mode is 3 amps (compared to about 1 amp for a K3). There may also be more heavily biased driver stages, etc.

Wayne
N6KR
Logged
KA4WJA
Member

Posts: 1098




Ignore
« Reply #87 on: January 23, 2015, 08:30:49 AM »

Wayne, Peter, et al,
The "lower efficiency" is typical of all marine HF transceivers, at least all that I've seen / measured in the past 40 some years...
The lower efficiency may be due in part to higher current drain contributed by non-transmit circuitry. The M802 manual says max in RX mode is 3 amps (compared to about 1 amp for a K3). There may also be more heavily biased driver stages, etc.

And, FYI, as I wrote earlier....the actual zero-signal transmit current draw is quite high, over 5 amps...
...here are the overall current draws of the radio on the various bands (subtract 2.1 amps, the current draw in receive, for transmitter-only current draw), at approx. 140-150 watts output...

The zero-signal, mic-keyed but no modulation, current draw was 5.4 to 5.5 amps on all bands, 160m - 10m...

Band      SSB-whistle      FSK Carrier
160m       25.6amps         25.6amps
80m         28.7                 28.7
40m         25.6                 25.6
20m         28.7                 28.8     
17m         24.7                 24.6
15m         29.3                 29.4
12m         24.4                 24.3
10m         24.6                 24.5

(FYI, it is spec'd at 30 amps, max current draw @ 13.6vdc....and all of these  figures are about typical for a 150-watt marine HF transceiver....and yes, that is more than your typical "100-watt" HF ham rig...)

All measurements were done near the radio, with a fairly new Klein clamp-on meter (which has proven to be within 0.1 amps of other meters I have tested it against), at ambient temp of about 70*F....with battery voltage of 13.7vdc, and voltage at radio of approx. 13.3-13.4vdc under full current draw of ~ 29amps...

I've actually not measure the PA nor driver current directly....but if you subtract about 2 amps (what the rest of the rig draws in receive), that leaves about 3.5 amps for the non-shared transmit stages (primarily the driver and PA stages), so figure about 3 to 3.2 amps of collector current for the pair of output devices, at zero signal...


And, BTW.....with 11 years of experience with my own M-802's, and > a dozen others, the current draws are all about the same.... ~ 2 - 2.1 amps on rec, 5.5 amps at zero-signal transmit, and 28 - 29 amps peak on transmit....
(the manual's spec of 3 amps on rec. is way high....but it's 30-amp spec on transmit is pretty close..)


I hope this helps...

73,
John,  KA4WJA
Logged
DL8OV
Member

Posts: 1057




Ignore
« Reply #88 on: January 23, 2015, 11:56:20 AM »

Here is some pure guesswork as to why we see differences here:

1) Different bias levels on either the driver or the PA (Wayne has checked the PA).

2) Different transistors, maybe higher power devices on the earlier units that are now unobtainable.

3) Different ferrites used on the PA input and/or output. I don't know what Minicircuits use in their wideband RF transformers but I would love to try some in a 100W HF amplifier.

4) Different drive levels, especially if they were using different transistors before.

5) How do they apply power to the drain/collector and has this changed? The correct way is through a separate bifilar transformer but a short cut is to use a tap on the output primary.

6) Changed supply decoupling arrangements between the old and new units.

7) Do they use emitter/source degeneration? If so, have the arrangements changed?

Cool Different components in the base/collector or gate/drain feedback.

9) Different bias circuit between the old and new units.

There are various ways to occupy the mind during a two hour trip to the supermarket and I have spent the time thinking of all the tricks used to clean up a power amplifier. Unfortunately I do not have an old or a new M802 to examine but I know some people who do  Wink

Peter DL8OV
Logged
N6KR
Member

Posts: 21




Ignore
« Reply #89 on: January 23, 2015, 03:41:52 PM »

In case anyone was wondering, we looked at the waveform and see no evidence of compression.

Wayne
N6KR
Logged
Pages: Prev 1 2 3 4 5 [6] 7 Next   Go Up
  Print  
 
Jump to:  

Powered by MySQL Powered by PHP Powered by SMF 1.1.11 | SMF © 2006-2009, Simple Machines LLC Valid XHTML 1.0! Valid CSS!