Does this need a Snubber?

[ericgibbs]

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

hi stu,
The dash post could be a single turn general purpose pot.
If the cable run to the dashpot is long it would be an idea to use a length of screened cable.
If you are Vsensing about 4.5V and a 6V2 ref diode is being used, I cannot see why the multiturn setpot is at one end.?
If you need to have the setpot higher, putting the dashpot in the top will make it worse.
For example:
say the Vsetpot is 5K and there is a 5K dashpot in the bottom.
The junction of the two pots will Vset be at 3.1V [half Vref], if you now trim the set pot so the wiper is halfway then the Vset will be at 4.65V, which is in the switching voltage region you require.

As you are adding the dashpot, like I said it would be advisable you add that 10K fixed resistor from the Vsetpot wiper to the LM393 input.

Its important to note, that the higher the resistive value of the dashpot the more coarse will be the 'user' adjustment
even though its greater range.

By experimentation you should be able to determine the optimum resistive values of the Vsetpot and the dashpot.


As a check measure the voltage at the top of the Vsetpot, it should be 6V2.

 

[Ubergeek63]

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

then you should be using a 200 ohm pot and a 4.8K in the place of the 5K pot.

 

[stuhagen]

Eric,

Thanks again for your input, always appreciated.

Just to clarify, my specific use for this circuit was needing the higher end of the Vsrc. But there were others interested in a wider Vsrc range, maybe even as low as 2.5v. So we decided to change the 5V1 to a 6V2. At this time, my test circuit board still is using the 5V1. I am happy in that it works great with the high precision 5K 10turn pot because I can turn it 2 full 360 degree turns for a V swing of 4.2 to 4.9V. This gives me a more refined setting for such a narrow V swing. Basically, it does put the wiper near the top at 4.9v.

But others would need more broader range. So for this specific application, having this dash pot makes good sense. Mostly because these precision pots are rather expensive. So I can set the 5K pot near the top, and then use a 1K or less pot to do the "finer" adjustment. This would be more of an "option" whereas others wouldnt need this. I will play around with a combination of pots to see what works the best.

So to answer the original question, the .001cap and the 10K resistor may solve the "dithering" issue? I have .01 caps laying around, but no .001. I wonder if the .01 would work as well?

Stu

 

[ericgibbs]

Eric,

So to answer the original question, the .001cap and the 10K resistor may solve the "dithering" issue? I have .01 caps laying around, but no .001. I wonder if the .01 would work as well?

Stu

hi,
The 0.01uF is a bit bigger than I would normally use, it could cause instabilty.

I would suggest you try it on the bench, with and without the 0.01uF, it will not harm the LM393.

 

[stuhagen]

Just wanted to post that the new circuit is made and tested. The dead band and the drop off is very nice now. Also, the chatter seems to be gone as well. The problem I am now having with this new version is that I can only go to 3.6V at the top end. Whereas before I could go a bit above 5V. So tonight I may try going back to a 5K Pot instead of the 10K one I have in there now. This is the first time I have used this 10K pot. Also, the only other thing that is different is the 6V2 over the 5V1. Not sure why that would do this. Unless the 10K resistor in the wiper is causing this?
So I will try changing the 10K pot, then back to the 5V1. If that still doesent get me up to 5V, then I may remove the 10K resistor. So any thoughts would be appreciated. Also, I did try the 1K dash pot in and out to see if that helped.

Stu

 

[ericgibbs]

hi stu,
With a 10K resistor from the pot to the LM393 and if you still have a 22K in the feedback, the 22K is pulling the Vset down.

As a test lift one end of the 22K out and check the Vset.

If required you may have to increase the 22K to say 150K when you have the 10K to the pot.

OK.?

 

[stuhagen]

hi stu,
With a 10K resistor from the pot to the LM393 and if you still have a 22K in the feedback, the 22K is pulling the Vset down.

As a test lift one end of the 22K out and check the Vset.

If required you may have to increase the 22K to say 150K when you have the 10K to the pot.

OK.?

No problem, I understand. I will first try the 5K pot. I would think that having no feedback (22K0) would not be good? Or is this just a test and a resistor would still be needed? In either case, I will try lifting it out if the 5K pot doesnt fix it. Would changing the 10K resistor in the wiper make any Vset changes?

Stu

 

[Willbe]

Some of the current that drives the relay may be going into your signal ground. You need a single point ground of zero impedance.

