Rohde & Schwarz SML Series Signal Generators - SML01 SML02 SML03 SMV03

[KungFuJosh]

I think this is the first time I've ever used the TinySA for something besides it's signal generator. 🤔

[KungFuJosh]

Well, I answered my own question about fan speeds. See attached.

At 12V, the stock fan is indeed twice as fast.
At 15V Noctua goes up to ~3558, which the stock fan beats at 8V.

So which one cools more efficiently?

According to my thermal camera, the peak temperature was 100F with the stock fan after hours of running. Currently, the Noctua is holding around 97F after a little over an hour. I'll update in a few hours to see how it looks with the thermal images.

The Noctua fan does not sound half as loud. I might use a dB meter app on my phone to compare, but we'll see. It sounds smoother for sure, but the overall volume change isn't amazing.

Thanks,
Josh

ETA: I tried two different apps on my phone, and both say the fans are about the same dB level. The stock fan has a higher pitch noise, that is more annoying than the Noctua fan, but they're definitely close in volume level either way. The Noctua fan has the softer/consistent Noctua sound, but is louder than I expected it to be.

[KungFuJosh]

Final verdict: Not worth the hassle, general waste of time...but still entertaining.

The pitch of the stock fan is slightly higher, but overall seems quieter, despite the piece of crap sound meter app I downloaded thinking they're the same level. I measured both fans in and out of the SML.

Difference in temperature performance is also nothing special in either direction. Max temp with the stock fan was 100.4F, and it was within half a degree over a 5 hour period with the Noctua fan.

I wonder if the voltage of the fan is regulated based on internal temp? That could explain why the Noctua was louder than expected if it was running higher voltage to make up for the temp difference. 🤷

[pdenisowski]

The stock fan has a higher pitch noise, that is more annoying than the Noctua fan ....

If you wait long enough, higher pitch noises won't annoy you nearly as much as they do now 

Thanks for all the great write-ups!

[KungFuJosh]

The stock fan has a higher pitch noise, that is more annoying than the Noctua fan ....

If you wait long enough, higher pitch noises won't annoy you nearly as much as they do now 

Thanks for all the great write-ups!

I got another SML03, and the stock fan is much better; quieter and no high-pitch noise. So it looks like replacing with a new stock model fan is the best option (besides getting older faster). 😉

[KungFuJosh]

Here's a couple photos of another SML03 I got this weekend. This one has the B1 OCXO and B3 Pulse Modulator options.

It also has a defective output unit. I put the output unit from my other SML03 in it, and it worked (then I put the original back in). I'm suspicious there might be something wrong with the B3 module also, but I will figure that out later.

If you're looking for an SML, pay attention to the photos. I've seen some advertised as having options that were clearly not installed.

Thanks,
Josh

[KungFuJosh]

Here are some photos of the allegedly defective output unit. Nothing jumped out at me for obvious issues.

[Martin72]

Is the red circle the output?
Then it should be some kind of relay.
The yellow mark looks like a suspicious solder joint.

[KungFuJosh]

Is the red circle the output?
Then it should be some kind of relay.
The yellow mark looks like a suspicious solder joint.

Good eye. Yes, that's the output, and it does look like a broken solder joint in the photo. I hope it is! I just reviewed the microscope video I made looking over the board, and that's the one spot I missed. 🤦

The reason I say I hope it is, besides being an easy fix, is that the coating they put on isn't smooth and creates misleading shadows.

Time to disassemble again. 😉

Thanks,
Josh

[Martin72]

I don't know if you can access it while it's running, but you could measure at the input of the “relais” to see if the set signal is still arriving correctly.

[KungFuJosh]

I don't know if you can access it while it's running, but you could measure at the input of the “relais” to see if the set signal is still arriving correctly.

It was the ugliest solder joint for sure, but it didn't look broken from other angles. I cleaned it up and resoldered it anyway.

The writing on the part didn't bring anything up for replacement. 5000-0266 C0T0-0815 M? Anybody know what this is?

ETA: My uneducated guess is that this is the only mechanical part on the output unit, and the most likely thing to die without any signs of stress. Despite looking much nicer/newer than the other unit, this one has currently 4410 hours of use.

[Martin72]

That's still brand new; my SML01 had over 100,000 hours on it.
(No  joke)

[KungFuJosh]

That's still brand new; my SML01 had over 100,000 hours on it.
(No  joke)

Good to know! 😉

I've decided to sacrifice parts from the 1st unit for the 2nd one. I can worry about fixing the 1st one or sell it for parts later.

