Vacuum pressure

[Al Yard]

Hi all,

What is a reasonable vacuum pressure to design to?   Is -50kPa a reasonable vacuum for picking parts?  Or should I plan on a greater vacuum ... or will it work reliably on less?

Thanks,

Al

[Michael Anton]

I guess if you know the mass of your heaviest part, and the nozzle diameter you would use with that part, then you should be able to calculate the minimum vacuum required.

[ma...@makr.zone]

I guess if you know the mass of your heaviest part, and the nozzle diameter you would use with that part, then you should be able to calculate the minimum vacuum required.

Yeah but don't forget to account for the maximum acceleration of your head when both axes speed up in a diagonal multiply by the distance from nozzle tip to your part's centre of mass (leverage over nozzle diameter). This will exert a lateral force on the part trying to pry it away from the nozzle, breaking the vacuum. Then there are Z+ acceleration Z- deceleration (changing the "g force" on the mass of the part), add vibration, jerk, surface properties, leakage variation, air friction ...

Just kidding :-)

No offence, but I think its much more complicated and calls for empirical values.

Nobody out there with solid practical experience?

I haven't placed large parts yet. But with the small parts/small nozzles, I found a value way above -50kPa to be sufficient. If nobody with more experience offers values I can go measure it precisely. Too high values tended to keep the part on the nozzle on discards (probably through valve leakage?).

This is on a Liteplacer mod with the standard small pump and valve. I have the MPXV5050VC6T1 sensor that goes to -50kPa and I am a good way above its range.

_Mark

[ma...@makr.zone]

This is on a Liteplacer mod with the standard small pump and valve. I have the MPXV5050VC6T1 sensor that goes to -50kPa and I am a good way above its range.

Well I should be more precise: I've got the pump on a hysteresis on a metal bottle to get a good "reserve of vacuum". The pump attached directly could go below the -50kPa limit of the sensor when the valve is closed. But the vacuum immediately "collapses" when the valve opens, even on a very good pick.

_Mark

[Cri S]

If you have a bigger reservoir you should limit the airflow.
Football needles works for this purposes.
On compressor you connect the tools after the pressure regulator, not directly.
With vacuum reservoir it should be similar.
This only works if you are using filters for the vacuum line, and for
instance, filters
works only if you blow out with pressure for separating the component
and nozzle and
at same time clean the filter.
If you don't blow out and have filter installed, instead of using big
reservoir, use longer vacuum tube and bigger inner diameter tube for
some sections
in order to overcome or smooth the pressure fluctuation of the pump.

2018-05-09 8:39 GMT+02:00, ma...@makr.zone <ma...@makr.zone>:
>
>>
>> This is on a Liteplacer mod with the standard small pump and valve

>> <https://www.liteplacer.com/finishing-step-14-vacuum-routing/>. I have
>> the MPXV5050VC6T1
>> <https://www.digikey.ch/product-detail/de/nxp-usa-inc/MPXV5050VC6T1/MPXV5050VC6T1CT-ND/951857>

>>
>> sensor that goes to -50kPa and I am a good way above its range.
>>
>
> Well I should be more precise: I've got the pump on a hysteresis on a metal
>
> bottle to get a good "reserve of vacuum". The pump attached directly could
> go below the -50kPa limit of the sensor when the valve is closed. But the
> vacuum immediately "collapses" when the valve opens, even on a very good
> pick.
>
> _Mark
>

> --
> You received this message because you are subscribed to the Google Groups
> "OpenPnP" group.
> To unsubscribe from this group and stop receiving emails from it, send an
> email to openpnp+u...@googlegroups.com.
> To post to this group, send email to ope...@googlegroups.com.
> To view this discussion on the web visit
> https://groups.google.com/d/msgid/openpnp/cb928f95-7dd1-4dce-b2ec-04db03e3b5ea%40googlegroups.com.
> For more options, visit https://groups.google.com/d/optout.
>

[Marek T.]

Practical information is that in my Philips/Yamaha machine it's designed to create 80-100kPa vacuum (converted from pressure). My vacuum sensors are -100kPa and shows level near 80% of their max range.
It doesn't say that 50kPa is not enough of course. Just only info how it was done by some old but professional SMT machine manufacturer.

[Al Yard]

O/K .... thanks.  That gives me a bit of a ball-park figure to work with.  This all started when I went to order a vacuum sensor, and realized I had no idea what range of vacuum pressure I was working with. Those little plastic vacuum pumps must pull a greater vacuum than I imagined.    I'm living at elevation 1000m, and the air pressure is only about 85 kPa on a low pressure day, so I'll plan on having about a -70 to -75 kPa vacuum at best.

