I repaired the blown frontend of my Siglent SSA3021X and was asked to provide a summary. So here you are.
Unfortunately, I did not take photos or video during the repair.
Original thread:
https://www.eevblog.com/forum/testgear/ssa3021x-input-blown/. There you find links to other threads and videos which are quite useful.
Hardware: Siglent SSA3021X HW revision 07.03.00 from 2016. There is no repair video for this old hardware.
Defect: Level display low 10 to 15 db, input impedance not 50 Ohms.
Teardown: Lots of screws but very easy. Follow this video:
https://youtu.be/-8fr_otW0q4 Main board starts here
https://youtu.be/-8fr_otW0q4?t=1317A partial schematic of the input section is attached.
Passive checks:
Tested diodes that are connected to V1, V2 inputs of RF switches. Found D7 to be defective.
Active checks:
To get access to the top side of the PCB, all coax connections on the back side have to be removed. Measured voltages at V1, V2 of the RF switches. V1 and V2 must be complement, if V1 = -5..8V then V2 ~0V or vice versa. If both voltages are the same (around 0V), the switch is certainly defective. Voltages at U3 and U4 are reversing when the input attenuator crosses 20dB.
If a signal (5 MHz at 0dBm) is connected to the input, the signal path can be followed with an oscilloscope. There should be nearly no loss from input socket to outputs of U3, with U3 RF1 active with attenuator >= 20 dB and RF2 active when attenuator < 20dB.
Replacement parts:
The RF switches marked 955C are Macom MASWSS0181TR-3000. They are available at Mouser
https://www.mouser.com/ProductDetail/MACOM/MASWSS0181TR-3000?qs=zQS7eqYV0KRx%252Bqryd%2FmxMQ%3D%3D&countryCode=DE¤cyCode=EUR. The chips I got are marked 952R, and from the data sheet the 95 is the type number and 2R is a date code. The diodes were suggested to be BAR63-02W, which are also available at Mouser, but were out of stock, so I ordered BAR63-02V which are slightly larger but usable
https://www.mouser.com/ProductDetail/Infineon-Technologies/BAR6302VH6327XTSA1?qs=K00xGehIljt7wdbutb%2FKnA%3D%3D&countryCode=DE¤cyCode=EURRepair process:
Use your favorite SMT rework process! I read the ceramic PCBs seem to be prone to trace lift off, so take care.
I never worked on such a PCB before, my approach was as follows and worked quite well. No destroyed traces.
- Cover the neighborhood of the parts to be replaced with Kapton tape
- Pre heat the PCB. I used an upside down flat iron which a thermo couple taped to the hot surface and set it to 120°C. Placed the PCB on the iron and waited until the top side reached about 75°C using another thermo couple taped to a ground plane near the part to be removed
- Heat the part to be removed with hot air 390°C until the part comes loose. That should take around a minute
- Remove PCB from iron and let it cool down a bit
- Apply flux to the pads and heat them with a soldering iron just until solder melts. That gives nice even surfaces which makes soldering easier
- Apply a little bit of solder to one corner of the new IC
- Apply flux to the pads
- Hold the IC on the pad with tweezers and tack-solder the tinned corner
- Apply more flux to the pads
- Solder the pins using low temperature solder. I used standard Sn62Pb36Ag2 which was available
- Check the connections under a microscope if possible and/or check with continuity tester
- Clean flux residues
Put the board back into the SA and test. It should work now. Be happy. Have a beer.
The SA probably needs a calibration. But AFAIK nobody knows how to do it. Mine is +/- 1dB up to 2 GHz which is good enough.