[Martin72]

Either way, you would then have an SML03 with OCXO reference that looks almost new and has very few operating hours, which is pretty cool....

[KungFuJosh]

Exactly. I already swapped it, and it's warming up now so I can calibrate it again (I pulled the battery to force recalibration and wipe the old data).

Before calibration, it outputs 100MHz as ~102.4MHz. After cal, it will show 100.0000MHz on the scope.

[KungFuJosh]

I've got the mix of parts in unit 2 now, and it's working great. However, I needed to remove the SML-B3 option (pulse modulation).

I haven't popped the B3 open yet, but something is wrong with it. After running the self calibration with B3 present, the output is too high. After correction, the output is still nearly 50mV too high. Removing the B3 and changing nothing else (not even recalibrating) puts the output damn close to 0dBm at ~222mV. The signal is also cleaner (mostly seems to affect lower range).

Any ideas what would cause this issue with the SML-B3 module?

Thanks,
Josh

[KungFuJosh]

Here's what the SML-B3 option unit looks like.

[KungFuJosh]

Just for fun, here are some photos of the PSU.

[KeBeNe]

Hi, sorry I haven't spoken for a long time.

As far as the pressure static of fans is concerned, there are usually diagrams that show the capacity of the fans and are not linear.
I have noticed that the Noctua fans react badly to high restriction, they get loud and the pressure drops sharply.

Regarding your SML03 and B3, the installation and conversion to B3 requires a factory calibration, as the module is installed in the RF signal path. (Service KIT software + NRVD + Z51)
This is why your levels are not correct.

Better repair the 3Ghz module, I think what you have marked is the RF coupler for the ALC.

The used switches 4227T have errors, replacement is very expensive, an exchange to HMC types is conceivable, work on it again entails a service kit calibration.

Unfortunately, even R&S no longer seem to have this software, at least here in Germany...


Servicemanual: https://manuals.repeater-builder.com/te-files/R&S/R&S%20SML_SHB_BD2_03%20Service%20part%202.pdf

[KungFuJosh]

Hi, sorry I haven't spoken for a long time.

As far as the pressure static of fans is concerned, there are usually diagrams that show the capacity of the fans and are not linear.
I have noticed that the Noctua fans react badly to high restriction, they get loud and the pressure drops sharply.

You were right anyway. I posted my results above- the Noctua lost. My second SML03's stock fan is much quieter, and not an issue.

Regarding your SML03 and B3, the installation and conversion to B3 requires a factory calibration, as the module is installed in the RF signal path. (Service KIT software + NRVD + Z51)
This is why your levels are not correct.

That's odd. The B3 came with the SML03 as a preinstalled option along with B1. I swapped mainboards, and I think the B3 was tamer with the one that was not calibrated for it.

Do you know if there's something in the mainboard that needs to be unlocked for the B3 to be used? I haven't tried the original mainboard yet, but with the replacement board, it said something about it not being available despite it being connected. I was hoping it would sense it like it does with the B1.

Better repair the 3Ghz module, I think what you have marked is the RF coupler for the ALC.

I don't know which part or photo you're referring to? Do you mean the part I have marked as Main Output Unit (which I think is actually an attenuator)?

The used switches 4227T have errors, replacement is very expensive, an exchange to HMC types is conceivable, work on it again entails a service kit calibration.

Unfortunately, even R&S no longer seem to have this software, at least here in Germany...

4227T sop8 appears to be available at aliexpress, but blocked to the US right now. It's okay; I swapped the attenuator out from the other unit, and it's fine. I'm going to sell the other unit (for parts or repair), and somebody else can fix that if they want it to work.

Your input and help is much appreciated!

Thanks,
Josh

[KeBeNe]

Hi,

If I understand you correctly, you have exchanged the mainboards, as the factory calibration is stored on the mainboard, in the flash, in the protected area, a factory calibration is required.
What I mean by this is that the mainboard with all the assemblies connected in RF path is a calibrated chain, if you dissolve it, it is good luck that it has the correct levels afterwards.
How are the levels measured?
Do you have a spectrum analyzer which is still accurate?
Measuring the values with an oscilloscope is one of those things, as the input of the oscilloscope can also have a mismatch.
The measurements are not optimal and you need more accurate measuring devices than an oscilloscope to test these measuring devices.

