HP E3610A Does this look worth fixing?

[rdl]

I got a HP E3610A power supply off eBay a couple of years ago. It was, of course, said to be in good condition. I opened the box when it arrived and noticed a cracked binding post. I got an adjustment from the seller, stuck the thing in a closet and never really checked it out until it was too late to get a refund. I pulled it out many months later and when I powered it up I noticed the voltage would jump to maximum before dropping to whatever it was set at. I figured it was probably capacitor problems, so I opened it up.

There's obviously some kind of leakage and evidence of a previous repair. Some of the traces have turned black. I'm just wondering if I should try replacing the caps or should I just chalk one up to experience and use it for parts?

[Alexei.Polkhanov]

If tracks on PCB are still intact it should be easily fixable. It is a linear supply - I love them for low noise and own 2 E3630A (3 output supply with ratio adjust). 
Full service manual is available :  http://www.physics.fsu.edu/users/Wahl/labmanuals/instruments/ps/AgilentE361xAManual.pdf. Hope you can find original diode bridge to replace that ugly contraption :-)

[rdl]

I scratched on the blackened traces a little bit and they seem to be intact under the black stuff. If anybody knows a good way to clean off the leaked out capacitor gunk let me know. It's going to be a bit awkward to get to that area for cleaning. Maybe I'll just run the whole thing through the dishwasher.

[edpalmer42]

What leakage are you referring to?  All I see is glue.  The two brown capacitors might be replacements.  It looks like one of them is spaced up from the board.  I don't think HP would ever do that.  There's also no reason for the glue to be where it is, but if the capacitor was from another piece of equipment, it might have glued two capacitors together to provide mutual support.

It's hard to tell, but I suspect that one or more of those three components shorted, perhaps took out the others, and caused excess current flow that caused the trace to get hot enough to burn the solder mask.  Normally, solder mask is impervious to capacitor goo; it only attacks the pads and any traces where the solder mask was damaged.  If that happened, you'd see solder mask lifted and/or peeled rather than burned.

Ed

[rdl]

Yeah, I think you're right. I took the board completely out and looking at the bottom I can tell both those filter caps are replacements. So is at least one of the fuses, and of course the bodged on diode bridge is not original. The brown stuff on the top of the caps is not capacitor gunk but some kind of dried up glue. The HP glue is grey and still rubbery. I wonder why they didn't use new caps for replacements.

So now I'm clueless as to what is causing the turn on voltage spike. I guess after the problems this thing has already had, it could be almost anything. Too bad because other than that it works pretty good.

[edpalmer42]

For lack of any better ideas, start with that repair job.  Check the replaced components for health, correct values and/or any other significant characteristics.  Look at the circuitry around the repair.  Maybe another component was damaged and the only symptom is the turn-on spike.  If you don't find anything there, you'll have to dive into the schematic and figure out how it's supposed to work.

I have a cheap variable power supply that had a turn-on spike due to really poor design.  The power section wanted to go to full voltage all the time.  It was up to the control section to drag it back down.  Problem was, on power-up, it took a fraction of a second for the control section to wake up and stabilize.  By that time, the output was heading for 50V.  The solution was to add a resistor in just the right place to disable the power section until the control section woke up and could override my new resistor.

Ed

[rdl]

Looking into this a little more, I noticed something interesting. The two capacitors that had been replaced are the main filter caps for the 7812 and 7912 regulators in the "Reference and Bias Supply". Whoever did the previous repair job used 470uf for both caps. However, in the HP schematic 470uf is specified for the 7912, but the one for the 7812 calls for only 330uf.

It doesn't seem like this substitution should cause any problems, but it doesn't make sense to me that HP would have used different values unless there was a reason. Then again, maybe it's just an old typo that was never corrected.

[edpalmer42]

Not only are the values different for the two caps, but so are the voltage ratings (35V and 50V).  The values are close enough that it doesn't matter.  The difference in voltage ratings is also irrelevent.  It's not like it's in a critical area of the circuit.  I can only guess that the component selection was somehow based on manufacturing efficiency or cost-reduction rather than technical requirements.

