Searching \ for '[OT] Designing power supplys' in subject line. ()
Help us get a faster server
FAQ page: www.piclist.com/techref/power/actodc.htm?key=power
Search entire site for: 'Designing power supplys'.

Exact match. Not showing close matches.
'[OT] Designing power supplys'
2000\02\21@202330 by

part 0 934 bytes
Ok so far, now, i select 24vac at the input (wich must be approximate 34vdc at the output of the rectifier). In order to mantein a good stability in high currents I know that the relation given by  w*Rl*C >= 40 where w = 2*pi*f  (50Hz), Rl is load resistance wich is 2.8ohm@10A and C is the capacitance.

So, i asembled the enclosed high current regulator circuit and "feed it" with the above mentioned power supply. Now, when I try to load the circuit the voltage falls to about 24 vdc when it was at 28 vdc with out load.

Any Ideas??

Important: The circuit below its been modify for availability reasons, as the LM195 (power transistor) i'm using the TIP142 and as the 2N... i'm using the TIP4X series (PNP OR NPN i don't remember, but is the same as the 2N..)

Any help will be apreciated, Mauricio Jancic

Attachment converted: wonderland:supply.jpg (JPEG/JVWR) (00018072)
What happens to the voltage at the Vin capacitor (right after the

Transformer rated to 24V @ 12A  doesn't mean it would holds 24Vac at
10A... this 12A specification mostly says that the transformer wiring
and magnetic flux stands to delivery 12A, not related to power delivered
(you can not trust it always).  Note that 24V @ 12A is around 288
Watts!!!... Are you sure this trafo is that big?

Probably you could find a switching power supply to delivery 24V @ 10A
cheaper than the just the trafo... much less weight at least... :)

What do you mean by; "..and a series 2.5 ohm NTC"  ?

Wagner

Mauricio;

You want a 280 watt power supply, right?  Well, you are experimenting with a linear power supply.  You are forced to accept the following parameters:

A bridge/resistor/capacitor filter is in reality a peak detector circuit (not an average or RMS circuit).  When there is no load, the capacitor charges to full peak voltage.  You have set that to the target.  When you draw a heavy current, you are not allowing the capacitor to keep the peak voltage on it.  If you look with a scope, you will see a big ripple with an average of 24V
and a peak to peak of probably 8VDC, at obviously 60 Hz (50? where are you?).  You will approach the RMS value.  0.707 times the starting value of 28 is 19.8.

The big filter cap on a linear supply will vary with load and line voltage variation.  You thus have to plan for the worst line and worst load.  Therefore, at no load and high AC line, you will have far too much voltage at the cap.  You must then follow this capacitor with a series pass regulator as you showed in your scanned picture.  Since you have to deliver at least 280W, you
must be prepared to burn off a lot too, a wild guess would be 100W.

An autotransformer is not ideal for this application.  They offer you no isolation whatever, which is dangerous, and are not exactly highly efficient.  You are better with a good old fashioned 'fillament' transformer, designed for the 380 or more watts that you need.

400W power supplies do not an easy project make.  In modern designs, you would definitely approach this design with a well designed off line switcher.  280 watts with 85% eff would waste 49 watts, a lot, but doable.  You would no doubt have to have a custom flyback wound for the project.  Again, not a good candidate for your first power supply.

I suggest you look into the off the shelf 'lab' power supplies which are available.  They are rugged, can withstand abuse, and have agency regulations assigned to them.  If you only need a few power supplies, this is by far the best way to go.

G'luck
Chris Eddy

Mauricio Jancic wrote:

{Quote hidden}

Yes, even that it is easy to "series regulate" the output Voltage with
very low power dissipation without load, it would require a small fixed
load to allow the regulator to do it, lets say a 5 Watts resistor at the
output, it means 200mA, 120 Ohms.
One important point for this somehow high output current, is to reduce
at maximum all the impedances as good wires and very well crimped and
soldered terminals.  The filter capacitor should be *in the middle of
the current path*, it means, +V and Ground should come FROM the
rectifier TO the capacitor and another set of wires FROM the capacitor
TO the regulator.

If you install the capacitor as an final element in a set of wires,
charging and discharging current would use different directions at the
wires, what is not good.  Capacitors this big ( 0.064F ) normally
doesn't have a low ESR, and in this conditions it could be better to use
2 or even 3 smaller caps (Lower ESR) in parallel, so you would have a
very low impedance at the filter.  Remember that during the time the
sine wave is below the cap voltage, the current to the regulator and
load would be coming just from the cap, and it is a nice 10 Amps... And
as you can guess, the sine wave is almost 85%± below the cap value
(depending on the ripple), so the cap impedance is what would rule about
noise, ripple and other nasty "insects".