Al

[Al Yard]

Well .... so much for me trying to pretend I'm an engineer and calculate what I needed for my vacuum system.  ;)

I was thinking about my little plastic vacuum pump and realized I didn't have a spare and it was likely going to give up one day.  I was looking for a cheap spare pump that would fit the existing bracket, and the pumps I found seemed to be only rated to pull about -50 kPa.  I wasn't sure if that would be enough and I didn't have a vacuum gauge to see what the pump was pulling.  Then I had a brain-wave! ... there are a bunch of aircraft parked around here, so I "borrowed" a Manifold Pressure gauge from one of them.  It measures in inHg and is in absolute pressure, but that is easy enough to convert.

I found my little pump could pull about -40kPa.  But when I hooked up the nozzle tip, etc, there were enough tiny little leaks (I think mostly in the nozzle coupling and the hose swivel), that the best I could get was -4 kPa with a part on.  Although I occasionally get a mis-pick where the part didn't pick up, I haven't had any problems with the parts falling off the nozzle .... even though I only have -4 kPa.  This was a surprise to me and tells me what vacuum pressure the pump can pull is of little importance, and what is more important, does it have enough flow to over-come the leaks in the system. 

So my project now is to try and reduce some of the leaks (I'm sure there will always be some leaks) and if I can get the pressure up to -10kPa with a part on, then it should be good and reliable.  Maybe I'll also get a pump with a little more flow too.

Al

[Eagle Media]

I have got a complete vacuum monitoring system.

[Al Yard]

O/K ... that sounds good .... can you tell me what your vacuum pressures are when you have a part on?   I'm still trying to get my head around just what is a desirable pressure.  I've been surprised with how little it takes to keep a part on the nozzle ... but I'm likely running less than ideal vacuum.

Thanks,

Al

[Marek T.]

-4kPa sounds like measure mistake, can't imagine what you can pick up with the vacuum like that.
If you have Juki nozzles with balls holders you can't avoid leaks, it's issue of this construction.

[Eagle Media]

AFAIR somwhere between 40-60. There was a page somebody made with all those numbers.

[Mike Menci]

how about using this sensors ?  https://www.ebay.com/itm/161308557994?rmvSB=true  for justifying pick up - rejecting? and visual vacuum pressure indication ? Can it be used with Smoothie?  where to connect on smoothie if usable!

Mike

[Mike Menci]

Parker Data sheets ; http://www.parker.com/parkerimages/pneumatic/cat/english/0801-3%20Sensors.pdf

[Marek T.]

There is some optional "Optional Analog Output 1 to 5VDC" which you probably need to connect to the Smoothie's termistor inputs

But the price $14 seems be little suspectly low for device like this...

[Mike Menci]

Those are used - still working - better this than a MPX sensor chip will cost me more! Loto !! 

Mike

[Mike Menci]

Good Used Working Condition. Removed from working service location. May have some cosmetic damage....

[Marek T.]

Clear, haven't paid attention that it's used item. Now I see that new are closer to $100.

If it's important for you to see the value - it seems be nice solution. If not - you can try also something like this, not expensive and new:

https://www.aliexpress.com/item/1pcs-XGZP6847-100-700KPa-pressure-sensor-transmitter-10KPA-20kpa-40KPa-100KPA-200KPa-500KPA-700KPa/32847133587.html?ws_ab_test=searchweb0_0,searchweb201602_1_10152_10151_10065_10068_10344_10342_10343_10340_10341_10696_10084_10083_10618_10304_10307_10820_10821_10301_10059_100031_10103_10624_10623_10622_10621_10620,searchweb201603_25,ppcSwitch_2&algo_expid=466bd8f5-3eaf-43f3-bfb6-55feb898a601-0&algo_pvid=466bd8f5-3eaf-43f3-bfb6-55feb898a601&transAbTest=ae803_1&priceBeautifyAB=0

[Cri S]

Smoothie board don't support 5V, only 3V.
Normally you convert it to 4-20mA and then from that to 0.66 to 3.3V .
The voltage divider is annother option, cheaper but less immune about noise.

> --
> You received this message because you are subscribed to the Google Groups
> "OpenPnP" group.
> To unsubscribe from this group and stop receiving emails from it, send an
> email to openpnp+u...@googlegroups.com.
> To post to this group, send email to ope...@googlegroups.com.
> To view this discussion on the web visit

> https://groups.google.com/d/msgid/openpnp/c4153438-6b7b-457c-9f8a-f96a9d2833fd%40googlegroups.com.

