HP E3610A - Sparking?

[alpha_rolf]

Hi,

The HP E3610A is my first lab power supply. Recently when I wanted to repair something and used it for the first time it was sparking when connecting the probes to the PCB.
Is this normal?
I set the current limit to 20mA and 14V. I used DuPont connector for the connection.

When I shorted them, they also sparked (This was just as a test).

I did another test with the normal crocodile clips and those did not spark.

Is this normal or do I need to repair something inside the PSU?

I was very worried that I damaged my PCB. Or am I doing something wrong?

As an additional question:
The lab PSU looks clean inside.

But do I need to change all caps before they maybe start leaking?

Or is this not necessary?

Thanks so much! 

Regards
Rolf

[ArthurDent]

Not enough information to give an answer. Please draw the circuit you're using. Generally you connect the power supply adjusted output to a device you're testing then turn the power on so you don't get a spark while making a connection.

Once connected do the meters on the supply indicate the correct voltage and current? If they do then connecting leads first then turning on the supply will eliminate this 'problem'. If the voltage drops and the current peaks you hay have a problem with the board.

Also is the supply output floating or connected to earth and are there any other devices connected to the board you're testing?

A schematic showing your connections and a complete description of what you are trying to do would help.

[mikerj]

The E3610A has a 470uF decoupling cap on it's output so you can expect a bit of a spark especially if the voltage is set high.  Beware of this if you are connecting something relatively fragile e.g. driving an LED in constant current mode, having the capacitor discharge into it from a high voltage will likely destroy it long before the PSU current limit kicks in. 

[alpha_rolf]

Yes I think you are right. I think the sparking comes from the 470uF cap. How can I drive e.g. an LED without destroying it? The PSU has no output on/off switch, so I am always getting sparks when I connect something. Is there a right way to use this PSU to avoid this? Sorry for dumb questions, but I have never used a lab PSU before. Or shall I remove the 470uF cap?

[Wallace Gasiewicz]

Perhaps put a high ohm resistor across the output maybe 100K or 1 meg. ("bleeder resistor") This should not hurt anything. If you try this let us know.

[pigrew]

Yes I think you are right. I think the sparking comes from the 470uF cap. How can I drive e.g. an LED without destroying it? The PSU has no output on/off switch, so I am always getting sparks when I connect something. Is there a right way to use this PSU to avoid this? Sorry for dumb questions, but I have never used a lab PSU before. Or shall I remove the 470uF cap?

With this supply, there is no great way. One way is to turn the supply off (power button) while changing the load.

You could also add a resistor in series with the output, perhaps around 500 ohm (be careful with its power rating, keep the output voltage lower than maybe 8V). This resistor will limit the current that goes to the LED, but the displayed voltage will be wrong since it'll include the voltage drop over the resistor. I'm assuming that you want to put the supply in constant-current mode.

Perhaps put a high ohm resistor across the output maybe 100K or 1 meg. ("bleeder resistor") This should not hurt anything. If you try this let us know.

The supply already has an internal 390 ohm resistor, to discharge the output capacitor.

[TheSteve]

If you want to connect the LED direct to the power supply output you will to use a series resistor. If you want to test an LED without a resistor the proper tool to use would be an SMU - Source Measure Unit.

[alpha_rolf]

Yes I think you are right. I think the sparking comes from the 470uF cap. How can I drive e.g. an LED without destroying it? The PSU has no output on/off switch, so I am always getting sparks when I connect something. Is there a right way to use this PSU to avoid this? Sorry for dumb questions, but I have never used a lab PSU before. Or shall I remove the 470uF cap?

With this supply, there is no great way. One way is to turn the supply off (power button) while changing the load.

You could also add a resistor in series with the output, perhaps around 500 ohm (be careful with its power rating, keep the output voltage lower than maybe 8V). This resistor will limit the current that goes to the LED, but the displayed voltage will be wrong since it'll include the voltage drop over the resistor. I'm assuming that you want to put the supply in constant-current mode.

Perhaps put a high ohm resistor across the output maybe 100K or 1 meg. ("bleeder resistor") This should not hurt anything. If you try this let us know.

The supply already has an internal 390 ohm resistor, to discharge the output capacitor.

Thanks. May I ask how they were used in the past? I know this supply is older, so back in the days, how did they avoid this "problem"?

Is there maybe a circuit I can build so that it works like a modern lab PSU? I don't know how modern PSU avoid this problem, but I guess that you can build something around it or?

A series resistor would indeed limit the inrush current. Is there another way to do a "soft start"?

[mikerj]

Yes I think you are right. I think the sparking comes from the 470uF cap. How can I drive e.g. an LED without destroying it? The PSU has no output on/off switch, so I am always getting sparks when I connect something. Is there a right way to use this PSU to avoid this? Sorry for dumb questions, but I have never used a lab PSU before. Or shall I remove the 470uF cap?

Turn output voltage to zero, connect LED, turn output voltage up until you hit whatever current limit you have set.  The 470u cap is there for a reason, removing it will hurt transient load performance.

[pigrew]

Thanks. May I ask how they were used in the past? I know this supply is older, so back in the days, how did they avoid this "problem"?

Is there maybe a circuit I can build so that it works like a modern lab PSU? I don't know how modern PSU avoid this problem, but I guess that you can build something around it or?

A series resistor would indeed limit the inrush current. Is there another way to do a "soft start"?

We used these supplies in my undergrad labs.  I've also used these supplies in semi-permanent test setups as a "system supply" where a board/device needs power. I've only used them as constant-voltage sources. I wouldn't normally use them in a situation where I'm constantly tweaking output parameters (I'd use something digitally-controlled, instead). The CC limit is mostly just to prevent the DUT from catching on fire or fuses blowing. Anything with a large output capacitor won't work well as constant-current source with respect to load transients.

Most modern supplies will have an "output-enable" button, so wouldn't need to be powered off to disable the output. Also supplies designed for CC operation will have a much smaller output capacitor.

As millerj mentioned, one solution is to turn down the voltage to zero prior to attaching the load.

I think you could add a front-panel toggle-switch to set the output to 0V. I'm not able to quickly visualize all the feedback loops, so I could be opposite and it would set the output to max voltage. It also could create some transient ringing during use, as I'm sure they didn't design the circuit to slew very quickly. See the attached drawing.

Anyway...  the optimal solution probably is to add some series resistance to your testbench, and then use a separate DVM to get the LED forward voltage.

[alpha_rolf]

OK! Thanks so much everyone! 

[pigrew]

Oh... I also found that there are some service notes for the E3610A relating to voltage transients at startup.

See this page.

[alpha_rolf]

Oh... I also found that there are some service notes for the E3610A relating to voltage transients at startup.

See this page.

Thanks. Do you refer to E3610A-04 or E3610A-02?

in E3610A-04: they removed the output caps C4 and C5
in E3610A-02: they replaced C13 (P/N 0180-4085,330uF 35V) with P/N 0180-4355 (470 uF 50V). Note: as far as I can see, this must be an Agilent thing. As in the old HP schematic, this cap has always been 470 µF 50V

Also, on some service notes it is written: "ACTION CATEGORY: IMMEDIATELY".
This looks a bit scary, does it mean I have to do all these actions now to be ok?