Vacuum Sensors Are My Nemesis

This one was quite a struggle. But in the end I VANQUISHED THEE, FOUL SENSOR...kinda. I picked a pump that's going to work well with the rest of the system, and found a way to monitor the vacuum inside the pneumatic line!
Thank you so much to PCBWay for sponsoring this video!
www.pcbway.com
Patreon:
/ stephenhawes

[@Graham_Wideman]

Stephen, in case you haven't discovered it already, I'll point out some problems with your schematic. The current source is supposed to set 100uA through the bridge. Your voltage divider R19 - R20 sets only 1.65V, which would cause U9A to set 1.65V across R23, setting only 82uA. (So you would get only 80% output compared to the reference example spec.) OK, but that 82uA passes through both R23 and the bridge, which is also 20k, with the resulting voltage across them of... 3.3V. In other words, this circuit design asks U9A to output 3.3V at pin 1. But LM358 (if that's what you're actually using) can't do that, because its output swing is limited to about +rail minus 1.5V, or about 1.8V. Even if you used an op amp that can swing rail-to-rail, you're asking it to sit at positive rail, so your "current source" is actually no better than just connecting the + side of the bridge to +3.3V. At any rate, these missteps screw up the ability to set the two bridge output voltages to approximately the middle of the useful range of the output opamp, which you need in order to get useful amplification (x50) at that stage. You managed to get some useful behavior by throwing in a pot to essentially tweak out the problems in the bridge area, but this is sticking a large bandaid on a problem that can be solved easily with (a) a rail-to-rail-capable op-amp (if you insist on running off only 3.3V), and recalculating the R19-R20 voltage divider so that it sets the bridge outputs at mid-rail (1.65V), and changing R23 so it sets 100uA. So about 6.5k for R23, and about 2.65V at the top of the bridge, so R19-R20 can be a ratio of 0.65 to 2.65, so maybe 6.5k and 26.5k. And since you would then be using the full bridge instead of half, you get twice the output and also insensitivity to temperature changes.

[@fdavpach]

When he said "general purpose op-amp" I thought the same but you have here an awesome explanation, nice work, also I really have to learn a lot more about op-amps

[@stephen_hawes]

Wow, thanks for the super in-depth info! I had a feeling someone would take a peek at the schematic and tell me what i screwed up! :D I still have a ton to learn about op amps, this has definitely helped me figure out what I need to spend time reading up on. I'm working through the Art of Electronics right now, it's been really helpful in the chapters I've gone through. In the meantime, I ended up going with a sensor that handles the signal processing for me, but this stuff will inevitably come up in future designs.

[@Graham_Wideman]

@@stephen_hawes Great, I'm glad to hear you have your hands on AoE. Something to grasp is that Op Amps actually vastly simplify analog electronics, essentially providing you with a small number of idiomatic circuits that function as Legos for "analog computing". That is to say, these Legos have a simple behavior where the output of each stage is a simple algebraic formula based on the input(s). It's helpful to know why that algebra works for each of the few op amp idioms, so that you're really convinced, and so that you know what you should also attend to in order to get that "ideal" behavior. The upshot of all that is actually there isn't really "a ton" to know. Just a few essential principles. In your circuit, you would have got the result you wanted if you realized how the first op amp sets the bridge current, that an objective was to maneuver the bridge output voltages to mid-rail so it has room to vary up and down, and that you need op amps whose outputs can actually swing across the voltage range of interest (not a universal trait when operating with such a low power supply voltage). So being able to pick out the basic specs on a datasheet would help too. Anyhow, I'm glad you found a solution, but regardless encourage you to get comfortable with the not-very-extensive basic suite of op amp idioms, and AoE can be quite a help in this regard, in addition to loads of material on the internet (some clearer than others, as usual).

[@electronash]

Dave did a great primer on opamps here. I learned a lot from it... kzbin.info/www/bejne/bXe8eadrjducgcU I still find analog stuff a bit daunting, but a bit of Ohm's law and opamp examples really helps.

[@justinclacherty9181]

Great explanation. You're probably also better going with an instrumentation amp for this kind of thing.

[@LucasHartmann]

You know... The power supply behind you was changing current wildly every time you sealed or broke the vacuum... That gave me a totally different idea on how to detect of you picked up a component.