Do you really want a nice filtering system?  Do you have space?  Install
a small 24V truck battery after the regulator, and reduce the size of
the cap... 2 x 12V batteries in series could end up in unbalanced
Voltage.  As the battery would be supplying your output voltage (24V),
you should regulate the Voltage TO the battery, not from the battery to

My best 5V power supply (low noise, high power outlet isolation, good
filtering, portable, etc., etc.) comes from a 6V battery @ 9.5A I ripped
off from one of those toy cars (Power Wheels), I just need to recharge
it once a month, during 30 minutes, and the recharge is not because the
power consumed, but because the battery discharge itself... :)

Wagner

Chris Eddy wrote:
[snip]
> The big filter cap on a linear supply will vary with load and line voltage variation.  You thus have to plan for the worst line and worst load.  Therefore, at no load and high AC line, you will have far too much voltage at the cap.  You must then follow this capacitor with a series pass regulator as you showed in your scanned picture.  Since you have to deliver at least 280W, you
> must be prepared to burn off a lot too, a wild guess would be 100W.
[snip]

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

>
>         So, i asembled the enclosed high current regulator circuit and
> "feed it" with the above mentioned power supply. Now, when I try to load
> the circuit the voltage falls to about 24 vdc when it was at 28 vdc with

Few linears can maintain regulation into zero load. (switchers too)
You'll likely need some minimum load to keep it in.

-----BEGIN PGP SIGNATURE-----
Version: PGP for Personal Privacy 5.5.2

iQA/AwUBOLIjvIFlGDz1l6VWEQJ4AQCfdfZqYgm/3F+bE5chs74fG2k43SAAnig9
TrTvhaYzm7HNOhHKLNfrHujA
=0ZG8
-----END PGP SIGNATURE-----

Mauricio Jancic wrote:

> Hi guys,
>         I have a cuestion for those Analogic electronics fans.
>         I must make a power supply for a proyect I'm currently on, it must be adjusted at 28 v, and it must draw about 10 amps with a 5% stability. I look at some motorola rectifiers desing manual and I ended in a 50 A bridge rectifier, with a 68.000uF GIGA capacitor and a series 2.5 ohm NTC. At the input i put a autotransformer wich can draw 12 A and is rated from 0 to 250 v.
>
>         Ok so far, now, i select 24vac at the input (wich must be approximate 34vdc at the output of the rectifier). In order to mantein a good stability in high currents I know that the relation given by  w*Rl*C >= 40 where w = 2*pi*f  (50Hz), Rl is load resistance wich is 2.8ohm@10A and C is the capacitance.
>
>         So, i asembled the enclosed high current regulator circuit and "feed it" with the above mentioned power supply. Now, when I try to load the circuit the voltage falls to about 24 vdc when it was at 28 vdc with out load.

What is raw voltage when the supply is loaded?  Vin?

{Quote hidden}

--
*
|  __O    Thomas C. Sefranek   tcscmcorp.com
(*)/ (*)  Bicycle mobile on 145.41, 448.625 MHz

http://www.harvardrepeater.org

-----Mensaje original-----
Enviado el:     Lunes, 21 de Febrero de 2000 11:06 p.m.
Para:   PICLISTMITVMA.MIT.EDU
Asunto: Re: [OT] Designing power supplys

What happens to the voltage at the Vin capacitor (right after the

Transformer rated to 24V @ 12A  doesn't mean it would holds 24Vac at
10A... this 12A specification mostly says that the transformer wiring
and magnetic flux stands to delivery 12A, not related to power delivered
(you can not trust it always).  Note that 24V @ 12A is around 288
Watts!!!... Are you sure this trafo is that big?

Probably you could find a switching power supply to delivery 24V @ 10A
cheaper than the just the trafo... much less weight at least... :)

What do you mean by; "..and a series 2.5 ohm NTC"  ?

Wagner

Wagner:

I mean that i have two wires from the rectifier, one goes directly to the negative side of the cap, the other one (+) is conected to an NTC [rated 2.5 ohm (OK?)] and the other NTC pin goes directly to the posittive pin of the cap. This alows the capacitor to charge "slow" and not draws a very big current when you connect the power supply.

