here's a new schematic. i feel safe posting it as there are other missing pieces not shown here, and some further tweaks need to be done.

can the 393 handle the fet gates? obviously the fet's will have smallest gate charge as possible, etc.

 

A PM sent to me...

what do you think of this one. i left out the 5v supply (which i already have a good design for), and i didnt add in everything around the 393. switch S2 needs to also control the 5v supply because when i turn this module off i want to disconnect the wires from the machine (yellow box). the idea is to switch pin B between A & C, like AB and then BC using PWM, etc. your thoughts?

http://s28.postimg.org/fdohy5aql/Untitled_2.png

I have attached the schematic from the above link.

Unless you use logic level MOSFETs and a CMOS 555, 5V for the power supply is a non-starter. Most MOSFETs need at least 10VDC to work correctly, and a hysteretic oscillator will not work at 5V for a conventional 555. I pointed it to you earlier, but bone up on the 555 Hysteretic Oscillator pointers. I have a couple of other links you might find handy I have written over the years, along with some notes to help design with a 555.

Bill's Blog
The 555 Projects

Now an then I get in a design mood. You hit me at the right time, it is getting rarer, though I'm started to write again. Call it clinical depression, but I'm getting away from it. The fact you are willing to try designing yourself has a lot to do with it.

Are you familiar with screen shots? It is how I pulled your schematic to repost it.

I will help you with a design you will be happy with, and try to teach you the why's and wherefore's along the way. I like to teach, call it paying it forward.

If you really want 5VDC I will make that a design constraint. You will need to get some logic level MOSFETs if so, or some other type of transistor (such as a BJT). 12VDC for the power supply is a better way to go, I have a junk box of wall warts, call me spoiled.

I don't have time to draw something at this moment, but I'll work on it after I see what you would like to do with the power supply.

 

i have been using the fairchild lm555cn with a good/stable/clean 5v supply for a long time, no issues, but it is close to the min 4.5v.

i can easily change the supply to 9v or 12v if need be, the ON MC78M09CTGOS-ND should suffice. i can then feed my supply input circuit (aka "power supply") 12v AC or DC from wall-wart, etc.

 

The 555 will work fine with 5V, but the hysteretic oscillator chokes on that voltage due to the 1.2V drop on the high voltage. It is a idiosyncrasy with that particular design. Just to keep it interesting I'll draw up a straight 556 circuit that might do what you want. You can find the designs off the two data sheets I attached earlier.

This design will run off 5VDC OK, it is based off this design in my cookbook...

Set the monostable for the shortest duration, then set the frequency, then adjust the monostable for the PWM % you desire. An oscope might be handy for this operation, but all long as you follow the sequence you may not need it.

Good enough?

 

the 555/393 is the better circuit for me. i can bump the supply voltage up.

 

OK, next attempt...

This circuit should combine the drive characteristics and simplicity of operation you want. It will operate 7VDC to 15VDC, 12VDC is optimal.

<whiny voice> "Are we there yet?"

 

So the point is to run the comparator output thru the other half of the 556 instead of additional transistors?

 

Output drive, 100ma drive both ways is nothing to sneeze at. A comparator can handle 16ma max, and this only sinking current. This with a 14 pin dip instead of a 8 pin dip, no extra parts.

 

Bill,
ah, very nice, and i have some 556's right here to play with. Q is though, if i were to buffer my fet drives with bjt would i get better/cleaner final stage output PWM vs letting the other 555 do the driving of the opto's directly? i like this 556 approach, less parts on my board, etc.

 

Not sure what you mean? The last illustration will have a very crisp and clean output, this drawing from Chapter 10, where I do use BJTs, was to make up for the deficiencies of the comparator.

It is a kludge, I like the approach I used in post #26 much better. Thing is, this is a how to article, so I don't feel the need to update it.

I might use the image as an alternate way of doing things though, you have helped me update this article in several ways (a good thing).

For the record, the 555 / 556 will do a much better job of driving MOSFETs than what is shown above.

 

with a 555 driver stage i can drive my opto's directly, no need for additional silicon, just as you have it in post #26.
i was comparing that to post #30 where a bjt buffers the fet drivers. so, #26 vs #30, which one provides the cleanest most accurate output looking at the drain of Q3 and ping 9 of the 556, etc.

 

I like the approach I used in post #26 much better. Thing is, this is a how to article, so I don't feel the need to update it.

I like that better, too, even if you still need the MOSFET for higher current. A few extra pins on the 556 is nothing compared to the transistors. The 555 can't hit the upper rail voltage, so that could be one negative for low voltage circuits.

Do reconsider updating fig. 10.4. I send folks there from time to time. It's not about need.

 

OK, next attempt...

This circuit should combine the drive characteristics and simplicity of operation you want. It will operate 7VDC to 15VDC, 12VDC is optimal.

<whiny voice> "Are we there yet?"

Dumb question Bill, should pins 8 and 12 on U1b both be tied together like you show? If it

 

The 7555 (a CMOS 555) will hit rail to rail, and is good for driving MOSFETs too, though it has no drive current to speak of (except at higher supply voltages).

 

A minor brain fart correction, forgot the reset pin on U1b...

Yes, pins 8 and 12 should be tied together.

 

The 7555 (a CMOS 555) will hit rail to rail, and is good for driving MOSFETs too, though it has no drive current to speak of (except at higher supply voltages).

intersil ICM7556
100mA, a respectable amount. i dont think i need fet driver for my opto's, i only need ~5-15mA, etc.

 

The 100 ma spec is only for 18VDC. At 3VDC power supply levels you are lucky to get 3 ma. A minor quirk of CMOS (all CMOS).

However, this design would work at 5VDC power supply using a CMOS 556, and would drive MOSFETs just fine.

Knowing the limits of chips is half the battle in designing.

BTW, I have updated the article, again.

 

well, instead of the 7556 maybe just a std lm556cn with a 9 or 10vdc supply, it can sink 200mA. if i can avoid adding a piece of silicon to the board i will, etc.

 

If (and it is an if) you use MOSFETs then use 12VDC, if not then 9VDC for the power supply is fine.

Can you tell the 555 is one of my favorite chips? People kid me about it regularly on this site. Don't mind it a bit though.

 

how do they say, a oldie but a goody...