[Marek T.]

But simple MPXV is also operating 0-5V and requires some voltage divider the same. Plus termistor inputs of Smoothie are on analog A-GND. I have divider made with transoptors powered from 3V of Smoothie, at occasion they separate GND (my MPXV are powered from 24V accessible on my head and having still other GND-24) and divide signal levels. Checked with oscilloscope that the noises level is the same and rising/falling edges are also the same on both sides of transoptors.

[Mike Menci]

Hello 

Could this  DC-DC Buck Converter 5-23V 12V to 3V 3.3V 5V 9V 12V 3A Step Down    be used ? 5V to 3V DC  https://www.ebay.com/itm/DC-DC-Buck-Converter-5-23V-12V-to-3V-3-3V-5V-9V-12V-3A-Step-Down-Power-Regulator/282994299384?hash=item41e3c651f8:g:7zYAAOSwHola6oH8

Mike

[Marek T.]

You need linear divider for measured signals. Something like:
0V on MPXV = 0V on Smoothie
2,5V on MPXV = 1,5V on Smoothie
5V on MPXV = 3V on Smoothie
Smoothie is not measuring from 0V but near 0,6V. MPX is giving 0-5V. But vacuum levels corresponding to values 0-0,6V are very small (like -10kPa or so so) and not very important.
If you ignore GND to AGND separation (not sure if reasonably), just simple two resistors divider will do the job.

[TheCunningFellow]

Use an In-Amp.

You can easily shift and scale the output to the range you are interested in.

If you have a uC with a DAC you can even change the range you are interested in on a per nozzle basis.

[ma...@makr.zone]

Hi

I'm not so sure, simple resistor dividers would do. At least not inside the specs of the MPX sensor.

See this build log post about the solution I chose:

https://makr.zone/vacuum-sensor/192/

It works at least in dry-runs:

https://makr.zone/pick-place-machine-first-simulated-small-test-run/66/

_Mark

[Marek T.]

Hi

First I'm using MPXV6115 sensor with built-in op-amp (there is a bit more output current than in 5050) to drive low current transoptors and it works well for me.
https://www.nxp.com/docs/en/data-sheet/MPXV6115V.pdf
First, sample with dividing resistor was mainly to explain kind of required working not particular circuit to built. Not buck converter. I don't know how about noises in this case when used higher value dividing resistors to don't overload sensors outputs, never tried and tested that circuit.

Second ad8495 found as better for this solution than Epcos. But maybe it's not important and Epcos is good as well.
http://smoothieware.org/temperaturecontrol#toc5

Third, Mark, 3.3V/4k7 is 0.7mA not 7mA or you took this current from some other calculation. However 0.7 is still more than 0.1 :-).

br
Marek

[ma...@makr.zone]

Third, Mark, 3.3V/4k7 is 0.7mA not 7mA or you took this current from some other calculation. However 0.7 is still more than 0.1 :-).

Thanks for pointing this mistake out, Marek. That's the danger of documenting things only some months after the fact. :)

I corrected the post to explain the correct reasoning:

https://makr.zone/vacuum-sensor/192/

I also added LTSpice simulation.

_Mark

[Marek T.]

Nice characteristic, maybe will solder it to compare with mine :-).
In my case I've found it works faster when replace 10uF on Smoothie into 1uF. But on your scopes it looks good (at this time scale at least).

[Mike Menci]

Hello
Nice and clear explanations!
Is it possible to monitor vacuum values from serial on Open PnP from Smoothie ?
Mike

[ma...@makr.zone]

Is it possible to monitor vacuum values from serial on Open PnP from Smoothie ?

Yes, just follow the guide here:

https://github.com/openpnp/openpnp/wiki/Setup-and-Calibration:-Vacuum-Sensing

These are my G-Code snippets:

         <command head-mountable-id="A1" type="ACTUATOR_READ_COMMAND">
            <text><![CDATA[M105 ; read temperatures (vacuum)]]></text>
         </command>
         <command head-mountable-id="A1" type="ACTUATOR_READ_REGEX">
            <text><![CDATA[^ok.*V:(?<Value>-?\d+\.\d+).*]]></text>
         </command>

[ma...@makr.zone]

Just updated my doc again to cover the reading-back in OpenPNP:

https://makr.zone/vacuum-sensor/192/

_Mark

[Marek T.]

