[MENDED] E4418B Power meter doesn't turn on

[nfmax]

There is a long-standing issue with these power meters that eventually they fail to respond to the front panel power on button. See, for example https://www.eevblog.com/forum/repair/breadcrumb-for-agilent-e4418b-repair-power-button-not-working/msg3468522/#msg3468522. It is associated with the front panel keypad FPCB, (https://www.eevblog.com/forum/repair/hp-agilent-e4418b-power-meter-keypad-flex-circuit-e4418-20007-needed/) and also, curiously with the memory backup battery (https://groups.io/g/HP-Agilent-Keysight-equipment/topic/31675429?p=Created%2C%2C%2C20%2C1%2C0%2C0)

After a year or two of normal operation, it struck my instrument too 

Guided by the various posts above, I dived into it. I found the CLIP, including circuit diagrams, posted online. I have excerpted some relevant parts in the attachment. I found that the power on/off state is maintained by flipflop U34-A, powered by VBACKUP. This is maintained by the backup battery, so the intention is that the instrument will turn 'on' when the mains is reconnected if it was 'on' when the mains was removed. In fact, the whole power switch circuit seems to have been designed to mimic exactly a mechanical switch. You can turn the instrument on & off with the mains disconnected (but why?) The power button momentarily connects VBACKUP, through a 51.1kΩ resistor R51, to the input of Schmitt trigger U33-A, which toggles the flipflop. U33 is also powered from VBACKUP.

With no mains applied, VBACKUP measured 2.86V, rather low. Interestingly, the voltage at ONSWITCH_1 (with the power button not pressed) was only 0.77V, too low to reach the threshold of U33-A, explaining why it wouldn't turn on. In fact, when the power button was pressed, it dropped down further to 0.72V, owing to the additional load from R52. This means that a current of about 41μA is permanently flowing from the battery through R51! Disconnecting the keypad* reduced this to zero, confirming a problem with the front panel. All voltages were measured with respect to the rear panel chassis ground connector, which takes a 4mm plug.

Connecting the mains supply increased VBACKUP to 4.64V, and ONSWITCH_1 to 1.19V, still not enough to turn it on, but the battery voltage, measured at its +ve terminal, rose to 3.02V. Momentarily jumpering a 10kΩ resistor between VBACKUP and ONSWITCH_2 caused the instrument to power up normally (with the battery self test passing!). In this state, VBACKUP measured 5.00V and ONSWITCH_1 measured a few mV higher. Now, the front panel button would turn the instrument off normally.

My diagnosis is that there is a leakage resistance, in the front panel FPCB, between ONSWITCH_1 and some circuit which is ~0V in the off state, but approximately +5V (or maybe more?) in the on state. Probably something to do with the LED indicators. With the instrument off, this gradually runs down the backup battery: there have been reports that replacing the battery can 'fix' the problem. In my case, this obviously isn't enough, and anyway it won't stop the discharge current.

What I did was to abandon the idea of being able to change the on/off state with no mains connected, in order to eliminate the battery drain. I removed R51, and connected a 4.7kΩ resistor between ONSWITCH_1 and +5VPSU (see attached image), which is the 5V supply from before the FET power switch controlled by U34-A: it is present whenever the mains is connected. The voltage at ONSWITCH_1 with mains connected was now about 3.46V, and the power on & off operation was restored 

This fix should also have removed the excess battery drain. The battery current can be measured by measuring the voltage across 1kΩ resistor R168, and in my case this was a satisfactory 110nA, low enough not to bother replacing the battery, which now measured 3.07V. The voltage drop across this resistor explains most of the voltage drop in VBACKUP observed above, when the mains was removed, the remainder arising presumably from the battery internal resistance.

This may not be the universal solution to the problem, but it should give some clues to anyone else in the same situation: much cheaper than replacing the front panel FPCB anyway!

Good luck!

(*) J14 is a latching connector: lever up the two ends of the top plastic part (with difficulty) to release the FPCB

[eb3frn]

Hi
same problem here, battery down and not turn on anymore also replacing the battery for a new one.l

After drop R51 and add the 4k7 resistor, it has worked again!

Thanks a lot!

73! Iban
eb3frn

[dxl]

Interesting. I had exactly the same problem a year ago. I lowered the resistance a bit, which made the voltage go above the threshold. This will decrease battery life in my unit, as i didn't made the modification like you did.

My unit still has the problem that it doesn't reset on power after it was switched of (it hangs in the bootloader). Luckily it resets if i do a quick power-off-power-on sequence (less than 2s). Maybe i'll try your modification and see whether it goes away.

The Reset issue is not caused by the Supervisor circuit - i replaced the MAX710, but that didn't help.

Thanks for your hint, i'll try it.

[nfmax]

Hi
same problem here, battery down and not turn on anymore also replacing the battery for a new one.l

After drop R51 and add the 4k7 resistor, it has worked again!

Thanks a lot!

73! Iban
eb3frn

You’re welcome!