Ed

[Kevin.D]

Hi there .
Fault like this shouldn't be to difficult to trace ,it's a very basic bench supply.
Usually somthing like this is caused by an auxillary supply which  is delayed/failing to  startup ,which then causes delay before the cntrl opamp's can start.
First check any replacment polarized caps on those 12V regs are fitted CORRECT polarity .
If the caps are fitted ok and you have DSO available then  you  next  confirm if it's an auxilary supply problem ,do this by  scoping  the output of  U1 on one channel  and the main supply output on another chan during startup ,and looking to see if U1 output remains high during startup ,and then falls conicidently with the main overvoltage spike . If this is so then it confirms you have a problem with startup of U1 ,so then you would scope each of the Aux supply lines that supply U1 during start upto to see which if any of them are having startup problems. Once you have narrowed it down then the rest is easy.

p.s  Don't forget when your connecting your two scope channels that the  +Vout on these supplies is your main Aux gnd .And so you  use this point as your single gnd on both scope channels (i.e dont short somthing out through your common scope gnd).

Regards

[Jamieson]

There are quite a few E3610A's up on ebay and about half of them look OK except they have broken buttons.  In some cases it's just the button cap missing.  But there are a few where a switch post looks totally broken off.  Are these parts (switch, caps) available to make a quick repair and save some $$$?

[rdl]

I still haven't found time to work on this power supply. I did take a look at the switches.

The power and range switch are Schadow Type NE-18. It seems they have become part of C&K. The CC set button is a different type and I couldn't find a name or number on it, but it has the same exact type of actuator, about 3.2 x 3.5 mm.

C&K do make some caps to fit these switches. It looks like none of them are quite the same shape or color as the original HP caps, but I think they'll work. Mouser has them in stock and they're not very expensive, so I'll probably put a bunch on my next order just to see what they're like.

The HP caps are made out of some strange plastic that's very brittle. The cracks tend to start from the inside, so even if you see one on eBay that looks okay, it may be just one push from splitting apart.

[LabSpokane]

The plastic is either polycarbonate or a PC/ABS blend. It is very strong, but does not like being under strain for extended periods of time and responds by cracking.  One tip is to *not* use any type of super glue or cyanoacrylate adhesive to repair them. That type of glue only makes them crack more.

[rdl]

I'm not so sure. I tried to plastic weld one back together using some MEK and it seem to barely dissolve, if at all. The exposed surface where it cracked open is very friable and if you scrape it comes off almost like a powder. It actually seems as if the plastic was overloaded with filler pigment when it was compounded.

[LabSpokane]

That could be since white has the highest pigment loading. The plastic could have been left in the barrel too long and/or too much regrind material was used. It also could be PBT, which would be unusual since there was no sublimation printing on those keytops. The latter would explain why MEK won't touch it. There are some expensive plastic epoxies that supposedly work on PBT/nylon. It is unusual to see stress cracking with PBT.

[rdl]

I needed some stuff from Mouser last week so I added a few of the C&K button caps. They come in four colors but Mouser only stocks the red and black, so I went with black. I wish I'd gotten a few red ones now, red might have looked good on the power button.

The C&K caps are not quite as big as the original HP/Agilent, but pretty close and it looks a lot better than an empty hole. I really wanted gray though, so I may try painting a couple of these while I look for someplace that stocks the gray ones.

The HP caps are about 12.25 w x 8.1 h x 9.0 d (depth into the hole), the C&K are about 12.3 x 7.0 x 10.5. The extra depth isn't noticeable, but you can see the difference in height when they're side by side. I think if all three were the same size then no one would notice the missing millimeter. The C&K caps are glossy and the HP have a textured matte finish. Again, wouldn't be noticeable if they were all the same.

So if you have one of these older power supplies with missing or cracked buttons, this is a decent fix. This is the F13 style cap for the NE-18 switch. The C&K part number for this exact cap (black) is F1301. They're 73 cents each, a cheap fix for a very common problem with these power supplies.

[macboy]

I needed some stuff from Mouser last week so I added a few of the C&K button caps. They come in four colors but Mouser only stocks the red and black, so I went with black. I wish I'd gotten a few red ones now, red might have looked good on the power button.

The C&K caps are not quite as big as the original HP/Agilent, but pretty close and it looks a lot better than an empty hole. I really wanted gray though, so I may try painting a couple of these while I look for someplace that stocks the gray ones.

The HP caps are about 12.25 w x 8.1 h x 9.0 d (depth into the hole), the C&K are about 12.3 x 7.0 x 10.5. The extra depth isn't noticeable, but you can see the difference in height when they're side by side. I think if all three were the same size then no one would notice the missing millimeter. The C&K caps are glossy and the HP have a textured matte finish. Again, wouldn't be noticeable if they were all the same.