Mark your schematic suggests that thermistor inputs of the smoothie are read to the same digital GND as 5V is what is not true. Analog port of this processor have separated ground AGND which is accessible on pins destined for thermistors only. And these 10uF are connected to this AGND not to GND of processor power.

If we want to talk about noises and really proper reading: AGND should be not connected to GND or analog port should be not read in acc to digital GND. At least in theory or in agree to smoothieware recommendation.
See discussion here:
https://plus.google.com/+ReneJurack/posts/hTpjbnZPuPC
Or maybe Arthur will tell something more wise about this, he's following openpnp groups so probably see this here.

[ma...@makr.zone]

Mark your schematic suggests that thermistor inputs of the smoothie are read to the same digital GND as 5V is what is not true.  Analog port of this processor have separated ground AGND which is accessible on pins destined for thermistors only

My schematic is just a simple simulation. Don't read too much into it. Physically the sensor is grounded at the thermistor connector AGND.

However having said that I do concede that my knowledge in such matters is limited. In fact the problem of grounding the machine properly with all three aspects of safety, EMI and ESD plus USB ground addressed at the same time is the topic of one of my planned posts with many question marks remaining.

One example: For the vacuum sensor I need to power the sensor and opamps. So I added the third connector from the +5V supply. The +5V supply is bypassed against GND not AGND. If I add a bypass capacitor on my sensor side PCB, I clamp the AGND to the +5V rail. Is this relevant? Should I add an inductor bead in series, like on the Smoothies AVCC? I frankly don't know. I figured the long wire to the sensor will be inductance enough.

Also I'm not sure if the GND vs. AGND question is really that relevant here. AGND and GND are connected near the MCU (see schematic below). The voltage difference on the few millimeters of AGND trace to the connector vs. the massive GND plane are most likely irrelevant compared to what is picked up with the two metres of wire to the sensor. The 10µF capacitor will simply smooth all that out. 

Also the ADC on the LPC1768 MCU seems to have well known issues that seem to be confirmed just by looking at the values reported even with a single thermistor connected very close to the MCU and no motors running etc. I think this shortcoming is much more relevant in the end.

_Mark

[Marek T.]

Ok, understand. I'm also not too experienced about the GND-AGND problems in this case. Just was curious your oppinion and wanted to know if you made as you made accidentally or intentionally.

I had more clear situation. In place where the sensors are mounted I have accessible only 24VDC to power them (lowered to 5V of course). This 24V is totaly separated from other supplies and using separated GND not possible to connect it to other grounds. Leading up to the sensors 5 or 3V from Smoothie was risky because it would be paralel to high voltage and high current wires (150V and 5A). So the only easy possibility I've found was to build optical separation with transoptors set with their current in the part of their characteristic. Transoptors are mounted on smal board directly on the termistors port of Smoothie.

[Mike Menci]

Hello, 

have a look at enclosed:

Mike

[ma...@makr.zone]

have a look at enclosed

 

Hi Mike

I don't know what you mean specifically. I can't see how these two papers are related (?)

The current loop interface is of course more immune against EMI but I think this is like "killing a mosquito with a bazooka".

I haven't seen excessive noise on this small PNP machine. We are talking about motor currents in the 1A range. With high frequency choppers and good local bypassing. PCBs with massive ground planes. Tiny parts and traces. All solid state switching. We have high inductance motor coils. Nice slow sinusoidal currents resulting, overlapping in two phases. Fast modern switching supplies that iron out all load variations for breakfast. Connected and grounded with short thick supply wires. I see no drooping or bouncing whatsoever on a scope. All wires are shielded or twisted with a ground wire.

I guess much of that ground bouncing "scare" might come from old times' electronics and/or large machines with massive unregulated AC or sparking brushed motors, sparking relays etc.. It might also include a lot of "reverse Murphy's law" in anticipation of less than ideal applications and wiring.

But then: I'm not the expert :) This is just observation plus conjecture of what might be the reason for such caution.

_Mark

[Marek T.]

I have relatively large machine (retrofited Philips/Yamaha) with massive unregulated but DC and sparking brushed motors 150V and 5-7A. And can't see some huge problems with noises. But the fact the machine is really reasonable smart designed with no money saving.

[Mike Menci]

This Parker Vacuum indicators work - but can not be used for vacuum pickup sensing and connected to T1-T4 on smoothie ... :-((

[Mike Menci]

Analog output  is optional  :-((

[Marek T.]

Is it problem or high cost to order it in optional version with analog ouputs?

[Mike Menci]

Well this was a cheep option - new cost over 100 usd per pc...

[Marek T.]

Uuuu, it pains :-(.

