There is pic related IC and its really rare. (a bbd delay, but it doesnt really matter which one)

There is pic related IC and its really rare. (a bbd delay, but it doesnt really matter which one)
Theoretically, could one remove the "case" and reverse engineer it? (And make schematics)
How hard and realistic is that?

Attached: MN3009.jpg (1100x768, 290K)

Other urls found in this thread:

siliconpr0n.org/wiki/doku.php?id=decap:epoxy
project5474.org/index.php?title=MN3009_Panasonic_43
twitter.com/SFWRedditImages

It's possible, in theory. Practically, if the Chinese haven't done it (coolaudio?) it's not going to happen. Not like there's labeled components in there.

siliconpr0n.org/wiki/doku.php?id=decap:epoxy

how do i make the chinks make a clone? also coudlnt you distinguish the components under a microscope? you wouldnt have the values, but you would have a rough schematic. everything else you can work out by building it with discretes or just simulating and trying out different values

Attached: here2.png (5496x7143, 461K)

Looks like die shots are available project5474.org/index.php?title=MN3009_Panasonic_43

If you can reverse engineer it from that and design a clone from discrete components, you could have a tidy little side gig selling them to pedal and synthesizer nerds

I wasnt aware that someone has already made so much progress on these. Thanks for the information.

Howard Davis successfully made a Panasonic MN3005 BBD chip clone. He’s famous for designing a popular vintage delay pedal(Memory Man) that used those chips. search for more info on how he succeeded.

Sounds really interesting. But the only thing I found when looking him up, is that he is the original designer of the mn3005. Your post is implying that there is more to find. Is there an interview with him or what exactly do you mean?

The reason they are expensive is there's no mass market for them, not that it's some big secret how they work. It's just a bunch of capacitors with buffers between them. Do you really have millions of dollars laying around to spend on masks and fab usage or are you just larping on the internet?

the only decent thread on Jow Forums

Attached: thumbsup1360825940361.gif (317x501, 2.41M)

No I havent. And I never said that I had. I talked about discrete components. SMD is cheap and pick and place machines are pretty affordable these days.

Thank you, glad you like it.

nigger just use an attiny. 256 stage digital BBD?

costs a dollar and gives you up to 16 bit precision at 256 stages.

These are available on eBay for less than 5 bucks.

Reverse Engineering it is not going to be easy unless you already have electronics knowledge. If you want to do it anyway, you will need a few of these because you'll damage many of them while decapping them with nitric acid.

Attached: s-l1600.jpg (900x636, 63K)

>design a clone from discrete components

The problem is a 256 stage bucket brigade device needs 256 capacitors and 256 FETs. The chips are rare, but still less expensive than what it would cost to assemble that out of discrete components.

Those are likely fakes. Chinese scammers etch a new label onto some dime a dozen chip of the same package. These are some fake MN3007s I got tricked into buying.

You don't need to reverse engineer anything. The internal circuitry schematic is right there in the datasheet FFS.

Attached: DSCN6278[1].jpg (2272x1704, 929K)

Attached: Capture.png (1080x566, 48K)

>the only decent thread on Jow Forums
and you just had to shit it up

>Mn3009 for sub $100
lmao no

But are you sure thats all thats inside? If I were the developer I wouldnt release my 1:1 schematic, but only a dumbed down version to stop others from directly cloning it. Still thats a great point to start from.
And exactly that precision is the problem. Analog is not perfect, and thats the ""great"" thing about it. There is a reason why analog synths are still being used, even tough digital ones are cheaper and more reliable.

>If I were the developer I wouldnt release my 1:1 schematic, but only a dumbed down version to stop others from directly cloning it.
If you are a second-source manufacturer you'd be anxious to show that your chip was EXACTLY the same as the original.
BTW, the original IBM3277 Display System used a shift register-based RAM for the screen buffer. The bipolar equivalent of your BBD.

I am pretty sure there is a datasheet

>If you are a second-source manufacturer you'd be anxious to show that your chip was EXACTLY the same as the original.
So you are saying that these schematics are probably correct?

>The bipolar equivalent of your BBD.
what does that mean?

