Aeroflex IFR3413 IFR3414 IFR3416 Signal generator

[G0HZU]

Apart from a few (unsuccessful?) repair threads, there doesn't seem to be much info on Eevblog about the IFR 3413, IFR 3414 and IFR 3416 sig gens from Aeroflex / IFR.

I've attached the datasheet below.

The aim of this thread is to discuss the good and bad points about these sig gens and how they compare to the competition. I'm a long time user of these sig gens having used many of them at work for about 20 years and I also have a few of them here at home.

At first glance, these sig gens look quite similar to the older Marconi 2023/4/5 analogue sig gens and they share a similar layout. However, the IFR 341x generator is a vector signal generator, and this allows almost any form of modulation to be generated up to about a 50MHz bandwidth using the I and Q inputs of the vector modulator.

The vector modulator(s) used in these sig gens have high performance, and this allows these sig gens to generate digital modulation with very low distortion on an adjacent channel. To maintain the low distortion through the whole signal path to the output, these generators also feature high performance amplifiers in the signal path.

This is great for manufacturers of modern wireless systems, as they can use these sig gens to produce spectrally pure test signals with various forms of digital modulation.

The power consumption is quite high compared to a classic analogue generator such as the Marconi 2024. These generators also run quite hot internally and they produce a fair bit of fan noise when fully warmed up. The amount of fan noise may depend on the version of the hardware. It appears the later models have improved thermal management so they run cooler and quieter and should be more reliable.

The 4GHz IFR 3414 typically consumes 125W when running. Compare this to just 70W for the Marconi 2024 and just 25W for the Marconi 2022. Both these sig gens produce significantly less fan noise than the 341x series.

These IFR 341x sig gens are generally quite expensive on the used market despite their age. This is probably because of the high performance they offer via the vector modulator and there may be regular industry demand for drop in replacements when they fail.
They were designed to be serviced via module swapping and so there isn't much info available for component level servicing. This makes it difficult for companies to get these generators repaired when they fail and this probably maintains demand for working examples.

The close to carrier phase noise is quite high on these sig gens because the vector modulator operates up at UHF and this is downconverted to lower frequencies when applicable. So the phase noise at (say) 14MHz is as high as the phase noise up at UHF. This makes this sig gen a poor choice if you expect low phase noise across the HF bands and into VHF. The phase noise really is quite high on this generator across HF and into VHF.

I can post up some performance plots and add info about reliability and what often fails on these sig gens.

Does anyone here have one of these generators? Has it been reliable and are you pleased with it?

[G0HZU]

If you begin to exploit the full potential of the IFR 341x sig gens, you can do some fairly advanced things using the vector modulator. For example, the waveform in the image below is an 8MHz wide NPR waveform that I set up using lots of tones transmitted at the same time.

If you transmit enough tones , the signal will look like wideband noise. If you then mathematically extract some of the tones from the waveform you can make a notch in the noise at an offset and width of your own choosing. This forms the basis of a noise power ratio generator. In the example below, I've generated over 800 (eight hundred) tones centred on 1.2GHz using the IFR3414 and I've also added an NPR notch 50dB deep and you can see this notch in the waveform below. All this is done on a PC and then the waveform is sent to the IFR 3414 via the vector modulator inputs.

There are over 800 tones  spaced at about 10kHz apart in the waveform below. This is why it is about 8MHz wide. Several tones have been deleted from the waveform in order to create the NPR notch. It's also possible to have many more tones than this. I tried again with 5900 tones for example. The more tones, the closer the waveform resembles white noise.

There is a dual arbitrary waveform generator option (option 5) for these sig gens and this is fitted internally on the main control board so this can be programmed to produce custom waveforms. However, I think the best performance will come from modulating it externally as in the example above.

I'm not sure how capable the dual arb generator is in the Siglent SDG2042X or the alternatives. In theory at least, it could be used to drive the I and Q inputs of the 341x vector modulator to provide waveforms like the one pictured below. A lot depends on how much memory the SDG2042X has and how easy it is to load waveforms into it. If it can do this, then the SDG2042X and an IFR341x would make up quite a versatile system.

