TSP #90 - Teardown & Analysis of an Anritsu MN9610B Variable Optical Attenuator

In this short episode Shahriar takes a detailed look at an Anritsu variable optical attenuator operating in the L- and C-Bands. The mechanical attenuator is defective and provides an opportunity for complete disassembly and examination of the free-space optics, mechanical components and electronics.

The electronics drive a pair of DC motors with potentiometer decoders in feedback along with operational amplifiers and a DAC. This allows the microprocessor to set the exact position of the motors based on calibrated values stored on EPROM. The free-space optics consists of two back-to-back disks with radial metal coatings. Depending on the position of each wheel, the coating limits the transmission of the light through the wheel causing attenuation. The entire housing is hermetically sealed to prevent degradation of the disks.

The Signal Path

[campbellmorrison]

I just love this optical stuff. I recently pulled a miniature variable attenuator apart and was amazed at its construction. Similar design but the beam was bent through prisms so the in and out were next to each other. The graded glass was moved through the beam by a miniature stepper motor and the glass position was achieved with a linear carbon track. The thread on the stepper motor was so small you couldn't see it with the naked eye, incredible must have cost a fortune

[soothcoder]

Aww! I wanted to see the broken attentuator pot's performance. Never mind - a very interesting video as always.

[InssiAjaton]

Highly interesting!!!
Up to a few weeks ago I did not have any optical instruments apart from some magnifying lenses. Oh, a slight correction — I have had a pocket type laser distance meter. But after seeing a youtube story about the wave length measuring instrument, I bought something similar myself. Looks fine, but is waiting for me to put in some more money in to get the necessary support software. Which brings me to my main point: Like (I bet) everybody seeing your collection of top notch instruments, I can only be envious. My lifetime collection of (now) antique instruments does not leave me any room for more modern ones, even if I could afford them.
Then a little comment about your postings. I watch everything that is less than one hour long. If the topic relates even remotely to my instruments, then I may watch even longer ones. But generally I have a time limit, like just one hour before I fall asleep. I also notice that most of the long presentations deal with something beyond my frequency range (i.e. more than my 1.5 GHz or 3 GHz spectrum analyzers or 300 MHz — unleveled — sweep generator).

[TKomoski]

I like any video you do, always informative and never boring. Cheers

[stivep]

15:14 Pretty known process in many cases similar method is applied to sintered molds (Sintering is the process of compacting and forming a solid mass of material by heat or pressure) However it is not always practical due to porousness of end product surface.
From mechanical standpoint there is no noble gas nor under-pressure.
The inner surface is first "dirty way" prepared, than cleaning solution with particle sealant is used to seal any porousness.
That also acts as anti reflective coating. Almost unnoticeable to an eye, but EDXRF does the good job here.
Mechanical surface on the top was prepared on CNC mill prior to that.
The device is placed in a clean room vacuum chamber with equal vacuum all around . Robotic screw-gun in placing the screw into the cover that is mechanically positioned on the top . The screw than is sealed and the assembly is moved out of chamber without the need to wait it to dry-out. I was told that this screw it is only important in manufacturing process, so the device can be placed back into the vacuum chamber and presented in full assembly for re-calibration.
Screw is removed to allow equalization of pressure all around, than the new screw is inserted.
Anrtitsu have made some changes later on into manufacturing process so actual assembly might look little different in later models .
Wesley

[daveblane]

Excellent as always. and that 7510 DMM! wow

[gnarlykoala]

Its cool for me to see Keithley equipment, i work at a calibration workplace at a industrial, instrumentation and electrical oriented company. We use a Keithley 2002 multimeter, to check the values we need in order to check/ calibrate the field instruments from our customers.

[joes]

Love that Keithley stuff! This video makes me want to upgrade my Keithley 2000.

[MrSparker]

Great! I have never thought that such a device exists.

[davidjereb]

The new lab is looking spectacular! I look forward to watching more of your videos.

[legionaire]

Fascinating video, I am not familiar with any optical equipment and really enjoyed watching this, keep up the great work!

[eek]

I may have missed you saying but the cylinders off the motors look like planetary gearboxes. maybe the oscillation was caused by mechanical wear. excellent mini teardown

[DAVIDGREGORYKERR]

That chip that you where not sure about is most likely a Texas Instruments DSP processor so there must be some memory chips associated with the chip as well as a EPROM that contains the DSP processing code.

[DeathTickle]

Thanks for the video Shariar ! That is a very nice instrument you opened up. I especially liked your comments regarding the calibration in relation to the feedback of the encoders. I was then wondering how is an instrument like this calibrated ? Would it have both the motors swept through a range while recording the attenuation ?

I also don't fully understand how the optical source gets reflected on the mirrored surfaces and end up to the output. In the video Shariar says the surface inside the attenuator is made to scatter light, but then doesn't the light hit all areas of the mirror instead of the desired one ? I'm not very familiar with optical instruments like this one.

[sefarkas]

Typical problem in Japanese DC motor servos. The pots on the bottom board most likely would have set gain and offset of the servos. A quick twist of these pots would have probably fixed it. But now you've let in the magic smoke to the attenuators. It's a nice box now for something new.

[Konecny_M]

I wonder if the internal surface of the optics aluminium box is just artifact of it being sand casted part, or if it was intentional part of the design. Given that they machined even outside surfaces they did not have to, I'd guess the second one.

[ralimtek]

Quick and dirty way to get rough closed loop control could be to take the inside controller from a cheap hobby servo, as its a ready made closed loop controller + h-bridge... :) (And damn easy to drive, abeit not highly accurate)

[DAVIDGREGORYKERR]

<#StarGazer>:DSP chips can do maths in just a few clock cycles like calculating angles or other things.

[rapsod]

Very interesting. Did they used metal vapor deposition for mirrors?

[rogeronslow]

I wonder if one disc is a coarse tune and the other a fine tune?