Maybe try open it and get access to some analog level signal. There is for sure some quite typical sensor with analog output and probably amplifier before digital A/D with display. Speculation but who knows.

[Mike Menci]

Well - I have two MPXV6115V ready to add to the heads... 

mike

[Mike Menci]

Marek /Mark

can you assist me on this one- 

You say 3V on smoothie - but I have 5V on smoothie trough this LCD adaptor PCB with 5V 1000mA regulator http://smoothieware.org/rrdglcdadapter

 If I ignore GND to AGND separation - I could connect my MPXV6115V to this 5V + GND and the readings would be OK from sensor ? Correct?

Thanks

Mike

[Mike Menci]

[Marek T.]

Hi Mike,

No. On the output of the sensor you have 0-5V range but Smoothie's thermistors inputs are 0-3V.

You need at least add resistors divider to protect Smothies inputs against the overload.

Try from the sensor output serial some 5k1 and 10k behind it to Gnd of Smoothie. Resistors' common point to the Smoothie input. I don't tell it's best solution but you won't burn Smoothies inputs at least.

br

Marek

[jdlv]

Hi,

to add a bit of difficulties, smoothie's inputs have a 4.7kΩ pullup to
3v3 and the MPXV6115V can only output a very limited current, 100µA if
I'm correct.
To overcome this I remove the pullup array resistor, change capacitors
from 10µF to 100nF and add a voltage divider made of 12kΩ and 18kΩ.
To me it was the simplest way but it require to hack the smoothie...

Joël

[Mark]

Hi Mike

Assuming you want to avoid building this...
https://makr.zone/vacuum-sensor/192/
... I'd recommend using an alternative ADC pin on the smoothie instead of hacking it.

I assume Smoothieware does allow you to configure any pin as a temperaturecontrol pin, though I haven't tried.

Spare ADC pins:

AD4 P1.30
AD5 P1.31

Normally used for click button and buzzer on a GLCD display.

http://smoothieware.org/lpc1769-pin-usage#lpc176x-adc-channels-and-pins

I assume you know how to find these pins on the smoothieboard expansion header and I assume you know how to configure the temperaturecontrol to use them.

WARNING: using these pins directly means you have no input protection etc. so be careful to connect the right pin and to use ESD precautions and proper grounding of the sensor.

I still very much recommend to build this:
https://makr.zone/vacuum-sensor/192/

It's not as difficult as it may look.

_Mark

Ref:
https://groups.google.com/d/msg/openpnp/UKI5PPMgu6M/eYXCT3ZjBgAJ

[Mike Menci]

Hello all, 

Thanks for your kind replys

Mark I would go for your - https://makr.zone/vacuum-sensor/192/  version but here it is not clear to me: 

- connector on Eagle cad sch on your web shows SL1 to smoothie thermistor connection to all  4 thermistors 

but your bellow post "Scraping together DIL through-hole, some 0805 smd parts...." is for one sensor only. 

I have two heads and 2 sensors MPXV6115V that goes down to -115kPa and its a bit confusing to me to understand above 

Would I be asking to much if I can ask to update your Eagle Sch above with dual op-amp and resistors , caps needed for twin sensors with output to Smoothie 2 thermistors only...

Thanks in advance  

Mike

[ma...@makr.zone]

On Wednesday, August 1, 2018 at 2:48:11 PM UTC+2, Mike Menci wrote:

Thanks for your kind replys

Mark I would go for your - https://makr.zone/vacuum-sensor/192/  version but here it is not clear to me: 

- connector on Eagle cad sch on your web shows SL1 to smoothie thermistor connection to all  4 thermistors 

but your bellow post "Scraping together DIL through-hole, some 0805 smd parts...." is for one sensor only.

This is a misunderstanding. The eagle schematic on my post is a detail from the Smoothieboard schematic (I updated the post to make it more clear). It's just showing the input circuitry to illustrate why you have to use the op-amp.

What you have to build is this:

I have two heads and 2 sensors MPXV6115V that goes down to -115kPa and its a bit confusing to me to understand above 

Would I be asking to much if I can ask to update your Eagle Sch above with dual op-amp and resistors , caps needed for twin sensors with output to Smoothie 2 thermistors only... 

The small sensor PCB is per sensor so you have to build two of those.

The assumption is that the sensors are at different locations on the machine and you want the signal to be amplified very close to the sensor to minimize noise. I guess one could build one PCB for two sensors, but that's really beyond my solution, sorry. It's really not that hard to build these two little PCBs.

_Mark

[Marek T.]