Has anyone tried modulating the IFR 341x series with the Siglent SDG2042X? I've been tempted to buy one but I don't know how easy it is to generate compatible waveforms and load them into the 2042X to generate custom waveforms with the IFR 341x sig gens.

[Mkpirulo]

Hi,

Thanks for sharing your experience with the 3410 series.

I just recently purchased an Aeroflex CD with helpful software for the 3410 series. What I was really looking for is the service manual, it was available on CD. I had a chance to purchase it four years ago before Aeroflex discontinued support and switch to their new name VIAVI Solutions.
Phase noise seems really good looking at these specs..

Michael

[G0HZU]

Hi Mkpirulo
The phase noise is fine up at UHF, but because it uses a downconverter from about 860MHz for frequencies below 375MHz, the phase noise at a carrier frequency of (say) 14MHz is similar to the phase noise up at about 1GHz. It really is quite noisy in this respect.

See the phase noise plot below at 14MHz to see what I mean.

I think there are at least two versions of internal and external metalwork for these sig gens as the early versions tend to run a bit hot and they produce more fan noise.

An ex Marconi engineer on another forum advised me that you can tell if you have one of the older units by looking at the fan grille cutouts. The fan grille is improved on the later versions for improved airflow.

I also spotted that the internal metalwork is different on the later units and the vent holes along one side of the chassis are also different on the later units. I've got several of these sig gens here, one is the older (hot) type and the others are the later version with improved thermal management.

The later units seem to run cooler and with less fan noise.

[Mkpirulo]

Hi,

Ok understood might not be a good choice for HF frequencies. But if your main focus is 440 on up then it’s not bad. Thanks for sharing your knowledge on these units. I have 2 3416 and 1 3413. One 3416 is the (hot) and one is the (cool) version: My 3413 is the (cool) version. Options on my (hot) 3416 are 03, 05, 09, 010.
I’m attaching pictures to show others the difference. Since I have both I’ll run some thermal couples in different locations to see what the temperature difference is in fact. I can do this with the 3413 as well.

I’ve worked on these power supplies, I understand their limitations, see my other posts on various forums. If your supply has SMDs under L12 then the fan that cools the supply must be running in standby, from the factory it does not. If your supply has two large ceramic resistors by L12 then your fan does not need to run when in standby. For those of you with the SMDs just plug the power supply cooling fan (with little adapter cable)into one of the two available fan ports near the center inside of the power supply. The power supply will control the fan speed when the supply needs more cooling. The 3410 series also controls the fan but it looks at the cpu board temp..

Hot version units have 612NGHH 33cfm, cool units have the 622HH fan 32cfm. These ratings are at 0 pressure, no grill. They need the integrated grill for RF leakage issues.


Hope this helps anyone trying to keep these units cool. My suggestion would be to leave the unit unplugged until you need to use it. Don’t leave it on all day especially if your using it at home or in your garage, much warmer then if it’s being used in an air conditioned environment. Dont mount them into a rack, they'll have no chance of sucking in cool air.

Michael

[G0HZU]

Since I have both I’ll run some thermal couples in different locations to see what the temperature difference is in fact. I can do this with the 3413 as well.

Over the weekend I did this with my 3416. This is an early (hot) version. I ran it for just over an hour from a cold start at a 25degC ambient and also at an 18degC ambient. I attached a thermocouple to the metalwork lid of the main RF board and placed it above the output section near the rear panel. I used thermal paste to get a reliable contact.

In both cases the sig gen settle in about 40 minutes. See the result below:

I've not yet tried this with the later (cool) version of the 341x generators.

I also had a look through the firmware code and it looks like this generator can run in various modes. It looks like the fans speed can be monitored remotely and it may also allow the fan speed to be set manually using remote commands.
I think that the temperature can also be monitored remotely.

I haven't yet worked out how to do this though. Probably via some form of remote terminal.

[G0HZU]

I don't know what the differences are between the firmware versions. The firmware may affect how the fans are controlled because my early (hot) 3416 starts up with the fans running at a very low setting (I can only just hear them).  Over the course of 40 minutes the fans gradually get louder and louder.

By contrast, my later (cool) 3414 generators start up with a higher initial fan speed. This doesn't change much over 20-30 minutes and as long as the later 3414 versions are kept away from other test gear, they seem to be able to run with tolerably low fan noise. If I put one on top of a stack of running test gear, it generates more fan noise after a while.