All the other people:

Thanks for your help. I'm reading the suggestions and ther will be usefull.

Mauricio Jancic

-----Mensaje original-----
Enviado el:     Lunes, 21 de Febrero de 2000 11:06 p.m.
Para:   PICLISTMITVMA.MIT.EDU
Asunto: Re: [OT] Designing power supplys

What happens to the voltage at the Vin capacitor (right after the

Transformer rated to 24V @ 12A  doesn't mean it would holds 24Vac at
10A... this 12A specification mostly says that the transformer wiring
and magnetic flux stands to delivery 12A, not related to power delivered
(you can not trust it always).  Note that 24V @ 12A is around 288
Watts!!!... Are you sure this trafo is that big?

Probably you could find a switching power supply to delivery 24V @ 10A
cheaper than the just the trafo... much less weight at least... :)

What do you mean by; "..and a series 2.5 ohm NTC"  ?

Wagner

As others have already said, designing a 280 watt supply is not child's
play.  Once you get it working, there is the question of how long it will
work before it fails.  If you don't get it right, chances are that it
won't be long.

If you only need a few (or one) and have the time, and can afford for it
to go up in smoke at the worst possible time, go ahead and try to make it

If you only need a few and your (or your employer's) livelihood depends
on them working well, buy them.  The few hundred dollars (US) they'll
each cost is worth the investment.

If you are going to sell these things, hire someone who knows what he's
doing.  In the long run, it will cost less than lost business and lost
customers because of poor reliability, and may even cost less to make.

With that said, I just looked through my Jameco catalog and found two
supplies that might do your job.  Part number 147053 is a 24V, 20A
supply, which may be adjustable up to 28V.  Price is US\$230-\$180
depending on quantity.  Part number 137402 is 24V, 12.5A and may be
adjustable to 28V.  Price is \$140-\$125.  You can check them out at
http://www.jameco.com or call at 800-536-4316.

In addition to what others have suggested, I'll also suggest that you
look at the resistor values on the transistor bases.  They may need to be
lower values (and higher power).  The LM195 that the circuit was designed
for has very high gain, which the TIP4x probably does not.  You have
calculated the capacitor value to give you 3V of ripple at V(in), which
is good.  You should measure it to make sure you got what you paid for.

Good luck.

Don

Dear Mauricio, it sounds strange to me, but this NTC 2.5 Ohms at 10A
current would holds 25V over itself. What was suppose to be your good
intention to reduce the current peak from the rectifier to the cap is in
true reducing it too much. I imagine that when you apply the load, this
NTC should heat a lot, isn't it?

To avoid the peak current, you should introduce a coil (shock) instead
of the NTC. The shock would avoid fast pulses or fast current changes.
I don't have the formulas to calculate this shock, but it should offer a
very little voltage drop at 10A *DC*.

uncharged when you switch on the power.  It is almost a short circuit to
the transformer and rectifiers... In those situations, you could use the
shock to reduce the HIGH peak charging current. A relay could switch
another resistor in series, but it would demand a good relay contact to

About the NTC yet, I can't understand how you have *any* voltage above
15V when you apply the 2.4 Ohms load at the output... so you see, you
are applying in real 5 Ohms to the rectifier, 2.5 from the NTC and 2.4
from the load, so in true you have a voltage divider, and only half

Apply the load, and measure the voltage DC before and after the NTC...
(or across the NTC).. tell me then.

Wagner.

Mauricio Jancic wrote:
{Quote hidden}

Mauricio, this NTC is working as a Inrush current limiter. And as it heats
its resistance will decrease until it reaches a stable value, usually mOhms,
and the voltage drop will be some mV. It will not work properly if it is
heated, you turn the unit off, the cap is discharged and you turn it on
again (its resistance will not return to 2.5R as faster). Note: make sure
the NTC you are using is suitable for this amount of power (you can find
different sizes).

Marcelo Y.

