What's new
What's new

RFQ - What am I doing wrong?

Wow! What a shit show 😮

Sounds like you got the mother of all blowies from an EDM salesman!

I have 3 EDMs, and 2 swiss lathes. Wouldn't touch you as a customer. What you don't understand what has been stated several times (particularly by DDoug) is that you asked for help and you AREN'T LISTENING!

I can tell from your print, as can many others in this thread you have no idea what you are doing. Not just as far as getting an RFQ, but as far as design work in general. Any designer, degreed mech E or otherwise (IMHO most Mech E's are lousy designers anyway) knows you cannot locate a part with any precision with setscrews. The necessary tolerance in the threads probably provide 1/2 deg of slop (shooting from the hip here), also cone point setscrews have no tolerance that I'm aware of concentric to the shaft, they are what they are.

Your print sucks, your toleranaces are assinine and borderline un-obtainable. When called out on it you get defensive, and say Joe Blow at told me a $500 dollar rotary on an EDM would take care of it. Anyone with half a brain knows that line is bullshit right out of the salesman's mouth and the point where I would walk away. A decent CNC rotary, which is what you would need, runs about $35K on top of your $100K machine that has enough travel to make your stupid part.

That said, you go drop 50K on a used machine Joe blow has, let me know how it works out for you. I'm sure in they covered EDM setup, electrode mfg, etc in your software engineering courses.

Personally, I think Garwood's camshaft idea has merit. But if the rest of the project is as f'd up as this part, it really wouldn't matter anyway would it?

You said you are from software engineering, and talk about terms you use for talking to your rubber duck. OK - We have a term in Machining and Design for when we talk to a dumb F&*K it's called a circle jerk.
 
I do appreciate the candor, and see my previous reply about what I was thinking with the set screws.

As for accuracy, again, I just don't see it being an issue with an edm. 0.00002" stepping on position accuracy, and no tool pressure. I thought there were two reasons to go EDM, hardness of materials, and accuracy. All the "perfect fit" "videos" on the internet. (I realize those are wire edm).

As for balance, yes, I have thought a lot about balance, and have made parts specifically to be able to balance the system to ridiculous levels. I've thought about resonance as well. As I said before, I've shrunk this to as small as I can think to make it for safety. Tolerances shrink too. I self-admit to sucking at mechanical engineering, but that isn't to say I haven't thought about the basics. Remember the cost quoted on the bearings? Yeah....I've thought about it, an awful lot unfortunately. I'm self-funded here and spent close to $500,000 so far. For a large company, this is chump change. And they certainly wouldn't quibble about $50k for a part. But I'm on a shoestring compared to that. For a self funded, one man research shop, I'm probably spending at the upper tier.

I really do enjoy the discussion, and don't mind people calling out stupid stuff as stupid, and hell, you don't know the whole design, or even what the project is, so you're right to make sure I've checked other aspects! It all helps. I'd hope someone who came this far would have done so, but you never know. The internet is full of insanity. I
No insult was intended or received.
Tone can very hard to read on the internet. We can still be friends :)

I’ll put it a different way.
“Why is no one responding to my rfq”
In a busy job environment a RFQ can be like a job resume.
An owners and managers job is to grind through the pile and select candidates that fit the company and will maximize profitability.
When a paper comes across the desk that does not look like a fit, or has splling/gammer/bottocorect errors it is filed away in the “not worth the headache” bin.
“Joey drill bit can handle that one. He takes on anything.”

This is likely the case with your Rfq’s.
And don’t take this the wrong way.
But the way the project is being explained puts up further red flags that there is a possibility the job may have snags or headakes.
Next candidate please.

I’ll apologize for myself and fellow people here, but “this” rfq is an industry trope we have all dealt with. Often too many times. I burned time with one this week. So the frustration is real.

For everyone’s amusement, I’ll share this week’s!

Here is a print. Need a prototype. 20 unique parts, super simple. Loose tolerance. Slap on a 40x20 vacuum chuck and mill some plastic.
Questions. Questions. What about this. That . I’m not worried about prototype price. Ok that’s a lot, what would small quantities look like. Bla bla bla. But seems very interested in moving forward. Gonna be $3000cad materials and labour. The odd local inventory prototype can be fun sometimes right? Help the local fellas :)
20 emails in 4 days. Ya. Short. Quick. But 20!
I’m an hour of unpaid labour in.
Finally tells me what it’s for so I better understand.
You remember those hockey tables you played with as a kid? With the big bubble dome over it? Like foose ball but hockey. He’s making “rink boards” for those that will accept slide in advertising panels……
That my friends is a “Joey drill bit job.”
Any jerk with a $2000 CNC router

My first email was blunt and crass specifically to avoid this type of thing. But it slipped past my bs defences this time…. That’s food off my table.

