Inverted input? (not inverting)

Thread Starter

shortbus

Joined Sep 30, 2009
10,045
You probably think I am anal, but your timing diagram does not agree with your schematic. See the first attachment. I realize you have the Q number designations on the same line, but those are not unambiguous, because they could refer to source or drain voltage or current, or gate voltage (which I now assume is what you were intending). I think you will agree that showing HIN and /LIN tied together on your schematic, and showing them as opposite polarity on the timing diagram, is potentially confusing to anyone who is not you.:D
I took the liberty of relabeling your timing diagram so that it is unambiguous, and hopefully is what you intended. See second attachment.
Ron thank, like I've been saying this is all new to me. I don't know the ins and outs of this stuff. I was meaning it to be the output/source voltage and input of the IRS2183. Sorry for the confusion.
 

Thread Starter

shortbus

Joined Sep 30, 2009
10,045
Praondevou, I have an unused 74hc74 in my logic circuit. If the frequency output of the SG3525, is twice that of the PWM frequency, couldn't the unused 74hc74 be used as a divide by two to drop the clock frequency back down to match the PWM frequency?

With out an oscilloscope is there there a way to check if the clock and PWM frequencies coincide?
 

praondevou

Joined Jul 9, 2011
2,942
Let's say RON is right and PWM is half of OSC.

The FF inside the 3525 and the 74HC74 are not related even though they come from the same oscillator frequency.
That means you can not be sure that the rising edge at Q of the 74hc74 coincides with the rising edge of the corresponding PWM signal.

Why not use the 74HC74 in your drawing but tie it's D and CL to GND. Then use PWM A to reset it and PWM B to set it. Q would go to U1A and /Q to U1B.

PWM A = High resets the FF and charges Bank B.
PWM B = High sets the FF and charges Bank A.

I can make a drawing later.
 

Thread Starter

shortbus

Joined Sep 30, 2009
10,045
Thank you all. I knew I had read it some where and found it. Instead of the SG3525, I'll use a SG3524. It is like the 3525 except there is no internal latch on the PWM flip-flop.

From the data sheet, page 7;

"The outputs may be applied in a push-pull configuration in which their frequency is one-half that of the base oscillator, or
paralleled for single-ended applications in which the frequency is equal to that of the oscillator."

To do this both outputs are tied with a diode "or", into one output.
 

Attachments

Thread Starter

shortbus

Joined Sep 30, 2009
10,045
Let's say RON is right and PWM is half of OSC.

The FF inside the 3525 and the 74HC74 are not related even though they come from the same oscillator frequency.
That means you can not be sure that the rising edge at Q of the 74hc74 coincides with the rising edge of the corresponding PWM signal.

Why not use the 74HC74 in your drawing but tie it's D and CL to GND. Then use PWM A to reset it and PWM B to set it. Q would go to U1A and /Q to U1B.

PWM A = High resets the FF and charges Bank B.
PWM B = High sets the FF and charges Bank A.

I can make a drawing later.
Doing that would put the PWM on all of the logic outputs, wouldn't it?
 

praondevou

Joined Jul 9, 2011
2,942
Doing that would put the PWM on all of the logic outputs, wouldn't it?
Not the way I thought. I still believe that you cannot guarantee that PWM will match the 74HC74 output, and this is what you need for the AND gate, right?

If you need to use the oscilator output then the starting point IMO has to be to know EXACTLY how OSC and PWM are related.

I would do it if I had a chip here, but I don't.
 

Jony130

Joined Feb 17, 2009
5,487
Jony, you measured this on a SG3525? Can you also measure PWM vs OSC?
The saw tooth vs PWM seems to confirm Ron's drawing.
Yes I can.
Here you have

saw tooth vs OSC



And

PWM vs OSC




So as you can see Ron drawing are correct.
And I also add LTspice model.
 

Attachments

Last edited:

Thread Starter

shortbus

Joined Sep 30, 2009
10,045
@Jony130, thank you very much for o'scoping the 3525. It made things clearer for me. And saved me from a failed project!

