My E4421B has thrown me a nasty curve ball today. It starts but when it gets to the splash screen I hear a loud click from one of the RF modules and it seems to reset into the splash screen. This cycle keeps continuing
Just as I wanted to make some measurements on my new HF differential probes.
Just what I need. Maybe it is karma. A few days ago I have bought a different RF generator which hasn't arrived yet. Hopefully that is in a somewhat workable condition.
Looks 5V power rail has an issue, fail at load, making reboot cycle. That is first thing i would check.
Then if 5V ok, remove all boards except digital and check it
The 5V looks ok with a DMM. I also see the diagnostic LEDs keep counting in a binary sequence. With all the RF modules removed the behaviour is the same. Maybe I need to try and reflow the SRAMs. There is a message about the RAM but it isn't clear if something is bad or not. It says:
Start = 0x300000
End = 0x3F9800
Errors =0x0
With the 'local' button pressed during power on, the E4421B won't start the firmware so the self test result stays on screen.
I made some progress. When it is cold and no RF modules are installed it will start succesfully. I tried using freeze spray but so far no joy.
Thermal camera?
Measure cold and warm current consumption on 5V rail
I'm not so sure this is 5V related. I'd expect the system to do a full reset (go through the bootloader) when the power fails. Instead it seems the application crashes and restarts itself. Yes, I already re-seated all memory chips in sockets on the main board.
The chassis allows for a whole bunch of extra boards (which aren't in my machine) so my assumption is that the power supply has way more power than necessary. I won't rule the PSU out but using freeze spray on it hasn't solved the issue. It also doesn't smell like cat-wee which is a tell-tale sign of a bad electrolytic.
There is a LED array indicating the status of the different power rails. If all of them are ON, then the power supply is OK. During boot, there is another LED array that indicates the boot step and once all tests pass, it counts up-down. If it stops with a specific patter, then you can read the hex value and look at the service guide.
All the power leds are on and the status LEDs don't stop at a specific pattern. The bootloader and self tests are OK. It is the application itself which crashes. The status LEDs invert after the bootloader is finished and the left 4 (with the display towards you) are showing a binary counting pattern. When it starts OK then all the status LEDs are showing a counting pattern.
I found a rather large ripple on the 12V supply (200mVpp). The capacitors in the PSU still measured OK though and replacing them didn't help. I'm going to sell the generator 'for parts'. With all the repairs I did on it and the hacked output amplifier it is not going to be worth much even if it works. I have a 'new' RF generator coming my way anyway.
This is not outcome we all expect
This thread is old but I think it is worth keeping any repair hints together.
Recently I acquired this signal generator and had missing columns on the bottom half of the LCD. After replacing the LCD, the problem persisted. While trying to map the LCD pins to the correct part on the motherboard, I used a continuity meter to map the pins. It shorted one or both HM628512BLFP-7 SRAM chips (I think they are called BootROM RAM) and the motherboard did not want to boot, resetting every 3 seconds due to the power supply being shorted. I then realized that the motherboard has a 3V lithium battery to keep some configuration alive and I passed the test probes over pins while they were powered by the Battery. When I removed the SRAM chips, the board was back alive, but stopped with LED diagnostic code 2, and the manual says it is checking for BootRAM RAM, which of course were removed. After soldering new SRAM chips, the motherboard booted correctly passing all the self diagnostics. The LCD problem is still there, and I traced the problem to the Video RAM (HM514260CJ6) next to the Chips F65510 VGA controller (The controller seems to be OK because I can measure signals changing on all 8 bits that goes to the LCD). I am waiting for the Video RAM chip to arrive and I will post the results
I bought the ESG mainboard from TK on ebay, to repair it. There were some more problems, most of them were easy to fix. One switching voltage regulator was shortet, and two transistors (display and green power LED) were shot.
But the lines in the display were still there.
