# PWM with SG3525

Discussion in 'Electronic Design' started by parkc23, Aug 13, 2004.

1. ### parkc23Guest

Hi, could anyone help me with SG3525?
I have a PWM circuit with SG3525 and need to modify the circuit so
that I can get ~0.2 Hz cycle period. This is very slow compared to
all of the circuits that I can get.
Probably, I can change the frequency by changing the capacitor or
resistor at Rt and Ct (pin 5 and 6), but I don't know the equation to
calculate those.

parkc23, Aug 13, 2004

2. ### leggGuest

On 13 Aug 2004 10:57:07 -0700, (parkc23) wrote:

>Hi, could anyone help me with SG3525?
>I have a PWM circuit with SG3525 and need to modify the circuit so
>that I can get ~0.2 Hz cycle period. This is very slow compared to
>all of the circuits that I can get.
>Probably, I can change the frequency by changing the capacitor or
>resistor at Rt and Ct (pin 5 and 6), but I don't know the equation to
>calculate those.

data at

http://www.microsemi.com/datasheets/SG1525A.pdf

The lowest frequency range specified for this device, using the
internal oscillator is 100Hz. The use of timing C values larger than
100nF isn't documented. I expect larger C values might be acceptible
if an external device was used to discharge it - a small mosfet driven
by an external signal to the sync pin might suit.

Bias current into the C timing pin is between 2 and 20uA, so this
limits the size of the timing resistor to ~150K, if charging time
isn't to be severely affected over temperature. This would require
~16uF timing capacitor to get down to near 0.2Hz.

Have you got frequencies and periods mixed up? For a 50Hz (.02uSec)
operation, with outputs operating out of phase, the oscillator needs
to run at 100Hz. (.1uF/30K or .02uF/150K)

The internal oscillator has a fairly simple internal structure,
allowing external clock drive.

As I recall, the 3525 comparator wass not latched. The 3525A version
latched, allowing only one output pulsewidth per timing period. This
is a usefull feature, so recheck your part number.

At the low frequency stated, you might get better results with
non-integrated components.

RL

legg, Aug 13, 2004

3. ### parkc23Guest

legg <> wrote in message news:<>...
> On 13 Aug 2004 10:57:07 -0700, (parkc23) wrote:
>
> >Hi, could anyone help me with SG3525?
> >I have a PWM circuit with SG3525 and need to modify the circuit so
> >that I can get ~0.2 Hz cycle period. This is very slow compared to
> >all of the circuits that I can get.
> >Probably, I can change the frequency by changing the capacitor or
> >resistor at Rt and Ct (pin 5 and 6), but I don't know the equation to
> >calculate those.

>
> data at
>
> http://www.microsemi.com/datasheets/SG1525A.pdf
>
> The lowest frequency range specified for this device, using the
> internal oscillator is 100Hz. The use of timing C values larger than
> 100nF isn't documented. I expect larger C values might be acceptible
> if an external device was used to discharge it - a small mosfet driven
> by an external signal to the sync pin might suit.
>
> Bias current into the C timing pin is between 2 and 20uA, so this
> limits the size of the timing resistor to ~150K, if charging time
> isn't to be severely affected over temperature. This would require
> ~16uF timing capacitor to get down to near 0.2Hz.
>
> Have you got frequencies and periods mixed up? For a 50Hz (.02uSec)
> operation, with outputs operating out of phase, the oscillator needs
> to run at 100Hz. (.1uF/30K or .02uF/150K)
>
> The internal oscillator has a fairly simple internal structure,
> allowing external clock drive.
>
> As I recall, the 3525 comparator wass not latched. The 3525A version
> latched, allowing only one output pulsewidth per timing period. This
> is a usefull feature, so recheck your part number.
>
> At the low frequency stated, you might get better results with
> non-integrated components.
>
> RL

Thank you so much!!!

The load is a solenoid (12V, 250mA) and I just don't know if a 16uF
capacitor can discharge it. Do you think it is OK?

As to the latching, does it generate multiple pulse width within a
period if it does not latch regardless of the circuit?

I tried the frequency equation on the data sheet
(f=1/Ct(0.7Rt+3Rd))but got a wrong number. Should I use a different
equation??

Finally, could you let me know the circuit with non-integrated
components? I am only a beginner in the electronics and don't know
much.

Thanks again.

cmp.

parkc23, Aug 14, 2004
4. ### leggGuest

On 14 Aug 2004 06:01:11 -0700, (parkc23) wrote:

>The load is a solenoid (12V, 250mA) and I just don't know if a 16uF
>capacitor can discharge it. Do you think it is OK?

The load in your circuit will not be involved directly with the timing
components.

If the SG3525 is used to drive a solenoid, directly, as a load, the
timing priods required will depend on the magnetic characteristics of
the solenoid itself and the physical characteristics of the
>
>As to the latching, does it generate multiple pulse width within a
>period if it does not latch regardless of the circuit?

Multiple pulse widths are possible if the feedback signal varies
within the timing period, unless the first pwm comparison is latched
and disabled until the next clock timing period.

>
>I tried the frequency equation on the data sheet
>(f=1/Ct(0.7Rt+3Rd))but got a wrong number. Should I use a different
>equation??

As long as your units are Farads, seconds and ohms, the result in Hz
should be pretty unambiguous. Post your numbers if you want your math
checked.

>
>Finally, could you let me know the circuit with non-integrated
>components? I am only a beginner in the electronics and don't know
>much.

