EP#438: Solder Slugs

Episode Transcript
Available transcripts are automatically generated. Complete accuracy is not guaranteed.
Parker Dillmann (00:11):
Welcome to circuit break from Macrofab, a
weekly show about all thingsengineering, DIY projects,
manufacturing, industry news,and mind reading contract
manufacturers. We're your hosts,electrical engineers, Parker
Dillmann. And Steven Kraig. Thisis episode 438. It's been a

(00:31):
while since we have not had aguest.

Stephen Kraig (00:34):
And we have a bunch of guests coming up
they're actually gonna be reallycool.

Yeah. Yeah. So this has got this is like a
little break for us.

Yeah. A little breather.

Yeah. Yeah. Speaking of breathers if you all
have time you listeners, I justposted a article called top 10
episodes for electricalengineers for the circuit break
podcast. Go check that out. Wehave this is kinda like another,
like, introduction, but for,like, engineers to the podcast.

(01:02):
So we have, like, the episodewith Bill Heard. We have the
Chips Act. We have probably 1 ofmy favorite episodes the,
mislabeled label mislabeledlevels of hell where Steven told
his story about

buying wrong hardware at your local hardware
store because people put it

in a wrong bin. Yeah. Yeah. We have product of
many imaginations. That's whenwe had Misha and, Chris Church
on, that's the CEO and the othercofounder of MacroFab on the
podcast to talk about supplychain complexities.
That was right around like themid I think it was like midway

(01:43):
2020 we did that episode. Thenwe have the idea fab does this
podcast change the world whichis I think my I say my all these
are like my favorite. I lovethese episodes.

That's why they're on the top 10.

That's why they're on top 10 list. Yeah.
It's hard. It's hard. I don'tthink I could pick 1 episode.

Gosh. Well, I mean okay. So out of this many
episodes, yeah, it's prettyhard.

Yeah. But, yeah, the idea of that podcast
if if you're a new listener,that's a collaboration between
the Idea Tank podcast and andand us, the circuit break
podcast, where we have Scott andEric, come on the podcast and
then they pitch us likereasonable ideas, Kinda like
Shark Tank and then we pitch,like, stupid ideas.

Just ridiculous ideas.

That that was where Steven came up with the
College of Average Joe's. Orabove average Joe's. Yep.

Yep. Yep.

Yeah. And then we have rogue Gerber's with
Jerry McDaniel who's ourdirector of customer care.
That's a really good episodethat talks about kind of like
how customer service works, butnot just that mac fab just like
in general. It's kind of coolepisode. This is an old old

(03:03):
episode your toaster is toast.
When when did that come out sayit's old it's only 335 it's only
a 100 episodes ago that episodefeels like another lifetime ago.

You know, okay. So that 1 reminded me of an
episode way back when when wewere talking about the the
toaster controller IC that wasdedicated that has the bagel pin
on it. And I I honestly thoughtthat's what you were going with.

No. Actually, the bagel pin didn't make this
list. That did not. Wow. No.
Then we have HyperSize partpackages. That's a really cool
episode where we talked aboutlike different, like it was it's
more of, like, not actually thefootprints, but, like, the part
packaging identification, like,markings and that kind of stuff.

(03:53):
Mhmm. And then datasheet lore,which probably came up with our
best idea ever that we neverdid.

Okay. Wait. Wait. I'm never did is a strong term
word. We haven't done yet.
We haven't done that. Be becausebecause I still think this is a
fantastic idea, and and if Iwere able to do this as a full
time job and, like, actually doit correctly, III think that
this would I I would do it in aheartbeat.

Yeah. But this but

this episode was all about how a a system for
determining how to pick theright part for your design.

And it's funny that is actually is your day job
almost now.

You you actually, what I think is funny is is
somehow it just kinda played,like, my career played that way.
Because we had this discussion,and now I I you're right. I
basically do that. Well,actually, I shouldn't say that.
It's not that I picked the rightpart.
I just tell people that theypicked the wrong 1. That's what
my job is nowadays. Actually, alittle short side story. So so

(04:58):
we got a new intern and everyweek, there is, on on 1 day a
week, there is, like, an internday where for 1 hour, somebody
on the team gets to go and talkto the intern about what they
do. And today was my my day.
So I got to I got to tell theintern what all I do. And and
what was funny was I had toactually sit and think

(05:20):
beforehand. I was, like, what doI do? Like, of course, I know
what I do, but, like, if I haveto summarize it to somebody,
what what do I do? And in a lotof ways, I I get to tell people
that they're picking incorrectparts a little bit more than I
get to tell people that they'recorrect, you know, choosing
correct parts.
So but, yeah, that that thatthat episode was fun and I don't

(05:44):
think that idea is is fully deadbecause it's just it's too good.

I'm surprised that no one's no 1 took that
idea and made it. There was somemovement with it, but this is
early on in the the slackchannel before we had that the
the, form at forum.macrev.com.You should go check that out.
Join our community. If you havea macro account, you already
have an account there.

(06:11):
You just log in. It just works.You don't need an account.

And and I've had some people reach out to me
actually about this idea, andand, you know, we we had some
conversations about it and howto make it a reality. And I and
the the the the difficulty therewas just figuring out, like, how
how do you give it legs? How doyou actually make this work? And
and I struggled a bit with withthe the basis of, like, just

(06:39):
really like, the beginnings ofit because because I know what
the end result kind of lookslike, but getting from a to b
was just very difficult on that.And just for the all of those
who are probably lost right now,this this idea was just creating
a system that that basically ishow do you pick a and then
choose whatever part categoryyou talk about.
How do you pick a resistor? Andthen really, like, just guiding

(07:02):
an engineer through all theintricacies and details, maybe
not all of them, but the theintricacies and details of
picking that part. And how doyou avoid all the traps, and how
do you avoid all the thedifficulties, and how do you go
on to your favorite vendor, pickany of them, and laser beam
narrow down to the exact partyou need, move on to the next 1.