Wagner wrote:

{Quote hidden}

directly to the negative side of the cap, the other one (+) is conected to
an NTC [rated 2.5 ohm (OK?)] and the other NTC pin goes directly to the
posittive pin of the cap. This alows the capacitor to charge "slow" and not
draws a very big current when you connect the power supply.
>>
>> All the other people:
>>
>>         Thanks for your help. I'm reading the suggestions and ther will
be usefull.
{Quote hidden}

Hi Wagner,
Thanks for your reply. I have been looking and I ended in connecting the load directly to the rectifier (with the cap of course!). Now I can draw about 8 Amp and I have 28.5v. I didn't try with 10A coz I don't have resistors that big (power) right now. So now, with out the regulator it works ok, I think that if sometime I'd like to use the regulator (wich I think is the right thing to do) I must buy a higher output voltage trafo.
Again, thanks Mauricio

-----Mensaje original-----
Enviado el:     Martes, 22 de Febrero de 2000 11:51 p.m.
Para:   PICLISTMITVMA.MIT.EDU
Asunto: Re: [OT] Designing power supplys

Dear Mauricio, it sounds strange to me, but this NTC 2.5 Ohms at 10A
current would holds 25V over itself. What was suppose to be your good
intention to reduce the current peak from the rectifier to the cap is in
true reducing it too much. I imagine that when you apply the load, this
NTC should heat a lot, isn't it?

To avoid the peak current, you should introduce a coil (shock) instead
of the NTC. The shock would avoid fast pulses or fast current changes.
I don't have the formulas to calculate this shock, but it should offer a
very little voltage drop at 10A *DC*.

uncharged when you switch on the power.  It is almost a short circuit to
the transformer and rectifiers... In those situations, you could use the
shock to reduce the HIGH peak charging current. A relay could switch
another resistor in series, but it would demand a good relay contact to

About the NTC yet, I can't understand how you have *any* voltage above
15V when you apply the 2.4 Ohms load at the output... so you see, you
are applying in real 5 Ohms to the rectifier, 2.5 from the NTC and 2.4
from the load, so in true you have a voltage divider, and only half

Apply the load, and measure the voltage DC before and after the NTC...
(or across the NTC).. tell me then.

Wagner.

Mauricio Jancic wrote:
{Quote hidden}

Hi Mauricio, do you connected the load directly to the rectifier output
or directly to the cap?  I understand you have the NTC between the
rectifier and the cap...  I want to know if the NTC was involved in the
current flow.

This is somehow a crazy idea. Imagine currents, at the secondary 28Vac,
primary 117Vac, ratio is about 4.17, take the 0.17 as losses, it means
that 10A at the secondary would be 2.5A at the 117V primary. Now, it
would be much easier to control the 117Vac @ 2.5A than the 28Vac @ 10A.
Suppose you connect a 400V x 8A triac (Q4008L4-ND Digikey \$2.00) in
series with the primary to the 117Vac. Suppose you connect this triac to
keep conducting all the time, using a resistor between Anode and Gate,
then you connect the output side of an opto-coupler to switch off (when
active) the triac.  The LED side of the opto-coupler would be fed by a
voltage comparator, like an LM358, the "dirty one".  Use an easy 5V
zener to feed a trimpot, center point fo to the LM358's (-) input, while
the (+) goes to a voltage resistive divider (3V) connected to the Cap +
side.  Whenever the cap Voltage is higher than the trimpot preset, the
LM358 will drive the opto-coupler LED, that will cut off the triac at
the transformer's primary. Without supply, the cap will loose Voltage
until the resistive goes below the 3V (trimpot preset) and the triac
will be free to conduct again, recharging the cap and so on.

VCAP
110Vac-------.   .-------.    .-------o--------o---------o----o Output
Neutral      |   |       |    |       |        |         |
3 C      ~ |    |+      |        R2 22k    R4 2k2
3 C     .----------.    | large  |         |
3 C     |          |   +++ CAP   o--.  .---o
3 C     |  BRIDGE  |   ---       |  |  |   |
3 C     '----------'    |    3k3 R3 |  | .---'
3 C      ~ |    | -     |        |  |  |   A  5VZener
|   |       |    |       |       Gnd |  |   |
|   '-------'    '-------o           |  |  Gnd
o-----.                  |           |  |
|     |                 GND   +VCAP  |  '------.
|     |                        |     |         |
---    R1 20k ?               .---.   |         R5 10k
Triac   V/A    |                     /   +|---' V*      |
|\    |             .------<     |             |
| '---o----.  Opto  |       \   -|----------->TRIMPOT
|          |        |        '---'            22k Ohms
|          V   <-_  V        LM358             |
|         ---      ---         |               R6 10k
Phase        |          |        |          |               |
117Vac-------o----------'        |          |               |
Gnd        Gnd             Gnd

Optocoupler should be a Zero Crossing Triac Output, p/n MOC3040IS-ND
Digikey \$1.60. The output of
the optocoupler should conduct in both sides of the sinewave.