I’m here to help him. But it’s 50% up front 50% on delivery. Preauthorized credit card hit in 7 days of no pickup.
Last email sent.

Your job is the extreme opposite. But, the same in a way Isn’t it?
Hopefully this example sheds some light from the other side.

Again. It’s the managers job to filter for profitability/efficiency And that hour I got lost in could have meant getting a quote faster to someone else that may have landed a better potential job. Winning a quote can be a race.
Make your documentation as good as can be. Easy to read and direct.
You did the right thing reaching out here. It will hone your approach. It takes a lot of hammer blows to forge a sword…

At least we on practical machinist choose to waste our time helping each other.
The poor people that need to sift quotes are materially harmed by a pile of crap Rfq’s with poor drawings. For jobs that are never gonna pan out anyways Time is money/food on the table
 
Last edited:
Heat shrink collars, the taper lock collars suggested earlier, glue collars, split shaft..... all allow placement radially and along axle. You can clock them with an indexer/rotary. How exact falls on you, not a part that you admit you can not measure. You can then clock to outside of collar- which is bigger. The set screw approach is not good for the accuracy you are asking- not just the price of axle.
You can also try heat shrink and glue in arrangements in house.
 
Some here may recall years ago a psychiatry professor was provoking responses to weird requests on this forum...material which had been designed by a team of head doctors to measure reactions in cranky old geezers.
That would be "Servicar Rider"
 
Ignoring all of the issues pointed out earlier... If you wanted this made as close to what you have on the drawing, I would send the shafts to a centerless grinding house to make blanks. Then, I'd find an EDM shop that has a rotary in their sinker. With the appropriate workholding, it COULD be made. Finding someone willing to take it on is another conversation.

Yes, machines and rotaries also have tolerances, so in all actuality, you're probably never going to find a shop willing to attempt a part like this with such strict tolerances - it's just too risky unless they have absolutely nothing else they could work on. The tolerances are not really able to be inspected based on what you have, so you need to start thinking about how someone could QC your parts before you even think of sending an RFQ.

If you want someone to be willing to make it? Take a good look at your tolerance stack with the other parts that interact with this one. Making one part ridiculously accurate isn't worth anything if the corresponding parts don't have similar accuracy. Can you find a way to ensure accuracy with another approach? Maybe something like locating flats or something? A shaft like this, with tapped holes and accurate locating surfaces, would be prime territory for a swiss lathe shop. Sometimes you can get them to make short runs, but it's cheaper than stumbling your way through it by teaching yourself how to use a sinker EDM.

If all else fails, talk to the shops that turned you down. Work with them to come up with what's realistically possible, and then modify your design to meet real-world expectations. I hope this helps...
 
Here's the biggest problem I see. Even assuming the part could be made as designed, and assuming super precision ground set screws will be used in super precision ground threaded holes in the mating part and everything turns out geometrically perfect, it still won't be balanced. Metal alloys are not homogenous; they'll have a little more carbon over here, and little more nickel over there, the grain structure will be a little coarser over here, etc., and that all effects local density. Toolholder manufacturers know this, and every holder, even if perfectly symmetrical, needs to be run through a balancing process in which the holder is spun to find its center of gravity, and material is removed to obtain that balance.

Most machinists have a touch probe on their machines these days. In order to function properly, the tip of the stylus must be concentric within .0001". So does the manufacturer grind everything in with hyper precision? No, they build in an adjustment mechanism (using standard, easily achievable tolerances) so the user can dial it in. If something gets bumped, and it's knocked out of alignment, you can dial it in again. Adjustability beats precision any day of the week.
 
If nothing else, the OP should make a single shaft with a couple of the magic divots and try putting shaft collars with pointed set screws onto it to see if the pressure of the set screw being tightened bends the shaft. I would not be surprised if a combination of the force and the outward loading of the wedge does bend the shaft at each divot.
 
Last edited:
Very funny to see designs like this get roasted. Every once in a while I will get an RFQ that I just know was made by an amateur and will never actually get built. I feel a lot of empathy towards OP as I recall being very invested in my early designs and now I cringe when I think back on it. If OP keeps working at it with this much passion, he will eventually get good.
 
There are a few things here that are glaring red flags. One alone you can usually overcome, two it’s getting tough. More than that and forget it.

