Does this need a Snubber?

[stuhagen]

This is from another post, but after testing today, the relay does oscillate a little if the voltage "hovers" at the Vsrc transition. Not too concerned because the odds of the voltage "hanging" are slim. So I was wondering if this circuit may need a snubber to prevent this damage when the relay does "chatter" or oscillate. Not sure how to figure out the values of R and C.
Maybe 1KO and .22uf?

Stu

 

[jpanhalt]

Where would you insert the snubber? Are you sure the problem isn't due to the comparator needing more hysteresis?

John

 

[Mikebits]

Where would you insert the snubber? Are you sure the problem isn't due to the comparator needing more hysteresis?

John

Agreed. Try increasing R3 reducing C2.

 

[ericgibbs]

For your application R3, 1M0 is to high a value, you require about 22K.

If you need to use a 1M0 hysteresis value resistor, you require a 10K resistor from pin3 to the wiper of RV1

 

[Mikebits]

I stand humbled and corrected...

 

[ericgibbs]

I stand humbled and corrected...

hi, whats with the change of persona.. [avator].

 

[Mikebits]

hi, whats with the change of persona.. [avator].

Lol, I changed my pic as a joke, and now I can't find my original pic. hehe

 

[GonzoEngineer]

I had the same problem once.....the advice you got is correct.

The comparator is the problem

Reduce the "Hystericalness"

I would also put a .1uF cap on the pot wiper to ground!

 

[ericgibbs]

Lol, I changed my pic as a joke, and now I can't find my original pic. hehe

hi Mike,
Thats more like it.!

Hi GonzoEng.
I also find that 1n0 or 10n0 across pins 2 and 3 helps with 'de-dithering'.

 

[stuhagen]

Thanks for the inputs. I have incorporated the above mentioned changes. Let me know if I missed something. As Eric mentioned, we needed the 22K for increasing the deadband. So in effect, by going back to the 1MO, but installing the 10K gets me the same deadband results?

So it looks like the whole issue is smoothing out things with the 2 Caps. Maybe I can try the 2 Caps first, then mess with the 10K to the wiper. If the combination of the 1MO and 10K to wiper removes this Hyster.....then I may do that as well.

Stu

 

[GonzoEngineer]

We like to refer to this as "poke and hope" engineering!

 

[kchriste]

As Eric mentioned, we needed the 22K for increasing the deadband. So in effect, by going back to the 1MO, but installing the 10K gets me the same deadband results?

I would go with a 100K to the wiper. This is because the deadband would vary too much if you just changed R3 to 22K and you would have very little deadband when the pot is at it's extremes. One thing to note is that pin 1 of the 393 only varies by 1.5V because of the loading of the 270Ω resistor and the Vbe drop of Q1. You need to account for this when figuring out the deadband.

So it looks like the whole issue is smoothing out things with the 2 Caps.

The caps will help filter out any noise on the inputs. They won't change the DC deadband though. I would have put C3 between pin 2 and ground myself.

 

[stuhagen]

I would go with a 100K to the wiper. This is because the deadband would vary too much if you just changed R3 to 22K and you would have very little deadband when the pot is at it's extremes. One thing to note is that pin 1 of the 393 only varies by 1.5V because of the loading of the 270Ω resistor and the Vbe drop of Q1. You need to account for this when figuring out the deadband..

So your saying to keep the 1MO "and" use the 100K in place of the 10K I added?
Just as a note...this circuits "range" is very minimal, and in fact the wiper barely moves off its top. In other words, there is less than 100 ohms between R2 (1K) and the wiper. The wiper will never get close to ground.

The caps will help filter out any noise on the inputs. They won't change the DC deadband though. I would have put C3 between pin 2 and ground myself.