Yeah. It's kinda like a choose your own
adventure for a part where,like, you go, okay, I need a
resistor for your example. Mhmm.And then it'll be like, okay,
you want a resistor, and then itmight be like and then I'll just
ask you like a question next,like, what is at this list the
most important thing that youcare about this resistor? Yeah.
And it will show, like, all theimportant features, so, like,

(07:47):
the resistance, of course, thepower rating, size, and then you
can also go in through, like,temperature code, like you can
list all these properties andand specifications that meet,
like, make a resistor. Mhmm. Butyou would have that for
everything. So, like, a a BJTtransistor. So because you can

(08:10):
kind of do this through likeDigikey and Mauser through their
sorting algorithm thing, butthis is would be more of a also
like it would tell you stufflike these are the things that
matter the most about aresistor.

Well or or or like you you you see a category
filled with information you haveno idea what any of the things
are. It could help guide you andtell you, oh, this means this
and this is this other thing,and you can easily read it and
go like, oh, I really care aboutthat or I don't care at all
about this, but it at leastequips you with the knowledge
and the information of what'sbeing presented to you. Mhmm. In

(08:49):
fact, it's funny. In in in my,discussion with with the
Internet work today, I askedthem, you know, okay.
So you you're in college, sayyou're going to get a capacitor.
What you know, in college, whatdo they teach you about? And he
goes, the value and maybe thesize of the package. And I was
like, great. That's fantastic.
That's exactly what they teachyou in college. And I was like,

(09:10):
what about all the otherparameters? Like like, all of
the other ones, we have to careabout those now. And and by care
about them, we have to maybemaybe 50% of them, we actually
don't care about them, but wehave to care enough to ask the
question, do we care aboutwhatever that parameter is.
Right?
And and, I would say tellinghim, like, a large portion of my

(09:32):
job is going through all ofthose parameters and being,
like, this is the ones we careabout, these are the ones we
don't. And, and I think this howto pick apart idea would
basically do that exact samejob. But but but really, really
0 in on all of the parametersthat you don't normally see, you
know, or you don't normally knowyou need to care about or a lot

(09:55):
of traps that screw us up whenwhen picking parts.

Mhmm. And then that was episode that was the
9th episode on that list. Right.Episode 10, on that list is,
chicken and tea leaves. Al'ssimulation recipe.
This is Al Williams. I don'tremember I don't think this is
the his first episode when hewas on the podcast, but this is
episode Al, Al Williams came onthe podcast to talk about FPGA

(10:24):
simulation. So I'm just likewriting Verilog or whatever, but
it's actually like thesimulation part of it. So it's a
really cool episode.

I've always enjoyed Al. He's he's been a
really fun guest. We need to gethim back on again.

You wanna shoot him an email?

Yeah. While we're recording. Yeah. Okay. So let's,
let's go ahead and move into, 1of our topics here.
So so 1 of the things we wereactually discussing at work
today, and I thought it would beinteresting to talk about
because I no longer work at acontract manufacturer, but
Parker still works at, Macrofab.It'd be really fun to talk about

(11:05):
the information that you canassume that your Centimeters
handles for you. And and it'sfunny because we were having a
conversation at work about thisexact topic, and I had a little
bit of a different perspectivethan a lot of the other people
at work, especially from ourquality department. Our quality
department, you know, blesstheir heart. They're they're

(11:27):
fantastic, but I think they havea lot of skepticism about other
people outside of their theirtheir walls.
They basically trust no 1, andthat's kind of what quality is
supposed to do.

That's their job.

That's their job. They don't trust anyone. So so
the 1 of the things we weretalking about is, your your
Centimeters, is it better foryou to assume that your
Centimeters can handle thereflow profile of their ovens,
or is it better for you todictate to your Centimeters,
here's what I need you to dowith your reflow profile? So in

(12:03):
other words, this this came upbecause we were talking about,
parts with bald, leads, likeBGAs, things like that, or any
of the packages that have oddshaped pads underneath, they can
have unique solder as the balls.The balls aren't always the
exact same solder or the exactsame flux that that your

(12:26):
Centimeters uses and they canhave a slightly different reflow
profile.
So the idea is should you relyon your Centimeters to
investigate every single partand make sure that the reflow
profile is correct for it, orshould you be trying to present
that information to yourCentimeters? And I'm curious to
know you what what your thoughtsare on that, Parker.

I have never seen a customer supply a reflow
profile for us. Okay. But I whatI do is when I I'm designing
stuff, and I noticed thatthere's a weird thing like that,
like because some manufacturelike, I've never seen, like, a

(13:11):
BGA actually specify anythinglike that, but I've seen
connectors that are surfacemount connectors that have,
like, a solder slug that's,like, on the lead, and it's,
like, it's, like, extra solderin the connector. Mhmm. I've
seen those parts have veryspecific, like, pace and and
flux requirements basically forthe the component, And if I was

(13:37):
designing something, installthat, I would make sure that
went along with my data package.
Like mice, I would make sure myCentimeters knew that. MacFab
has a feature for that. It'sthe, the the, there's a part
where you can have you can seeall of your components on your
board and you can click the partand you can add a part note so

(13:58):
you can get

this kind of Hey, this one's special.

Yeah. This one's got a weird thing on it.

Yeah. Yeah. Yeah. It's got solder slugs. Okay.
So I'll give you a bit of anexample. I've ran into a part
that from the manufacturer, thepart the the the solder on the
pads are 95 5 lead tin. So it's95% lead, 5% tin. That's a

(14:23):
fairly exotic solder. And and sothe metallurgy that actually is
required to solder that to theboard requires a fairly unique
reflow profile.
So so at the you know, in ingeneral, there's 2 kinds of
reflow profiles. There's justleaded and non leaded, and I
don't wanna say just, but theythey fit into those 2

(14:46):
categories. And then withinthose categories, there's a lot
of parameters that yourCentimeters can change about
their oven to fine tuneeverything. Right? But with this
particular solder, which was 955, it it required it required a
a reflow standard that thatdidn't seem to fit nicely into

(15:06):
either 1 of those.
And so my in my opinion, I dothink it is the contract
manufacturer's responsibility tolook at parts and to identify if
there are any unique thingsabout that. Now I don't III

100% agree. What I was saying though is it's
always nice to be nice topeople, and if you it's it's 1
of those withholding informationonly hurts you as as a designer.
Right? Correct.

