A guest post by bagho20.
We have a simple, real-world million-dollar problem, and maybe the power of
the CH brain can solve it.
The problem: We manufacture a product. Let’s call it a robotic snow shovel. The shovel is pushed by two
DC powered electromechanical linear actuators. The problem is that we are forced to use actuators that are not
manufactured with sufficient consistency to all push at the same speed. Some are slightly slower. In the product,
they are side by side about 6 inches apart pushing on the blade of the shovel. Since they often don’t push exactly
the same speed, one actuator does all the work, draws too much current, binds the shovel, and wears out too soon.
How do we fix it?
This is a real problem that I must and will solve, but I was wondering what others might come up with. There are
better and worse solutions due to costs, reliability, etc, and the best solution is the simplest and cheapest, but we’re
just brainstorming at this point. Think outside the box, find a new paradigm, you have to learn to love yourself
before you can love anyone else, be the change, give it 110%, It’s a no-brainer, everything happens for a reason,
there are no dumb questions, etc. Now take it and run with it….. right up the flagpole.
I’ll let you know what solution we eventually prototype and use. Maybe it will be what you come up with.
78 comments:
Match the actuators.
Wire them in series.
Well, if you don't, you can always sell a service contract with it, provide new jobs/work for repairmen (and women, sheesh), makers of the actuators, etcetera, etcetera... You know, sort of the Microsoft wait of doing business. Push it out half-assed then 'fix' it. Multiple times.
Sorry Bag. Just being a smart ass. I haven't had my caffeine fix yet.
Why not one larger centrally located actuator with the push bars extended left and right in a T shape?
Explain "forced". Is that due to cost? Management dictate, what?
Why two actuators? Why not one, or three?
Why push? Is there a reason the shovel cannot be pulled?
I could go on, but I would need some more information about the environment the blade works in, the cycle time, why hydraulics are not a good idea, things like that.
Agree with Beta. Alternatively, bench test the actuators before installation to find matching pairs.
Or change to a hydraulic system.
Add two (or more) actuators per blade, so that at any given time you are more likely to have at least two actuators doing the work simultaneously, causing less strain.
Are there horses being towed by the snow shovel? You could let them out, hitch them to the front of the shovel, and have them pull when needed until the actuators recover.
Sometimes the old ways are the best ways.
Well since you asked to find a new paradigm, move to Florida. Then you won't need the snow shovel.
The "fat man" A-bomb posed a similar engineering challenge at Los Alamos. It comprise a sub-critical mass of plutonium which needed to be compressed symmetrically in order to go critical. "Shaped" explosives surrounding the core were used but there were months and months of trial and error to get it right. If Cedarford were here, maybe he could give some tips. I always got the feeling he was a WMD expert of some sort.
So THIS is Where All the Sexy Talk Is.
I just came from a different website?
And it had an embedded YouTube video?
It was from an episode of Seinfeld?
It was the scene where Mr. Lippman confronts Elaine about all the exclamation points in her edit job?
I thought it was funny?
So I came over here and stole the idea?
But I changed the exclamation points to question marks?
WMD?
Rh is probably right.
Unless, Evi L still make the most sense.
That was fascinating, M. Bat.
Never knew there was an actually 'breed' of those dogs.
I like this post. A nice problem to work on.
I was also thinking that testing and matching would be the easiest solution. Maybe there is another way.
Sixty and others raise good points. But an off-the-top-of-my head solution would be to use one rotating motor, either centrally-located or directly driving one side, and to connect the two pushers together with a timing belt. The linear force would come from the pushers being either ball screws (expensive but accurate/reliable) or cheap Acme screws with maybe plastic nuts. You'd have to shield the screws that do the pushing, but that can be done pretty easily. You'd also need some kind of limit switch.
If you buy cheapo parts, all the stuff to make this would be pretty inexpensive. Look up Techno's drive components, for example, to judge relative costs. There are a zillion companies in the actuator and drive component business.
I may be practicing engineering without a license, as well as without as many brain cells as I would like, but, hey, it's free on the internet and worth every penny.
One big design question is, how fast and over what distance do you want to push? If you're looking for something like vibratory motion to keep the shovel clear, then my screw solution wouldn't work.
Sorry, I would first have to understand the problem to provide a solution.
Not that bright.
I'm assuming they are simple linear actuators without any feedback of any kind. If that's the case, then RH's suggestion with one caveat. I would wire them off the same power supply, but on their own separate rail so that each actuator is getting the same power off the same supply. Then you could run them initial baseline by which to gauge how they are running. See their performance. You might have an actual problem in how each one is processing power. Maybe the drives themselves while being identical are exhibiting different characteristics. If you have a meter, you can hook them up and see if one is consuming less current than the other, which might be causing your issue, if voltages are identical.