[@decee1157]

That is actually not stupid idea to just sense the current! And it can be done with just a simple INA214 and a 10mOhm shunt! Used it a few times before and it works really well! If you read the comment Stephen then just remember to place the INA214 close to the shunt. There are great layout examples in the schematic or HMU if you need some help with it Using 10mOhm shunt and the INA214 you will get a 1:1 so 100mA is 100mV

[@mikesnapper9001]

that is actually pretty smart solution

[@CarlBugeja]

Great video! 🙂 Small tip - there are some MCUs that have an integrated analogue comparators which allows you to read the reading directly

[@xxportalxx.]

Hey another favorite up and coming tech diy channel!

[@NonExistChannel]

Oh, that's what it's there for. Never came up with an application for an analog comparator

[@stephen_hawes]

Oh wow, that's rad! Didn't know that was a thing! I'll have to keep an eye out for that, would have saved me a ton of heartache!

[@MattBilsky]

@@stephen_hawes depending on how small the differential signal voltage is you may be able to get away with a cheap board like the HX711 (typically used for strain gauges). Or if you have a few signals to measure, play with the ADS1256 which can read 4x differentials and has PGA (I have a couple tutorials for them if you need). Lastly, the teensy has a PGA built in as well. Or go old school and use an instrument amp to convert to single ended to read with standard ADC on MCU

[@MrGatlin98]

If the vacuum pump bogs down when you cover all the holes, couldn't you just measure the increased current draw to the motor to detect when you've picked something up? This seems a lot more simple than trying to add a pressure sensor.

[@jakie996]

Hi Stephan, take a look at the HX711, it is able to interface to the Wheatstone bridge and it has also a 24 bit adc, a bit overkill but it should do the work, it is the one that is used with strain gauge, and having this pressure sensor a similar working principle it should work, also it give you directly a digital signal to work with! Very good job!

[@ilicz]

I was about to suggest the same... :)

[@powil4ss]

Alternative to HX711 is NAU7802..which you can buy from Digikey.

[@stefankachaunov396]

Same here! The HX711 uses it's own weirdo protocol with clock and data, but there's an Arduino library for it and the generic breakout board costs about a buck. Maybe worth a shot!

[@syber-space]

Love the little update. Nice ElectroBOOM ending too. Keep it up!

[@kyhldk]

Why not just meansure the current going to the pump? When it's on "load" it will draw more current With that you can tell if it has picked up a thing or not Easy

[@stephen_hawes]

hmmm, that's a pretty cool idea! might also be useful for other things that could be controlled with the FETs on the board. ill look into that!

[@kelna2]

Was about to say the same thing but saw this comment. Way simpler to do this, and you don't need to buy a special vacuum sensor

[@DEADB33F]

Just made a similar comment, but it will draw *less* current, and RPM will *increase* when the inlet is blocked (not bog down as someone else said). Try it with your vacuum cleaner at home. Block the hose and RPMs will rise quite sharply and if you have something to monitor the power drawn from the socket you'll see a drop in power consumption. ....it's fairly counter-intuitive, but pump is doing less work and moving less air so motor is under less load (so will spin faster). Same happens with a box fan or PC fan if you lay it flat on a surface. --- Oh yeah, with a diaphragm pump the opposite will be the case (but that's not what we're using here).

[@lunacyworks]

@@stephen_hawes while this is a cheap way, a kinked hose will send a wrong signal. And if you plan on having more than one vacuum actuator in action (multiple pick heads) then it will be an issue.

[@fredo514]

Hey Stephen, have you looked at the Honeywell pressure sensors? sensing.honeywell.com/sensors/amplified-board-mount-pressure-sensors They have the drive and amplification integrated and they have analog out and digital I2C. I’ve used them in the past and they’re super easy to work with!

[@xxportalxx.]

Op amp circuits are designed using nodal with the assumption that the two inputs draw no current. However if I were you I'd stick to common designs and use spice simulators or at the very least Google a website called falstad circuit sim

[@GISED_Link]

Use spice design and / or ask discord community ...

[@AndyMcBlane]

12K subs? Good job!

[@JirkaKlimes_]

He is so underrated, he should have like 120k subs!