Yeah. Yeah. And I was just about to make the
point. I think the bestsituation is you both tell them
and they're responsible. Yes.
I understand. In other words,like, yes, if there is something
unique, flag that somehow. Andand I think 1 of the solutions
that we were talking about, andI think this is totally
reasonable. What what what let'ssay there's 2 situations or

(16:03):
solutions. What you justmentioned where you can add a
part note in the Macrofabplatform, that is probably by
far the best way to do it,especially for Macrofab.
But, you know, if there was ifif if all you had was a
fabrication drawing, perhaps anote on there saying, this this
assembly includes parts withunique reflow, requirements and

(16:27):
then call out the REFDEZ of thatpart.

Yeah. I wouldn't that's fine. I would
actually make the note on, like,the bill of material you're
handling because you're handlingthis off to like a Centimeters
that doesn't have a factoryplatform so and I know from past
experience that a lot of CMsdon't even read the assembly
documents, so, putting that onyour bill of materials is

(16:53):
probably a better bet becausethey are going to have to

read that. They have to read that. Yes.

So putting that note there in, like, another
cell at the end. Right?

Yeah. Yeah. Yeah. Totally. Oh, yeah.
Like, if there's just a cellthat says notes or

Yeah.

Whatever. Put it in big bold. Yeah. It's it's
fine putting it

in the assembly dock. It's just I know from past
experience with other CMs that alot of times they don't read
assembly docks. Right. Right.We've talked about that before
on the podcast.

The the the part about this that is difficult is
so the responsible engineer fora board. Let's say that they're
creating they design a board andthey create a bill of materials
for it. For the most part, thethe design engineer is not an
expert in manufacturing. Theyusually those they usually don't
mix. And so it's not necessarilythe design engineer's absolute

(17:46):
responsibility to include thatbecause it's not.
They may not know that that is aan exact Yeah. Exactly. So
that's why I still think it isthe contract manufacturer
because at the end of the day,it is their responsibility to
provide a properly assembleddevice, but I totally agree with
you. Like, it takes if you doknow there is a unique thing, it

(18:09):
takes a few extra seconds andit's just that much more in
insurance to getting a properlyassembled device. So

now I kinda wanna build, like, a board that
has, like, all these traps onit. Like, it has parts that you
can't you parts that you can'twater wash, parts that have
weird reflow and like

see how put them all on 1 board to make it as
hard as possible

that's hard as possible to make

to make And and actually, you just mentioned it.
That was the the next thing isis water washable. Is, so that
is a note. Can can you assumethat your Centimeters is going
to scrub your bomb to look tosee if parts were water washable
or not? And and the answer is noand also yes.

(18:55):
No as in similarly with the lastsituation, you should note that.
If you know 1 of your parts isnot water washable, you should
absolutely flag that with yourCentimeters. But so so they
should be checking, but also putthe note there. And I remember
back at at, Maccab when I wasworking there, we did bomb

(19:17):
scrubs for water wash or not. Wewe absolutely did do that.
I remember doing plenty of thatactually. And and the a lot of
times, you can just look at aboard and and know if it's water
washable or not. Right? Butyou'd see a part where, like, I
don't know. There's a part witha hole in the top of the of the
the the package.

(19:38):
You go

Yeah. We need to do that. Relays that had
vents. Yep. Or

Microphones, MEMS stuff, pressure sensors, things
of that sort.

Yep. Oh, and you're talking about MEMS
devices, a lot most of thosecan't be ultrasonic cleaned.
Ultrasonic clean isn't really athing for, I mean, it is still
for PCBs, but it's much out offavor nowadays. It used to be a
big thing in, like, thenineties, but once MEM
components started kicking offthe scene, ultrasonic cleaners

(20:10):
just absolutely destroy the themechanics inside of a of a MEM
sensor. So

Well well, so also my buddies over at NASA did
a huge study on ultrasoniccleaning of electronic
components, and they found thatit, degraded capacitor terminal
connections to the actual Thatmakes sense. Of the capacitor.
And so NASA has outlawedultrasonic cleaning for any,

(20:35):
electronic assembly. And so in alot of ways, because of them, we
just don't do it anymore. Eventhough that kinda sucks because
ultrasonic cleaning is awesome.
Yes. It's probably my favoritejob.

It's probably my favorite way to clean car
parts. Yeah. I just put it inthe cleaner and then I hit a
button and then, like, 2 hourslater, I walk by and I pull a
clean part out.

Oh, it's it's fantastic. Yeah.

Yeah.

That and, oh, gosh. My father and I, when we
were working on bikes back inthe day, he used to have a big
jug of carb cleaner where youjust take the carb apart, you
put it in there, go to bed, wakeup in the morning, pull it out,
and it's perfect. Yeah. I Idon't remember what it was
called. It was like carb dip orsomething like that.
Yeah.

It's carb dip. It's made by companies. This was
a b.

That stuff was was amazing.

Barium? No. It's not barium, but it's like
DER Herb dip.

Yeah. It's something like that. Didn't
smell good and it looked like ithad been cancer.

Berryman Chem dip carburetor part cleaner.

Yeah. Chem dip. That's what it was.

It comes in a paint can. Yeah.

It's a it's a paint can, but it has a little
basket that goes in there. Yep.Yeah.

If I if if I'm working on something that can
actually a carburetor thatactually could fit in there, I
use that.

Yeah. Your carbs are usually a lot way bigger
than the 1 that I've picked on.Yeah. I love it.

But ultimate, ultra cleaner does pretty good
work on a on a carburetor. So

Yeah. Oh, I this this was just a hair before I
met you, Parker, but at mybachelor party, this is a total
tangent here. But, at mybachelor party, we spent 3 days
out of my buddy's land, whichyou went to, once or twice. But

(22:24):
we went up there, and we we wentto split logs on my buddy's old
log splitter, and it wouldn'tfire up. So we spent half a day
cleaning a carb and drinkingbeer.
And you know what? That was thatwas an amazing bachelor party.
That was, like, just so muchfun. And we totally got it
working, you know. So I

wanna do that right now.

It wouldn't well, wait until it's cooler outside.
Okay. Going into summer out outto the land out there was pretty
rough.

No. I'm just saying, like, drink beer and
clean a carburetor.

Oh, yeah.

It's like it's like therapeutic. It is.