1. As expressed, your drawing is really really hard to interpret at first glance. Understandable for your first time, but it does put you at a disadvantage.
2. It’s also poorly dimensioned in terms of defining what I think you really want.
3. The tolerances, despite a somewhat reasonable approach to trying to determine what you could get, are either unrealistic, or eye wateringly expensive.
4. The features, as loosely described, don’t sound like they are going to do what you want, and that’s a recipe for time consuming drama. Machinists tend to be drama averse. They also tend to be poor estimators of just how much time=money it will cost them. As a result they are doubly averse.
5. You are averse to hiring qualified help, and seem to be giving off vibes that you think someone will steal your idea. If you are, file a provisional patent and get an NDA going with your contractors. If you’re still worried, hire a bigger firm. They have way more to lose in reputation than they stand to gain from stealing or leaking your idea. The risk with a big player is that they’ll give you their B or C team engineers because they don’t see you as a good long term revenue source.
I’ve spent several years of my career making things for people that I don’t think are going to work. I don’t mean my part won’t work, just that the larger plan won’t. Sometimes it does, often it doesn’t, but my part does its purpose. I mention this because despite seemingly being a Beltway bandit, you give off a lot of the same vibes as the tech companies I work with every day. Bright, motivated, hard working, and utterly incapable of building hardware for the same reasons the old hardware companies are incapable of building software. It’s unfortunate because a lot of the ideas are quite good, they just devote resources to the wrong places, which means good ideas die.

Hire someone with a decade of NPI design experience to at least spend a couple hours with you going over the design. Redesign it as appropriate for the quantity you need now. Sometimes an expensive to make architecture (and I don’t mean tightening tolerances) is cheaper due to being low risk. Once you prove the concept, then go back and make it scaleable. This particular one seems like it wants the next assembly level up to be one piece or tuneable/actively aligned, but that’s a shot in the dark without knowing a lot more.

I’m probably not the right guy because I’m busy and expensive, but less busy and less expensive equivalents exist.
 
There are a few things here that are glaring red flags. One alone you can usually overcome, two it’s getting tough. More than that and forget it.

1. As expressed, your drawing is really really hard to interpret at first glance. Understandable for your first time, but it does put you at a disadvantage.
2. It’s also poorly dimensioned in terms of defining what I think you really want.
3. The tolerances, despite a somewhat reasonable approach to trying to determine what you could get, are either unrealistic, or eye wateringly expensive.
4. The features, as loosely described, don’t sound like they are going to do what you want, and that’s a recipe for time consuming drama. Machinists tend to be drama averse. They also tend to be poor estimators of just how much time=money it will cost them. As a result they are doubly averse.
5. You are averse to hiring qualified help, and seem to be giving off vibes that you think someone will steal your idea. If you are, file a provisional patent and get an NDA going with your contractors. If you’re still worried, hire a bigger firm. They have way more to lose in reputation than they stand to gain from stealing or leaking your idea. The risk with a big player is that they’ll give you their B or C team engineers because they don’t see you as a good long term revenue source.
I’ve spent several years of my career making things for people that I don’t think are going to work. I don’t mean my part won’t work, just that the larger plan won’t. Sometimes it does, often it doesn’t, but my part does its purpose. I mention this because despite seemingly being a Beltway bandit, you give off a lot of the same vibes as the tech companies I work with every day. Bright, motivated, hard working, and utterly incapable of building hardware for the same reasons the old hardware companies are incapable of building software. It’s unfortunate because a lot of the ideas are quite good, they just devote resources to the wrong places, which means good ideas die.

Hire someone with a decade of NPI design experience to at least spend a couple hours with you going over the design. Redesign it as appropriate for the quantity you need now. Sometimes an expensive to make architecture (and I don’t mean tightening tolerances) is cheaper due to being low risk. Once you prove the concept, then go back and make it scaleable. This particular one seems like it wants the next assembly level up to be one piece or tuneable/actively aligned, but that’s a shot in the dark without knowing a lot more.

I’m probably not the right guy because I’m busy and expensive, but less busy and less expensive equivalents exist.
This post needs to be pinned and referenced for all queries from "inventors".
 
This post needs to be pinned and referenced for all queries from "inventors".
Agree! I also had (ok, still have) customers who are "inventors". I helped developing their products or parts of them and some of them I had my doubts whether they would work.

One project this year in particular was a nightmare. The customer came to me with this highly complicated device which had like 200 small parts inside which were all mechanically linked and this part was like only 3 inches wide. The customer's drawings were not manufacturable, so I charged them for re-designing and tolerancing every part and making the parts, but I would not be involved in assembly of the product and whether it would work or not (because of my doubts but I didn't want to tell them that).

One and a half month went by, and 80% of the parts were done with design stage, and I got an email from customer with their new revision of every frigging part. They told me while I was doing work for them, they decided to go to a professional firm and they redesigned the whole thing, so all of my work were obsolete. I dropped them immediately; good thing I took payment for designing upfront.

Another project was a product that I helped designing the whole thing from a hand-drawn concept customer gave me. After the first batch was done and sent to customer, I learned that the product was actually just a rip-off of another well-known product in the market. I stopped doing business with them immediately.

So, be careful of customers who bring you their inventions and it appears they have not taken proper steps to make their inventions a reality and instead come straight to a machine shop to get them made.
 








 
Back
Top