Pages: [1] 2   Go Down

Author Topic: Transceiver Intermodulation Distortion ( IMD )  (Read 1006 times)

HAMHOCK75

  • Member
  • Posts: 1297
Transceiver Intermodulation Distortion ( IMD )
« on: September 13, 2022, 03:38:42 PM »

Recently I had a project where it would be nice to have a transceiver with low IMD.

This is a link to a Rob Sherwood article in which he describes the Collins 32S3 was the cleanest transmitter he has owned measured at -36.5 dBc or about -42 dB PEP. He also includes the Yaesu FT1000MP Mark V which measures -42 dBc or -48 dB PEP.

http://www.madisondxclub.org/NC0B_Rig_Test.pdf

This is a pdf of the performance of many modern rigs.

https://www.remeeus.nl/hamradio/pa1hr/productreview.pdf

It appears that there is no shortage of solid state rigs that can match or better the 32S3 performance.

This is another test of the IMD performance of the Yaesu FT1000MP Mark V introduced 22 years ago in 2000 using its class A output amplifier.



This is the IMD performance of the Yaesu FT1000MP Mark V using its class AB output amplifier.



Another difference in the test results is that the Collins 32S3 is a 80-10M transmitter. The Icom 7300 covers 160-6M. In the chart the 7300 has a -30 dB PEP rating which is not great until you realize, the -30 is due just to 6M. From 80-10M it can match or exceed the 32S3 performance as shown in these tests.

https://www.ab4oj.com/icom/ic7300/7300notes.pdf

Logged

KT4WO

  • Member
  • Posts: 425
    • homeURL
Re: Transceiver Intermodulation Distortion ( IMD )
« Reply #1 on: September 14, 2022, 04:34:23 AM »


THANK YOU for posting this!!
I have been jumping up and down trying to tell folks for
YEARS to check the TX IMD before they buy!!

One local said " the TX side does not affect me"
Not joking :(
Logged

KT4WO

  • Member
  • Posts: 425
    • homeURL
Re: Transceiver Intermodulation Distortion ( IMD )
« Reply #2 on: September 14, 2022, 04:40:52 AM »

"One local said " the TX side does not affect me"
Not joking :(  "

BTW...This is the same guy who runs "ESSB" and his signal
is 10Khz wide. He thinks he is a "radio DJ" or something.

Also-  My ELAD FDM-DUO "claims" -39 TX IMD but I have not checked that.
Logged

WA6NUT

  • Member
  • Posts: 25
Re: Transceiver Intermodulation Distortion ( IMD )
« Reply #3 on: September 14, 2022, 07:11:30 AM »

Rob Sherwood's article "It's Time to Clean Up our Transmitters" (QST, November 2019, pp. 38-41) addresses IMD.  His article mentions adaptive predistortion (APD) as a solution to IMD.  APD not only reduces IMD in the transceiver, it also reduces IMD when an external amplifier is used.  Currently, only Apache Labs offers APD (PureSignal), but Elecraft and Flex Radio have announced plans to incorporate APD in their products in the future.

When I enable APD (PureSignal) with my Apache Labs ANAN-10E transceiver and Ameritron AL-811 amplifier, splatter on adjacent frequencies disappears.  It really works!

At the end of his article, Rob Sherwood states, "Transceiver receive performance today is excellent, so the next order of business should be to make a concerted effort to improve the transmit side of the equation."  I agree.
Logged

W9IQ

  • Member
  • Posts: 8866
Re: Transceiver Intermodulation Distortion ( IMD )
« Reply #4 on: September 14, 2022, 07:25:41 AM »

Predistortion will be mainstream in a few years. Icom's announcement that it will be turned on in the 7610 to support their new linear will help drive other manufacturers.

The processors in a modern transceiver are comparatively idle during transmit. That is a lot of available processing power for the pedistortion feature. Kudos to Icom that they foresaw this and included the necessary hardware connections to their processor chain.

- Glenn W9IQ
Logged
- Glenn W9IQ

God runs electromagnetics on Monday, Wednesday and Friday by the wave theory and the devil runs it on Tuesday, Thursday and Saturday by the Quantum theory.

VK6HP

  • Member
  • Posts: 1222
Re: Transceiver Intermodulation Distortion ( IMD )
« Reply #5 on: September 14, 2022, 09:20:39 AM »

I think it's great that manufacturers are responding to pressure and producing cleaner transmitters in the course of the last decade.  It's worth noting that for conventional systems (no pre-distortion) it took about 50 years from the time the 32S-3 was designed until the cleaner solid transmitters of the past decade began to emerge (noting a few honourable exceptions along the way).  No-one needs reminding that there were many shockers in the interim.