Yeah. I I like I like getting my hands nasty and
and grimy and dirty working onsomething out in the garage.
And, you come in smelling likeoil and and gas, and it just
feels like you did something,felt like you worked. You know?

Is it how I made this joke I think like last
week to you, Steven, which islike, you know how hard you
worked is how many shirts youwent through.

That's true. Yeah.

So right now it's like 4 shirt days outside
in Houston. So Oh, that's that'shot it is. Yeah. Anyways, back
to, Yeah. So so so mind readingCMs.

1 more bit of information that I think this 1
is a hot topic because I've I'veseen engineers be way on both
sides of the aisle.

And this 1 this 1 ticks ticks my my anger anger
a little bit. Panalyzing PCBs.Let your Centimeters do it.

Yeah. I totally agree. I am in the same pay boat
as you. I honestly I can see areason why you would want to
penalize your own PCB if youwere designing, say, fixtures or
something that made sense onyour side where you wanted to
receive boards in a panel andyou depanalize them or or, like,

(24:31):
say if you had, like, aprogramming jig where you could
put, like, a whole batch in anddo it all at once. But but for
just standard boards, let yourCentimeters penalize them.
It usually makes their life waybetter. And and frankly, in my
experience, they're way betterat penalizing than you are.

Like And it's gonna be cheaper for you. Yeah.
Yeah. Typically. Yeah.
Because the CEM is not going topenalize them in a way that
costs more money. Now I agree ifyou have pre existing tooling
this is the thing though you cantell your Centimeters be like
hey, I have pre existing toolingand here's my board outline and

(25:12):
border locations and soldierslet them penalize it. It saves
you a bunch of work.

Yeah. A ton of work. Yeah. Yeah. And, and on
top of that, like, the waypenalizing works in every EDA
tool is different, and you canget wildly different results
and, you know, knowing how to dov scores or or mouse bites,
properly, your Centimeters isgonna know what works for their

(25:40):
machines.

Yep. What machines they've got. Do they
have a pizza cutter style vscore machine, or is it a do
they have a fancy laser cutter?Who knows? Right.
I mean, you can

ask them And and and I've seen I've I've seen
people penalize stuff in reallyawkward ways where, you know,
they'll do, like, a 1 by 8array. So it's this big long
line of boards, and then maybethat's too long for the machine,
and they can't build the firstor the last board on that array.
And if you just done, like, a 4by 4, or I'm sorry, AAA 2 by 4

(26:14):
array, then they could have goneon done all 8. And and so, like
you were saying, that it can itcan it can bite you in the butt
if you do it. Now I think youcan work with your Centimeters,
That's another option.
Like, they can tell you the thethe parameters and you can if
you still have to be the 1 whopanelizes things, maybe you can

(26:35):
work together to do that. At theat the same time, I've I've run
into this in the past. You canwork with your Centimeters, to
panelize to a particular numberper panel. Like, let's say, you
wanted or or let's say yourCentimeters was, like, hey. I
wanna do a 2 by 2 array.
Well, then you know it makessense for you to purchase in

(26:56):
multiples of 4. So so don'tdon't buy 3 boards when they're
panelizing on a 2 by 2 array.Right? Because then you're just
throwing away PCB real estate.So this really ends up being a
conversation with yourCentimeters and but but for the
for the most part, I absolutelyagree with you, Parker.
Let your Centimeters do it. Youcould just give them the

(27:19):
Gerbers, and they'll handle therest. Yep. Because even you can
get into, like,

copper thieving and stuff like that that happens
on the, fabrication side for theboard when it gets penalized up.
So Mhmm. Making sure it works intheir process.

Also, remember, your board doesn't necessarily
go through 1 machine. It goesthrough multiple. So it had it
goes through paste, it goesthrough reflow, it goes through
your pick and place, and itmight also go through, some kind
of through hole processing atthe same time. So your
Centimeters is gonna know therequirements for all of those
machines. You may know therequirement for only 1 of them.

(27:59):
So it just makes way more sensefor your Centimeters to handle
that. And I think that is out ofall of the things in this list
that we've talked about. So thereflow profile, the soldering of
unique, BGA packages, partswashing, and panelizing, I think
this is the 1 thing that you canabsolutely rely on your
Centimeters to do. All of theothers, it makes sense to help

(28:22):
guide them with extrainformation. This is the 1 where
you can be, like, here's myGerbers, penalize them in the
way that makes the most sensefor your manufacturing line.

Yeah. Again, the only gotcha is if you have
pre existing fixtures. But thebest way to handle that is be
like, hey, here's my preexistingfixture, and this is how it's
designed. And then they can makea panel that fits that and also
fits their processes. Right.
Right. Like how big the railsis? How big is the spacing
between anything that like,little locating pins? Where does

(28:53):
the panel fiducials go? Right.
Right. Yes.

In fact, I've ran into it once where, we had
somebody penalized something andthe rails were they would be
accepted by our machine, but thelaser that senses the board as
it comes into the machine wasjust over a particular cut out.
So, all of the boards would besensed at the wrong location

(29:17):
when they came into the machine,and that was a situation that we
wouldn't have run into if we haddone the panelizing our self. So
that you know, did were theboards able to be made? Yes. But
we had to do extra work to getaround that.

Someone had to stick their finger in there.

Yeah. Yeah. Stop now. So it's little it's just
little things like that. I mean,we we literally had to readjust
our laser sensor on our machinebecause of 1 customer's
panelization that they did.
And they did a totally okay jobpanelizing it. It just didn't
function great with our machine.Mhmm. So I don't know.

(29:58):
Hopefully, that's some extrainformation.
The the the general rule ofthumb is give your Centimeters
information, but you can trustthem.

Yep. And if you disagree, fight us
atform.macafab.com.

I like everything is fight us. Alright.

Go for it. Go ahead. Oh. Well,

I was gonna say let's move on to personal
projects.

Yeah. Yeah. Yeah. Props fan. Project I've
been working on for a long timeand threw away because the, the
use case went away, I guess.
But now it's back. So what propfan was was an automotive
controller for a fan that youknow the temperature sensing PWM

(30:43):
control all that good stuff, andit used the parallax propeller.
This is a long time ago. Oh,yeah. And, this is like start of
the podcast era.
And I've restarted the project acouple times, but now I am fully
on building this. I got likemost of the schematic done over
the last

weekend. Well, I I do have a question for you
real quick. So you said thisuses a parallax propeller.