[@wizrom3046]

Hi, your PNP vids have been interesting. 👍 I used silicone surgical tubing on my machine, it is rubbery soft are easily rotates the 180 degrees, and it easier for the machine to move it than plastic tubing. You can buy smaller sizes, sometimes also used for surfboard legstraps or slingshot rubbers. Definitely try the Honeywell sensors, with inbuilt amp giving calibrated 0 to 3v output. Finally for the pump I use an aquarium airstone pump, rigged to suck not blow. It is very quiet and can run forever without overheating etc. Also, your vacuum solenoid to turn the vacuum on/off should be a SPDT type so it releases the component by letting air in. They are common in 12v fuel injected cars and motorcycles. 🙂

[@schroedingershat7912]

Measure the motor current instead, when it loads it'll draw more current. Should be good enough for this purpose.

[@DEADB33F]

How about monitoring the pump's motor current? (or RPM if you can fit a hall effect sensor somewhere on the motor) When there's a part attached and vacuum drawn the current should decrease and RPM increase due to less airflow through the fan. --- EDIT: Should have read the comments. Seems this has already been suggested a few times.

[@pinolec]

Look at MPX5050DP. No opamps needed.

[@joshuatelson5645]

Thoroughly enjoying these bud, though I will never build a pic in place machine, I have wanted to build the MPCNC with the prusa you recommended (my favorite tool in the house). Nice tip of the hat to electro boom!

[@stephen_hawes]

Thanks dude! :D Glad you're diggin' the Prusa, I've been thinking of buying one myself for a while. If you build the MPCNC I'd love to see your progress!

[@animo3d]

Hi Stephen, I'm using the MPXV7025 Series Vaccum sensor, No need for Op-amps, Direct output form 0 to 5V!!! a little more expensive bit is so much worth the expense...

[@RoterFruchtZwerg]

I bet you have some BME280s lying around. Add some hotglue and they woud have probably worked fine 😄

[@RoterFruchtZwerg]

or just ignore the pressure and measure the current of the motor 🤔

[@arnauddurand127]

The "Programmable-Air" uses a similar or the same gauge pressure sensor and is controlled via Arduino. The schematic is available here: github.com/Programmable-Air/PCB/blob/master/programmable-air-v0.5/schematic.pdf

[@nils2698]

Progress on the vacuum sensor looks promising! Keep on rocking!

[@applicablerobot]

Opamps are awesome, but they require full circuit analysis which isn't always a quick process...

[@MattInglot]

"I really like digital, I don't like analog at all" You and me both brother 😂 I've been forcing myself to get better. It's a journey. Art of Electronics is a tough read ;)

[@Roobotics]

INA333 is a great part for your wheatstone bridge, btw! one resistor to set the gain, GND referenced signal out.

[@stephen_hawes]

sweet, thanks for the part number! i'm looking into using chips from that INAxxx family, seems like it'll be more straightforward than doing it discrete.

[@adeliokilcullen263]

Fave KZbinr rn

[@joshblake775]

I've used the ADS1247 ADC from TI for similar resistive applications - it can be configured to output a fixed current that can be used to drive a resistive load (in my case I was using a PT100 thermistor) in a 'ratiometric' approach. Accuracy will be down to how accurate your reference resistor is (ie pick a 0.1% tolerance resistor and you've got very high accuracy). You can communicate with the chip over SPI, set interrupt triggers on the chip itself (ie once it detects a certain level, it triggers an interrupt pin that can be used to signal back to the MCU). I guess in the bigger picture, the same ADC can also be used for high resolution sampling of thermocouples or other analogue devices. Just a thought but it's an easy implementation that gives you a ton of functionality (and a digital output).

[@joshblake775]

Take a look at 2.4 (Three-Wire RTD Measurement, Low-Side Reference, One IDAC Current Source) in this TI doc for rough idea on reference design: www.ti.com/lit/an/sbaa275/sbaa275.pdf ... I dropped the input impedance resistors and smoothing caps on the line side and it worked well for my purposes! You could basically just drop the part across those pins and you'd have your digital output.

[@poptartmcjelly7054]

What about the MPX4250?? From their datasheet: Integrated Silicon Pressure Sensor On-Chip Signal Conditioned, Temperature Compensated and Calibrated 0.2-5V output.

[@Pott_AV]

nice Job stephen

[@BLOCKsignallingUK]

Hi Stephen, you could consider the XGZP6847 in the -100 to 0kPa form which has a 4.5V to 0.5V analogue output. Costs $8.