Other pins are also 0-3V not 0-5, and only ADC pins have 10uF capacitors.

[Mark]

>Other pins are also 0-3V not 0-5, and only ADC pins have 10uF capacitors.

You are back to the alternative ADC pins, right?

 

AD4 P1.30  
AD5 P1.31  

 

Yes, but these can then be simply driven by a high value resistor divider without having to put the solder iron to your Smoothieboard.

 

Again, for robustness I do recommend the proper solution with op-amp.

 https://makr.zone/vacuum-sensor/192/

 

_Mark

 

 

[Mike Menci]

Hello, 

So for two sensors it can be as enclosed sch?

[Mark]

Hello, 

 

So for two sensors it can be as enclosed sch?

Yep, except R2, R8 should be 47k (instead of 4.7k).

 

(and R4, R10 is probably meant to be 100k, instead of 110k)

 

_Mark

[Marek T.]

Resistors that I've mentioned was not intended to solder onto Smoothie but between sensor and Smoothie, so there is not any hacking.

Probably option with OA is best but I don't have it and it works.

Mark, what is rising time when vacuum grows from 0 to -80-110kPa?. I guess you have just valve between vacuum pump and nozzle and the vacuum pump is constantly working.

I have some problem with this, my rise time is too long as for my expectations, but on the moment still not found if the problem is my sensor circuit or ventouri injectors.

[Mark]

> > So for two sensors it can be as enclosed sch?

> Yep, except R2, R8 should be 47k (instead of 4.7k).

> (and R4, R10 is probably meant to be 100k, instead of 110k)

 

Afterthought: You can simplify this by using just one dual op-amp.

 

Leave out the 1B and 2B Opamps and connect sensor output directly to resistor divider. I only used the two stage design because I had half a dual op-amp “to waste”.

 

 

_Mark

[Mark]

>Resistors that I've mentioned was not intended to solder onto Smoothie but between sensor and Smoothie, so there is not any hacking.

 

I responded to this quote by Joël:

 

> To overcome this I remove the pullup array resistor, change capacitors  from 10µF to 100nF and add a voltage divider made of 12kΩ and 18kΩ. To me it was the simplest way but it require to hack the smoothie...

 

And we already discussed some weeks back why hacking it is needed to get a fast response.

 

_Mark

[ma...@makr.zone]

On Wednesday, August 1, 2018 at 6:39:11 PM UTC+2, Marek T. wrote:

Resistors that I've mentioned was not intended to solder onto Smoothie but between sensor and Smoothie, so there is not any hacking.

Mark, what is rising time when vacuum grows from 0 to -80-110kPa?. I guess you have just valve between vacuum pump and nozzle and the vacuum pump is constantly working.

I have some problem with this, my rise time is too long as for my expectations, but on the moment still not found if the problem is my sensor circuit or ventouri injectors.

I have a reservoir on the pump on a hysteresis. It takes a long time to raise the vacuum. Can't measure the time at the moment.

Once it has reached the setpoint level, it works quite OK. In this video you can hear the hysteresis working (~7 seconds).

https://makr.zone/pick-place-machine-first-simulated-small-test-run/66/

_Mark

[Mike Menci]

Thanks to all and you Mark 

and Good  I hope I will be programming smoothie for vacuum soon! 

PCB enclosed for others that might need it 

Thanks

Mike 

[Marek T.]

You're right, I have also decreased these capacitors...

I have situation that from moment of ventouries switching, some 50-60ms changes nothing about vacuum rising. Then when starts rising it goes quite fast (some 20ms 0->max). Total like 100ms. Must turn ventouris for non stop ON and add some serial valve to track it. You can see movie how my nozzles picks up in the "automatic pick retry" thread.

[Mike Menci]

I missed this note- Leave out the 1B and 2B Opamps and connect sensor output directly to resistor divider. I only used the two stage design because I had half a dual op-amp “to waste”.
So the Board needs to be modified accordingly.
Mike

[ma...@makr.zone]

Do you make the PCB yourself?

_Mark

[Mike Menci]

Hello, 

I could use some standard ones with holes and just add TH parts..

Rev PCB enclosed

[Mike Menci]

Hi 

I found one Op amp CA1458E - will this do the job? 

https://www.mouser.com/catalog/specsheets/intersil_fn531.pdf

Mike

[Mike Menci]

Or TL072CN   ?? 

Mike

[Mark]

>I found one Op amp CA1458E - will this do the job? 

 

I don’t know. I found no minimum supply voltage spec and all specs are given for +/-15V = 30V so I guess this rather old model won’t work on 5V.

