Shop OBEX P1 Docs P2 Docs Learn Events
Oh...my...Voltage regulator pinout woes — Parallax Forums

Oh...my...Voltage regulator pinout woes

grindelgrindel Posts: 68
edited 2010-03-24 12:58 in General Discussion
The 3.3V regulator that I bought from here: www.mouser.com/Search/ProductDetail.aspx?R=LD1117V33virtualkey51120000virtualkey511-LD1117V33 (mouser) has the tab hooked to Vout.

Why in the world would you want the Heatsink to be hooked to Vout?

Screenshot-DATASHEET%20SEARCH%20SITE%20%7C%20WWW.ALLDATASHEET.COM.png

It's always the little things that get you...Just thought I would share so something like that doesn't waste 4 hours of anyone else's life.

Comments

  • Peter JakackiPeter Jakacki Posts: 10,193
    edited 2010-03-20 07:12
    It's not like they hook it to the heatsink on purpose but because the substrate of the silicon chip (remember it's silicon) rests on the center leg which is part of the lead frame it is stamped out of. In all regulators and transistors this substrate is connected to the center and usually has a leg for this. Since this part is surface mount it makes more sense to skip the redundant leg and leave room for larger pcb pads for the other two. The center tab needs the intimate connection to the copper to help remove the heat.

    Meter check your TO220s and you will see the tab is the same as the center pin. Never has it been any different.

    ▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
    *Peter*
  • sylvie369sylvie369 Posts: 1,622
    edited 2010-03-20 12:16
    After getting bitten by this once, I made up a document with a table of all of the Vregs I use, their pinouts, current capacities, dropout voltages, and my preferred sources, and anything "special" about them.
  • Erik FriesenErik Friesen Posts: 1,071
    edited 2010-03-20 13:37
    Get a different regulator. I got bit by this once too, and decided thereafter to make sure the datasheet specs show the normal pin 1 in, pin 2 ground connected to tab, and pin 3 out.
  • grindelgrindel Posts: 68
    edited 2010-03-20 20:19
    The problem is all in my inexperience. Both VRegs that Parallax sells are center ground and ground at the tab. It seems like a bad idea to leave the tab at 3.3V since there is a pretty good chance it is going to be a big open terminal to short things against
  • PJAllenPJAllen Banned Posts: 5,065
    edited 2010-03-20 21:45
    You can short to ground, too.· Ground is no less electronically active than some positive voltage.·

    You shouldn't have expectations, or guess, or feel; you should read the datasheet and know what you're doing.
  • ElectricAyeElectricAye Posts: 4,561
    edited 2010-03-21 14:04
    I feel your pain. I encountered the same thing with the very useful MOSFET IRF3708. I had two IRF3708's with the heat sinks screwed on and setting atop a wooden table. I didn't realize anything was wrong until the point where the two heat sinks just so happened to be touching started doing their rendition of a monkey being crushed under a bed of nails. The resulting squeal turned rabid every guard dog in the neighborhood. As PJ suggests: never presume anything. Always read the data sheet with care.
  • Peter JakackiPeter Jakacki Posts: 10,193
    edited 2010-03-22 01:03
    grindel said...
    The problem is all in my inexperience. Both VRegs that Parallax sells are center ground and ground at the tab. It seems like a bad idea to leave the tab at 3.3V since there is a pretty good chance it is going to be a big open terminal to short things against

    You sound ("It seems like a bad idea") as if they have a choice or that they forgot to connect the tab somewhere else, it doesn't work that way. It is the way it needs to be if you understand how silicon ICs are constructed. I have attached the block digram and a photo of the die (silicon).

    The humble bipolar TO220 has the silicon substrate mounted to the center tab and leg which is one piece. Due to the planar construction of the collector, base and emitter it turns out that the part that rests on the metal tab is the collector, the same goes for MOSFETs and their drain. This also is advantageous as the heat from the substrate can be coupled through the tab to dissipate the heat.

    Anyway, this is the real world, we have to fit in with it, not it with us, presume nothing, always read the datasheet.

    ▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
    *Peter*
    lm1117 diagram.png
    763 x 527 - 24K
    lm1117 die.png
    520 x 759 - 359K
  • grindelgrindel Posts: 68
    edited 2010-03-24 02:46
    The only reason I was surprised is that I hadn't yet come across a TO220 that wasn't center/tab ground It makes sense to me to have the ground on the center/tab since ground is usually also joined to the chassis if it is a metal container so the container can be used as a heat sink. Also it seems like I have seen them joined to a ground plain on a PCB, again for use as a heat sink.

    Anyway, thanks for the information.

    datasheets before, after, during...
  • ElectricAyeElectricAye Posts: 4,561
    edited 2010-03-24 03:36
    grindel said...
    ...It makes sense to me to have the ground on the center/tab since ground is usually also joined to the chassis if it is a metal container so the container can be used as a heat sink. Also it seems like I have seen them joined to a ground plain on a PCB, again for use as a heat sink.
    ...

    It seems to me that it ought to be that way, too. But I got slapped here on the forum for saying the very same thing a while ago. As for reading data sheets: I almost got "burned" today because I had not paid attention to the mechanical drawing on a data sheet. Stupid me, I thought all 24 DIPs were one standard size. Oops.
  • PJAllenPJAllen Banned Posts: 5,065
    edited 2010-03-24 12:58
    Lots of times stuff does get attached to the chassis and so on, as you note.

    Nylon washers and·fiber bushings, silpads, mica insulators, thermate·and the like are used to keep things electrically isolated while maintaining thermal conductivity.

    That's how it's done.
Sign In or Register to comment.