Does not have

this anymore. Oh, does not. Well, okay. But well,
that Before it did. Yes.
Before it did, it does not now,but you had bought a prop 2. I
did buy a prop 2.

It wasn't for this project though.

Have you turned it on?

Yeah. I turned it on.

Oh, okay. So you so you've played around with it?
Yeah. Yeah. I turned on when

I first got it.

Okay. Yes.

Years ago at this point. Yeah.

Yeah. Yeah.

Yeah. And so, this is so there's a an
automotive fans for electricfans that cool down your your
radiator. For the past, I wantto say 6 or 7 years now, almost
every single OEM out there hasmoved towards brushless fans.

(31:52):
Mhmm. Because it simplifies alot of the control schemes
because now you you basicallymove your drivers into the fan
housing into like the motorhousing and they got heat sinks
so so you actually keep themotor driver like the MOSFETs
cool.
And It also supplies everythingon the other end where like now

(32:12):
you just have to supply it powerin a PWM signal, right? So you
don't have to you have to havelike a big MOSFET driver like
you know actually pulsing a DCmotor and everyone's been moving
towards us you can actually getthese fans like almost every
single vehicle has them now and,so they're easy to get, you can

(32:36):
go to like AutoZone and you canbuy these fans, and these aren't
like your old school, like,dinky little fans that they
would put on, like, hot rods andstuff back in the day. No. These
are 800 watt brushless fans thatare pulling like 3, 500 CFM
after it's gone through aradiator. So that's like with

(32:57):
its, what do they call that?
The static pressure. So theyhave a insane static pressure
where they can just draw throughthese fans, and they're designed
to be run all the time. They'relike a 100% duty cycle fans.

(33:17):
These are like insanelyoverbuilt fans. They're just
beefy.
Yeah. Beefy fans, And they rangefrom I think you can go down to
like 450 watts up to like 875.175 watts is like the range you
can get these things in

also they're intended to run at like what's
the temperature of an engine bay

200 degrees. Yeah. Yeah.

So so yeah. There's they're meant to be run
hot and put away wet.

Yeah. Yeah. Exactly. So I I've I've been
working with I have 2 of thesefans I've been working with,
just so I can get around or juststart playing with the ones a
500 watt, ones an 850 wattmotor. Because they have both
different all these usedifferent PWM schemes.
They all center around like a100 Hertz for the PWM, but the

(34:09):
different scaling of like whatthe actual pulse width is. Some
companies like to do like anormal like 10% pulse width is
10% fan speed. Something likehave a linear scale, some are
like the exact opposite Jeep isthat way Jeep likes to be like
80% is low speed and it rampsdown kind of weird. So if you

(34:33):
get 1 of these you have to playwith like what is it scaling,
for the ramp. So basically I'mdesigning a controller for this.
There's a couple people that dothis out there, but I want to
build an open source 1 that alsohas some extra features. And
we'll get into the extrafeatures later, though. So the

(34:53):
first the first part of thepuzzle is how do you power this
thing? But the the fan is easy.You connect it to 12 volts.
Right? Boom. Done. You can'tit's actually like an 8 gauge
wire too because it's if youthink like 800 watts at 12
volts, you're at like 70something amps. And so yeah you
need an 8 gauge wire, and it'sactually like that's like the

(35:15):
max size all these they all usethe same connector too.
We talked about this before butthey all use the same kind of
connector. I wish I rememberwhat the part number was for
that connector. I'll have topost it because it did take a
while to find what like thecomputer side of that connector
is, but the max size of wire youcan fit in there is an 8 gauge

(35:37):
like TXL style automotivejacketed wire. But, to actually
power the the controller, longlong long time ago, I found a,
white paper from analog actuallyit was from, I think it was from

(36:01):
Maxim before they got boughtout. No.
It's by Linear Tech. That'sright. It was before, it was a
white paper by Linear Techbefore they got bought by
analog. K. And so like if youopen up the PDF that's in this
URL it has all like linear techcolors.

Yep. Yep. Yeah. Classic.

Yeah. So they use an lt8672 for circuit
protection and that's abasically an ideal diode
controller, which is a fancyword for a MOSFET controller.
Right. With reverse currentprotection so it's got like
sensing circuits so it can seeif it's reverse biased. It can

(36:45):
clamp on transient voltages andthat kind of stuff.
I would if if anyone'sinterested in doing any kind of
like automotive devices, atleast check out this white paper
because it has a lot of, like,edge cases that happen in
automotive. Like, what happensif the battery gets
disconnected? Because that thebattery is, like, what's used to
regulate a lot of these systems,and you take away the regulation

(37:08):
parts. Mhmm. And then you havethis like, alternator just goes
going crazy with its outputright and spiking all over the
place because it doesn't haveanything to level out its
output.
Anyways go check this out. We'llput a link in the show notes,
but what I'm using also for isto control, the actual powering

(37:31):
of the rest of the circuit. Youdon't wanna leave this thing
connected to battery all thetime. You want it to turn off
when you turn the switch off.And so I'm actually going to use
the enable pin on this chip.
So when you put your key in, andthen turn on the key, we call
that, like, ignition voltage orswitched voltage. So the switch

(37:54):
voltage will go into the enablepin and when enable goes high it
will actually open up that idealdiode in quotes and allow So

it's it's a load switch.

Yeah. Yeah. It's a load switch. Yeah. But
it's also I'm able to dualpurpose it basically.
Yeah and of course like that's12 volt switch you also have to
make sure you're not gonna blowit up so it has like TVS
protection and all that goodstuff. And then the next thing I
wanted to make sure to do was, Iwant to be able to see like the

(38:31):
temperature that the fancontroller is seen. Because a
lot of fan controllers thatpeople like for auto at the
market, it's just a box thatsits under the hood. So you
don't actually really know whatit's doing. You just hope it's
doing what it should be doingand, it's reading the
temperature correctly and it'sturning on at the right time.

