| Electronics > Repair |
| R&S SM300 vector sig gen repair |
| (1/2) > >> |
| Microbit:
Wondering if any here have had experience with repairing Rohde&Schwarz SM300 vector sig gen? Mine suddenly output approx. 20 dB lower than normal and after a while output 0... I found it was the classic fault AGB3303 "gain block" that failed, combined with the output multiplex AWOO2R2. Now I get about 7-9 dB low output everywhere. This shows it's not the HCM307 step dividers. The remaining problem is that it is now obvious I somehow managed to put RF into it by accident: The "sniffing" detector circuit still seems faulty. The input post-low pass filter (3 GHz) has a series 47 ohm resistor (fried) into 2 series diodes, measuring short... Now the actual problem. These 2 series sniffer diodes route into a SOT-23 part marked "CCTUB". I think it's a bipolar but unsure. Went through marking codes and cannot find that code - or part thereof. Anyone has any idea about this SOT-23? |
| TurboTom:
Only now I noticed your inquiry: I repaired two of these generators and I'm fairly familiar with the output stage and its typical faults after a "reverse power event" ;) The small diodes are most likely BAT62-02W, the larger -03W variety may also fit the footprint with some tweaking. The SOT-23 with the "CCTUB" marking is an SST4393 JFET. I never found this to be broken. If it actually is in your case, finding a spare may be difficult and you may have to revert to a similar replacement like the MMBFJ113. Moreover, you may want to check the output filter components, sometimes the tiny inductors used there get fried. Unfortunately, I haven't got any specs for those, I once replaced one with a 1nH chip inductor which must apparently have been approximately right, considering the instrument is working within specs again. What I find most surprising in these siginal generators is the use of solid electrolyte tantalum coupling caps in the gain blocks as coupling elements without any microwave ceramics caps as bypasses. It seems to work anyway... |
| Microbit:
Thanks TurboTom! I had kinda given up a bit (unusual for me :-) ) and was going to put it on Ebay soon for $AUD600.... But I haven't sold anything there yet, so I've been avoiding it :-) In the interim I bought an R&S SML01 which I'm very happy with, I find it quicker to use (I currently service more than I design). Nevertheless, I might consider having an attempt given your kind advice. I got a bit tired of the tedious unscrewing and re-screwing :horse: And about the tantalums, yes! I wondered too. I would have thought there would be something in parallel to grease the wheels a bit at 1 GHz and up? First time I opened my RSZ300, it even had a tantalum rolling around, a bypassing one near some TTL logic on the other side. I am not sure which potentially fried inductor(s) you referred to? I thought the output filter was purely a stripline-style 3 GHz corner one? Finally, would you say the shorted series BAT62 would cause what I see, an average of 8 dB low in output? It doesn't quite make 13 dBM, it starts clipping at about 8 dBM IIRC. I would have thought a faulty "sniffer" for ALC would cause a higher than normal output level, rather than lower? Thanks again for reviewing! |
| TurboTom:
I attached a photo of the affected area with the components that I had to replace marked with red dots in the more severely damaged unit that I had been working on. the mentioned inductor is the brownish 608 SMD component where two identical ones are located directly next to each other. The BAT62 diodes didn't measure completely defective but they differed a lot in threshold voltage so I replaced them just because they appeared suspicious. I was quite surprised that the output driver MMIC (AnaDigics AGB3303) was still intact (at least apparently), but I replaced it anyway. As you told already, disassembly and reassembly of the unit is such a PITA that, once in there, it's reasonable to replace everything that may appear suspicious. In order to test the MMICs, I carefully supplied some voltage to the feed-through inductor in the upper left corner of the photo. This way, I could check the bias current sources for the two gain blocks (Hittite HMC313 pre-driver and the aforementioned AGB3303 driver). When powered up, the two LEDs (used as voltage "references" for the current sources built around the three BC807W "5C" transistors) should illuminate. Be careful with the supply voltage since I don't know where else this voltage is being fed. IIRC, I stayed below 10V on that rail. But this permits checking the bias voltages at the MMICs and comparing them to the figures in the datasheets. If you get the low output signal on all power settings consistently, the fault could also be located in the first AW002R2 RF switch and possibly the BAR63-03W PIN diode next to it. But it could also indicate a problem in another module of the "RF deck" which would be much more difficult to diagnose because the other cavities are soldered shut. If the level is correct below -25dBm CW, the problem is likely located in the power amp section. If the block diagram that can be found in the "Service Manual" (...) is correct, all the circuitry with the diodes, the JFET, the OPA137 SOT23-5 opamp and the second AW002R2 should just protect the instrument from reverse power events (not too successful as it seems, unfortunately) and it shouldn't affect its output level accuracy. I hope these thoughts make sense to you and they may help you to repair your SM300! Cheers! :) |
| Microbit:
Thanks! I'll start to prepare to tear it all open one more time and see where it goes with your advice. The reason I got concerned about the CCTUB part: The series 47R resistor was totally fried open circuit, I couldn't read the print anymore. Learned it's 47R from pics of this area. Now, considering that 47R fried o/c and both lil BAT series diodes measure near short, I am wondering where the current went while doing this damage. The only potential DC path I (thought) I see is to the JFET cctub? (IIRC). No idea also what the weird little 4 pin white part is connecting to the diodes, some sort of bypass, filter or integrating thing? I didn't think current found its way to gnd or vcc rail through that? I'll do the surgery once I get the little BATs and maybe I could try that apply of Vcc on that upper left feedthrough.... I wondering though, when I initially looked for a 3303 PA replacement, it seemed most of those parts can't take more than 6 V Vdd? Finally, I'm not sure which 2 LEDs you're referring to in the current source test? You refer to 3 x 5C bipolars but I can only see 2 in pic? Oh, when I get at that -25 dBm crossover point in output, the output level jumps like 4 dB or so instead of ~ 1 dB. The output was low under -25 dB too, by the same amount pretty much. I must say I don't even really understand how ALC then works here, esp. with switching the AHB3303 in & out..... how can the output level be sniffed at like -80 to -127 dBm etc? Normally the ALC is done at full output level and then it's just a matter of step attenuators. Unless the 2 BAT diodes that are fried are not ALC but something eise? But if so, where is the output level being sniffed/,controlled as the 4 HCM attenuators are varying the level 🤔 🙄 😳? Totally confused. I'm completely unfamiliar with the concept level of this sig gen and the output level is controlled.... I will await your comment, and in the interim order the diodes and also the JFET I think. I'll most certainly keep in touch as to where it's going. Thanks so much, 73. |
| Navigation |
| Message Index |
| Next page |