Hypersensing with THCAD - better way to do ohmic sensing

More
18 Apr 2020 07:26 #164425 by phillc54

On a dry table it takes about 0.04mm of movement on contact to gain a full 24 volt signal on my testing so any lag from proposed overtravel is not significant at all.

I wasn't thinking about lag, that is irrelevant. I was think about the distance of the overrun. You mentioned 3 cycles of trending up and 3 cycles of trending down. Then I got that wrong and said 6 cycles but it is only an overrun of three cycles but that IS overrun and folk will need to take that into account and make sure their float travel will allow for it.


SO we are going to get more efficient at this. Those 30 cycles are currently inherent in most relay configs used by Plasmac that are in use.

I don't know, I am getting confused here.

How many servo cycles does it take plasmac to back up say .03mm? (say from 24 volt back to 5 volt to turn off a relay?)

But you said earlier that they turn on at 8V so it only has to go from 8V to 5V not 24V to 5V

Please Log in or Create an account to join the conversation.

More
18 Apr 2020 07:58 #164426 by rodw
but the overrun travel while stopping will likely take it up to 24 volts so its got a long way to get back to 5 volts .

Lets just say for example we set the threshold at 22 volts and the overrun takes that through to 24 volts as we have full contact and the floating head has taken up that travel. So then you back off and as contact breaks (float switch travel taken up fully), the voltage will start to fall from 24 volts. If we get back to <22 volts AND we have had a downward trend leading up to that servo cycle, then ohmic2.comp will turn off the probe signal. So the state changes on the first servo cycle we are < 22 volts on so there will be no overrun at all as we have basically eliminated switch hysteresis. I like the idea, but I have to be in the mood to code it.

Please Log in or Create an account to join the conversation.

More
18 Apr 2020 08:07 #164427 by phillc54
Now you have totally lost me...

There is no overrun on probe up (well there is 0.001mm)

I have no concern with overrun and voltage, only the physical ability of the machine to be able to stop before something gets mashed which will happen if the overrun exceeds the maximum float movement.

You may save 10 ~ 20 milliseconds per probe but who knows, at the moment it is all guesswork.

Please Log in or Create an account to join the conversation.

More
18 Apr 2020 10:25 #164437 by rodw

Now you have totally lost me...

There is no overrun on probe up (well there is 0.001mm)

I have no concern with overrun and voltage, only the physical ability of the machine to be able to stop before something gets mashed which will happen if the overrun exceeds the maximum float movement.

You may save 10 ~ 20 milliseconds per probe but who knows, at the moment it is all guesswork.


But there is overrun that you don't consider. Look at the specs for an Opto22 input module which I think is a popular choice mentioned several times on the forum documents.opto22.com/0451_Std_Dig_DC_Inputs_data_sheet.pdf
It has a turn on and turn off delay of 20 milliseconds. So for 20ms the system is moving and Plasmac does not know.....

Their output modules turn on and off < 1ms but they turn on at 18 volts and off at 1 volt so they are not going to work with water around as they will never turn off.

Yes right now it might be guesswork but its informed guesswork given the amount of time I've spent studying halshow and halscope to get this far.

My concern is to improve the ohmic algorithm so it works reliably. We have the advantage that we can control the on and off voltage, not just accept what a relay manufacturer builds into his design..

Please Log in or Create an account to join the conversation.

More
18 Apr 2020 10:41 #164438 by phillc54
Yes, accurate and consistent probing is what we should be looking for, who cares if you save some milliseconds per probe. I would think that comparators would also be with a look. Do you know what John uses, he seems to have reasonably consistent results.

Please Log in or Create an account to join the conversation.

More
18 Apr 2020 11:41 #164443 by tommylight
Well i think i was silent long enough about this...
Find a 1KOhm resistor at 1W ( or 2x 2.2K @0.5W, or 4x 4.7K @0.25W, in parallel ), and attach it to the THCAD5 input, that should make all your water problems go away. It is to sensitive and that is very good, but not when water is involved. That will make it require more current passing through the system to trigger ( or change measured values in this case )
470Ohm/2W should do even better, just a bit over 5 mA going through it.

Please Log in or Create an account to join the conversation.

More
18 Apr 2020 12:31 #164447 by rodw

Yes, accurate and consistent probing is what we should be looking for, who cares if you save some milliseconds per probe. I would think that comparators would also be with a look. Do you know what John uses, he seems to have reasonably consistent results.


John uses a relay circuit using opto22's. He's being doing it for along time and has a process down pat but I know triggering pre-flow is part of that.

Well i think i was silent long enough about this...
Find a 1KOhm resistor at 1W ( or 2x 2.2K @0.5W, or 4x 4.7K @0.25W, in parallel ), and attach it to the THCAD5 input, that should make all your water problems go away. It is to sensitive and that is very good, but not when water is involved. That will make it require more current passing through the system to trigger ( or change measured values in this case )
470Ohm/2W should do even better, just a bit over 5 mA going through it.


I don't think you can do that as it will change the scaling so full scale won't be 24 volts anymore. PCW did wonder if 24v was enough. Its still all a journey of discovery. The stuff we now know and have documented is pretty amazing. Stuff thats not publicly available anywhere.

Please Log in or Create an account to join the conversation.

More
18 Apr 2020 12:40 - 18 Apr 2020 12:41 #164450 by tommylight
Yes, you can do that and it will not move the scaling, 24V will still be 24V, just the current to get there will be higher, hence lowering the sensitivity to water.
Humour me, give it a try.
I have 4 THCAD's wandering the globe somewhere collecting viruses, still did not get them, so i can not try that yet.
I did say, parallel to the input, not in series, did i ? Guess not.
Last edit: 18 Apr 2020 12:41 by tommylight. Reason: More info

Please Log in or Create an account to join the conversation.

More
18 Apr 2020 13:45 #164456 by Uthayne
Tommy, is this what you're saying?
Attachments:

Please Log in or Create an account to join the conversation.

More
18 Apr 2020 14:12 #164460 by tommylight
Yes, but if you have that 390K resistor, connect it before the resistor.
Attachments:

Please Log in or Create an account to join the conversation.

Moderators: snowgoer540
Time to create page: 0.223 seconds
Powered by Kunena Forum