Benchtop Power Supply Options

On 5/25/2014 12:41 PM, Jurd wrote:

>
> 1) Is the ripple due to C1 and C2 being unable to keep up with the 120hz
> swing, and

er...  60hz.

-J

Jurd scribbled thus:

> On 5/25/2014 12:18 PM, John Larkin wrote:
> 
>>
>> That doesn't help when you're building a power supply. You may as
>> well just connect the rectifier caps to the main linear regulator.
>> That's better, actually; a ripple canceler ahead of the regulator
>> makes things worse.
>>
>> The issue is energy storage. 120 times a second, the transformer
>> output goes to zero volts. If you want to keep powering the load
>> then, the energy has to come from somewhere, and in this case it's
>> the filter caps.
>>
>>
> 
> So I guess my questions (that I was afraid to ask previously) are:
> 
> 1) Is the ripple due to C1 and C2 being unable to keep up with the
> 120hz swing, and
> 
> 2) Would increasing the value of the output cap (C3) help at all with
> this?
> 
> If not, what kinds of design features of a linear regulated power
> supply am I looking for?
> 
> Thanks.
> 
> -J

The amount of ripple is a function of the value of the smoothing cap and
load current.

-- 
Best Regards:
                     Baron.

"Baron" <baron@linuxmaniac.net> wrote in message 
news:lltc5b$rlc$1@dont-email.me...
> Jurd scribbled thus:
>
>> On 5/25/2014 12:18 PM, John Larkin wrote:
>>
>>>
>>> That doesn't help when you're building a power supply. You may as
>>> well just connect the rectifier caps to the main linear regulator.
>>> That's better, actually; a ripple canceler ahead of the regulator
>>> makes things worse.
>>>
>>> The issue is energy storage. 120 times a second, the transformer
>>> output goes to zero volts. If you want to keep powering the load
>>> then, the energy has to come from somewhere, and in this case it's
>>> the filter caps.
>>>
>>>
>>
>> So I guess my questions (that I was afraid to ask previously) are:
>>
>> 1) Is the ripple due to C1 and C2 being unable to keep up with the
>> 120hz swing, and
>>
>> 2) Would increasing the value of the output cap (C3) help at all with
>> this?
>>
>> If not, what kinds of design features of a linear regulated power
>> supply am I looking for?
>>
>> Thanks.
>>
>> -J
>
> The amount of ripple is a function of the value of the smoothing cap and
> load current.

Radio hams tend to stick with linears because switchers radiate harmonics, 
pretty sure I've seen the reservoir cap equations on a ham site somewhere.

The N1HFX site has various tutorials - that might be one of them. 

On Sunday, May 25, 2014 10:41:46 AM UTC-7, Jurd wrote:


> 1) Is the ripple due to C1 and C2 being unable to keep up with the 120hz  
> swing, and

Not keeping "up with the 120 Hz swing"  is called filtering.  Those
are filter capacitors.  That is their sole function!

> 2) Would increasing the value of the output cap (C3) help at all with this?

That's not generally a good idea; the current  limit is upstream from C3,
bigger C3 causes sparks to be fatter if there's a short, but otherwise does
very little (and it harms the LM317 if it gets too big).  There's a recommendation
to add  a diode if you need large (1000 uF or so) C3.  

<http://www.ti.com/lit/an/snva517b/snva517b.pdf>

Instead of searching only for transformers, look for complete power supplies
(every recent item will have switchmode regulated supplies).  A 30V 1A
power brick for a discarded printer would serve this purpose well.

On Sun, 25 May 2014 12:41:46 -0500, Jurd
<guitardorkspamspameggsandham74@gmail.com> wrote:

>On 5/25/2014 12:18 PM, John Larkin wrote:
>
>>
>> That doesn't help when you're building a power supply. You may as well
>> just connect the rectifier caps to the main linear regulator. That's
>> better, actually; a ripple canceler ahead of the regulator makes
>> things worse.
>>
>> The issue is energy storage. 120 times a second, the transformer
>> output goes to zero volts. If you want to keep powering the load then,
>> the energy has to come from somewhere, and in this case it's the
>> filter caps.
>>
>>
>
>So I guess my questions (that I was afraid to ask previously) are:
>
>1) Is the ripple due to C1 and C2 being unable to keep up with the 120hz 
>swing, and

Yes. It's an energy storage issue.