 

[stuhagen]

Last night I changed the 22K to 150K and the Vset now can go to 6V. The noise is gone. I think the reason is at the transition the Vset jumps up immediately by about .2v I'm guessing why the chattering is gone is because the Vset and Vsrc are moving and not lingering.
Still trying to get the dashpot to work. With a 5K or a 10K Vset pot, the only value that really works for the dash pot is 5K. I can get about 1.5V swing in this configuration. Which for a specific application, this would be OK. Problem is I would have 2 pots. Better to remove the Vset pot and replace it with a specific voltage divider array and the dashpot.

 

[ericgibbs]

Last night I changed the 22K to 150K and the Vset now can go to 6V. The noise is gone. I think the reason is at the transition the Vset jumps up immediately by about .2v I'm guessing why the chattering is gone is because the Vset and Vsrc are moving and not lingering.
Still trying to get the dashpot to work. With a 5K or a 10K Vset pot, the only value that really works for the dash pot is 5K. I can get about 1.5V swing in this configuration. Which for a specific application, this would be OK. Problem is I would have 2 pots. Better to remove the Vset pot and replace it with a specific voltage divider array and the dashpot.

Fully setup the Vset and dashpot for optimum operation when using a 5K Vset pot, remove the Vset pot and measure with an ohm meter from the wiper to each end of the pot, the measurements should give the two divider resistance values.

 

[stuhagen]

By reversing the usage of the *dashpot* to Vset pots, I found that the Vset pot with a value of 2.2K and replacing the dashpot with a fixed resistor value of 4.7K produced the best Vset range. My total voltage range with the Vset pot at 2.2K is from 3.5v to 5.2v. Since the Vset pot is not required to swing all the way from 0v to 5.v, the linear sweep on the pot is much more refined. I think this will eliminate the need for a high priced 10turn Bournes. I also changed the resistor from pin 3 to pin 1 to 47K. So I will use the Vset 2.2K pot as the *new* dashpot, and the fixed 4.7K to ground.

I have a question about the overall switching. When the Vsrc is below the Vset is the relay supposed to be "on" (energized)? So when the Vsrc goes above the Vset, the relay turns off. Is this correct? I was getting confused on which pin taps to use on the relay whether the N/O or the N/C for my solenoid. I think this is the case because that is why you inverted the LM 2&3

Stu

 

[ericgibbs]

I have a question about the overall switching. When the Vsrc is below the Vset is the relay supposed to be "on" (energized)? So when the Vsrc goes above the Vset, the relay turns off. Is this correct? I was getting confused on which pin taps to use on the relay whether the N/O or the N/C for my solenoid. I think this is the case because that is why you inverted the LM 2&3

Stu

IIRC you wanted the relay to de-energise when Vsrc exceeded Vset.

The original circuit I posted, the relay energised when Vsrc exceeded Vset, thats why we reversed pins 2 and 3.

 

[stuhagen]

IIRC you wanted the relay to de-energise when Vsrc exceeded Vset.

The original circuit I posted, the relay energised when Vsrc exceeded Vset, thats why we reversed pins 2 and 3.

Ya, that is what I thought. But in reality, it really doesnt matter right? I mean whichever way you do the pinout of the LM, it could be reveresed by using the N/O or N/C taps on the relay. If it went back to the original circuit, then I would be using the N/C "off" state on the relay (as the 12v feed). As it stands now, I am using the N/O side of the relay because it is switched that way. Does this make sense?

Stu

 

[ericgibbs]

Ya, that is what I thought. But in reality, it really doesnt matter right? I mean whichever way you do the pinout of the LM, it could be reveresed by using the N/O or N/C taps on the relay. If it went back to the original circuit, then I would be using the N/C "off" state on the relay (as the 12v feed). As it stands now, I am using the N/O side of the relay because it is switched that way. Does this make sense?

Stu

As you are using a single pole change over relay, providing you use the correct pair of contacts you can get:
Open when energised or closed when energised,,, or vice versa.

 

[stuhagen]

As you are using a single pole change over relay, providing you use the correct pair of contacts you can get:
Open when energised or closed when energised,,, or vice versa.

I agree......but........would the same senerio be if the pins on the LM were reversed? Or is this important to do it a certain way based on my needs.
In other words, if the LM pins were reveresed, then you could switch the relay contacts to suit.

Stu