Voltage Regulator Question

[beb101]

I have a MAX883 5v regulator on a breadboard hooked up according to the attached application circuit from the data sheet. I am using a 3uf capacitor on the output because that is all I happened to have at hand. The input is from the Wii battery (6v) discussed here,

http://www.picaxeforum.co.uk/entry.php?74-Battery-power-for-projects-inexpensively

My question relates to the output. I am measuring 5.1V at the output, but I would have expected it to be closer to 5.0V.

Is this sort of tolerance typical from a voltage regulator?

Baxter

 

[srnet]

My question relates to the output. I am measuring 5.1V at the output, but I would have expected it to be closer to 5.0V.

I would expect it to be closer to 5.0V also.

How do you know your multimeter is accurate ?

 

[AllyCat]

Hi,

The data sheet (page 2) seems to specify a range of 4.75 to 5.25 volts over most of the working range, but line regulation, load regulation and temperature only appear to contribute a few hundred millivolts, so I would say it's probably well within specification.

But anyway, I wouldn't normally expect better than 1% accuracy with a "typical" (non-precision) regulator and a 1% multimeter accuracy which is consistent with your measurement (2% overall).

Cheers, Alan.

 

[beb101]

Thanks,

Actually, the measurement is 5.09V.

How do you know your multimeter is accurate ?

My meter is a Fluke and generally they are more accurate than the $5 Chinese models. However, my Chinese meter also reads 5.09V.

Baxter

 

[boriz]

What's the SET pin do? Can you tweak the output? (Can't view the PDF it crashes my reader)

 

[MPep]

As per Fig2 from the datasheet, page 9, you can use SET to adjust the output voltage.

 

[srnet]

My meter is a Fluke and generally they are more accurate than the $5 Chinese models. However, my Chinese meter also reads 5.09V.

Well even so, I would not start from an assumption that the multimeters are correct and its the regulator that is wrong.

You can buy precision voltage referances, accurate to 0.1% or so, which you can use to keep a check on your multi meters.

 

[hippy]

Is this sort of tolerance typical from a voltage regulator?

I don't know about the MAX883 but in my experience it's rare to have a regulator give an exact voltage output. The ones on the AXE401 Shield Base seemed particularly good but that wasn't using a calibrated meter.

Most electronics don't require an exact supply voltage so it doesn't matter as long as the supply is within the component's operational limits. For most PICAXE operation at up to 5.5V is acceptable.

 

[Dippy]

I'd go along with Alan's and srnet's comments.

Hippy is right... unless you wish to do some fussy ADC work?

Sadly, without a good reference ,or calibrated equipment, you're simply going round in circles with 2 unknowns.
I too would trust a fresh Fluke 100x more than a $5 China special.
Have you 'scoped to see if any noise is impacting on measurement?

As a slight aside, I had to do some PIC low-budget ADC circuits using a LP2950 5V regulator. I sampled 5 circuits and all regulators were giving better than 4.98 to 5.03V measured using 2 (recently calibrated) Flukes (2 different models) at ~25oC.
That was good enough for my particular App.

 

[beb101]

I thought I would close the loop on this regulator. My conclusion is that it is not a very good regulator. I rewired the circuit, but this time I added a 4.7K pullup to pin 1, LBO\, per the data sheet. I did not have this in my previous circuit. The output still reads 5.09V as before.

When testing the OFF feature with a jumper to ground, the output slowly decays, taking of the order of 20-30 sec to get to 2.0V. However, if I remove the jumper after the output starts decaying, the output spikes to about 5.9V and then stabilizes at 5.09V. This is probably due to contact bounce. Moreover, the regulator will not latch off unless the output has decayed to around 0.5-1.0V before removing the active low sgnal. Reducing the output filter capacitor to 1uF shortens the off decay time as expected, but the regulator will not latch off with this value.

Perhaps the 2.2uF datasheet specified capacitor is critical. This is just one of many regulators and not worth fooling with any longer.

Baxter

 

[Goeytex]

I thought I would close the loop on this regulator. My conclusion is that it is not a very good regulator. I rewired the circuit, but this time I added a 4.7K pullup to pin 1, LBO\, per the data sheet. I did not have this in my previous circuit. The output still reads 5.09V as before.

I think this conclusion is based upon expecting this very nice regulator to perform beyond its intended purpose. 5.09v is well within the published specification for this regulator.

When testing the OFF feature with a jumper to ground, the output slowly decays, taking of the order of 20-30 sec to get to 2.0V. However, if I remove the jumper after the output starts decaying, the output spikes to about 5.9V and then stabilizes at 5.09V. This is probably due to contact bounce.

If it takes that long to decay, then you are probably testing the output with "no load" other than your meter which is not a good testing method. You should have some kind minimum load on the output. If you look at the charts Maxim uses a 1K load resistor for most of their tests.

Moreover, the regulator will not latch off unless the output has decayed to around 0.5-1.0V before removing the active low sgnal. Reducing the output filter capacitor to 1uF shortens the off decay time as expected, but the regulator will not latch off with this value.

Again, with no load on the output, all bets are off.

Perhaps the 2.2uF datasheet specified capacitor is critical.

Not at all. The datasheets clearly states that 2.2uf is simply the minimum value.
From the Datasheet:

An output filter capacitor is required at the MAX882/MAX883/MAX884 OUT pin.
The minimum output capacitance required for stability is 2.2μF. The filter capacitor’s size
depends primarily on the desired power-up time and load-transient responses. Load-transient
response is improved by using larger output capacitors. The output capacitor’s equivalent
series resistance (ESR) will not affect stability as long as the minimum capacitance requirement
is observed. The type of capacitor selected is not critical, but it must remain above the minimum
value over the full operating temperature range.​

So from the datasheet it is clear that transient response is actually improved by using a larger
output capacitor and that 2.2uf is the bare minimum for good regulation.

Another thing to consider from the datasheet is that this regulator is specifically designed for a battery
supply and that it is optimized for that.

From the datasheet:

The MAX882/MAX883/MAX884 are designed to
achieve low dropout voltages and low quiescent currents
in battery-powered systems. However, to gain these
benefits, the devices must trade away power supply
noise rejection, as well as swift response to supply
variations and load transients.....​

That seems pretty clear. This device is not intended for used with a cheap wall wart as a supply or where
fast response to supply or load transients is needed.

This is just one of many regulators and not worth fooling with any longer.

For your particular application that may be true. However the Max883 can be a
very good choice for the applications that it was designed for.

 

[beb101]

Hi Goeytex,

Thanks for your comments and tips. Your point regarding the load resistor is well taken; I did not have a load resistor in the circuit, so I will give this another try.

Ok, here are the results for 3uF output capacitor, 1K load resistor, all other connections as previously noted:

Vout = 5.08V  for Vin =6.01V (battery) ---> OK
Pin 7 (On_Off/)
  Output at 5.08V, Pin 7 to ground ---> instant off (latched off) ---> Vout = 0 ---> OK
  output at 0V,    Pin 7 to Vin    ---> instant on  (latched on)  ---> Vout = 5.08V ---> OK

The regulator now works. The nice thing about this Forum is that it is loaded with with people who know what they are doing and willing to help. I now know to always test a regulator with a load (and also, to read the datasheet more carefully).

Thanks again,

Baxter