[@jonathanwinegar1221]

Check out Honeywell's TruStability Sensors. I have used the Analog output ones with a lot of success, but they also have digital I2C and SPI versions. (kinda pricey tho, they're pretty accurate)

[@randomname3894]

You can get similar sensors as a module on aliexpress with analog out. I’m using the green round ones.

[@electrodonkey]

Hi Stephen, i really enjoy your channel and projects! Please have a look at Infineon's offering of pressure sensors. Some of them provide internal compensation and digital outputs. KP2xx series if i remeber correctly!

[@Nigel-Mac]

I don't really need one, but man do I NEED one

[@djkiss1999]

You can use a venturi effect vacuum generator instead of the vacuum pump

[@webslinger2011]

Paul from Learn Electronics taught some stuff on op amps and how to build a couple of circuits on it.

[@TheRainHarvester]

I tried the 180 degree strategy. But the open PNP guys didn't want to support the software. They want 360° rotation. I changed the code, but they update the code so much, that I gave up.

[@Tenacious21]

Could you measure the current draw of the vacuum pump itself in order to tell if it's got a part picked or not? I imagine it would draw more when the motor is under load.

[@powil4ss]

Nau7802 differential! that's all you needed

[@mortaldrumming]

Great videos, man. I was wondering when you were going to use you desktop CNC, and you did in this video. Did you make it your self or where did you get it? What parameters are you using to get such a clean finish? I want to build one for milling PCBs too. Thanks in advance. Keep up the good work!

[@cesarmardueno8502]

It would be excellent if you showed us how you do the pcb in the CNC, do you use flatcam?

[@jurevrbancic2264]

Hi How about SDP series sensor from senserion. I have worked with them an they are fabulous little things. No analog stuf, just I2c digital signal.

[@mehrdadfeller]

Hey stephen, is it possible to point to where you purchased those vaccum pumps?

[@奚牛]

hello，I can successfully upload your firmware to ARDUINO, ramp1.4, or connect with pnp software, but I can’t control the motor movement. I hope to get some help, and what went wrong, or you can issue a simple issue of how to upload the firmware Control the video.

[@tablatronix]

What kind of vacuum tubing are you using ? It looks flexible, the thick wall stuff I have is hard as a rock. FYI Sparkfun just released a really nice pressure sensor breakout board. I hate opamps, lol I spend the money on an instrumentation opamp, so worth it.

[@sciencoking]

Should've just used a MEMS barometer :p lots of them go down to vacuum

[@Graham_Wideman]

Hey Stephen -- if you're going to venture into unfamiliar areas of electronics, I highly recommend you get a copy of the famous "Art of Electronics". You seem to be a smart guy -- it's tragic not to equip your smarts with actual knowledge :-).

[@xxportalxx.]

Something I wish I learned earlier myself, sometimes formal training even in the form of a book will save you months to years of struggle

[@cdyoutoob]

Steven, You can get 360 degree rotatable fittings from Festo that doesnt leak like a sieve. Don't confuse with the Orientable type. Pages 11-14 www.festo.com/cat/xdki/data/doc_engb/PDF/EN/FUNCTION-FITTINGS_EN.PDF Also think you should use the smaller diameter tubing reducing volume of air being moved and increase time for vac etc. Possibly PUN-CM-4-SW. Its antistatic as well.

[@mastermoarman]

Sence that thing is a basically a load cell cloud you use a load cell amplifier? I know there are some that are i2c capeable with like 12bit adc on them. The hx711 is a popular one

[@blackhunter2310]

Hi stephen. Really like your vids. You could also use a barometric sensor like the bmp388. Btw is there a discord channel or something in where we could think along with your struggles?

[@syber-space]

The PnP channel gets used sometimes.

[@pioneer1943]

Which OpAmp are you using? From your description I think it's an LM358. Try using a TL072 :)

[@uint16_t]

Don't be discouraged by opamps! My first opamp experience burned blisters to my finger. It gets better.

[@stephen_hawes]

:D thanks! all the advice from you kind folks has put me on the right path! ill be diving back into them soon!

[@randycarter2001]

Op-amps are easy. Try doing it with BJT's. R6 in the Omron user manual for the sensor is the biggest gain control part. The bigger R6 is the more gain you get. Have you looked at the MPXHZ6250A, MPXH6400A, or MPXV6115V pressure sensors? I'm partial to the MPXV6115V.