The  basic relationship is

C*V = I*T

or C = I*T/V

So if the load current is I=1 amps, and the caps have to power the
load for, say, T=6 milliseconds, and you want V=2 volts of droop
between recharging the caps, ie 2 volts p-p ripple,

C = 1*0.006/2 = 0.003 farads

or 3000 uF.

That's approximate, but probably good enough.





>
>2) Would increasing the value of the output cap (C3) help at all with this?

No. The big caps should be on the rectifier side of the regulator.

>
>If not, what kinds of design features of a linear regulated power supply 
>am I looking for?

Lots of issues.

What voltage and current range?

How accurate is the output voltage?

Does it go down to zero volts?

How much output noise/ripple do you want?

Adjustable current limit?

How much output capacitance? If too much, shorting the output can make
a fat spark and maybe zap things.

Metering?

Heat sinking.


It would be educational to design and build a power supply, but not
cost-effective.

In article <YLqgv.275239$dT1.192494@fx12.am4>, 
gangprobing.alien@ntlworld.com says...
> 
> "Baron" <baron@linuxmaniac.net> wrote in message 
> news:lltc5b$rlc$1@dont-email.me...
> > Jurd scribbled thus:
> >
> >> On 5/25/2014 12:18 PM, John Larkin wrote:
> >>
> >>>
> >>> That doesn't help when you're building a power supply. You may as
> >>> well just connect the rectifier caps to the main linear regulator.
> >>> That's better, actually; a ripple canceler ahead of the regulator
> >>> makes things worse.
> >>>
> >>> The issue is energy storage. 120 times a second, the transformer
> >>> output goes to zero volts. If you want to keep powering the load
> >>> then, the energy has to come from somewhere, and in this case it's
> >>> the filter caps.
> >>>
> >>>
> >>
> >> So I guess my questions (that I was afraid to ask previously) are:
> >>
> >> 1) Is the ripple due to C1 and C2 being unable to keep up with the
> >> 120hz swing, and
> >>
> >> 2) Would increasing the value of the output cap (C3) help at all with
> >> this?
> >>
> >> If not, what kinds of design features of a linear regulated power
> >> supply am I looking for?
> >>
> >> Thanks.
> >>
> >> -J
> >
> > The amount of ripple is a function of the value of the smoothing cap and
> > load current.
> 
> Radio hams tend to stick with linears because switchers radiate harmonics, 
> pretty sure I've seen the reservoir cap equations on a ham site somewhere.
> 
> The N1HFX site has various tutorials - that might be one of them. 

 There is no amount of caps you can add that will remove the ripple 100%
and expect to be able to get full use of the supply..

 The best you can do is use a group of smaller caps to combine a larger
value. Caps have ESR ( series resistance) and large types have more of 
this.
  It's like putting a resistor in series with a perfect cap.

 So, if you were to use a few smaller onces combined and use low ESR 
types you would do better.

 You can get large caps with very low ESR, expect to pay.

 Also, hope that your transformer will handle the inrush currents while 
charging up. Some supplies employ soft start circuits.

Jamie

On Sun, 25 May 2014 12:41:46 -0500, Jurd
<guitardorkspamspameggsandham74@gmail.com> wrote:

>On 5/25/2014 12:18 PM, John Larkin wrote:
>
>>
>> That doesn't help when you're building a power supply. You may as well
>> just connect the rectifier caps to the main linear regulator. That's
>> better, actually; a ripple canceler ahead of the regulator makes
>> things worse.
>>
>> The issue is energy storage. 120 times a second, the transformer
>> output goes to zero volts. If you want to keep powering the load then,
>> the energy has to come from somewhere, and in this case it's the
>> filter caps.
>>
>>
>
>So I guess my questions (that I was afraid to ask previously) are:
>
>1) Is the ripple due to C1 and C2 being unable to keep up with the 120hz 
>swing, and

---
No.