[rplabs]

Hi,

I implemented your fix by dropping the R51 and adding the 4k7 resistor with the idea to reduce the battery drain but it seems it is generating another issue.
I also replaced the battery and it shows 3.31V and also replaced the flexible keyboard PCB.

I applied the GPIB fix to revert it back to E4418B as in the application note and I tested to make sure it keeps the settings. All OK.

I didn't use the power meter for one week and I unplugged it from mains. After one week I powered it on and it has the Battery error. I opened the power meter and measured the voltage of the battery and it has 3.31V. No drain in one week...

I did more tests and it looks like it keeps the correct settings and it doesn't gives any error only if unplugged for a short period of time. If unplugged from mains more than 10 minutes it loses the settings and it shows up as E4418A and also it gives the battery error.

I noticed there are multiple motherboard versions, could it be that your fix is only for one particular revision? There is also a service note for these EPM power meters where Agilent said to remove one resistor because of excessive battery drain but from what I can tell it is not for my board revision. I don't have that resistor in the place where they indicate it.

I might reverse your fix since I need to run the GPIB commands at every power on to change it to E4418B.

EDIT: I found the issue why it was not working. I checked your troubleshooting guide and did more measurements on the board and I saw that I didn't have battery voltage in some places where I should have. Guess what? The jumper near the battery was removed by previous owner. I bought this power meter as faulty and it was clearly visible that someone was there before me. I installed the jumper and now everything seems fine.

Another note regarding the 2 type of motherboards that I saw in existence, the older one (I think it is older) that doesn't have the IC near the riser connector must have the jumper near the battery fitted. The newer one that has the IC near the riser connector should not have it installed. From what I saw the new motherboard comes with a jumper but it is only installed on one pin not on both.

I hope it helps others.

Best regards!

[perdrix]

I'm replying to this thread even though it is old because I had apparently almost identical symptoms, but ...

With no mains applied, VBACKUP measured 2.511V and the voltage at ONSWITCH_1 (with the power button not pressed) was 1.298V (measured with a meter with a nominal input impedance of 10M\$\Omega\$).

The OP thought that this was caused by a permanent leakage current through R51 caused by a problem with the front panel assembly.

This didn't feel correct to me, and a study of the front panel schematic didn't reveal any plausible paths for that sort of leakage.

So I measured the resistance of between the right hand end of C62 and the via to the left of R51 (effectively across R51, but including traces and solder connections).  I got a reading of  5.2M\$\Omega\$ which varied quite a bit.

Re-touching the solder connections to R51 brought temporary relief, and all seemed well, but not for long: Only a few minutes later the power button was unresponsive again.   I think this is a bad resistor so I've ordered a replacement.

David

[Seiji53]

My E4418B also turns on when I plug in the power cord while holding down the power switch. I removed R51 and added a 4.7 kΩ resistor. However, while I could turn the power off at any time, I was only occasionally able to turn it on. I replaced the 4.7 kΩ resistor with a 2.2 kΩ resistor and added a 3.3 V zener diode between one end of it (the power switch side) and GND. I confirmed that the power switch operates normally for both On and Off. The zener diode is there to prevent latch-up in the SN74HC14 used for the power switch input.
With R51 installed, a multimeter cannot take an accurate reading because current is flowing from the battery.

[hp3310a]

[...] I replaced the 4.7 kΩ resistor with a 2.2 kΩ resistor and added a 3.3 V zener diode between one end of it (the power switch side) and GND.
[...]

Are you sure this isn't shorted at the zener diode? I can see no gap at all.

[Seiji53]

There is no short circuit. This is because the voltage, which was 5 V before the zener diode was installed, is now approximately 3.1 V (measured value). If this line were shorted to GND, the power supply would not turn on or off. The actual zener diode used is a Toshiba 02CZ3.6 (SMD). The voltage is on the low side because the current flowing through the zener diode is low. (Rated current: 5 mA)
I’ve cleaned off the flux and other residues from the attached photo to make it easier to see.

[Seiji53]

The reason the power won’t turn on is that “silver migration” causes a current leak from the power switch circuit, resulting in a voltage drop that prevents the switch from operating.
Applying current to the switch will temporarily resolve the issue. However, applying voltage accelerates the “silver migration,” so there is a high likelihood that the power will fail to turn on again afterward. If the power does not turn on, you can temporarily resolve this issue by holding down the power switch while plugging in the AC power cord. To fully repair the unit, the front switch plate must be replaced.

[hp3310a]

There is no short circuit.

Sorry, I didn't see  GND on the other side.

[perdrix]

I think this is a bad resistor so I've ordered a replacement.

The replacement 51k resistor arrived and was installed.  Meter now works again.

David

[Seiji53]

This is a report on the results of changing the wiring connection to the power switch.
1. No parts were replaced or added.
2. Made two cuts in the circuit wiring.
3. Added two jumper wires.
Please refer to the attached photos for details on the modification.
I have not verified whether this modification will resolve all power switch malfunctions.
I confirmed that the voltage across R51 was 2V before the modification and improved to 2.9V afterward.