4
\$\begingroup\$

I have been having problems with 2N7000 mosfets that perform perfectly on the breadboard, but when moved to a perfboard start failing. I'm on my third iteration now, replacing all mosfets, between iterations having eliminated possible other sources for failure. On my third attempt I've been probing each separate mosfet for shorts and testing them with the uC between soldering the next. After soldering the third of eight in total, all three suddenly failed. This time they failed having not a shorted gate-drain, as before, but turning the gate ON/OFF has no effect, current now always flows from Drain to Source. I'm quite desperate at this point. :(

  • I've always been very careful with ESD, always handling the mosfet by the plastic, not the pins. Last run I moved my soldering iron to a grounded socket.
  • Before starting handeling components and soldering, I always ground myself by touching the ground prong on the wall socket.
  • I've reduced soldering temperature to 250 C.
  • For testing purposes I'm switching an 1mA LED, at 24V (I verified the current).

My setup is as follows: A microcontroller* provides 0/5 volt to the gate of the mosfet (100ohm resistor on that). On the Drain, there's the LED with a current limiting resistor. Source is connected to ground. Basically there are no other components.

schematic

simulate this circuit – Schematic created using CircuitLab

Since everything is OK on the breadboard, and things start failing after soldering, but never after the first mosfet. I'm fairly confident that something goes wrong during soldering. What could I possibly do wrong here?

*) To be precise, an PCF8547 connected over i2c to an 5V Arduino.

EDIT: I verified that the wall socket is actually grounded.

UPDATE: I decreased temperature to 250C, no luck. I'm fairly sure that I'm not a source of ESD as I'm grounding myself throughout the soldering process, before and after. My suspicion is now on the cheap (but grounded) soldering iron. I'm going to replace it with a soldering station soon, hope that helps. Any additional insights/tips are welcome, I'll keep monitoring this topic and post new findings when the soldering station is in. I may also get some ESD workplace stuff like a strap, I have still to check what's available in this department.

EDIT & UPDATE: Decreased temperature further down to 250C. Now waiting for delivery of two new soldering irons.

\$\endgroup\$
  • 2
    \$\begingroup\$ Include a schematic instead of describing it in text, edit your question and use the build-in schematic drawing tool. I doubt the soldering is the issue. I solder at 330 degrees Celcius and in the last 30 years I do not remember destroying a component by soldering it for too long. \$\endgroup\$ – Bimpelrekkie Aug 14 '17 at 12:10
  • 1
    \$\begingroup\$ Since they are working very comfortably on breadboard, I don't blame the ESD here because the environment would be similar esd wise. Are you sure about the components and voltage levels being the same on breadboard and perf board? \$\endgroup\$ – Whiskeyjack Aug 14 '17 at 12:11
  • \$\begingroup\$ @Bimpelrekkie. Yeah, I'll add a diagram now.. gimme a few minutes ;-) \$\endgroup\$ – svenema Aug 14 '17 at 12:24
  • 1
    \$\begingroup\$ Wrap a copper wire around all pins before you kill more of these. \$\endgroup\$ – ajeh Aug 14 '17 at 17:57
  • 2
    \$\begingroup\$ @svenema I mean that literally. That is how you protect MOSFETs from overvoltage during soldering: by shorting its pins. \$\endgroup\$ – ajeh Aug 17 '17 at 13:45
2
\$\begingroup\$

What kind of soldering iron are you using?

If you use a soldering iron powered directly from mains, instead of a proper soldering station, its tip will be connected, for safety reasons, to the earth ground of the electrical system of your building.

Although MOSFETs like 2N7000 are fairly rugged, compared to discrete parts of 20 years ago, they might get damaged by phantom voltages coming from the ground connection of your electrical system.

\$\endgroup\$
  • \$\begingroup\$ Cheapo Toolcraft/Conrad, direct to mains iron. I noticed that the tip was connected to ground. That's good or bad? An upgrade to a more professional station is high on my wishlist ;-) \$\endgroup\$ – svenema Aug 14 '17 at 12:57
  • \$\begingroup\$ It's one of these: conrad.nl/nl/… \$\endgroup\$ – svenema Aug 14 '17 at 14:00
  • 1
    \$\begingroup\$ The 2N7000 is older than 20 years, and not rugged at all. Many other FETs add a Zener to the gate. \$\endgroup\$ – CL. Aug 14 '17 at 14:26
  • 1
    \$\begingroup\$ Is there a better protected/more modern drop-in replacement for this FET that is not SMD? \$\endgroup\$ – svenema Aug 14 '17 at 14:33
  • 2
    \$\begingroup\$ @CL Well, 20 years ago the 2N7000 was a relatively new thing, at least here in Italy, especially in the hobby market (low power, vertical "power" mosfet). The majority of other MOSFETs were easily killed by simply touching them without an ESD protection wrist band. I never managed to kill a 2N7000 by simple handling or soldering with a soldering station. So, yes, compared with older MOSFETs 2N7000 is fairly rugged. Yes, I know that handling that device without proper ESD countermeasures could degrade its performance, but for hobby work that is usually not a big issue. \$\endgroup\$ – Lorenzo Donati -- Codidact.org Aug 14 '17 at 22:30
1
\$\begingroup\$

I bet you overheat it with to low of a temperature forcing you to apply the heat for a longer time. I use 370C and you just have to quickly touch the pad, pin and solder so the heat doesnt transfer upp into the casing when soldering small lite fragile components like this. You can't blast it with several seconds of heating. If you use led free solder the situation becomes even worse.

\$\endgroup\$
0
\$\begingroup\$

Excuse me for my English.

Although this thread is old but I want to mention something that may help someone with similar problem. I had similar problem with current flows from source to drain when gate voltage was 0V on IRLML2502 and I solved the problem by disconnecting the soldering iron from main socket during soldering the transistor. This problem is observed only with infineon manufactured mosfet and not others like IOR ,that was very odd to me!!

\$\endgroup\$
  • \$\begingroup\$ thanks, will try that! -- I had kinda parked this project, will resume it somewhere in the next year. replying to old threats on StackExchange can definiatly be useful :) \$\endgroup\$ – svenema Dec 24 '19 at 6:19

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.