The ripple comes about because of the rectifiers' output voltage
falling below the capacitor voltage as the capacitor discharges into
the load.
---

>2) Would increasing the value of the output cap (C3) help at all with this?

---
No.
---

>If not, what kinds of design features of a linear regulated power supply 
>am I looking for?

---
For a mains ripple-free output from the [linear] supply, all that
needs to happen is for the voltage on the reservoir cap(s) never to
fall below the headroom threshold of the regulator.
---

>Thanks.
>
>-J

---
Yer welcome.

John Fields

On Sun, 25 May 2014 12:43:20 -0500, Jurd
<guitardorkspamspameggsandham74@gmail.com> wrote:

>On 5/25/2014 12:41 PM, Jurd wrote:
>
>>
>> 1) Is the ripple due to C1 and C2 being unable to keep up with the 120hz
>> swing, and
>
>er...  60hz.
>
>-J

---
Since the negative-going portion of the sinusoid is flipped by the
bridge and slipped into the space between the normally
positive-going half-sines, the time between adjacent peaks is
halved, making the ripple frequency twice the mains frequency.

John Fields 

On Sun, 25 May 2014 17:07:49 -0400, "Maynard A. Philbrook Jr."
<jamie_ka1lpa@charter.net> wrote:


> There is no amount of caps you can add that will remove the ripple 100%
>and expect to be able to get full use of the supply..

---
Sure there is.

All you have to do is make sure that the ripple valleys are high
enough to give the regulator the headroom it needs to provide a
ripple-free output at the supply's rated output current.
---

> The best you can do is use a group of smaller caps to combine a larger
>value. Caps have ESR ( series resistance) and large types have more of 
>this.
>  It's like putting a resistor in series with a perfect cap.

---
So what?

To overcome that problem all you have to do is raise the raw supply
voltage enough to compensate for the ESR.
---

> So, if you were to use a few smaller onces combined and use low ESR 
>types you would do better.

---
True
---

> You can get large caps with very low ESR, expect to pay.

---
True again, but so what?
---

> Also, hope that your transformer will handle the inrush currents while 
>charging up. Some supplies employ soft start circuits.

---
The problem isn't with the transformer, since all you really have to
worry about is not to exceed the fusing current of the wire in the
windings for a few seconds, at most.

The real problem lies in not letting the magic smoke out of the
rectifiers during turn-on, and that's easily side-stepped by sizing
(overrating) the rectifiers properly.

 

On Sun, 25 May 2014 18:33:16 +0100, "Ian Field"
<gangprobing.alien@ntlworld.com> wrote:

>
>
>"John Larkin" <jjlarkin@highNOTlandTHIStechnologyPART.com> wrote in message 
>news:e794o9tj2im9tflvlbisagg29g1crl7468@4ax.com...
>> On Sun, 25 May 2014 17:53:03 +0100, "Ian Field"
>> <gangprobing.alien@ntlworld.com> wrote:
>>
>>>
>>>
>>>"Jurd" <guitardorkspamspameggsandham74@gmail.com> wrote in message
>>>news:llrkq5$hr4$1@news.albasani.net...
>>>> On 5/24/2014 8:48 PM, John Larkin wrote:
>>>>> On Sat, 24 May 2014 20:06:45 -0500, Jurd
>>>>
>>>>>
>>>>> That bridge configuration will in theory charge the caps to 1.41 times
>>>>> the RMS voltage of the transformer secondary, because a sine wave has
>>>>> a peak voltage 1.41x its RMS.
>>>>>
>>>>> In real life you'd typically get more DC than that at light loads and
>>>>> less at heavy loads. And the "DC" will have ripple, which makes the
>>>>> voltage dip at 120 Hz (100 Hz in the hinterlands).
>>>>>
>>>>>
>>>>
>>>> Ah thanks.  Good to know about the ripple, as that's certainly something
>>>> I'd like to avoid.  Back to the Googling board!
>>>
>>>Search under "active ripple cancelling" - pretty much just an emitter
>>>follower with some bias and a not quite as huge electrolytic as you'd need
>>>on its own.
>>
>> That doesn't help when you're building a power supply. You may as well
>> just connect the rectifier caps to the main linear regulator. That's
>> better, actually; a ripple canceler ahead of the regulator makes
>> things worse.
>>
>> The issue is energy storage. 120 times a second, the transformer
>> output goes to zero volts. If you want to keep powering the load then,
>> the energy has to come from somewhere, and in this case it's the
>> filter caps.
>
>I never said don't use reservoir caps - an unregulated emitter follower with 
>a heavily decoupled base does its best to follow the insignificant ripple on 
>its base. 