Another opto could be the PS3601LNEC-ND a dual inverted photo diodes
(\$1.15 at Digikey).

Trimpot Voltage swing from 1.13V to 3.63Vdc, with the circuit above it
will regulate VCAP from 8.66 to 27.84 Vdc.

Change R5=22k, TRIMPOT=10k, R6=33k, Trimpot will regulate from 19.6 to
25.52 Vdc.

The zero-crossing opto would give you smooth noise and triac will work
cool (I guess), so you probably would need a small heatsink (if any)...
pretty different from the linear regulator with transistors, etc...

Problem is; your Voltage regulation would be as fast as 16ms.
Changes can be made to change phase at the triac, still regulation at
16ms.

It also requires a fixed resistor at output as a small load to help the
regulation.

I don't know how this circuit will behave at power on... perhaps needs
some fix to avoid 358 latching up, because voltages at the LM358 would
be crazy, so its output.  Paul? anything you see? probably a 9V zener in
series to the LM358 Vcc to make it works only after Vcap goes above 18V?

Wagner.
http://www.ustr.net

Mauricio Jancic wrote:
>
> Hi Wagner,
>         Thanks for your reply. I have been looking and I ended in connecting the load directly to the rectifier (with the cap of course!). Now I can draw about 8 Amp and I have 28.5v. I didn't try with 10A coz I don't have resistors that big (power) right now. So now, with out the regulator it works ok, I think that if sometime I'd like to use the regulator (wich I think is the right thing to do) I must buy a higher output voltage trafo.
>         Again, thanks Mauricio

Hi,
>>      Hi Mauricio, do you connected the load directly to the rectifier output
>> or directly to the cap?

Directly to the cap.

>> I understand you have the NTC between the
>> rectifier and the cap...  I want to know if the NTC was involved in the
>> current flow.

Well... I solve it in a diferent way... I conect the coil of a 28v (or 24v) relay to the cap. The NC connector of the relay makes all the circuit current to pass trough the NTC untill the cap voltage is the required to open the NC contact and close the NA contact wich conects the capacitor directly to the rectifier. I know this needs a "hard" relay that supports all the circuit currents, but I think this is, some how, a easy solution that complies with all the requirements, anyway yours is a good Idea but I know that the "relay idea" will work at once and thats what I need right now.

Mauricio

-----Mensaje original-----
Enviado el:     MiŽrcoles, 23 de Febrero de 2000 09:24 p.m.
Para:   PICLISTMITVMA.MIT.EDU
Asunto: Re: [OT] Designing power supplys

Hi Mauricio, do you connected the load directly to the rectifier output
or directly to the cap?  I understand you have the NTC between the
rectifier and the cap...  I want to know if the NTC was involved in the
current flow.

This is somehow a crazy idea. Imagine currents, at the secondary 28Vac,
primary 117Vac, ratio is about 4.17, take the 0.17 as losses, it means
that 10A at the secondary would be 2.5A at the 117V primary. Now, it
would be much easier to control the 117Vac @ 2.5A than the 28Vac @ 10A.
Suppose you connect a 400V x 8A triac (Q4008L4-ND Digikey \$2.00) in
series with the primary to the 117Vac. Suppose you connect this triac to
keep conducting all the time, using a resistor between Anode and Gate,
then you connect the output side of an opto-coupler to switch off (when
active) the triac.  The LED side of the opto-coupler would be fed by a
voltage comparator, like an LM358, the "dirty one".  Use an easy 5V
zener to feed a trimpot, center point fo to the LM358's (-) input, while
the (+) goes to a voltage resistive divider (3V) connected to the Cap +
side.  Whenever the cap Voltage is higher than the trimpot preset, the
LM358 will drive the opto-coupler LED, that will cut off the triac at
the transformer's primary. Without supply, the cap will loose Voltage
until the resistive goes below the 3V (trimpot preset) and the triac
will be free to conduct again, recharging the cap and so on.