All I am trying to do is eliminate the relay from oscilating at the transition Voltage. It only does it if the 2 voltages merge and "hang" Like I said, It most likely will never be this way, but if it does, I was hoping that adding some component that reduces this. Eric suggests the cap between 3+ and 2-, and you suggest the cap from pin 2- to ground. Also, once the wiper is set, it won't change really, So the only thing that goes up and down in Voltage on a constant level is what comes in to Pin 2. And even that usuage range is small, like 4.2v to 4.9v that's the "control" range. Anything below is useless, and the circuit really goes out of its range at 4.96v


Stu

 

[kchriste]

So your saying to keep the 1MO "and" use the 100K in place of the 10K I added?
Just as a note...this circuits "range" is very minimal, and in fact the wiper barely moves off its top. In other words, there is less than 100 ohms between R2 (1K) and the wiper. The wiper will never get close to ground.

All the more reason to use the 100K. The impedance looking into the zener will be quite low and since you'll only have 100Ω between it and the wiper, the hysteresis would be minimal even with a 22K feedback resistor.

Eric suggests the cap between 3+ and 2-, and you suggest the cap from pin 2- to ground.

I prefer it to ground because then it doesn't interfere with the hysteresis window. When the comparator switches, the voltage at pin 3+ is supposed to shift so that the output stays latched until pin 2- moves to the new set point. The capacitor causes pin 2- to momentarily move in step with pin 3+ which defeats the hysteresis at the point in time that you really want it: at the transition point.
If there was no hysteresis, then a cap between pins 2&3 makes sense. With hysteresis, it makes no sense to me.

 

[Ubergeek63]

For your application R3, 1M0 is to high a value, you require about 22K.

If you need to use a 1M0 hysteresis value resistor, you require a 10K resistor from pin3 to the wiper of RV1

Inconsistant-that will only work at the high end of the 5K pot. He might need more feedback but definitely needs a 50K after the pot and should have filtering on the input if it is remote.

Dan

 

[stuhagen]

Inconsistant-that will only work at the high end of the 5K pot. He might need more feedback but definitely needs a 50K after the pot and should have filtering on the input if it is remote.

Dan

As mentioned above.......I am "at" the high end of the pot, with maybe only 100 ohms away. So my "range" is like 0-100ohms off the top of RV1.
So I "was" going to add in the 100K from the wiper to Pin 3, and try the 1MO or the 22K. Oh....and a .1uf cap from the RV1 wiper to ground, and ("before" the 100k.)

Stu

 

[ericgibbs]

Inconsistant-that will only work at the high end of the 5K pot. He might need more feedback but definitely needs a 50K after the pot and should have filtering on the input if it is remote.

Dan

hi Dan,
To set the record straight, the circuit that has been posted on this thread, is NOT the final circuit which I provided.

 

[stuhagen]

hi Dan,
To set the record straight, the circuit that has been posted on this thread, is NOT the final circuit which I provided.

For the record, this "is" Eric's final circuit. What I "may" include now is a 100k resistor from the wiper (R4) to Pin 3+, plus maybe a .1uf from the wiper to ground if all is in agreement.

To re-iterate, the usefull range for this circuit's operational range is from 4.2v min to 4.9v maximum. This in turn puts the wiper near the top end at all times. Where 4.9v at the maximum end puts the wiper's resistance to Pin 3 at almost "0" ohms.

Stu

 

[ericgibbs]

This is how I would see the final circuit, which includes the 1K dashboard pot.

Note the dash pot goes in the bottom end of the main 10K/5K setpot.

Turn the dashpot to its centre of rotation, trim the setpot to give the required voltage operating point. The dashpot should give some 'user' control.

 

[stuhagen]

Eric, good idea on the dash pot. 2 questions though, 1) should the dash pot be a multi turn precision? and 2) since my range is at the "other end" of the pot up top, shouldnt this dash pot be up there instead? Not toward the ground side. I am between 0 ohms and 100 ohms off R2 for my range. I am having a hard time seeing how I can get to near "0" ohms in this manor.

Stu

Stu