(38:53):
I've been burned by so many ofthese aftermarket controllers
that I'm like, no. I want toknow if my controller thinks it
should be on or thinks it shouldbe off and what temperature it's
reading. So I wanna know thisinformation. So I'm like, okay.
I need a screen that tells methis information.
And so I'm gonna go with acrystal fonts cfa533 TFHKC so

(39:18):
this is a 16 by 2 characterdisplay that's automotive spec'd
but it also has a like, buttonspad on it. So you can go through
menus or you can settemperatures or stuff like that.
You can set all that stuffthere, and it talks I square c.

So there's a problem. That's con that's
convenient.

Well, it's convenient, but this thing is
inside the cab. My PDMcontroller is under the hood.
Sounds like you need to godifferential signal. Yes.
Because I'm going to use a PCA9615, which is an I square c to
differential signal converter.
And I actually use these in thepast and these work great. So
what that does is basicallytakes your your clock and your

(40:03):
data signal and converts it to adifferential signal over 2
twisted pairs. You run thatthrough Ethernet, and then it
goes on to the other side, andyou have another 1 of these PCA
96 fifteens, and it converts itback to I square c.

What do you know what protocol it's talking over?

On the actual, differential signal? Yeah. Don't
know, don't care. You did that.Like, you

just buy 2 of them, plug them together.

And it works and it works great. Yeah. It's
like, we were experimenting themfor DEF CON projects and they
work great.

Now you got me raising an eyebrow. What are you
using? What are you needing tosend over long cables?

Uh-huh. Maybe I'll

you'll have to kill me. Yeah. Yeah. It'd be
cool. That that was a

couple years ago we were experimenting with
it.

Oh, okay. Cool.

Yeah. Nothing current right now, I don't
think.

Gotcha.

So I'll have an ethernet cable between the 2,
makes it super easy to wire upbecause you could pass power,
for the screen and ground, butthen I got to thinking, I'm
like, oh, what if I put a USBport on my display? So I can
plug if so I don't even have togo under the hood to actually,
like, let's say, oh, I want tolook at the serial terminal, the

(41:20):
COM port on my my PWMcontroller. Well, now I can just
hook it up into the display.

That's cool.

Yeah. And I just well, I'm like, oh, I need
a flash of firmware instead ofhaving, like, go under the hood
and open up the box and becauseI'm gonna use a waterproof
enclosure. Instead of having todo all that I can just plug into
the

Right into the display. That's cool.

Yeah. So the big feature I want that no 1
else does for these kind of fansis I want to read how much
current the fan is pulling.Mhmm. And I'm gonna use, like,
an Allegro, you know, passthrough chip to do that. Mhmm.

(42:02):
And this is the reason why, issometimes it's the controller
that fails, sometimes it's thefan that fails, sometimes it's,
you know, something else. Right?Like a relay gets stuck or or
fails to close or something likethat. K. So what I wanna know is
like, hey, if the fan's on, Ishould be pulling some current.

(42:24):
Right? This 850 watt fandepending on what speed it is.
Right?

But but any speed should be a a fairly significant

Should be something. So I actually wanna
plot that, like, at I if I'm,like, not moving, how much
current am I pulling? Like, atdifferent ranges. And that's 1
of the things I wanna do is,like, can you detect how fast
your car is going? Like how muchair is moving through your
radiator?
Can you detect like if you'recommanding let's say 80% fan

(42:53):
speed and you're driving 60miles an hour, can you actually
detect that on the load?

I don't know if you could do that reliably.
You'd have to take a lot of datato be able to plot all of this
and then figure

it all out. But the thing is I wanna be able to
do that.

So I

want I wanna experiment with this. That's
actually another thing is, like,I wanna have speed control
because some, some applicationsare like, oh, if you go above 45
miles an hour or whatever, youwanna make sure you're not
running the fan at all. Becausesome, some car packaging, you

(43:30):
don't want to run the fanbecause the fan inhibits flow
when it's running. Right. And soyou need to be able to test
speed.
And so you I'll be able to needI need a way to, like, detect,
like a speedometer signal, whichis like a spike pulse at 12
volts. But what I also want tobe able to do is I want to
experiment with GPS control. GPSspeed control, and, I've been

(43:55):
playing around with that a bitnight. I I need to work more on
it because I can't get fastlocks. Like, it needs to be able
to, like, lock on within, like,ideally within maybe 30 seconds
of the key turning on.

And

I haven't figured that part out yet. So if
anyone out there knows, like, acheap GPS units that can be
under the hood of a car, now youhave like metal on top of it
which sucks for GPS but to beable to like sense the speed
because that's all it needs. Itdoesn't need location. It just
needs to have, like, real it hasto have, absolute speed. Doesn't

(44:33):
it doesn't even need velocity.
I just need to know how fast itis moving. Sure. Yeah. And then
I want to do like multipletemperature inputs possible, be
able to control off based offyour your AC compressors on or
off stuff like that. So I waslooking and I have all these

(44:55):
like analog inputs, analogoutputs, I have digital inputs,
digital outputs, and I'm likeokay, I need to be able to
protect these low level stuff.
And so, I came over some ideasfor that, but I was like you
know what? There's a couple opensource like engine managements
hardware out there calledSpeeduino and Megasquirt, and it

(45:17):
was like, I wonder how they doit. That was way over
complicating my circuitry. SoI'll post the schematics I found
online, the the open sourceschematics for these these 2
devices. But like for likeanalog sensor inputs, you know,
you have an analog in pin andwhat they do is they first pull
up the pin because most sensorsonly, can pull down.

(45:40):
Right. And then it goes youbasically have a low pass pi
filter. You have like a 1nanofarad, and then it goes to
like a big chunky resistor, andthen you have a 1 microfarad on
the back end of that, and thatgoes into your analog in pin.
And that's enough. Some somecircuits actually put a TVS

(46:00):
diode there on the, like, rightat the connector.
I'm gonna do that, but mostdon't do that even do that. Like
Yeah.

That 2 point They just rely on the filter doing
just Yeah.

The filter does it good enough job at it. Sure.
Yeah. And then there's likechips like the SP 72 1, which is
a SCR which is a siliconcontrolled rectifier, TVS diode
array which has like semi activesuppression of transient
voltages. It's kind of coolstuff.

I have not heard of that.