From experience with my 32S-3 and other contemporary systems, including a TS-890S, I would note that the high-order IMD roll-off of the 32S-3 is matched by very few ham transmitters (use the 9th order product as a quick comparison in the table referenced by Hamhock75 and note the -76 dB with respect to PEP for Rob Sherwood's 32S-3).

I've used my 32S-3 quite a bit as an exciter in IMD tests and characterized it in that role.  I don't doubt that Rob's unit is typical and well within specification but would note that my measurements of transmitter IM3 are e.g. 4 dB or so better on 20 m, with IM9 commensurately improved.  I've noticed that 516F-2 power supply performance, as well as overall alignment and driver/PA tube health, have roles to play in the exact figures measured.

For the more contemporary side of my shack, I too look forward to mainstream pre-distortion capable exciters which will allow modern LDMOS power amplifiers, including my KPA1500, to finally beat the 1960s Collins transmitter + PA lineup in the IMD stakes.

73, Peter.   
Logged

HAMHOCK75

  • Member
  • Posts: 1297
Re: Transceiver Intermodulation Distortion ( IMD )
« Reply #6 on: September 16, 2022, 05:24:51 AM »

This brief excursion into IMD was a bit revealing that it may not have been the transition to solid state that brought poor IMD. It appears it began a decade earlier between 1960 and 1970; the prime mover being sweep tubes and the likely influence of a certain CB mentality that prized power output above all else. 

During that decade, companies like Collins with the 32S3 and Heathkit maintained IMD specifications in their popular HW101, SB102, SB401, etc ham gear at -30 dB relative to two tones.

Yaesu with their FT101 series and the FLDX400 state better than 30 dB. In their FTDX400 ( introduced in 1967 ), 401, 560 series, it drops to better than -25 dB. I suspect these are PEP numbers because I happen to own a Tempo One which is really an FT200. The Tempo One manual does say -30 dB ( P.E.P. ).

The other reason is I feel pretty certain Yaesu specifies PEP is that VK6HP actually measured one of the FTDX4xx series near it's rated output power finding the IMD was around -18 or -19 dBc if I recall correctly.

Swan with the 500c and 700cx specify IMD down approximately 30 dB in 1969. I suspect PEP.

Drake with their TR4C specifies -30 dB below PEP.

« Last Edit: September 16, 2022, 05:33:12 AM by HAMHOCK75 »
Logged

VK6HP

  • Member
  • Posts: 1222
Re: Transceiver Intermodulation Distortion ( IMD )
« Reply #7 on: September 16, 2022, 06:22:31 AM »

That's correct: I did indeed measure IM3 at about -20 dBc (-26dB wrt PEP) for the FTDX560 at 300 W PEP output, with new (old stock) 6KD6 sweep tubes.  Interestingly, the FTDX560 is a close contemporary of my 32S-3, the latter being a "round emblem" unit built in 1969 (estimated from the dates on the crystals).  So, very good and very bad performers essentially overlapped.  The Yaesu is at least a fine looking radio and has now gone to a local collector who wanted it on his shelf and who promised (for the sake of us all) to use it for CW, or at least no more than 100 W PEP.

At the other end of the 50 year period being discussed, it's also interesting that there was no single technology break-through that led to progressively better IMD specifications: it's likely just a case of pressure and reward leading to improved design using well-established topologies and principles, many of which had long been applied outside ham radios. To some extent that's true even of contemporary pre-distortion techniques but, in that context, ham radio does provide something of a sand-pit for designers in terms of the need for versatile, robust systems suitable for a range of hardware and users.

73, Peter
Logged

HAMHOCK75

  • Member
  • Posts: 1297
Re: Transceiver Intermodulation Distortion ( IMD )
« Reply #8 on: September 17, 2022, 02:44:17 PM »

Quote from: VK6HP
it's also interesting that there was no single technology break-through that led to progressively better IMD specifications

There is one technology that has been grinding forward mostly driven by demand from the defense industry. The number of devices to build an amplifier at a given output power continues to decline. At the 1.2 kW class, the Mercury IIIS is down to a single device. The older MFJ Ameritron ALS-1306 uses eight devices grouped as four matched pairs.

The Mercury IIIS device is actually two devices in one package, but likely matched from the same wafer. Before these type of devices, individually packaged devices had to be hand selected for match.

My study of the 2KL amplifier shows that even hand matched devices are far from ideal. I suspect that wafer matched devices are likely better for distortion. The Icom 7100, 7300, and 7610 all use versions of these single devices.



The gain difference at the collectors ( blue and yellow traces ) is pretty obvious with the above matched devices.



With tube rigs, its not as easy to measure a lower power as with solid state. Did you even see what the IMD was like with the 32S3 if the output power was lower?