Become a snowbird. Just don't move back north till you are sure it's all melted.
There is no problem that cannot be solved by avoidance.
john, we think alike. I am getting sick of these Madison winters.
Matching components for production is usually an expensive pain in the ass.
But the design questions we don't know about are:
1. What is the intended travel?
2. What is the intended force?
3. What is the cycle time?
And the biggest questions of all:
5. What is the budget? R&D? Tooling and setup?
6. What is the total manufactured cost per unit?
7. What margin do you expect to make and how soon in the product life cycle?
These kinds of questions strongly interact with engineering decisions.
My solution is to eliminate the need for the product - eliminate snow - encourage global warming - take long, carbon-spewing jet flights, rent big four wheeling hummers, jets skis, do some bon fire keg parties, etc. In short: go on vacation.
Hubby engineer said use a rack and pinion connection.
You can almost never answer engineering questions without knowing the budget.
JAL's solution is excellent.
Piezoelectric sensor where each actuator contacts the "shovel." Adjust current flow to each actuator to match the force applied between them. Probably too exotic and expensive. Would have the bonus of knowing when the "shovel" is pushing against something it shouldn't and can't move (such as a boulder, or Obama's ego) and then shutting it down.
Because I'm a tooling guy, I always think in terms of screws and motors. Rack-and-pinion would lower parts count and maybe solve the misalignment problem. Elegant.
Can you make everyone pay, even if they won't be using it? That way you will be able to keep the price lower for the people who do need one.
TTBurnett said...
Because I'm a tooling guy, I always think in terms of screws and motors. Rack-and-pinion would lower parts count and maybe solve the misalignment problem. Elegant.
Also, ripe for double entendre.
Clearly this isn't actually a snow shovel.
There's not enough information here to come up with a good solution.
If you want help and not just my stupid questions, feel free to email me either SolidWorks or STP files of the assembly. Or email me an NDA first, and then send the models after I return a signed copy.
This is a geek magnet, not a snow shovel.
Lots of great ideas. And we have come to a lot of the same ones and are experimenting with them right now.
To Sixty Grit, we are "forced", because these are the only actuators currently manufactured with appropriate specs. We have looked into making our own, but the numbers just don't support it.
Like all our stuff, the entire product is American made, except for one small component, the little motor inside the actuators, which is made in China, and is the source of the problem. I don't know if that's why the don't work well, but it's one reason we make everything here in the U.S. of A.
This problem will likely be solved long term with a single larger actuator that's being developed right now, but we need a two unit solution till that comes on line.
The budget is simply the solution can't really cost anything. Maybe a couple bucks top. This is often the case in manufacturing. If you want to spend to fix something, you need to innovate some other cost out.
I wish we could just tell our customers they have to pay more from now on, like the government does, but we don't have a monopoly, and our customers aren't dumb enough to vote to increase their own costs because I have an awesome pant crease.
Ah, NDA's. You know you're getting into real engineering. Personally, I prefer IGES files. Same file will open nicely in both SolidWorks and Rhino. STL's are often a real tangle in Rhino.
But please don't send me anything, as I have enough to do.
If you want to spend to fix something, you need to innovate some other cost out.
The story of my life.
"feel free to email me either SolidWorks or STP files of the assembly."
How about on the back of a nice cocktail napkin. All our engineering drawings are kept in a system of napkins, sticky notes and faint memories. The important stuff we tattoo right onto the bodies of employees.
Seriously though Pasta, thanks for the offer, but I think we are good. It looks like the one mostly like to be used is what rhhardin said in the first comment and others said later, and which was also my suggestion to the R&D Department: match up the actuators. It's almost free, requires no new parts, and can be done immediately. We are running some trials with that today, as well as some other mechanical and electrical ideas.
Oh I really love the idea behind the post-
A hey --"let's fix this!"-- thread.
Sorry I got nothing about the actual content however.
We use about 6 different CAD programs in house. I'm pretty non-controlling and let different people use whatever they like, but for the most part we don't design virtually. We find in our relatively low-tech market, with most products being able to fit on a desk that it's usually faster to just prototype it in real materials, test and repeat.
People can get lost in their drawing programs, and sometime they are not heard from again for weeks. We often need a solution and working prototype shipped in 24 hours. Mostly, we shoot from the hip, empty the magazine, and hope to hit something.
Can you switch the actuators on a rotating type device so they work off alternate supplies?
Can you switch the actuators on a rotating type device so they work off alternate supplies?
