3

In the 7905 negative voltage regulator datasheet, it didn't mention that it needs a minimum load current to operate.

The 7905 that I have is outputting -7 V but adding a ~2 mA load causes the voltage to drop to nominal voltage -5 V. Input is -15 V.

Does a 7905 need a minimum load or is it broken?

TonyM
  • 21,742
  • 4
  • 39
  • 62
ElectronSurf
  • 2,185
  • 3
  • 24
  • 61
  • How comes your 7905 can have 0 mA load in your application? Are you using it as a voltage reference or something or is it when you have a connector unplugged? – TonyM Jun 08 '20 at 13:02
  • @TonyM the 7824 regulator in the same circuit without any load is outputting 24 V, I thought 7905 should do the same too. it's going to be connected to a load (op amps). – ElectronSurf Jun 08 '20 at 13:32
  • As shown below, the spec'd conditions are when drawing 5 mA to 1 A. You need to look at the load on your regulator and ensure that it'll never be lower than 5 mA. Posting your full schematic will make this question much easier to answer and better for others referring to it in the future. The schematic editor here is a breeze to use. – TonyM Jun 08 '20 at 13:37
  • @TonyM Added the schematic, BTW with those large capacitors at the input and output, do you think that I should add the input/output diode or it's not required? – ElectronSurf Jun 08 '20 at 13:42
  • 1
    Thanks for that, you've made it an excellently-presented question now, upvoted :-) The input bridge rectifiers protect against accidental reversing of the DC supply. I don't like them and don't design them in, as the power dissipation through the diode drops is usually unacceptable. More efficient protection can be done with a reverse-biased diode across the supply and a polyfuse in the supply, or a pass FET. If your connectors are polarised and a reversal is unlikely/impossible, ditch the bridge rectifiers like a shot. They also make EMC much worse for switching loads. – TonyM Jun 08 '20 at 19:54

2 Answers2

9

Yes. -5 mA as stated in the datasheet, although not separated and verbatim.

Too high output voltage is common for 78xx and 79xx series linear regulators with too low load.

enter image description here

winny
  • 13,064
  • 6
  • 46
  • 63
  • In general, you'll need to replicate the given test conditions (clearly stated here, sometimes relegated to a footnote) if you want to achieve the associated parameter. In effect, the test conditions are instructions on how to use the device. – Graham Nye Jun 08 '20 at 12:02
3

No, normally it will operate properly with no load, though the load regulation is not guaranteed. There is a slight internal load provided by the internal voltage divider, which is more than good enough at or near room temperature.

Note that the pinout is different from the positive regulators, and that output capacitance is absolutely required for stability, unlike their 78xx positive cousins. Chances are you’ve got one of those two things wrong. If it is oscillating it may appear to have poor regulation unless you have an oscilloscope to see what is really going on (and your load could be damaged),

Spehro Pefhany
  • 376,485
  • 21
  • 320
  • 842
  • 1
    The input capacitance is large (4.4mF) and output capacitance is large too (2.2mF) because the load is a bit away from the supply. the output is not oscillating with or without minimum load. and the wiring is correct. – ElectronSurf Jun 08 '20 at 13:29
  • @ElectronSurf, do you have any high-frequency capacitance on the output, such as a 100 nF capacitor in parallel with the 2200 uF already there? – TonyM Jun 08 '20 at 13:34
  • @TonyM No, just a electrolytic one. should I add one? – ElectronSurf Jun 08 '20 at 13:35
  • 1
    @ElectronSurf Yes. The simplistic view looks like several capacitors will behave as a single one of equivalent capacitance would. In truth, the separate capacitor paths respond differently to different frequencies. Put one on, it'll take no time. But I don't expect it to solve the ouput voltage problem. – TonyM Jun 08 '20 at 13:40
  • 1
    The original manufacturer’s datasheet also says “ The bypass capacitors should be mounted with the shortest leads, and if possible, directly across the regulator terminals”. Worth a try with a small 0.1-1uF ceramic at the leads. – Spehro Pefhany Jun 08 '20 at 13:40
  • P.S. since you have relatively large output capacitance (as disclosed in the comments) if you have added a large-ish Schottky protection diode it might leak enough to raise the output voltage out of regulation. – Spehro Pefhany Jun 08 '20 at 13:50
  • @SpehroPefhany Just soldered a 0.1uF SMD capacitor right on the output and ground pin of the regulator, it didn't had any effect on the output voltage. and there's no diode connected to the regulator not for reverse polarity and not from input to output. – ElectronSurf Jun 08 '20 at 13:55
  • @SpehroPefhany The thing is; I used 7905 before several times and it never needed a minimum load current, I always read -5 V from it's output without any load. maybe it's different from manufacturer to manufacturer? – ElectronSurf Jun 08 '20 at 13:58
  • Part was procured through reliable channels and has no history of the input voltage being shorted to ground (or equivalent heavy load on input supply at turn-off) during operation? – Spehro Pefhany Jun 08 '20 at 13:58
  • @SpehroPefhany No, I just bought it and soldered there. I even was careful not to heat the device too much when soldering... – ElectronSurf Jun 08 '20 at 13:59
  • I agree, something seems wrong here. I would not be surprised to see it go out of regulation at 150 C but at room temperature is troubling. – Spehro Pefhany Jun 08 '20 at 14:00
  • @SpehroPefhany Should I replace it? I don't want to get to trouble later on. I tested it with 500mA load though and voltage dropped just ~400mV. – ElectronSurf Jun 08 '20 at 14:01
  • Did the parts come from reliable channels? I would replace it with another from a different source. – Spehro Pefhany Jun 08 '20 at 14:03
  • @SpehroPefhany I'm living in Iran; I get what I get. Can't order parts from "reliable" sources. the seller said it is original but hey who knows... – ElectronSurf Jun 08 '20 at 14:05
  • 1
    Okay, got it. I guess one with different pedigree is the best we can do. – Spehro Pefhany Jun 08 '20 at 14:06
  • 1
    @SpehroPefhany Thanks for the guidance, appreciate it. – ElectronSurf Jun 08 '20 at 14:08