So if you have one of these older power supplies with missing or cracked buttons, this is a decent fix. This is the F13 style cap for the NE-18 switch. The C&K part number for this exact cap (black) is F1301. They're 73 cents each, a cheap fix for a very common problem with these power supplies.

For the similar E3620A, the grey caps had a similar problem, in that they would split and fall off due to a bad batch of brittle plastic. Agilent had a program to provide free replacements to all affected units (based on serial number). I had a few, so I ordered up the replacements, which were promptly next-day couriered to me. I only replaced the ones that had cracked so I have several extras. Unfortunately, they are square, not rectangular like the ones on the E3610A.

I checked for a similar service note on Keysight's site but there is not one. They do however list several service notes intended to improve the performance due to a variety of issues from oscillation to overshoot. Unlike most HP/Agilent supplies, I don't think the E3610A deserves its good reputation.

[rdl]

Replacement buttons painted and installed on the Agilent E3610A. Much better than empty holes. Also, a good comparison shot to show the differences between the older HP and the newer Agilent style. I still think a red power button might look nice, but will have to wait until my next order from Mouser.




I agree that the E3610A does not deserve it's good reputation. I doubt I'll buy another.

[HighVoltage]

I still think a red power button might look nice, but will have to wait until my next order from Mouser.

Can you order fitting buttons from mouser?

I have a few of these older and newer PSU's and they all work really great.

[rdl]

Can you order fitting buttons from mouser?

Yes, you can.
They're not exact matches but they're close. See my earlier post for details.
Mouser only stocks the black and red. You would have to go elsewhere for the gray.
I just bought some black ones and painted them.

I have a few of these older and newer PSU's and they all work really great.

Have you checked them with an oscilloscope for overshoot on a cold start?
It seems to be a flaw with the 3610s, maybe not for others in the series.
They are fine after they are turned on.

[sync]

Have you checked them with an oscilloscope for overshoot on a cold start?
It seems to be a flaw with the 3610s, maybe not for others in the series.
They are fine after they are turned on.

My E3610A doesn't overshoot. In reality the ramp up was too slow. So i changed C10 from 6.8uF to 1uF to make it faster.

[HighVoltage]

Thanks for the info on the buttons
Every time I needed one before, I ordered it directly from Agilent.

In regards to overshoot, I never noticed that
I do have a few of this series and will measure them shortly, to see if I missed something.

Just used one today.
I needed to measure a low resistance of 3 mOhm but with 1 A current
Set the PSU to 1.000 A and read the mV voltage across the resistor = mOhm
These power supplies have been proven very reliable to me over the last 10 - 15 years

[rdl]

I've just been poking around inside mine. C10 is the cap that gets replaced in "Service Note" 03. There is also a trace to be cut and a jumper to install. It looks like this change was incorporated by the time my Agilent was built, it has the 6.8 uf tantalum. My HP version has a .001 uf ceramic disc.

The service note only referred to a "sporadic overshoot caused by an intermittent voltage control", not a turn on overshoot, so I didn't pay much attention to it. My Agilent has turn on overshoot also, but it's nowhere near as bad as the HP. As I've said before, I guess I just have bad luck.

Thanks for the info on the buttons
Every time I needed one before, I ordered it directly from Agilent.

Well, my HP 3610 has a cracked binding post and when I checked last year, they wanted $13 plus shipping for a replacement, so I was afraid to ask about buttons

[nowlan]

I just bought a e3610a on ebay. The CC button is cracked unfortunately.
You guys dont recommend super glue.
Is the black button from mouser best option?
Did anyone get a price from agilent?

Im surprised this is such a common fault.
Mind you, a heat sink fin is bent so who knows what life this thing had before hand =)

[LabSpokane]

I just bought a e3610a on ebay. The CC button is cracked unfortunately.
You guys dont recommend super glue.
Is the black button from mouser best option?
Did anyone get a price from agilent?

Im surprised this is such a common fault.
Mind you, a heat sink fin is bent so who knows what life this thing had before hand =)

Try using epoxy.  It may be salvageable. I'd be interested in seeing a photo of where the button is broken.

[HighVoltage]

Epoxy usually does not stick well to this type of plastic.
I have found that an acrylic glue, as liquid as water, works best on these type of plastics and really welds the parts together.

Look at my fixing of the cracked Keysight E3649A here:
https://www.eevblog.com/forum/repair/brand-new-keysight-e3649a-psu-repair-and-teardown/