Unknown, it floats. If you add a 10Mohm to ground it will either go 100% or 0% if the wiper fails.

i am thinking 10M on each comp input.

 

You could accomplish that at the pot, perhaps.

But I wouldn't bother until you analyze what the consequences are. I mean, if the comparator "scratching" from one state to another as you turn the pot is very bad, then by all means protect against that. But if the pot is rarely moved and a blip wouldn't do serious damage even if it happened, then there's not much incentive to protect against such a thing.

Wiring the "dead" leg of a pot is free, easy and good form. But it's far from a mandatory practice. It's just something you do because you have a choice, and there's a small reason to choose one side of that choice. I don't think most designs specifically address the type of intermittent pot failure we're talking about. My 2¢.

 

You could accomplish that at the pot, perhaps.

But I wouldn't bother until you analyze what the consequences are. I mean, if the comparator "scratching" from one state to another as you turn the pot is very bad, then by all means protect against that. But if the pot is rarely moved and a blip wouldn't do serious damage even if it happened, then there's not much incentive to protect against such a thing.

Wiring the "dead" leg of a pot is free, easy and good form. But it's far from a mandatory practice. It's just something you do because you have a choice, and there's a small reason to choose one side of that choice. I don't think most designs specifically address the type of intermittent pot failure we're talking about. My 2¢.

the two pots will be used on a regular basis. i will be using quality sealed pots, but i might as well protect against the possibility.....

 

The only input that needs the 10MΩ is the one connected to the wiper.

 

The only input that needs the 10MΩ is the one connected to the wiper.

the - of comp might float due to the dpdt switch i have in there.

 

so, in the pic below

• you see the cap & resistor choices via that dpdt switch. by math a 68u(C1) cap w/ 150ohm(R4) and R6 pot is 10k, should yield a freq range of about 1-68Hz. my scope is showing ~30Hz on the high side. why?
  
• pin 1 of LM393AN swings 0-5v, but pin 9 of 7556 only swings 0-3.12v. i see 3.12v only when pin9 is connected to the opto. when not connected to opto i see 0-5v. why do i see 3.12v when connected to the opto? this is only seen when opto led #2 is connected (opto pins 3-4).
  
• i am also a tad confused as to why the PWM out of comp pin 1 is reversed on pin 9 of 7556.

 

ok, so, having the 2nd half of 7556 drive the two opto led's (two led's back to back with 7556 in the middle) didnt work. so, looking further it looked like i needed to isolate the drive for each opto led, so i had to frankenstein a np fet soic i have and i am driving the fets directly from the comp pin-1 output.
the np fet soic is logic level #MMDF2C03HD by ON. this ON ic has very little gate charge so i dont see any reason the comp cant handle about 25nC through 100ohm. with the switch back to np fet driver my circuit works as designed, and thus i only need one 7555 as i am not using the 2nd half of 7556. i am letting my circuit do some burn-in at 196Hz.

output on the opto side is very clean square wave, some very minor rounding at the corners (very small radius in 80us, and full rise & fall taking 0.2ms). now just need to figure out why 68u w/ 150ohm with 10k pot doesnt get me the freq range i was expecting.

 

so this frequency equation. using 10k pot R1 w/ 150ohm R2 and 68u. i calculate 1-68Hz, but i am only seeing 1-29Hz on my scope. why?
0.7/(0+150)*68u = ~68Hz
0.7/(10k+150)*68u = ~1.01Hz

Here is another variation in Chapter 10...

 

Are you using a conventional 555 and a 5V power supply? The equations go a little wonky for 5V, reread the hysteretic oscillator article. It has to do with the output characteristics of a 555.

If you need to reverse the polarity of the inputs of the comparator to invert the PWM signal.

If needed I can show you a more classic 555 circuit that will work with 5V and be more predictable. The triangle wave is all we are after.

 

i am using ICM7556IPDZ
the freq range i have changed to my liking. the voltage issue i noted has to do with fact that the two opto leds were connected in series from +v to gnd. i have now added my np fet driver between comp and opto and it all works the way i want it to. i should be doing functional testing next week after i get an Amphenol connector.

 

Just remember, CMOS powered by low voltages tend to have greatly reduced drive, I've use this on my circuits. A 7555 powered by 15V can sink 100 ma, at 3V it is 2ma. This could be the problem.

I'll find or draw up a correction in a bit. Stepping out for now.

 

Yep, I think the problem is the 150 ohms for the high speed. The CMOS 555 can't drive much. That's why the timing is ok when the 10K is in the circuit. I suspect you can see it change as you adjust the pot lower.

 

Yep, I think the problem is the 150 ohms for the high speed. The CMOS 555 can't drive much. That's why the timing is ok when the 10K is in the circuit. I suspect you can see it change as you adjust the pot lower.

ok, you are talking about the freq adjustment? i switched to 22u & 500ohm, this gets me into an acceptable range with 10mA as the surge as the cap charges and discharges. with 150ohm and 68u its pushing 33.3mA (well, it would if perhaps the supply was higher than 5v).

but to note, the high side freq using 500ohm and 22u still doesnt match the math, i am still seeing less than the math.

looks like i am pushing 7556 too hard?
At a supply voltage of 4.5V or more, the ICM7555 and ICM7556 will drive at least two standard TTL loads
which is about 1.6mA x 2 = 3.2mA. so using output to sink/source the timing circuit needs to be done carefully, no less than 1500ohm in that RC leg. what is the cmos doing above 3.2mA ? but the tech sheet says max output current is 100mA?

as for post #71, the opto LED is running at 10mA, perhaps too much for the 7556 output given the supply of just 5v. but now that i have np fet driver everything seems to be working ok. and now i have full isolation of the opto led's.

 

The 7555 and 7556 at low voltage have a definite drive problem, which is probably what you are running into. In other words, what Ronv said.

Try this design for the triangle generator instead...

It should just drop in, and does not depend on the power supply voltage in the slightest.

Or you could use 12VDC instead.

 

i am thinking of just std 555 in post #26 using 9-12v.
i dont fully understand the 7555 spec sheet. can these cmos IC's outputs drive (sink & source) more than 3.2mA ??

The 7555 and 7556 at low voltage have a definite drive problem, which is probably what you are running into. In other words, what Ronv said.

Try this design for the triangle generator instead...

It should just drop in, and does not depend on the power supply voltage in the slightest.

Or you could use 12VDC instead.

 

If this is for your Tig pulser, why not use the 12V output from the amphenol connector?

 

If this is for your Tig pulser, why not use the 12V output from the amphenol connector?

yikes..... i already stated why..... this module will be isolated from the machine. the opto side is doing exactly what i want it to do, which basically moves the wiper from one side to the other, which is exactly the same thing as cycling the mechanical pedal quickly with your foot.

having the module completely isolated from the machine is a nice safety feature in terms of isolating machine circuitry from module circuitry, a BIG plus imho. the downside is that the module will need a wall wart to power it.

 

re: post #71, i bumped the supply up to 9v and i saw no changes in the frequency ranges using 22u/500ohm or 68u/150ohm, both are still shy of the math on the high side of the freq range, etc.

 

Just for the halibut try 12 V and a regular 556.

Like I said, more than half the battle designing is knowing the limitations of your chips.

 

Just for the halibut try 12 V and a regular 556.

Like I said, more than half the battle designing is knowing the limitations of your chips.

well, given the opto led isolation when using np fet between comp and opto, i no longer need 556.... a 555 or 7555 will do. i will try a 555 @ 12v to see if the freq ranges work any differently.