Author Topic: LM723 Regulator, DC output waveform at 5Amps  (Read 4531 times)

0 Members and 1 Guest are viewing this topic.

Offline fivefishTopic starter

  • Frequent Contributor
  • **
  • Posts: 440
  • Country: us
LM723 Regulator, DC output waveform at 5Amps
« on: October 17, 2015, 02:51:49 am »
Hi all, this is a datasheet standard schematic of linear regulated PSU using LM723 + external transistors.

I just noticed that my DC output at 5Amps look weird, the downward dips starts at around 4.7Amp, and by 5amps, it gets bigger and looks like this.  (AC coupling on scope)

Does anyone have a clue what might be happening?


ADDENDUM:
Currently, I have (4x) 1000uf filtering caps before regulators. I haven't tried it yet, but do you think I probably need more caps?
« Last Edit: October 17, 2015, 02:57:11 am by fivefish »
 

Offline fivefishTopic starter

  • Frequent Contributor
  • **
  • Posts: 440
  • Country: us
Re: LM723 Regulator, DC output waveform at 5Amps
« Reply #1 on: October 17, 2015, 03:17:59 am »
Never mind... I think problem is solved.

Added an extra 1000uf cap for a total of 5000uf before regulator. Now output is clean and no more unexpected dips @ 5.0Amp load.
Though voltage Dips start showing a little bit at 5.5Amps, and adding another 1000uf caps (total: 6000uf) fixed that issue again.
So it looks like I need 1000uf of filtering caps pre-regulator for every 1 Amp of output load.





 

Online Ian.M

  • Super Contributor
  • ***
  • Posts: 12805
Re: LM723 Regulator, DC output waveform at 5Amps
« Reply #2 on: October 17, 2015, 03:31:09 am »
Insufficient headroom to maintain regulation.  If you are using an ordinary bridge rectifier, replacing its diodes with Schottky ones to reduce the voltage drop will give you a bit more margin.  Boosting the bulk capacitance can be problematic - it increases the duration of the inrush current surge and also increases the stress on the bridge rectifier diodes.
 

Offline fivefishTopic starter

  • Frequent Contributor
  • **
  • Posts: 440
  • Country: us
Re: LM723 Regulator, DC output waveform at 5Amps
« Reply #3 on: October 17, 2015, 03:41:33 am »
I think I have sufficient headroom. 25VDC regulated down to 16VDC. 
I'm using MURF860 diodes rated at 8Amps... Vdrop is 1.5V, but that still leaves me with 22VDC, enough headroom for a 16VDC out?

Luckily, I designed the PCB with enough clearance around caps. I'll be able to upgrade to an 18mm dia. cap, and replace my existing 16mm dia. caps with something else. 1500uf to 2200uf apiece comes in 18mm dia package (with same lead spacing).

Here's a photo of PSU I'm working on.  I've load tested it for several hours at +/-5Amps, though it will do higher up to 6Amps.

« Last Edit: October 17, 2015, 03:45:31 am by fivefish »
 

Online Ian.M

  • Super Contributor
  • ***
  • Posts: 12805
Re: LM723 Regulator, DC output waveform at 5Amps
« Reply #4 on: October 17, 2015, 04:28:32 am »
Scope the unreg rail feeding the LM723 and the pass transistor on one trace and the output on the other.  I suspect you wont be pleased at high load currents.  Also, its worth doing brownout testing with the mains supply reduced to its lower tolerance limit if you are empirically refining the design.
 

Offline PChi

  • Frequent Contributor
  • **
  • Posts: 264
  • Country: gb
Re: LM723 Regulator, DC output waveform at 5Amps
« Reply #5 on: October 17, 2015, 12:57:49 pm »
A crude rule of thumb is 1000 uF per Amp minimum which is now met.
It's worth checking the ripple current through the capacitors by measuring the ripple voltage and estimating the ripple current. It might have an effect on the capacitor life.
 

Offline DimitriP

  • Super Contributor
  • ***
  • Posts: 1288
  • Country: us
  • "Best practices" are best not practiced.© Dimitri
Re: LM723 Regulator, DC output waveform at 5Amps
« Reply #6 on: October 17, 2015, 02:10:31 pm »
Never mind... I think problem is solved.

Added an extra 1000uf cap for a total of 5000uf before regulator. Now output is clean and no more unexpected dips @ 5.0Amp load.
Though voltage Dips start showing a little bit at 5.5Amps, and adding another 1000uf caps (total: 6000uf) fixed that issue again.
So it looks like I need 1000uf of filtering caps pre-regulator for every 1 Amp of output load.



There is an awful lot of excitement  (aka noise) on that trace.    The LM 723 has an output noise voltage of 2.5uV (microVolt) where is the rest of the noise coming from ? Is it lower with a lighter load ?
   If three 100  Ohm resistors are connected in parallel, and in series with a 200 Ohm resistor, how many resistors do you have? 
 

Offline dom0

  • Super Contributor
  • ***
  • Posts: 1483
  • Country: 00
Re: LM723 Regulator, DC output waveform at 5Amps
« Reply #7 on: October 17, 2015, 02:27:18 pm »
A crude rule of thumb is 1000 uF per Amp minimum which is now met.

Yes, it's a very crude rule.

A much better way to calculate required filter capacitance is by beginning with the maximum load current and permissible ripple voltage, then calculating the required capacitance directly (t=10 ms, for a two-pulse rectifier at 50 Hz. Per definition of Farad: C = (I * t) / U).

For low ripple voltages (<10 % of nominal output) you need to use an even bigger transformer than the usual rule of thumb for your rectifier configuration (single [~2.5] vs bridge [1.6-1.8] vs center-tap two diode [~1.4]), because cooper looses in the transformer rise to the square of current, which becomes larger the smaller the ripple becomes.
,
 

Online Kleinstein

  • Super Contributor
  • ***
  • Posts: 14069
  • Country: de
Re: LM723 Regulator, DC output waveform at 5Amps
« Reply #8 on: October 17, 2015, 03:31:53 pm »
WIth 1000 µF per amp of current you still get quite some ripple. So a little more capacitance could help. However more capacitance also makes the current flow more pulsed.

Because of the pulsed current the power factor is rather poor. So the 6.8 A rated trasformer may not deliver much more that 3.5 A DC current at rated load, especially if the capacitor is large. So more capacitance will increase the load to the trasformer and diodes. Peak current will be way more than 8 A.

Some of the LM723 circuits need quite some headroom. In this case it would help to have separate diodes and capacitor for the supply of the LM723 - though this will likely not fit the existing layout.
 

Offline fivefishTopic starter

  • Frequent Contributor
  • **
  • Posts: 440
  • Country: us
Re: LM723 Regulator, DC output waveform at 5Amps
« Reply #9 on: October 17, 2015, 06:25:30 pm »
Thanks for the input guys.

Quote
There is an awful lot of excitement  (aka noise) on that trace.    The LM 723 has an output noise voltage of 2.5uV (microVolt) where is the rest of the noise coming from ? Is it lower with a lighter load ?

It's probably from my scope leads and environment noise (LED lighting, etc.) I don't have differential probes to really get the exact actual value.

I've ordered some higher value, low impedance caps (18mm dia) with higher current ripple rating. Will replace my current 1000uf general purpose caps and see how it goes from there.
 


Share me

Digg  Facebook  SlashDot  Delicious  Technorati  Twitter  Google  Yahoo
Smf