Yeah. I've seen that used in in some of these
motor controllers or enginecontrollers I should say, so I
would use those. That seems tobe like enter like industry
standard for automotive design.So why make a new standard? Just

(46:54):
copy it.
Right? But I'm really interestedin the current jaw. I want I
want to really see, like, evenif I'm not even, like, detecting
how fast I'm absolutely going,like, can I detect that my fan
shouldn't even be on? Right?Because, like, I'm combining,

(47:14):
like, the engine's running superhot, and so the fan is running
really hard and but I'm, like,not pulling that much currents.
That might be, like, a situationbe, like, hey, something's,
like, weird. Right? You shouldprobably, like, pull over and
figure out what's going on.Like, something's not working
correctly. They're, like, itcould be, like, hey, I'm
commanding the fan to run andnot pulling any currents.

(47:38):
Like, so something's up with thefan or, like, the current fan's
pulling way more power than itshould. Like, I wanna know these
edge cases so then I don't endup, like, you know, blowing up a
motor.

What what what's 1 of your cars? I mean, I know
you're making this generic, butwhich 1 of your cars are you
wanting to put this in? This isgonna go in the red jeep. So you
okay. Is it possible is thereany, actually, I don't know this
about cars.
Do cars report what gear they'rein electrically?

Modern cars, yes. The red jeep, no. Because I
was about

to say, you could just take the gear and correlate
it to an RPM and then that wouldgive you a speed.

You can just read your speedometer. So most
most cars, your your speedometersignal comes off, like, the tail
end of your your, transmission.Actually, even newer cars that
just goes through thetransmission controller. It just
has an internal sensor on theoutput. Anyways, you can just

(48:40):
read that square wave and

And that's good enough. And that's good enough.

I mean that's how the car knows how fast it's
going. So It's good enough for

the car. Yeah.

And so to be able to hook up that sensor,
because I thought about using,like, 1 of those, like, donut
rings for the current, and I'mlike, nah. I wanna use, like, an
onboard sensor. It's more waymore accurate too. So I need to
be able to hook big chunky wiresup to the board. I didn't really

(49:12):
want to do like screw terminals,like a big screw terminal
because, those kinda tend tovibrate loose and that kind of
stuff, and I'm like, you knowwhat?
I want to do, like, battery likelike a like a crimped lug on the
end of the cable, And so I needto put that on the board, and I
didn't really want to do, like,a contact and, like, pass a bolt
through and then have to, like,have a nut on the backs of the

(49:34):
PCB. So I found like s and tstyle threaded inserts are a
thing like Keystone makes them.There's a couple other
manufacturers up there too, butthey none none of them make 1

(49:55):
that's big enough to handle an 8gauge wire, and so I found a
company called Lugs Direct Andthe moment I found, like I think
Steven and I were talking 1night, and I found this website.
I just couldn't get over thefact that it's just Lugs Direct.
I love the name.

They do 1 thing. Well, I mean, they do a lot, but
clearly, they do Lugs. Lugs. And

and they have a lot of different styles like SMT
and through hole threadedinserts that you can connect big
wires up to, and so I found somethat work. I'm gonna go through
hole because it makes sense forthis part And then the howls the
entire thing. So, like, the partthat's inside the cab doesn't

(50:40):
really matter. It's gonna staydry. I'll probably just make
like a 3 d printed enclosure forit.
But for under the hood it needsto be waterproof and needs to
handle heat. It needs to alsohelp reject heat if possible.
There's not a lot that's goingto be going on inside this box,
but you know it's gonna getheated up.

Oh, for sure. So I want to

go with an aluminum enclosure and I think
I'm gonna go with what's calledan EX52 by Polycase because it's
a waterproof extruded aluminumenclosure So I can actually take
the end plates off, I canmachine the end plates with the
ports I need, and then installthe right waterproofing

(51:22):
connectors on the ends of it,and then I can assemble the unit
and install it. Right.

And doesn't it have, the the internals of the
case has places for you to slidea PCB into it?

Yes. Yeah. So you can slide a PCB, like, in
the middle of it. Right.

Right. That's convenient.

Yeah. And so the the PCB will just just fit
snugly inside and be good to go.

Oh, also, it has it has some mounting flanges.

Yes. That was that was the big thing is
finding a waterproof extrudedaluminum enclosure that had
mounting flanges.

Yeah. These are convenient. Yep. So Also, it's
not expensive and neither arethe flanges, or not too
expensive, I should say.

Yep. Yeah. I was actually pretty happy with
it. Nice. So I haven't orderedit yet but the specs make me
happy.
I'm gonna wait till I order I'mgonna order the board first
before I order the enclosure. Sowhat's left to do on the board
is I need to finish up some ofthe circuitry on the schematic

(52:25):
and then do the routing. Andthen I have to design the board
that goes in the back of thedisplay that goes in the cab.
That's it. Nice.
Feels like I'm over halfwaydone. I got I got all the hard
stuff, I think, done andeverything feels like it's gonna
work. Like, the the the thetheory of operation seems to

(52:48):
like, I have all the rightparts, And they're all connected
in the right order. Right.

Now you just need to execute on it.

It's only most I'd say over halfway executed
too.

So Well, I mean executed as in, like, you gotta
have it in your hands and youactually have to, like, turn a
fan. Turn a fan. Yeah.

Yeah. Yeah. A 100%. Because I've made, like,
my Sam d 21, dev board. I madeit turn I've turned this fan.

So Yeah.

It it works. Now it's all the other stuff it
has to do and and pay attentionto. Yeah.

And care about it. Yeah.

And so I am gonna make it open source, so
people can iterate on it andadd, like, oh, I want to be able
to control 2 fans cool putanother PWM you know protection
circuit and run out of the case.Mhmm. So I I wanna make it I
should have there should beplenty of board real estate on

(53:45):
the on this enclosure, orallowed by the enclosure, I
should say, so that people caneasily add, like, another
whatever into it.

Well, there's not a huge amount of things going
on. So No. I think, yeah, Ithink you should be able to do
that fairly easily.

Yeah. Yeah. It might get complicated inside if
we add, like, a GPS unit, anoptional GPS to read speed. That
might be the more complicatedthing, like how do you get the
GPS antenna in there? Or like isthere a way you don't need a
really good lock with GPS justto get speed readings?