---
So what?

It pays for it by costing more, lowering the output voltage to the
load, and dissipating power.

Version 4
SHEET 1 880 1248
WIRE -768 -16 -992 -16
WIRE -368 -16 -624 -16
WIRE -256 -16 -368 -16
WIRE -368 16 -368 -16
WIRE -256 16 -256 -16
WIRE -256 112 -256 80
WIRE -160 112 -256 112
WIRE 16 112 -160 112
WIRE -768 160 -768 -16
WIRE -624 160 -624 -16
WIRE -992 176 -992 -16
WIRE -368 208 -368 80
WIRE -368 208 -432 208
WIRE 16 208 16 112
WIRE -160 224 -160 112
WIRE -432 256 -432 208
WIRE -368 400 -368 208
WIRE -160 400 -160 288
WIRE -160 400 -368 400
WIRE 16 400 16 288
WIRE 16 400 -160 400
WIRE -368 432 -368 400
WIRE -256 432 -256 112
WIRE -992 528 -992 256
WIRE -768 528 -768 240
WIRE -768 528 -992 528
WIRE -736 528 -768 528
WIRE -624 528 -624 240
WIRE -624 528 -656 528
WIRE -368 528 -368 496
WIRE -368 528 -624 528
WIRE -256 528 -256 496
WIRE -256 528 -368 528
WIRE -768 656 -992 656
WIRE -368 656 -624 656
WIRE -256 656 -368 656
WIRE -368 688 -368 656
WIRE -256 688 -256 656
WIRE -256 784 -256 752
WIRE -160 784 -256 784
WIRE -64 784 -160 784
WIRE 96 784 32 784
WIRE -768 832 -768 656
WIRE -624 832 -624 656
WIRE -992 848 -992 656
WIRE -368 880 -368 752
WIRE -368 880 -432 880
WIRE -160 880 -160 784
WIRE -128 880 -160 880
WIRE -16 880 -16 848
WIRE -16 880 -64 880
WIRE 96 880 96 784
WIRE -432 928 -432 880
WIRE -160 944 -160 880
WIRE -16 944 -16 880
WIRE -368 1072 -368 880
WIRE -160 1072 -160 1008
WIRE -160 1072 -368 1072
WIRE -16 1072 -16 1008
WIRE -16 1072 -160 1072
WIRE 96 1072 96 960
WIRE 96 1072 -16 1072
WIRE -368 1104 -368 1072
WIRE -256 1104 -256 784
WIRE -992 1200 -992 928
WIRE -768 1200 -768 912
WIRE -768 1200 -992 1200
WIRE -736 1200 -768 1200
WIRE -624 1200 -624 912
WIRE -624 1200 -656 1200
WIRE -368 1200 -368 1168
WIRE -368 1200 -624 1200
WIRE -256 1200 -256 1168
WIRE -256 1200 -368 1200
FLAG -432 256 0
FLAG -432 928 0
SYMBOL diode -352 80 R180
WINDOW 0 26 68 Left 2
WINDOW 3 30 0 Left 2
SYMATTR InstName D3
SYMATTR Value ES1D
SYMBOL ind2 -784 144 R0
WINDOW 0 -43 40 Left 2
WINDOW 3 -43 75 Left 2
SYMATTR InstName L1
SYMATTR Value 50
SYMATTR Type ind
SYMATTR SpiceLine Rser=,1
SYMBOL voltage -992 160 R0
WINDOW 3 24 104 Invisible 2
WINDOW 123 0 0 Left 2