I've seen various firmware versions (4.06, 4.07, 4.10 and 4.12) and I assume the same example of firmware can be loaded and run in all 341x platforms. However, I've not tried this.

The fan noise and high phase noise and relatively high power consumption plus the lack of features (compared to other Agilent vector sig gens) means that I rarely use any of the 341x generators I have here. The RF levelling vs frequency is very good and so is the accuracy of the levelling over a 120dB range. The AM and FM modulation accuracy and quality is very good as well. So I do use them sometimes.

[Mkpirulo]

Nice work.

Connect a set of small wires from the plug on both fans to see if both are controlled at the same time. Fan voltage starts around 6vdc at least measured with a voltmeter. Or scope it could be pwm%. Would be interesting to see if the fans are controlled independently of each other.
My (hot) 3416 has some of the electrolytic caps under the display ribbon cable slightly discolored, it most have gotten hot at some point. The grill on the (hot) ones appear to be more restrictive.

[G0HZU]

Yes, I see a definite difference in airflow between the two versions. I tried fitting the outer chassis cover from a later model to the old version of the 3416 and it ran a bit cooler and quieter.

At the moment, the only way I can easily monitor the fan speed is to listen to the rear of the unit using a microphone feeding a PC soundcard. By using some spectrum analyser software on the PC I can see a peak that corresponds to the fan speed. It slowly increases in frequency as the fans speed up. It also appears that the PSU fan runs at a different speed if I move the microphone to the fan PSU.

At some point I plan to try and reverse engineer parts of the firmware because it appears that there are a few diagnostic tools built into it. It may be possible to launch these tools at bootup and have some extra diagnostic controls via a PC terminal. Many years ago I did the same with an Agilent VNA and an Agilent scope and found some useful (undocumented) stuff in their code.

At first, I'm going to try and find as many versions of the firmware as possible and look for differences and see how the password system is configured. There appears to be an FTP mode with a password and it would be nice to hack into that.
Some Agilent sig gens support a kind of web page via FTP and maybe these Aeroflex sig gens have something similar. It may be possible to hack into them and access a few hidden things via this interface.

I've also noticed that my 3416 runs cooler when used above 4GHz. The power consumption is lower by a few watts and the fan speed slowly decreases to a lower point when the 3416 is used above 4GHz. The sig gen uses a different output stage above 4GHz. The output PA device is made by Macom and it costs about £350 to replace. Hopefully, these parts are reliable! They do look to be well bolted down although there was no thermal grease between the 6GHz PA metalwork and the main RF metalwork. This was a surprise.

[rickells]

 There are a limited number of  retrospective  forum questions/answers about the Aeroflex IRF 3410 - 3416 Signal Generator series
  ( 3410, 3412, 3413, 3414, 3416 ) on the Groups io Marconi forum,
      ( but these are subsequenty blocked by the Moderators, who prevent further discussion ).

  Search there for Aeroflex 3410, & individually for 3412, 3413, 3414, 3416
i.e.
          https://groups.io/g/Marconi-Test-Instruments/search?p=Created%2C%2CAeroflex%20%203416%203414%2C20%2C1%2C0%2C0&q=Aeroflex%20%203416

https://groups.io/g/Marconi-Test-Instruments/search?p=Created%2C%2CAeroflex%20%203412%2C20%2C1%2C0%2C0&q=Aeroflex%20%203413

Note that,  ... sadly, the Moderator / Group owners there "Lock" all discussion about these Aeroflex IRF Signal Generators due to their  "... not being Marconi content "  or older than 60 days . 

Glean what you can from the above, but all discussion on Aeroflex IRF Signal Generators of this Series have been obfuscated by those particular Moderators. 

[G0HZU]

To be fair, I did initially ask on the Marconi forum if it was OK to post stuff about IFR and Aeroflex and I was told that it was fine by one of the moderators. However, this may not be an opinion shared by everyone on that group...

Eevblog is probably the best forum to discuss test equipment because it will get loads more views and it's easier to post up technical content via attachments etc. The thread will also stay open for updates and it will hopefully be easy to find with a search engine.