 

_Mark

 

 

[Mark]

> Or TL072CN ??
> Mike

No, this one has it in the data sheet:

> VCC Supply voltage 6 to 36 V

https://www.mouser.ch/datasheet/2/389/tl072-957235.pdf

_Mark

[Michael Anton]

CA1458 parts are basically a dual 741 op amp.  These are a very old designs, and are only meant to be run on fairly high voltage dual supplies, and not on single supplies like most modern parts.  The common mode input voltage range is fairly restricted and doesn't come near the negative rail, making it pretty useless for running on a single supply.

[Michael Anton]

The common mode range on these only reached within 3V of the negative rail, so watch out for that.  They will have limited use on low voltage single ended supplies, with ground referenced signals.

[Mike Menci]

Thank you,
Any recommendation ?
Mike

[TheCunningFellow]

The 741 is 1/2 a century old this year.

50 years.

Go to digikey - parametric search the opamps.  Select rail to rail and min single ended supply of 5v or less.  you will get a few 1000 options better than a 741/ca1548

[Mark]

>Go to digikey - parametric search the opamps.  Select rail to rail and min single ended supply of 5v or less.  you will get a few 1000 options better than a 741/ca1548

 

Cheapest is 0.29€ at 1 piece.

https://www.digikey.ch/product-detail/de/microchip-technology/MCP6002-I-P/MCP6002-I-P-ND/500875

 

_Mark

 

[SMdude]

You want something that is rail to rail output.

Ad8029(or one of its family members should do the trick.

[ma...@makr.zone]

You want something that is rail to rail output.

...and rail-to-rail input (or at least to the GND rail). That's often marke(te)d as "RRIO"*.

The Ad8029 is RRIO, but at 125MHz bandwidth it is even more overkill than my 10MHz MCP6022 :-).

The Ad8029 is also shockingly 7 times more expensive than the cheapest MCP6002-I/P ... still only 1.94€ ... ¦-D

_Mark

(* Note that Digikey does not always mark them correctly as "RRIO", the MCP6022 is not marked as RRIO but the data sheet says so)

[SMdude]

I have a heap of 8029 in my junk box!

Been a while since I looked at the pressure sensors, but yes you are right, they do need a RIRO amp.

Another good alternative would be a current sense amp as these normally go beyond the rail.

[Michael Anton]

Yeah, rail to rail in and out are good parts to standardize on.  Other than that, decide what input offset you can tolerate (though you could mostly calibrate it out if you wanted to), and you should be able to find plenty of good parts.  As suggested MCP6022 would probably work well.  If you want something even less expensive, I'm sure the TLV9002IDGKT would work fine.  It has a GBW of 1MHz, but since these are used as buffer amps, and have a gain of 1, this is still way more bandwidth than required.

[Vinay Dand]

Consider LMV324, though it is Quad op-amp, 'V' in LMV324 is for Rail-to-rail output.

This is a low cost fit for application that works in 5V single supply range. 

[Mike Menci]

SMT - PCB available - I have some leftover if someone wants - Made in Europe (not China)  https://aisler.net/mike_58/my-project-repository/mpxv-vacuum-sensing-smt/sharing 

Mike

[Mike Menci]

[Mike Menci]

Hello 

So Sensors installed but setting up OpenPnp problems: 

I do not find under Gcode - Settings - any Vacuum request or Vacuum report Regex ,... 

Where are this coming from ?? - Video from Jason does not work for me ??

Please help?

[Jason von Nieda]

Config system for vacuum changed a little while back, see: https://github.com/openpnp/openpnp/wiki/Setup-and-Calibration%3A-Vacuum-Sensing

Jason

--
You received this message because you are subscribed to the Google Groups "OpenPnP" group.
To unsubscribe from this group and stop receiving emails from it, send an email to openpnp+u...@googlegroups.com.
To post to this group, send email to ope...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/openpnp/c0baf4ff-24cc-4d7c-b932-9249bcaba358%40googlegroups.com.
For more options, visit https://groups.google.com/d/optout.

[Mike Menci]

Thank you Jason,
Would help others to update instructions - link there to video is wrong one - and lead to proper one.

Mike

[Jason von Nieda]

Hi Mike, what page are you seeing the wrong link on?

Jason

--
You received this message because you are subscribed to the Google Groups "OpenPnP" group.
To unsubscribe from this group and stop receiving emails from it, send an email to openpnp+u...@googlegroups.com.
To post to this group, send email to ope...@googlegroups.com.