WINDOW 39 0 0 Left 2
SYMATTR Value SINE(0 170 60)
SYMATTR InstName V1
SYMBOL ind2 -608 144 M0
WINDOW 0 -46 46 Left 2
WINDOW 3 -62 79 Left 2
SYMATTR InstName L2
SYMATTR Value .229
SYMATTR Type ind
SYMBOL cap -176 224 R0
SYMATTR InstName C1
SYMATTR Value 2000&#4294967295;
SYMBOL res -640 512 R90
WINDOW 0 0 56 VBottom 2
WINDOW 3 32 56 VTop 2
SYMATTR InstName R2
SYMATTR Value 1G
SYMBOL diode -240 496 R180
WINDOW 0 -21 69 Left 2
WINDOW 3 -52 -4 Left 2
SYMATTR InstName D1
SYMATTR Value ES1D
SYMBOL diode -384 432 R0
WINDOW 0 -23 -5 Left 2
WINDOW 3 -50 70 Left 2
SYMATTR InstName D2
SYMATTR Value ES1D
SYMBOL diode -272 16 R0
SYMATTR InstName D4
SYMATTR Value ES1D
SYMBOL res 0 192 R0
SYMATTR InstName R1
SYMATTR Value 100
SYMBOL diode -352 752 R180
WINDOW 0 26 68 Left 2
WINDOW 3 30 0 Left 2
SYMATTR InstName D5
SYMATTR Value ES1D
SYMBOL ind2 -784 816 R0
WINDOW 0 -43 40 Left 2
WINDOW 3 -43 75 Left 2
SYMATTR InstName L3
SYMATTR Value 50
SYMATTR Type ind
SYMATTR SpiceLine Rser=,1
SYMBOL voltage -992 832 R0
WINDOW 3 24 104 Invisible 2
WINDOW 123 0 0 Left 2
WINDOW 39 0 0 Left 2
SYMATTR Value SINE(0 170 60)
SYMATTR InstName V2
SYMBOL ind2 -608 816 M0
WINDOW 0 -46 46 Left 2
WINDOW 3 -62 79 Left 2
SYMATTR InstName L4
SYMATTR Value .229
SYMATTR Type ind
SYMBOL cap -176 944 R0
SYMATTR InstName C2
SYMATTR Value 2000&#4294967295;
SYMBOL res -640 1184 R90
WINDOW 0 0 56 VBottom 2
WINDOW 3 32 56 VTop 2
SYMATTR InstName R3
SYMATTR Value 1G
SYMBOL diode -240 1168 R180
WINDOW 0 -21 69 Left 2
WINDOW 3 -52 -4 Left 2
SYMATTR InstName D6
SYMATTR Value ES1D
SYMBOL diode -384 1104 R0
WINDOW 0 -23 -5 Left 2
WINDOW 3 -50 70 Left 2
SYMATTR InstName D7
SYMATTR Value ES1D
SYMBOL diode -272 688 R0
SYMATTR InstName D8
SYMATTR Value ES1D
SYMBOL res 80 864 R0
SYMATTR InstName R4
SYMATTR Value 100
SYMBOL npn -64 848 R270
SYMATTR InstName Q1
SYMATTR Value 2N2222
SYMBOL schottky -128 896 R270
WINDOW 0 32 32 VTop 2
WINDOW 3 0 32 VBottom 2
SYMATTR InstName D9
SYMATTR Value BAT54
SYMATTR Description Diode
SYMATTR Type diode
SYMBOL cap -32 944 R0
SYMATTR InstName C3
SYMATTR Value 10000&#4294967295;
TEXT -752 136 Left 2 !K1 L1 L2  1
TEXT -984 1232 Left 2 !.tran 10
TEXT -752 808 Left 2 !K2 L3 L4  1