After analyzing the protocol of the data transfer to the display, I was short before taking my logic analyzer. I even replaced the RAM chip a second time. But then I did some testing with removing signals to the display, and pulling the 8 lines to H one at a time. No result on bit 8.
Problem was a bent contact in the connector of the board to the LCD display! Looking back, it was obvious, because there came no signal for these lines out of the board. With black screen, alll signals were constant L. Pulling one of the 8 bits H there were white lines. But not on bit 8. It just had no contact to the display. But I was also sure that the problem came from the display controller or a bad display RAM chip. Key to the solution was to understand the display transfer protocol, which explained whyonly the lower half of the display was faulty.
This thread is old but I think it is worth keeping any repair hints together.
Recently I acquired this signal generator and had missing columns on the bottom half of the LCD. After replacing the LCD, the problem persisted. While trying to map the LCD pins to the correct part on the motherboard, I used a continuity meter to map the pins. It shorted one or both HM628512BLFP-7 SRAM chips (I think they are called BootROM RAM) and the motherboard did not want to boot, resetting every 3 seconds due to the power supply being shorted. I then realized that the motherboard has a 3V lithium battery to keep some configuration alive and I passed the test probes over pins while they were powered by the Battery. When I removed the SRAM chips, the board was back alive, but stopped with LED diagnostic code 2, and the manual says it is checking for BootRAM RAM, which of course were removed. After soldering new SRAM chips, the motherboard booted correctly passing all the self diagnostics. The LCD problem is still there, and I traced the problem to the Video RAM (HM514260CJ6) next to the Chips F65510 VGA controller (The controller seems to be OK because I can measure signals changing on all 8 bits that goes to the LCD). I am waiting for the Video RAM chip to arrive and I will post the results
I bought the ESG mainboard from TK on ebay, to repair it. There were some more problems, most of them were easy to fix. One switching voltage regulator was shortet, and two transistors (display and green power LED) were shot.
But the lines in the display were still there.
After analyzing the protocol of the data transfer to the display, I was short before taking my logic analyzer. I even replaced the RAM chip a second time. But then I did some testing with removing signals to the display, and pulling the 8 lines to H one at a time. No result on bit 8.
Problem was a bent contact in the connector of the board to the LCD display! Looking back, it was obvious, because there came no signal for these lines out of the board. With black screen, alll signals were constant L. Pulling one of the 8 bits H there were white lines. But not on bit 8. It just had no contact to the display. But I was also sure that the problem came from the display controller or a bad display RAM chip. Key to the solution was to understand the display transfer protocol, which explained whyonly the lower half of the display was faulty.
I am glad you found the problem with the board and the display is now working. Thanks for the follow up.
New (second hand) E4421B arrived and was told it "works good" by reputable dealer. Not so, the reference is faulty, being too high in frequency by about 200-300 hz as well as the external Ref says its unlocked! Damm.
Went whole hog in testing in stages and finally tracked down to the 10Mhz TCXO diagnosed as faulty. Dismantled and added a 1.5pf across xtal to tune. Reassembled and tested A1 and external ref now works 100.
Free run testing and the 10Mhz does vary about 3-5hz over cold to hot conditions. At least external Ref is fully operational at all times.
Problem solvered!
Unit destined for a fishing sinker is now pride of place in workshop.
This is the link to Mike's ( Mikeselectricsstuff ) teardown video (as requested per PM):
Nice find! I guess the output amplifier is still working? This was the biggest headache for the unit I used to own.
Hey guys!
I'm having problems with my 4421B, it simply doesn't activate the modulation, I hear the sound from the relay, but it doesn't work. It's like trying to activate modulation, but it doesn't, I've already done some tests:
1) I used the external modulation input and it worked well;
2) The OCXO is in order, 10mhz perfect and signal without imperfections.
One detail, about a year ago, my time base stopped working and I replaced it with the same one (ebay) and everything went back to normal, I don't have the service manual, if someone could make it available it would be of great help.
Note: The signal in blue and the attempt to modulate in FM (dev=100khz)
Hug!