To view this discussion on the web visit https://groups.google.com/d/msgid/openpnp/d262064c-0e45-469a-af8e-15b7544fe9c1%40googlegroups.com.

[Mike Menci]

Jason 

the first video link on this page _ https://github.com/openpnp/openpnp/wiki/Setup-and-Calibration:-Vacuum-Sensing 

There is a video that shows how to migrate settings from the old system to this system at https://www.youtube.com/watch?v=FsZ5dy7n1Ag.

But this is the same link as yours - that means I am having- looking at new -(not old) configuration, 

- I do not have  & I do not find under Gcode - Settings - any 

- Vacuum request or 

- Vacuum report Regex ,... 

Where do I get this or where are this coming from ??

Mike

[Mike Menci]

Jason 

I think the problem is old system to new system - I do not have "old" system installed - I am not migrating from old to new system therefore this commands in settings are not there ? 

Mike

[Jason von Nieda]

Hi Mike,

The video is intended to serve as a guide for migrating from the old to new settings, but it also serves to show how to setup the new settings. The only difference is that you don't copy your values from the old style, you just put them in the new.

The instructions at https://github.com/openpnp/openpnp/wiki/Setup-and-Calibration:-Vacuum-Sensing (below the video link) are accurate for the current software, and mirror the video aside from copying your values from the old way.

So, to be clear, the old way was with the VACUUM_REQUEST stuff, and the new way is just with an Actuator. If you follow the instructions on that page it will guide you through setting up an Actuator, and that is the correct method. In the video you can just ignore the part at the beginning about VACUUM_REQUEST.

Jason

To view this discussion on the web visit https://groups.google.com/d/msgid/openpnp/806afd4f-8afa-4892-b287-a538fb359d93%40googlegroups.com.

[Mike Menci]

Hello Jason, 

I followed (I think ?) the second part of the video and instructions and I got to the testing stage but I am having an error on Java - timeout reading from sensor: 

-see the screenshots enclosed 

Any tips where I made a mistake ?

Thanks

Mike

[Marek T.]

Why in your log there is no record luke (COM3 vacuum_1.......)?
Have you configured properly what the string is needed to be sent to controller to read the sensor and proper regex for it?

[Mike Menci]

I think so! 

See enclosed; 

[Marek T.]

So this "vacuum_1" should be visible as directed to comport but is not visible in log.
Where do you send this "vacuum_1", to Smoothieboard to read the inputs? Do you have "vacuum_1" defined in Smoothieboard as serial code to read an input?

[Mike Menci]

Hello Marek, 

Smoothie is configured ;  on pin 0.23   for N1   and pin 0.24 for N2 

temperature_control.nozzle.enable true #
temperature_control.nozzle.thermistor_pin 0.23 #
temperature_control.nozzle.heater_pin nc #
temperature_control.nozzle.designator V #
temperature_control.nozzle.min_temp -10 #
temperature_control.nozzle.max_temp 300 #
temperature_control.nozzle.thermistor EPCOS100K # see http://smoothieware.org/temperaturecontrol#toc5
temperature_control.nozzle.runaway_heating_timeout 0 # it seems to crash
temperature_control.nozzle.runaway_range 0 # a rapid delta in temp triggers a halt - 0 disables it
temperature_control.nozzle.readings_per_second 320 # smoothieware takes the median out of the last 32 measurements, so read frequently

[Mike Menci]

Marek I used this page for my settings of Smoothie; 

https://makr.zone/vacuum-sensor/192/#comment-8

[Marek T.]

I'll send you tomorrow my Smoothieboard config and machine that works.

[Mike Menci]

Thanks Marek for your support!

Mike

[Marek T.]

Pleasure Mike.

See attached machine and vac settings for Smoothie. You can put this file to the SD card as separate file but add to the main config file an extra line containing just:

include vac

Run your openpnp with my machine and see names of actuators, their gcodes in driver and regexes.

I use ad8495 thermistor form not Epcos, without any histeresis or other that used by Mark. Maybe Mark's thermistor choose is better as Marks looks like really smart guy :-), but my one works on my machine really good.

Use them for beginning then later you can change thermistor type and its definitions if you feel it's needed.

[Marek T.]

Find lines in my machine containing strings "middle="6.0" and delete this declaration.

Probably your jar will not start when see this.

W dniu niedziela, 19 sierpnia 2018 23:53:53 UTC+2 użytkownik Mike Menci napisał:

[Mike Menci]

Mark
I have some spare PCS - if you need one let me know -?

[Melitonas Mikaločius]

Mike,

is it possible to get two pcs?