I recently picked up a Global Specialties PW-2120 Sweep/Function Generator at a local electronics shop for $30. They had it in back as broken, but didn’t have the time to repair. At the time they only mentioned that the knob was broken (it was, the whole PCB had “popped” off the back of the rotary encoder. But, some prodigious application of hot glue fixed that). The bigger issue is while the buttons and everything worked, there was zero output. Changing to any frequency, sweep, or wave shape resulted in nothing on the outputs. More interestingly, the frequency counter function worked perfectly. So that wasn’t it.
Getting out the FLIR, I was able to see that the CLC522 (variable gain amp), and the 5V regulator (7805) were both getting pretty warm. More interestingly, checking the resistance between the +5v and gnd, there was only 500ohms. That’s not right (or, at least, doesn’t seem right). Visual inspection showed something a bit worrying. U1 on the main board was missing. Now, it COULD be an add-on chip, but that seems unlikely for a chip labeled “U1”.
So, first thing’s first. I contacted GlobalSpecialties to ask for schematics or a technical manual. First response I got back was the general manual, which I had already found. So, emailed them back, and the next day received a nice zip file of all the schematics! Awesome!
First thing I looked for was U1 on the main board. Turns out it’s a MAX038CPP, a “High-Frequency Waveform Generator”, that is Obsolete. Crap. eBay had one for ~$8 that I ordered. Sadly, that doesn’t explain the high 5V regulator output (explained by the 500ohm near-short to ground). So something clearly was partially shorted out. I removed the 5V regulator, and hooked up my bench power supply to that rail. Drawing about 420mA. No wonder it was getting so hot (5V rail is fed off a 9V transformer rail, so dropping that 4V at 420mA means the 5V regulator was dissipating about 1.7W).
So, the MAX038CPP showed up, and on install and power up, things still didn’t work, and the new chip got HOT! I shut it back off, and pulled the chip, and all the others in sockets. Then I checked the resistance between the -5V and +5V rails… 270ohms. crap. So, back to power rails. I should have known. At this point, my thought was to just remove all 4 Voltage regulators (±5V, and ±15V), hook up my as-yet-unsold 4 output system power supply, and power all 4 rails with that to see where the draws are. Also, up the limits on the amperage, and look for heat signatures with my FLIR. There has to be a shorted IC that is on both the +5V and -5V rails (which, is actually a lot more than I thought there were when I started going over the schematic). I know it’s not the front panel, as that’s entirely unhooked. I’m guessing it’s an Op-Amp, or some other IC (or I suppose possibly a couple of the couple dozen tantalum caps on the board… god I hope not).
Next step, then, was to desolder all the IC’s from the board that were looked to both the ±5V rails. Thankfully I didn’t do that, as it would have taken some time. I happened to be looking at some JFET information with regards to the dual JFET in the unit (U441), and something caught my eye at 250-500ohms resistance. Huh. I unsoldered the part, and my ±5V rails showed a much more reasonable resistance both to each other and ground. Crap, so my meter ohms voltage was enough to turn on the JFET. But, the unit tested fine with my component tester, so the JFET seems to be fine. It’s in the Frequency counter section of the unit, which I tested to work correctly.
After that appeared to fix the voltage rails, I ordered a new MAX038CPP from another vendor (also in China, so probably another knockoff), and replaced the socket that the IC went in. I did notice that a few of the “sockets” were distorted, and may have caused bad connections, but I’m fairly confident the first MAX038CPP I got was indeed bad. I confirmed this by wiring up an example circuit given in the MAX038CPP datasheet. The -5V rail was partial short to ground (pulled a 1A -5V rail down to about -2V, so the chip was dissipating 15W).
Two weeks later, the second MAX038CPP showed up, and I dropped it into the example circuit, and fired it up. Perfect 20Mhz sine wave! Awesome. I moved it into the PW-2120, and powered the unit on. It worked! But, the 7805 5V regulator was still getting pretty hot (62C or so after 30 minutes of running). I didn’t overly like the heatsink it had, so I modded an Aavid 530002B02500G (2.5″ tall) heatsink to fit around the main 5V filter cap, and that got the temperature down to about 45C after 60 minutes. Much better1. I also replaced the heatsink on the -15V regulator, which, for some reason runs hotter than the +15V rail. I also purchased a replacement rotary encoder for the unit. Stock is the Clarostat 500EN-209P-1B, but the Clarostat 510E1A48F209PB works just fine. So, total repair cost was about $20. I think at this point I’m going to turn around and sell it, but we’ll see. I except to be buying a modern Rigol Function Generator shortly, so I don’t have a real good use for this. It was mainly the challenge and interest of repairing it. =)
Thanks for the read!
- I also purchased a drop in DC-DC converter that’s a drop-in replacement for a 7805, but I’m unsure if the switching noise will impact the performance of the unit [↩]