(54:22):
Because I know they make GPSbasically speedometers for old
jeeps and old vehicles. So youdon't need like the old
speedometer cable anymore. It'slike how does that thing get a
lock so fast? I don't know. I'veexperimented with, like, a cheap
GPS unit on my desk and, like,it takes like 5 minutes for it

(54:44):
to lock on.
I'm like that's not reallyacceptable.

Right. That doesn't help you very much at
all.

No. So Well, cool. It's fun to

see an old project come back to life.

Yeah. And this 1II need. Because I because I
what what happened is on the redJeep, I burnt up my last
aftermarket controller, and I'mjust like fuck this shit. I'm
building it my

own now.

I'm done buying off shelf ones, and I'm like I
am going to build 1 and use 1 ofthese badass OEM fans that is
designed for what I need it fornow.

Well, now you can just burn up your own
controller.

Well, the great thing is, like, it doesn't
actually have. It shouldn't burnup at all because it doesn't
have any control circuitrybesides the signal. Right.
Right. The the Jeep has reallyweird aerodynamic problems that
mean that you have to run a fana lot.

Challenges. Don't you say problems?

Yeah. It has challenges with aerodynamics.
It's really hard to force airinto the radiator. I'll put it
that way. It's interesting asmodern jeeps also still have the
same problem.
I think it's the problem withjust having a lifted vehicle is
you get a lot of high airpressure under the vehicle,
which is not good for airflowfor your for your fan. I mean,

(56:04):
they got there. Getting or justgetting air through your
radiator because most time youhave a low pressure zone
underneath your car. Yeah. Andso your car has a high pressure
where the radiator is

at. But it wants to

drink air. And it wants to drink air and it
will then that hot air goesunder your car. Right. That is
not a thing on a Jeep.

Also, the original designers, III think
they they used bricks as their,inspiration for for designing
the car. So I don't know likethese kinds of things didn't
seem to be forefront in theirmind when originally designing
these these jeeps. They they hadthey had other criteria they
were looking at.

Yeah. Other criteria. Yeah. And so, like, if
you look at modern jeeps thatare using these 850 watt fans
Yeah. They're running themthey're not running them a 100%
through, it's like a 100% oftime, but they're running them
on the freeway at, like, 40%.
So they're actually runningthem. Mhmm. So I'm, like, good.

(57:03):
So and there's no widespreadproblems with these at all,
like, everyone's just like,yeah. It's like, honestly, the
most reliable jeeps becausethey're not overheating all the
time.
Because they have now a fanthat's able to it. I'm like
good. I'm gonna retrofit thatinto mine and make that work. So
and the great thing is like it'san OEM fan and so you only have

(57:25):
to pay you you can pay OEM andaftermarket prices. You don't
have to go buy a specialty fannow.
So it's like a quarter of thecost. Yeah. So I'm look I'm
looking forward to it. Like,people need to be at least in
the automotive world need tolike, for hot rods need to
really start looking at thesefans. And I'm hoping to provide

(57:46):
a controller that, like, enablespeople to use these fans because
they're awesome.
They levitate.

They'll they'll float around on their own?

They'll float around and hover.

Yeah. Wow.

It's how much power how much They're just
beefy. Well, yeah. Because youhave 850 watts of motor and
they're thick because they,because it's a it's a brushless
motor and how they can design itso they can inset the motor more
into the hub. Mhmm. And so theycan have basically steeper
blades.
Pitch angle? Pitch angle, andthey have more torque than a

(58:21):
standard brushed motor. And soyou can really get the air
moving that static pressure up.Right. So looking forward to it.
I'm I'm gonna try to work onthis over the, 4th July break.
That's my goal. Just work on it.So maybe I might not get to,
like, fully routed by then, butI'm hoping to get, like,

(58:43):
schematic done, parts put on theright spot on the board, like,
have a border outline that Iknow will fit in the enclosure
That's it. Alright.
And if you wanna talk about thiscontroller, hit us up at
form.macfed.com. Let me know howwhat feature you want it, like,

(59:06):
if I didn't mention a featureor, like, the GPS. I really
wanna know about the GPS becauseI'd love to have that as a thing
you can do if you need thatfunctionality. Yeah. Ideally,
you shouldn't need it because ifyou start moving fast enough
more air comes in and it coolsdown the temperature and then
the temperature sensor goes,okay, I can turn off the fan.

(59:26):
That usually how it works, butsometimes you have edge cases
where like like, in my case,aerodynamic problems. But some
people might have packagingproblems where, like, there's a
engine right behind the fan, sothere's not a lot of airflow to
begin with. So, like, when thefan the fan doesn't know to turn
off because it doesn't hit that,like, tripping point to turn off
when it's driving down thefreeway. There's a lot of edge

(59:49):
cases, so having speed controlis a good thing. Right.
Yeah. Let us know.Form.macrofab.com. And thank you
for listening to circuit break.

Stephen Kraig (01:00:02):
We were your hosts, Stephen Craig And Parker
Dohlman. Take it easy.

Parker Dillmann (01:00:08):
Do you wanna do the outro this time?

Stephen Kraig (01:00:10):
Thank you, yes, you breaker for downloading our
podcast. Tell your friends andcoworkers about the circuit
break podcast from Macrofab. Ifyou have a cool idea, project,
or topic you want us to discuss,let Parker and I and the
community of Breakers know. Ourcommunity where you can find
personal projects, discussionsabout the podcast, and
engineering topics and news islocated atforum.macrofab.com.

All Episodes

Episode Transcript

Popular Podcasts

On Purpose with Jay Shetty

Stuff You Should Know

Dateline NBC

.css-15opob5{left:0;position:absolute;top:0.8rem;} All Episodes

.css-14f5ked{margin:0;word-break:break-word;display:-webkit-box;-webkit-box-orient:vertical;box-orient:vertical;-webkit-line-clamp:2;overflow:hidden;}EP#438: Solder Slugs

Episode Transcript

Popular Podcasts

.css-r6mb8g{margin:0;word-break:break-word;display:-webkit-box;-webkit-box-orient:vertical;box-orient:vertical;-webkit-line-clamp:1;overflow:hidden;}On Purpose with Jay Shetty

Stuff You Should Know

Dateline NBC

All Episodes

EP#438: Solder Slugs

On Purpose with Jay Shetty