You indicated originally that you were modifying an existing piece of
hardware. If you have a schematic of this device, it might be quicker
if you can copy it to ABSE.

RL

legg, Aug 14, 2004
5. ### parkc23Guest

Thanks again.
I attach the link for the circuit.

http://home.att.net/~wzmicro/3525.html

If magnetic properties of the solenoid as a direct load affect the
operation of PWM, how can I isolate each? (maybe, some kind of
interface??)
Also, if latching is necessary, can I directly replace SG3525 with
SG3525A?
I am still confused if I can go upto 16uF to get 0.2Hz or higher Rt.
What about the circuit with non-integrated components?
I can choose any circuit since I have not started building any yet.
Thanks.
cmp.

parkc23, Aug 14, 2004
6. ### R.LeggGuest

(parkc23) wrote in message news:<>...

>> with the timing components.
>>
>> If the SG3525 is used to drive a solenoid, directly, as
>> a load, the timing periods required will depend on the magnetic
>> characteristics of the solenoid itself and the physical
>> characteristics of the mechanically-coupled armature's load.
>>>
>>> As to the latching, does it generate multiple pulse width
>>> within a period if it does not latch regardless of the circuit?

>> Multiple pulse widths are possible if the feedback signal
>> varies within the timing period, unless the first pwm comparison
>> is latched and disabled until the next clock timing period.

> I attach the link for the circuit.
>
> http://home.att.net/~wzmicro/3525.html
>
> If magnetic properties of the solenoid as a direct load affect the
> operation of PWM, how can I isolate each? (maybe, some kind of
> interface??)
> Also, if latching is necessary, can I directly replace SG3525 with
> SG3525A?
> I am still confused if I can go upto 16uF to get 0.2Hz or higher Rt.
> What about the circuit with non-integrated components?
> I can choose any circuit since I have not started building any yet.

I suggest that you do not have sufficient grasp of the finction of
the 3525 PWM function, yet, to modify the original circuit.

I suggest that if you want to learn about the 3525 and pulse-width
modulation, then you should build the circuit as described, with the
components suggested in the reference and then use it to turn the
solenoid on and off by adjusting the rheostats intended originally to
alter motor speed. You could then figure out how to command the 3525
to turn the solenoid on and off at your intended low frequency rate by
manipulating the appropriate analog or digital control line with a
cmos 555.

The 555 could be used to control a simple bipolar of mosfet switch to
do the same thing, without going through the PWM controller, if
switching the solenoid is your only aim.

http://www-s.ti.com/sc/ds/tlc555.pdf

RL

R.Legg, Aug 15, 2004
7. ### parkc23Guest

I guess I need more homework before asking questions. Could you
recommend a good book for the PWM and 555, not too difficult? I am
reading a book by Malvino (Electronic Principles), but this touches
those relatively briefly.

By the way, the reason I need a solenoid control is that I want to
control a 3-way solenoid valve where fluids A and B coming in and a
certain ratio of mixture A and B going out by changing switching
between A and B by simply turning a potentiometer. However, the
solenoid valves are not designed to operate at that high frequency.
That's why I need a very slow PWM.

Would it be possible to do this just by 555 timer without using PWM?
Can I get a circuit?
Thank you.

parkc23, Aug 16, 2004
8. ### Ken SmithGuest

In article <>,
parkc23 <> wrote:
[...]
>Would it be possible to do this just by 555 timer without using PWM?
>Can I get a circuit?

Ascii art:

1N914
-->|-- ------------
! ! ! CMOS 555 !
\ !--------!thr Q!------- Out
Out -->/ ! ! !
\ !--------!trg !
/ ! -----------
100K ! !
--!<--!
!
---
--- 1U
!
GND

Is this what you mean. As you turn the pot from end to end the duty cycle
of the Out signal varies from about 0% to about 100%. You just need to
make it so that Out being high turns on A and out being low turns on B.

--
--
forging knowledge

Ken Smith, Aug 16, 2004
9. ### parkc23Guest

(Ken Smith) wrote in message news:<cfpa7d\$nic\$>...
> In article <>,
> parkc23 <> wrote:
> [...]
> >Would it be possible to do this just by 555 timer without using PWM?
> >Can I get a circuit?

>
>
>
> Ascii art:
>
>
> 1N914
> -->|-- ------------
> ! ! ! CMOS 555 !
> \ !--------!thr Q!------- Out
> Out -->/ ! ! !
> \ !--------!trg !
> / ! -----------
> 100K ! !
> --!<--!
> !
> ---
> --- 1U
> !
> GND
>
> Is this what you mean. As you turn the pot from end to end the duty cycle
> of the Out signal varies from about 0% to about 100%. You just need to
> make it so that Out being high turns on A and out being low turns on B.
>
> --

Yes, this is exactly what I want!!
Could you get me a more complete circuit? (if it's better for you to
attach file, you could email me as an attachment.)
What's coming in and what's going out? I guess I hook up the
solenoid(12V, 250mA) directly to the pin out, but not sure how I
configure the pot in the input side.
In my humble electronics book, there is an example of astable
operation, but it only operates the duty cycle of 50~100%. Here is
the equation for the duty cycle.
D=(R1+R2)/(R1+2R2)
So, theoretically, the shortest duty cycle only goes down to 50%.
Is it possible to get a range of 0~100% duty cycle in your circuit?
Thanks.

parkc23, Aug 16, 2004