It would also be nice to provide somewhere that lists lots of common faults and how to fix them. I've fixed quite a few minor faults on these sig gens already.

The PSU module is definitely a weak area and I've repaired three PSUs now.

The components on the RF board are prone to failure because it runs so hot. I've seen SHF-0189 MMIC failures and failed electrolytic caps on the RF board for example.

The display backlight can fail and I converted one of my 3414 sig gens to LED backlighting with good results.

The little 12V standby PSU module can also become intermittent. These are potted inside so they are very difficult to work on if they fail.

[G0HZU]

I recently looked through my old images of the insides of my IFR 3416 and I now think it is possible to mount the entire driver/RF/6GHz boards vertically together on the back of the control board.

There is a spare/unused connector poking out of the side of the chassis and I wondered why it was there. It is part of the driver board. Having looked at how there is a similar connector on the back of the control board, I think it is possible to rotate and remount the RF boards vertically on the control board in a kind of 'service' mode.

This will presumably allow access to the 6GHz board and the driver board with the sig gen still powered and running.

This will greatly help with faultfinding, especially with the 6GHz board.

Has anyone tried this? Is it listed in any service manuals as being possible?

[G0HZU]

If anyone has access to any service docs can they also confirm if there is an additional 'engineering' interface that can be accessed? Looking through the firmware I think there is an engineering menu and also an FTP interface. I don't yet know how to activate these but maybe it is described (or at least hinted at) in some service docs?

However, in this case, there may already be diagnostic software available for the Aeroflex sig gens. This isn't the same as the DIAG commands listed in the programming section of the manual. I think there are other things that can be monitored and controlled via a different interface. I think the fan speed can be monitored and even controlled remotely for example.

I've done similar things with car ECUs where the manufacturer has hidden a diagnostic port that would have been used by their engineers. They remove all trace of the port hardware but left the code intact. So it was possible to reverse engineer the port requirements and reactivate the engineering port and then reverse engineer all the commands by looking through their code. This allowed real time monitoring of all the variables in the ECU as it was running as it permitted direct access to all the memory addresses in RAM.

[G0HZU]

Here's an image showing a comparison between the repaired display (converted to LED backlight) and an IFR3414 with the original backlight.

When viewed with the naked eye, the original CFL display is slightly brighter, but that is because I have run the LEDs at a fairly low current. The datasheet suggests the LEDs will last up to about 100,000 hours at this current.

It's fairly easy to convert the display to the LED system once the original CFL tube fails. The main niggle is that the display is quite fiddly to remove as there are lots of screws holding the front panel together. The alternative is to buy a replacement display for approx. £100 to £130 from China or Japan.

The LED backlit display is the top display in the image below.

[Mkpirulo]

Hi,

Looks good. Could you provide us the LED PN that you used. Also include a link to the China units.

Thanks,
Michael

[Mkpirulo]

Hi,

From my experience those guys seem to be ok with IFR 3410 series talk, these units are close enough should be no issue.

Do you have a source for the RF mmics?

Thanks,
Michael

[G0HZU]

Hi,

Looks good. Could you provide us the LED PN that you used. Also include a link to the China units.

Thanks,
Michael

See below for an image of the display used in the IFR3414. This shows the part number on a sticker.

The last time I looked these were available from several sources but it looks like many are now sold as the sources have vanished. I posted this info on another forum a while back so maybe some service/repair people have bought them all.

For the backlight I just used two 3mm ultra bright white LEDs and used resistors to bias them at about 10mA from a 5V supply. It is just possible to cram these diodes in the space vacated by the failed CFL tube in the original display.

Note that the display is a touch screen display but the touch part of it is embedded in the front panel glass with its own touch screen interface. The actual LCD module is just a basic LCD that sits behind the 'touch' glass.

[G0HZU]

The SHF-0189 MMICs are long obsolete but there are various vendors online. I suspect many of the parts available now will be clones made in China. I don't know how well they will function compared to the original.

I did a lot of design work with the SHF-0189 and the 289 and 589 many years ago so I have quite a few genuine parts kept safe here. I'm probably going to need all of them as the SHF-0189 is also used in some of my Agilent sig gens and four SHF-0189 MMICs have already failed over the years in these sig gens. I think I've got less than twenty genuine SHF-0189 parts left now.