We have pretty low-tech products made from metal and plastics. We use laser cutting, machining, press brakes, welding and injection molding to do almost everything.
When we encounter a design challenge, we come up with a couple ideas that can be started immediately and get people prototyping it. At the same time someone else will be working a computer designed solution. Someone else will be using cardboard and paper clips, and someone else just pencil and paper. The ideas come in at different times, and you never know who will win. Sometimes the hare, sometimes the tortoise.
Don't know a lot about this, but could you use a current limiter to force all actuators to a minimum spec? Don't know how much they cost, how much amperage you are using, etc.
I am with TT on this one - my CNC has the Y-axis gantry riding on rails 6 feet apart, driven by stepper motors, one on each end, using a pinion engaged with a 10' long rack on each side.
X-axis is simpler - one motor, one rack and pinion. Same with Z-axis.
But the Y has to be coordinated or else the whole gantry will bind.
The whole thing was built with common parts, nothing special, I bought it cheap and it has run for 14 years, rock solid.
I think it boils down to getting some feed back from the system or else using more expensive pieces to move the blade.
The ideas come in at different times, and you never know who will win. Sometimes the hare, sometimes the tortoise.
Sounds like you have brought a lot of fun into your company, using competition in a fun and healthy way.
The actuators are wired in parallel (same voltage supply) from a digital controller that controls what angle it's at based on input from a number of sensors. No, it's not really a snow shovel, but works similar.
"Sounds like you have brought a lot of fun into your company, using competition in a fun and healthy way."
That's job 1. We just make sure we make money so we can keep doing it.
Also, as with custom furniture orders and additions to houses, there is never enough budget.
Fuck those cheap-ass customers. I got your budget right here!
If the difference between the actuators is small, putting coil springs between actuator and plate would dampen the differential. Properly sized,they might be enough to minimize the problem.
If you want to make a fix in the middle of a run, matching actuators is a no-brainer, cost-wise. If you're in the R&D stage, then more engineery solutions are the way to go. I like the way bagoh operates, though: Sketch and try it out. 3D modeling is almost as much a time-waster as blogging. I still do most of my work in 2D, and only when basic designs and tolerances have been worked out, then it goes into SolidWorks. We still have a seat or two of Pro-E, and I have my own Rhino, but SolidWorks is the best for most of the real world. I'm still trying to figure out how to open a file in Pro-E, despite months of classes back in the day.
Corporate management often likes 3D modeling, because they have been convinced by salesmen that everything will be more concentrically armulated with the latest encabulation analysis package. People who actually have to make stuff usually are with bagoh: Sketch the damn thing, make it on the shop floor, and see if it works. Dilbert and Wally will still be deep in collaboration on BOM and routing issues, talking levels and work orders, while the guys who do the work will have the stuff actually going out the door.
And, Chickelit, you have a dirty mind.
JAL said...
Hubby engineer said use a rack and pinion connection.
That would do it too, but I'm taking his question to heart based on the sketch he provided. A lot of times people look at something like that and say, "How can I improve upon this..." That is my initial impulse as well, but I didn't see that in the request for help. So I given what is there, how do you get that to work. If it doesn't work or if there is a problem, what is the the problem and how can it be fixed to get him to that state. Once those issues are eliminated and it still doesn't do what it needs to do given that everything works like it should, then how do you make improvement as minimally as possible to get to work.
bagoh20 said...
The actuators are wired in parallel (same voltage supply) from a digital controller that controls what angle it's at based on input from a number of sensors. No, it's not really a snow shovel, but works similar.
Okay, so it's getting the same voltage from the same supply, but are both actuators getting the same current as well?
bagoh20 said...
We have pretty low-tech products made from metal and plastics. We use laser cutting, machining, press brakes, welding and injection molding to do almost everything.
When we encounter a design challenge, we come up with a couple ideas that can be started immediately and get people prototyping it. At the same time someone else will be working a computer designed solution. Someone else will be using cardboard and paper clips, and someone else just pencil and paper. The ideas come in at different times, and you never know who will win. Sometimes the hare, sometimes the tortoise.
Bag, can you die cut small parts made form material like Poron that has adhesive backing on it already? If you can, I'd like to send you something that I'd like to get a quote on.
Geeks speaking in tongues.
Forget plastics. The future is in hamsters.
Rabel said...
If the difference between the actuators is small, putting coil springs between actuator and plate would dampen the differential. Properly sized,they might be enough to minimize the problem.
I'm thinking star trek... but why star trek technology to shovel snow when we are going to have a lot of illegals becoming legal and a lot of them probably looking for employment.
