EB5AGV's Workbench: Agilent E7401A repair

An interesting collaborative repair

Agilent E7401A

Lately, my main work has me so busy that I have little time to work on my own equipment, and even less time to write about it. But I enjoy doing both things so I am back with another repair story. This time it is an Agilent E7401A EMC Analyzer, which is a kind of 9kHz to 1.5GHz spectrum analyzer with some special functions for EMC (Electromagnetic Compatibility). I bought it in non working condition (nothing new...), with just the information given on the eBay listing, which was not too extensive. But, well, I wouldn't be myself without accepting the risk of a new repair ;-)!


Unit as received

The information I got from eBay was that the unit did not power up and showed just a white display. The unit had the always interesting tracking option (1DN) and also with the HPIB option (A4H). Here you have a couple pictures from the eBay listing:

Agilent E7401A

Agilent E7401A

And, yes, once received, the unit did exactly that:

Agilent E7401A


First attempt to fix: RAM

Well, all modern test equipment has some kind of firmware on it so, the first thing to try on a non-starting unit is to reload the firmware. I was almost sure that had been already tried by the seller (who was a professional one) but, anyway, I should test it to be sure it was not such a simple fix...

So I managed to re-create the floppy disk set (using and old P-IV laptop, which still has a floppy and real serial and parallel ports, running Windows XP) and tried to install the new firmware. It went through all the process perfectly, which ended with a message telling that the installation firmware was verified and that it was going to write it to flash memory. Once that ended, I was instructed to cycle power... and that is: the unit did the same than when I have first powered it, just showed an all-white display and the floppy kept doing some periodic noises.

Agilent E7401A

Agilent E7401A

Agilent E7401A

So I opened the E7401A and found that, according to the Service Manual, the status LEDs indicated a RAM failure. Once I removed the module, cleaned its contacts, and restarted with same result, I tried to locate on my stock a replacement 32MB 8x32 72pin EDO RAM, but I had none in stock, so I looked for one.

Here you have an inner view and the memory modules:

Agilent E7401A

Agilent E7401A

This is the CPU BOARD with the RAM and FLASH memory modules inserted:

Agilent E7401A

Some days later I got a couple RAM modules but they didn't do the trick. The unit stopped, showing this LED-coded error:

Agilent E7401A

Doing some further checks I was afraid the CPU BOARD was defective... and that was a big problem!


Second attempt to fix: CPU

As I found out that the CPU board was the culprit of the malfunctioning, I tried to find a used one. It didn't show up, but then I found one that was visually identical, and was defined as an E4401-60206-01-0101-00376. So I went for it.

This is the CPU board I bought:

Agilent E7401A

And these are the two CPU boards side-by-side:

Agilent E7401A

The replacement CPU board worked better that the one I had, which was not able to even boot on analyzer mode. This one did, but was shown as a E4411B:

Agilent E7401A

So the system booted, but then asked for calibration (System Alignments, Align Now, All required):

Agilent E7401A

Which never ended. It kept showing an Aligning ADC message)

Agilent E7401A

I tried to reinstall firmware and, in fact, I updated it to a more recent version. But it still thought it was an E4411B. I thought that the difference with an E4401B or E7401A board should be a matter of programmable devices on the board as, otherwise, it was exactly the same than the E7401 board and so it should be similar to the E4401B one. But then, was there a way to modify / reprogram the E4411B CPU board so it acts as an E4401B?. There was a suspicious PALS labelled port but, of course, I had not the binaries nor the tool to program them. I was stuck!


Third attempt to fix: CPU remake

Some months passed and I continued looking for possibilities... Looking in detail to the two CPU boards, I found that there were two large programmable devices which had a different Agilent label:

Agilent E7401A

If I was able to swap them, perhaps the E4411B board could be converted to a E7401A board... But I had not the needed tools to do that. Then I recalled that a repair colleague, Jose García, had a nice lab with the needed tools, so I talked with him and sent the two boards. He did a wonderful job, even saving the discarded ICs, just in case. Here you have the E7401A ICs soldered on the E4411B board:

Agilent E7401A

Agilent E7401A

So I installed the remade CPU and, after some infinitely long seconds, got this:

Agilent E7401A

It kept thinking it was an E4411B... and, guess what it did after that?

Agilent E7401A

And, looking into more detail, there was no doubt:

Agilent E7401A

So it seemed as I had done nothing...

But I am hard to defeat!!!


Fourth attempt to fix: CPU remake II

I kept thinkig there should be something which kept telling the unit it was an E4411B instead of the E7401A I wanted it to be... So I took the Service Guide and studied one by one the ICs it had. And then I found a tiny programmable IC on the other side of the PCB!

Agilent E7401A

So, on the E4411B board, I replaced it by the E7401A one:

Agilent E7401A

Agilent E7401A

And then...

Agilent E7401A

YES!!!. At long last!. Finally, the unit acted as a E7401A!

Agilent E7401A

A nice surprise was to find these interesting options:

Agilent E7401A

But... would the unit work now?

No, it didn't yet!


Fifth attempt to fix: the hidden problem

I had done some advances but the E7401A was not working yet... One interesting thing is that the E7401A worked in colour instead of the monochrome E4411B. This is a trace it has stored in memory from a previous owner:

Agilent E7401A

I did lots of things: reinstalling the latest firmware (A.14.06), installing an older one (A.11.00), removing optional PCBs... But anything helped. There was a hidden problem!. Of course, there was a possibility which I didn't want even think about: the replaced ICs could have been damaged on the process. It is not the same to replace an IC with a new one than to reuse an IC. But I was very confident on my colleague work. So there should be another cause for the remaining problem.

I must admit that I spent a full day doing lots of things and, at 1:00AM, I was about to let it go. But then I did a silly thing: I removed the power supply module:

Agilent E7401A

Inside it, easily accessible by the lower aperture, I found a plug-in PCB, which I took out:

Agilent E7401A

I cleaned the contacts of that PCB as I do on memory boards, and reinstalled it. Then I put back the power supply and... I was able to calibrate the unit!

Agilent E7401A

After the recalibration, the error message was gone and a trace was shown!!!. I was elated. And pretty happy!

Now, each time the SA is powered up, it performs some calibration:

Agilent E7401A

Here you have some samples of the E7401A working as a conventional SA:

Agilent E7401A

Agilent E7401A

Agilent E7401A

Agilent E7401A

Success at last!!!


Go back to WORKBENCH INDEX