To all, due to seeing how the output of the SG3525 really looks, it wont work for this project. When used in the DC speed control circuit it did work. In that circuit it doesn't matter that the PWM output is actually two separated wave forms. In the circuit I'm doing now it does make a difference.

So now I'll go back to plan 'A'. :) Before finding the SG3525 circuit, I was going to use one of Bill Marsden's 555 PWM circuits, This was plan 'A'.
 

Attachments

Last edited:

Thread Starter

shortbus

Joined Sep 30, 2009
10,045
@Praondevou, Won't work like it's supposed to.

1. Each PWM output is only good to ~ 49%.

2. The PWM should be from ~ 5% to 90%. If there would be a disruption in the PWM during the discharge, like the 3525 does to get more than 49% PWM, the spark would stop and not have enough voltage to reionize the gap.

3. The only thing that should be a PWM is the output on Q3, for each cap bank.

4. Cap bank 'A' should be charging when cap bank 'B' is discharging, and vise-versa.
 

praondevou

Joined Jul 9, 2011
2,942
@Praondevou, Won't work like it's supposed to.

1. Each PWM output is only good to ~ 49%.

2. The PWM should be from ~ 5% to 90%. If there would be a disruption in the PWM during the discharge, like the 3525 does to get more than 49% PWM, the spark would stop and not have enough voltage to reionize the gap.

3. The only thing that should be a PWM is the output on Q3, for each cap bank.

4. Cap bank 'A' should be charging when cap bank 'B' is discharging, and vise-versa.
Ok, I guess I do not understand the requirements.

Let me see if I get this right now. You want:

1. Two PWMs that do not overlap
2. Both can be from 5% to 90%, if one is 5% then the other is 85%
3. you cannot use the 3525 because there is a pause between both PWM, i.e. you cannot just OR them
 

praondevou

Joined Jul 9, 2011
2,942
Unless you want to go with the 555 solution you could also use a UC3842A which has typically max duty cycle of 96%. Very common model.
 
If you tie the two inputs together, and drive a half bridge with the outputs, the half bridge output will be low when the inputs are low, and vice-versa.
 
Last edited by a moderator:

Thread Starter

shortbus

Joined Sep 30, 2009
10,045
Ok, I guess I do not understand the requirements.

Let me see if I get this right now. You want:

1. Two PWMs that do not overlap
2. Both can be from 5% to 90%, if one is 5% then the other is 85%
3. you cannot use the 3525 because there is a pause between both PWM, i.e. you cannot just OR them
1. Two PWMs that are the same but separate, one for 'A' and one for 'B'. This is what the 'AND' gate after the 74hc74 does in the logic.

2. Both must be the same. At a preselected percentage. See number 1.

3. Correct, the pause between the PWMs is something that I didn't know was there. I can 'or' the signal but the pause would disrupt the signal in the logic. ( The data sheet shows this being done, but they don't say there is a pause. I assumed that when one 3525 output stopped, the second one seamlessly started.)
 

Thread Starter

shortbus

Joined Sep 30, 2009
10,045
Unless you want to go with the 555 solution you could also use a UC3842A which has typically max duty cycle of 96%. Very common model.
The UC3842A doesn't have an osc output. How would it be used to drive the 74hc74?

I'm so dumb at the overall electronics, I took a long time to learn what little I know concentrating on just this project, that it will take awhile to let this sink in.

I really can't put into words how much I appreciate you being patient and helping me.
 
Last edited:

praondevou

Joined Jul 9, 2011
2,942
I'm a little slow. Why exactly do we need the OSC output? Because you want a fixed charging time for the cap?

Can we not just:

- charge the cap bank whenever the PWM is OFF (discharge is OFF)

From what I understand you want the PWM up to 90%. That means the cap bank has to be charged always in 10%, wether the PWM is at 5% or 90%.

Why can the charging voltage not be applied longer than 10%? Doesn't change anything because the cap bank charges in 10% time but greatly simplifies your design.
 
Top