[@smgvbest]

Stephen, I understand your digital vs analog wows very well but in all honesty you cant get away from analog and the opamp is a very important item to learn and worth the time. Graham Wideman gave a great piece of feedback for you below check out check out w2aew, he does some fantastic tutorials Also eevblog has some great opamp discussions, #600 is really one of the best to understand them i've seen. I like your videos allot, I love seeing peoples thought processes an you share them instead of editing them out. I do think that coupling the vacuum line directly to the motor is going to come back and bite you in the long run though with rotated parts as those little stepper have little torque.

[@guillermocobaafonso409]

A wheatstone bridge is not neceserly build of two extensiometric gauges the definition you gived wasnt wrong neither wasnt good, as long as you only nead 4 normal resistor to build a wheatstone bridge which btw is a quite common way of signal aconditioning for all the resistive sensors. And for the sake of electronics learn at least ideal op amp they are the easiest thing in the world of electronics. And for those type of builds you wont need to know whats happening inside

[@guillermocobaafonso409]

However love your videos dude you are awesome and that pick and place machine iss pffff

[@palliyil]

For prototyping purposes, you could try programmable air: www.programmableair.com/ which is an open source arduino compatible board that can both control and sense vacuum. plus a couple of valves thrown in too. The kit comes with all the things for you to build the pneumatics for the pick and place, right away. But, since, i dunno when their next batch will be up, you might not be able to buy one. (I got mine during its first crowdfunding campaign, and have been trying to learn from your channel, the journey towards building a pick and place machine myself). It is open source. So you could just see what parts are used there and use the libs for it too. github.com/Programmable-Air. Also, there are ADCs with differential amplifiers built in, like the ADS1115 (16 bit) or HX711 (24 bit), with which you can do your strain guage measurements. You can make constant current sources with LM317s too, which is generally something that you'd probably have a better intuitive understanding of.

[@palliyil]

It looks like it has the same sensor you're using. And uses an instrumentation amplifier

[@mrechbreger]

You mention that the rotary joint is leaking air, where does it leak air? is it the adapter below or the screw part of it? There's sealing material available which is normally wound around the screw part. de.aliexpress.com/item/33033430576.html I did not see that on your rotary joint, can you double check that?

[@josedaniel9609]

Try to 3d printed the vacuum

[@byronwatkins2565]

Change R29 and R31 to 500k for 10 times more sensitivity.

[@andrii3810]

Yo bro, could you please send the link where to buy your pcb vise? Thanks!

[@SDelduwath]

Hakko C1390C Omnivise PCB Holder, it is a pricey paperweight ($50-70) but is still my favorite tool for general board and wire holding.

[@filipcvetanovski2545]

KZbin really does not like you. I got a notification for the video 9 HOURS after upload.

[@chahilpatel2993]

Just use load cell amplifier 🤦🏻‍♂️🤦🏻‍♂️

[@TetedeMort666]

I love your channel and I want to see more videos .... So please ! Put some shoes when you manipulate electronics, soldering iron and oscilloscope ! Safety first ! ;)

[@Orlopzi]

Vacuum sensors suck. Pressure sensors blow.

[@xaesthetics1769]

Woah I'm really late! Haha, nice video!

[@xaesthetics1769]

Whoops! I meant early!

[@thatonesnowboarde]

Maybe this one could be an alternative to that sensor? kzbin.info/www/bejne/bmSvpq2Aia6Vd9k

[@RMW360]

is that servo motor horn hanging on your neck kzbin.info/www/bejne/bJ-8k4mKlNaYZ9E :p

[@michaelbishton9439]

Does something like this I2C pressure sensor work for you? www.adafruit.com/product/3965

[@p_mouse8676]

Love your videos and your can do mentality. But saying that you hate analog electronics, made me a little sad. It's a fundamental part of working with electronics. Understanding at least some basics would also be very beneficial for yourself. So I would recommend getting at least some nice books or follow some channels that explain the basics. In the end it's not so hard at all. More important is to experiment with those as well, since you're clearly a do'er.

[@stephen_hawes]

Thanks for your thoughts! I've changed my tune since then, I was quite salty coming off of a few days of frustrating debugging, but all of y'all helped me realize what I was doing wrong and not to completely dismiss analog because of a few frustrations. It's just the next thing to learn! I've been diving into the Art of Electronics as of late, super helpful.