If you want to spend to fix something, you need to innovate some other cost out.
The story of my life.
--------------------------
Yes. Now if only the Federal government (and all others) would understand THIS.
"Okay, so it's getting the same voltage from the same supply, but are both actuators getting the same current as well?"
Yes, they are on both fed from a single source like in the drawing. They are considered one motor to the controller. They may draw different currents, because they are slightly different, and that's the problem.
Sorry Meth, we don't do die cutting. We don't have the presses for it any more. If you wanted laser cutting of metals we could, but any plastics require us to change air cleaning filters, etc, so we avoid it. Wish we could though.
nd: Well, I WAS speaking in tongues, at least somewhat planted in cheek.
Here is a classic engineering video, nearly ready to transcribe for your next presentation to management.
Why are you trying to build better things when you could be hiring lobbyists to rack up some bailouts? Has the government not done enough to make it clear to you what they prefer?
TTB, hilarious!
Good question freeman.
The snow melts, lobbyists are for ever. (i'm kidding)
It's not cheap but i'm sure the government will buy it.
funny.
If this were a Snap Circuit situation I think the answer is in resistors and capacitors. Each actuator has one or the other.
In this examples the actuators are doing the deciding on their own about which actuator does the most work based on their own inconsistencies. So make them consistent by controlling electrical flow using resistors and capacitors before it flows through the actuators. Make each actuator work in accordance to pressure sensors if necessary, rather than allowing them to draw what they will and do with that what they can. Force them to work harder by holding back flow from the workhorse actuators by arrangement of capacitors and resistors. I think. I don't have a Snap Circuit toy myself, so I don't know for certain, I'd have to ask a child that has already completed a Snap Circuit Jr.
We are also exploring the possibility that the slower actuators could attend some motivational seminars, maybe make a Star Wars video to teach them how to work as a team. That kind of stuff. If that doesn't work, maybe a sweat lodge retreat with crystals and a pyramid motif.
One possibility would be get the actuators unionized, then they would all be required to go as slow as the slowest ones.
bags is inhaling the nitrous oxide today?
Hey you, you're getting all sorts of free consultations ... people pay real money for this! (All kinds of low posters and lurkers showed up.)
Treat the winner to dinner if they can make out there to you.
We need to take a collection for Chip so we can give him a Snap Circuit kit for Xmas. (We can purchase it through Insty and spread the joy around.)
... to go along with his hieroglyphic texts and his Julia Childs spatulas. And his Exacto® knife.
bagoh20 said...
"Okay, so it's getting the same voltage from the same supply, but are both actuators getting the same current as well?"
Yes, they are on both fed from a single source like in the drawing. They are considered one motor to the controller. They may draw different currents, because they are slightly different, and that's the problem.
Sorry Meth, we don't do die cutting. We don't have the presses for it any more. If you wanted laser cutting of metals we could, but any plastics require us to change air cleaning filters, etc, so we avoid it. Wish we could though.
Okay, bummer though. I need a die cutter that can die cut small foam parts. Either way, if you are having different current draw from the same linear actuators that are slightly different models, then you should try to be consistent and swap them out and make sure they are identical models and see what happens.
A small charge produced from a Gilbert Chemistry Set can displose of the whole problem.
"Treat the winner to dinner if they can make out there to you"
Sure, if she's single. I could use a nice nerdette in my life.
They still sell matched 6L6's for amplifiers, which presumably they do by testing and sorting.
Open up a matched actuator business.
"bags is inhaling the nitrous oxide today?"
Long ago, I used to have a friend that had access to tanks of nitrous, and he would have parties with balloons full being passed around. It was fun, and low calorie.
Once I had a dentist that over did the nitrous on me in the chair. I was tripping out. I like a nice buzz as much as the next head, and it was feeling pretty awesome, but I was actually getting scared. He had a TV up in the corner and it was playing The Six Million Dollar Man TV show. That sound he makes when he starts showing of for the girls was going right through my soul. I almost pussied out, and asked him to stop it, but I hung in there till the end. I figured if I was gonna fly too high for once, this was a good safe place to be doing it. They're doctors or something.
"Open up a matched actuator business."
I knew legalized Gay marriage would lead to this eventually.
ActuatorHarmony.com
Call the vendor who makes the actuator. Getting a coordinated action from multiple units is probably something they deal with on a regular basis.
I was frequently suprised at how helpful industrial sales and technical reps were compared to the retail/consumer type. Even for small customers.
It could be any one except me.It is legitimate that when you work for anyone he should pay you in exchange.
Thanks
Bruce Hammerson
Hydraulic Hammers
Post a Comment