« Last Edit: September 17, 2022, 02:48:06 PM by HAMHOCK75 »
Logged

VK6HP

  • Member
  • Posts: 1222
Re: Transceiver Intermodulation Distortion ( IMD )
« Reply #9 on: September 17, 2022, 08:41:26 PM »

It's certainly true that good on-wafer matching helps in many topologies, e.g. push-pull high power linear amplifiers. I seem to recall LDMOS devices being mainstream in many telecoms applications in 2010 or so. I bought my KPA1500 in 2017 and recall re-reading Langford-Smith and others on device transfer function matching accuracy to reconcile the residual even-harmonic suppression I observed in the pre-filtered output of the unit.

I don't have a lot of 32S-3 data at low outputs but looking at a plot for 40 W I see IM3 at about -45 dBc, IM5 a few dB lower, and not much else visible on the power scale and averaging I was using to simultaneously observe the KPA1500 output.  (Most measurements were done with a drive power attenuator, for safety).

One thing to watch is that some older radios such as the FTDX560 have quite a large amount of audio and later stage distortion, which actually contributes noticeably to the RF low-output intermodulation spectrum. That was the case with the FTDX560 - something I hadn't seen before. There are some vestiges of carrier and minor components in the 32S-3 output, but nothing like the FTDX560.

73, Peter.



Logged

KA4WJA

  • Posts: 1601
    • HomeURL
Re: Transceiver Intermodulation Distortion ( IMD )
« Reply #10 on: September 19, 2022, 01:15:22 PM »

Hamhock75,
1)  Welcome to the team! :)  (okay, maybe my humor isn't great....what I mean is, welcome to the party of hams that are concerned about our transmitters' IMD!)

There's a wealth of info and data on this, including IMD scans of the FT-1000MkV in Class A (w/o any ALC action), 32S-3, K3, etc. etc...
I welcome you all to have a look!

And, yes....there is more to this than just the PA...
Tube or SS...good linear RF tubes or sweep tubes, etc...
The power supplies, the output matching, etc., all do play a role. 
{and, ironically just 2 nights ago, talking with some friends on 80m....we discussed some "in-band" distortions that are inside the rigs passband, and the ability of the ANAN's PureSignal to clean them up as well....so, maybe if you all could design/implement a pre-distortion module for those old '560's, etc, you may breath new life into them!  hi hi :) }


2)  And, yes Glenn....I agree that pre-distortion should be "mainstream" in a few years.

Although, how many years will it take to before pre-distortion rigs represent an over-whelming majority of the rigs on our HF bands? 
That's a question where there may not be a definitive answer to?
But, my guess is that it will many, many years....maybe even a couple decades....{as this isn't like the issue (incompatibility) as when the amateur HF community moved from AM DSB to suppressed-carrier SSB...}

Predistortion will be mainstream in a few years. Icom's announcement that it will be turned on in the 7610 to support their new linear will help drive other manufacturers.

The processors in a modern transceiver are comparatively idle during transmit. That is a lot of available processing power for the pedistortion feature. Kudos to Icom that they foresaw this and included the necessary hardware connections to their processor chain.

- Glenn W9IQ

Maybe I'm wrong, but while I applaud Icom's move, as we've been waiting for this from them (and Flex, Elecraft, etc.) for over a decade now....I don't see the world of amateur HF comms changing to pre-distortion rigs for quite a while.  :(

And, even after the new version of the 7610 and PW-2 are released, how long before other manufacturers release their pre-distortion rigs...rig/amp combos?
Now, even if these are all released quickly, how long before pre-distortion rigs become the majority of rigs on-the-air?
Again, I suspect a long time...

Even looking at the explosion of the '7300 onto the market as example, it's been over 7 years and while they sold >20,000 in the first 2 years, and double that since then....even if you double that again in the next 4 - 5 years, they'd still be a minority on-the-air...
Just saying that I love that Icom is doing this (and, I have admired Apache Labs / ANAN for doing this years ago!), it's just not going to be all splatter on the bands will disappear in the next year or two. 



3)  Also, in regards to the new version of the 7610 and PW-2....I may be wrong here, but it appears that the 7610 cannot take in (sample) an analog RF sample, so while its software/firmware can be changed/updated to use the DSP processor to implement pre-distortion, it seems (as of now) that the only amp that can be placed inside the pre-distortion loop is the PW-2, as it will have a digital-sample of the amp's output and will be able to send that to the '7610 (via its ethernet cable?, etc.?) to allow the 7610's DSP to use that sample info?

Now, is this a precursor of what is to come from Icom?
Will they debut new rigs that will have analog RF sample inputs?  Etc?
Maybe...