Regarding the options, I have never had an SML, but it may be a similar procedure to the SMT/SME/SMIQ, i.e. the modulation type is activated with a key.
You can look in the menu to see if they have the item "installation", this is for installing options, these can be entered by deactivating Protect Level 2, at least with the big brothers.

It would be interesting for me to know whether an SML02 can be turned into an SML03 using a factory key.
With the SMIQ02B and SMIQ04B you "only" need a factory key for the 3Ghz or 6Ghz version to turn an 02B into an 03B and an 04B into an 06B.

I have edited your picture to explain the modules, the OPU1 is on the mainboard, the OPU2 is the doppler/trippler module.

[KungFuJosh]

Hi,

If I understand you correctly, you have exchanged the mainboards, as the factory calibration is stored on the mainboard, in the flash, in the protected area, a factory calibration is required.
What I mean by this is that the mainboard with all the assemblies connected in RF path is a calibrated chain, if you dissolve it, it is good luck that it has the correct levels afterwards.

The SML series does a self calibration to determine levels. It generally works quite well, but not when the B3 module is installed. The mainboard sets the correct levels, but the B3 boosts what is actually outputted.

How are the levels measured?
Do you have a spectrum analyzer which is still accurate?
Measuring the values with an oscilloscope is one of those things, as the input of the oscilloscope can also have a mismatch.
The measurements are not optimal and you need more accurate measuring devices than an oscilloscope to test these measuring devices.

Measured on:
SDS2504X HD oscilloscope
TinySA Ultra
Keysight V3500A RF Power Meter

All devices are agreeable (TinySA required minor level adjustment at 1st frequency, then matches levels across the range within reason given its limitations).

Regarding the options, I have never had an SML, but it may be a similar procedure to the SMT/SME/SMIQ, i.e. the modulation type is activated with a key.
You can look in the menu to see if they have the item "installation", this is for installing options, these can be entered by deactivating Protect Level 2, at least with the big brothers.

There's nothing in the service manual about installing anything except hardware. Hardware is detected when its installed, and the options for it appear on screen. I'm going to swap mainboards back again to experiment with the B3 module.

For example- either mainboard ignores the OCXO if not installed. Either motherboard also gives an oven cold warning on startup when the OCXO is installed.

Either mainboard doesn't have any pulse options on screen when the B3 module isn't installed. The Pulse menu appears when B3 hardware is installed.

It would be interesting for me to know whether an SML02 can be turned into an SML03 using a factory key.
With the SMIQ02B and SMIQ04B you "only" need a factory key for the 3Ghz or 6Ghz version to turn an 02B into an 03B and an 04B into an 06B.

I have no idea on this one. The module says 2/3GHz. It would be funny if there's something like a jumper hiding in there to limit the module.

I have edited your picture to explain the modules, the OPU1 is on the mainboard, the OPU2 is the doppler/trippler module.

Thanks. I labeled "main output unit" that way because that's how it's labeled in the service manual. The attenuator is just a part of it. Though you're right, it probably is the hard to find switches that need to be replaced in the defective unit.

Thanks,
Josh

[KungFuJosh]

Here's some examples. The only changes are the hardware attached.

1st Unit (didn't come with B1 or B3 options) with B1 OCXO connected:



B3 hardware not installed:



B3 hardware installed:



Thanks,
Josh

[KungFuJosh]

I just put unit 2 back together with unit 2 parts except for the main output module, and 2/3GHz module from unit 1. B3 is installed. It's not warmed up yet, but output is now showing slightly under 0dB rather than being high.

All I did was clean it a little more, and wiggle connections. I bet that wiggling helped. 🤔

I'll let it warm up, and test the levels again. If it's still low, I'll run the self calibration and see how it looks. Hopefully, running calibration doesn't mess it up.

Update:
Warmed up, and ran 'Calibrate all' again.
-0.49dB at 10MHz through 500MHz, -0.75dB at 1GHz, -1.5dB at 3GHz using the V3500A power meter.

That's dead on with the scope showing 211mVrms at 10MHz:




That must have been some special wire wiggling. 🤷

Thanks,
Josh

[KeBeNe]

Please look under utilities, there you will find the item Install for SME and SMIQ.

Yes, if you install the B1, you do not need a key, this is the case for all devices, but it is for PM, even if the HW is installed, for the other RF generators this is the option B11(HW). this also includes software keys.

Calibration, the self-calibration has nothing to do with the factory calibration.
The factory calibration is carried out over the entire RF chain and sets a reference point; this reference point is used as the target for the user calibration.