>Analog is not perfect, and thats the ""great"" thing about it

add random or predefined noise or loss to give it "character"

you probably won't even be able to replicate identical characteristics if you don't use the original toolset.

with digital, you can easily calibrate losses manually, if you want.

>tfw realizing that high fidelity "audiophile" quality is actually a scam

Attached: wat.png (853x482, 506K)

>The problem is a 256 stage bucket brigade device needs 256 capacitors and 256 FETs. The chips are rare, but still less expensive than what it would cost to assemble that out of discrete components.
His expensive is this chip? Caps and FETs can be bought for under $0.01
Either way, a discrete version is probably going to be shit due to all the stray conductances everywhere

Get an original, do precise measurements on signal and distortions, replicate it in software. I don't see why it can't be done. Given enough input data one can precisely define what sort of analog characteristics a given circuit has.

Chiefly, the problem lies in finding discrete components that match the distinct specification of the FET and capacitors found in the integrated circuit. I'm assuming there is no technical specification available for the individual FETs found in the IC, so one would have to measure and compare. What sort of input impedance and frequency response do the transistors of the IC have?

i think this entire thread is an X Y problem or an outright contradiction.

OP wants audiophile grade gear - expensive inefficient old-timey crap by definition - but doesn't want to spend the 5 bucks it costs.

>how do we make chinks make clones of this
>those are shit because they are chink clones!!! do not want
What do you want?

>>those are shit because they are chink clones!!! do not want
when did I say that

Coolaudio makes reverse engineered ships already

>Fake IC's don't exist on ebay
you clearly haven't been at this long

see I never said anything about audiophile, noone in here did. And a mn3009 costs around $100 per piece, not 5. (that is because they stopped manufacturing them in 19?? and the demand is high,not because they are special)

I know, but not this one

>>you probably won't even be able to replicate identical characteristics if you don't use the original toolset.
Nobody tried it yet, so I don't know about that

>with digital, you can easily calibrate losses manually, if you want.
TAL Audio made a digital version of the juno chorus, which sounds nice, but is nothing compared to the "real" one. And unless you are someone like Urs Heckmann who has years of experience and does makes a living of making analog emulations of synths, you wont come close to a perfect, or even believeable result.
Analog modeling COULD theoretically work tough (Like the Roland Cloud or JU06, where I personally cant tell the difference between analog and digital). Still, you somehow gotta fit this in a 8 pin dip ic, and desu, I think ttun it would be cheaper to replikate a real one)

>I never said anything about audiophile, noone in here did.

I assumed, because you're either retarded, or want to sell to retards.

>And a mn3009 costs around $100 per piece, not 5.

maybe. I just checked ebay

>(that is because they stopped manufacturing them in 19?? and the demand is high,not because they are special)

yeah they stopped manufacturing them because we have digital shit now, and because the professionals know that BBDs are a meme at best

if you want it done really professionally looking, I would use atmel chips and simply repackage them so the chips match.

if you just need to replace a single chip, I'd mount it on a proto board and just plug it in place.

> but is nothing compared to the "real" one

I bet they don't know 'bout machine learning :^)
this is a perfect job for RNNs.

sounds like an interesting niche, too bad I'm swamped with other shit.

>chips match

pins match

>I assumed, because you're either retarded, or want to sell to retards.

>be me
>own a small synth repair shop
>one day someone brings a old juno 106
>the chorus is broken
>try to replace "the" ic
>buy a replacement online
>realize that I am retarded because of that
makes sense, doesnt it?

option A) have the customer pay for it? maybe? is that such a novel idea you couldn't think of that?

option B) look for alternatives!

Just buy a mn3209. You're not going to be able to reverse engineer it if you're asking Jow Forums. You would (probably
) need extensive knowledge on analog ic design ( which most universities teach at 3rd and 4th year, and at the graduate level electrical engineering).

I don't know exactly how this bbd works, but looking at the schematics and frequency response, it seems like the mn3209 or other 256 stage bbds would work as a replacement.

If you wanted to make a discrete version, many things don't transfer completely from an IC to discrete (e.g, the capacitors may be parasitic).