VCAP
110Vac-------.   .-------.    .-------o--------o---------o----o Output
Neutral      |   |       |    |       |        |         |
3 C      ~ |    |+      |        R2 22k    R4 2k2
3 C     .----------.    | large  |         |
3 C     |          |   +++ CAP   o--.  .---o
3 C     |  BRIDGE  |   ---       |  |  |   |
3 C     '----------'    |    3k3 R3 |  | .---'
3 C      ~ |    | -     |        |  |  |   A  5VZener
|   |       |    |       |       Gnd |  |   |
|   '-------'    '-------o           |  |  Gnd
o-----.                  |           |  |
|     |                 GND   +VCAP  |  '------.
|     |                        |     |         |
---    R1 20k ?               .---.   |         R5 10k
Triac   V/A    |                     /   +|---' V*      |
|\    |             .------<     |             |
| '---o----.  Opto  |       \   -|----------->TRIMPOT
|          |        |        '---'            22k Ohms
|          V   <-_  V        LM358             |
|         ---      ---         |               R6 10k
Phase        |          |        |          |               |
117Vac-------o----------'        |          |               |
Gnd        Gnd             Gnd

Optocoupler should be a Zero Crossing Triac Output, p/n MOC3040IS-ND
Digikey \$1.60. The output of
the optocoupler should conduct in both sides of the sinewave.

Another opto could be the PS3601LNEC-ND a dual inverted photo diodes
(\$1.15 at Digikey).

Trimpot Voltage swing from 1.13V to 3.63Vdc, with the circuit above it
will regulate VCAP from 8.66 to 27.84 Vdc.

Change R5=22k, TRIMPOT=10k, R6=33k, Trimpot will regulate from 19.6 to
25.52 Vdc.

The zero-crossing opto would give you smooth noise and triac will work
cool (I guess), so you probably would need a small heatsink (if any)...
pretty different from the linear regulator with transistors, etc...

Problem is; your Voltage regulation would be as fast as 16ms.
Changes can be made to change phase at the triac, still regulation at
16ms.

It also requires a fixed resistor at output as a small load to help the
regulation.

I don't know how this circuit will behave at power on... perhaps needs
some fix to avoid 358 latching up, because voltages at the LM358 would
be crazy, so its output.  Paul? anything you see? probably a 9V zener in
series to the LM358 Vcc to make it works only after Vcap goes above 18V?

Wagner.
http://www.ustr.net

Mauricio Jancic wrote:
>
> Hi Wagner,
>         Thanks for your reply. I have been looking and I ended in connecting the load directly to the rectifier (with the cap of course!). Now I can draw about 8 Amp and I have 28.5v. I didn't try with 10A coz I don't have resistors that big (power) right now. So now, with out the regulator it works ok, I think that if sometime I'd like to use the regulator (wich I think is the right thing to do) I must buy a higher output voltage trafo.
>         Again, thanks Mauricio

Interesting. And what about a mixed concept?
You can command your primary triac by means of an optotriac which LED is
someway between the input and the output of a linear voltage regulator, so
as to keep that input-output difference at a rather low level, say 4-6
volts. It could be useful for an adjustable voltage power supply. The TRIAC
is a pre-regulator, and you get pretty much the characteristics of a linear
regulator whithout the high power dissipation when you need low output
voltage at high current ratings.

Do you understand the concept?
Do you understand my english?
Is this too way [OT]?
Who turned the lights off?!

Marcelo Fornaso

----------------------------------------------------------------------------
---------------------------------------------
{Original Message removed}
Hi Mauricio hola:

I've seen those NTC in PC switching power supplys, but at the primary side
He visto esos NTC en fuentes de PC, pero del lado del primario...

Regarding your relay solution: it's certainly not much elegant, but if it
works for you it's ok.
Con respecto a tu solucion a reli: no parece muy elegante, pero si te
funciona eso es lo importante.

Nevertheless I've got a suggestion: why don't you connect N.O. contacts
across your NTC, so as when your cap reaches voltage enough, the contacts
close and you got two paralell paths for your current. As I see there is no
need to open the NTC path, and it will help if your relay contacts are no
perfect.
Sin embargo tengo una sugerencia: porqui no conectas los contactos N.A. en
paralelo con el NTC, de manera que cuando el capacitor alcanza la tensisn
suficiente, los contactos se cierran y tenis dos vmas en paralelo para la
corriente. Me parece que no hay razsn para desconectar el NTC, y puede
ayudar a los contactos del reli.

Regards to all.
Besos y abrazos para los gauchos.

Marcelo Fornaso
http://www.sysameri.com/marcelo/

{Original Message removed}
The only difficulty he might have is the transformer will not like having it's
power turned on and off like that.  You may need extra circuitry to avoid the
resulting surges and noise...