But, here again, this isn't going to mean that all the splatter on-the-air is going to disappear in the next year or two.  :)



4)  Okay, Hamhock75, et al...here's some more info/data for you all...
Have a look at these threads...

Perils of Hi-Power and some IMD scans
https://www.eham.net/community/smf/index.php/topic,137162.msg1273707.html#msg1273707


Better IMD Please (where you'll see that many have been talking about this for decades, and talking about pre-distortion now for > 10 years....and some saying a decade ago  that Elecraft, Flex, etc. will have pre-distortion "soon")
https://www.eham.net/community/smf/index.php/topic,79120.msg556484.html#msg556484


Lost scans (tinypic!), but lots of data / numbers
https://www.eham.net/community/smf/index.php/topic,97093.msg1053647.html#msg1053647
https://www.eham.net/community/smf/index.php/topic,101398.msg821031.html#msg821031


Tube vs. SS, etc....
https://www.eham.net/community/smf/index.php/topic,100600.msg809639.html#msg809639


And, some posts on IMD, rig choice, etc....
https://www.eham.net/community/smf/index.php/topic,128776.msg1194749.html#msg1194749
https://www.eham.net/community/smf/index.php/topic,128776.msg1195427.html#msg1195427
https://www.eham.net/community/smf/index.php/topic,128776.msg1227431.html#msg1227431
https://www.eham.net/community/smf/index.php/topic,128776.msg1227433.html#msg1227433




Hope this helps with the discussion.

73,
John,   KA4WJA
« Last Edit: September 19, 2022, 01:23:25 PM by KA4WJA »
Logged

HAMHOCK75

  • Member
  • Posts: 1297
Re: Transceiver Intermodulation Distortion ( IMD )
« Reply #11 on: September 19, 2022, 04:28:13 PM »

Quote from: VK6HP
One thing to watch is that some older radios such as the FTDX560 have quite a large amount of audio and later stage distortion, which actually contributes noticeably to the RF low-output intermodulation spectrum.

I inherited a Yaesu FLDX400. It also had a microphone amplifier issue. I had to do some work on it described below to get it working properly,

https://forums.qrz.com/index.php?threads/yaesu-fldx-400-nl-6s.585383/page-5#post-5220850

I plan to test the FLDX400 for IMD in the near future. The topology of amplifiers may favor tubes especially as power goes down. The push pull design has to deal with cross over distortion that does not happen in single ended tube designs

Quote from: KA4WJA
Hamhock75,
1)  Welcome to the team! :)  (okay, maybe my humor isn't great....what I mean is, welcome to the party of hams that are concerned about our transmitters' IMD!)

Thank you for that wealth of information. The abundance of your information reminded me of where information is missing. I noticed that although Yaesu has always had an IMD specification when they arrived on these shores, Kenwood did not have an IMD specification until the arrival of the TS830S which was introduced around 1980.

The big surprise is Icom. I cannot find an Icom specification for IMD for any of their rigs right up to the present.
Logged

VE7DQ

  • Member
  • Posts: 325
Re: Transceiver Intermodulation Distortion ( IMD )
« Reply #12 on: September 19, 2022, 04:42:39 PM »

Push-pull design in the FLDX400?  Where?
Logged

HAMHOCK75

  • Member
  • Posts: 1297
Re: Transceiver Intermodulation Distortion ( IMD )
« Reply #13 on: September 19, 2022, 05:00:39 PM »

Quote from: VE7DQ
Push-pull design in the FLDX400?  Where?

Nowhere. The FLDX400 is a tube design.
Logged

VK6HP

  • Member
  • Posts: 1222
Re: Transceiver Intermodulation Distortion ( IMD )
« Reply #14 on: September 20, 2022, 04:32:32 AM »

Quote from: VK6HP
One thing to watch is that some older radios such as the FTDX560 have quite a large amount of audio and later stage distortion, which actually contributes noticeably to the RF low-output intermodulation spectrum.

I inherited a Yaesu FLDX400. It also had a microphone amplifier issue. I had to do some work on it described below to get it working properly,

https://forums.qrz.com/index.php?threads/yaesu-fldx-400-nl-6s.585383/page-5#post-5220850

I plan to test the FLDX400 for IMD in the near future. The topology of amplifiers may favor tubes especially as power goes down. The push pull design has to deal with cross over distortion that does not happen in single ended tube designs


You have a fighting chance with the FLDX400 mic circuit. The FTDX560 circuit is plain odd, with the mic gain in the wrong place and an ineffective overall feedback loop.  There were a few QST articles to improve it but with poor results.  It looks like most people used a very high output mic and called it good enough.  As a teenager with a second-hand FTDX401, I was blissfully unaware of any of its limitations and just used the crystal mic the previous owner threw in. 
Logged
Pages: [1] 2   Go Up