Author Topic: Question to all forum members - project direction (participation requested!)  (Read 13439 times)

Artur D'Assumpção

  • Moderator
  • Hero Member
  • *****
  • Posts: 661
  • Karma: +23/-0
  • Sintra, Portugal
Hi guys,

I am addressing to all of you asking for your comment on the following subject. I can't stress enough the importance of your contribution since we've reached a critical stage of the project and have to make crucial decisions, some with no returning point.


As you may well know from recent news we have finally concluded our Beta 2 version for the VCA consoles. We're are still testing it on a real console, but we are very confident we've finally reached a very good & high quality hardware design, which, after much testing in a real production environment,  will eventually be tagged as the final 1.0 version. The next phase, starting in the following weeks, will port this same architecture into the Ultimation systems. 

Through the several beta versions we've worked on we were able to try out several architecture approaches, some less complex in terms of hardware but which didn't delivered the performance and features we were hoping for. This latest design is the best of what we could achieve being restricted to trough-hole technology. Despite this limitation, no corners were cut and it fully accomplishes our initial goals, meeting all the hardware/feature requirements we've planned. Unfortunately this comes at the expense of a slightly higher component count and real-state usage (card size), when comparing with the earlier simpler designs or if we've opted instead for SMD technology.


- please refer to the images attached below -


The VCA I/O card ended up with a component number close to 210 (including connectors). It may seem much, but in our experience it's a very easy and straight forward card to assemble, since most components are very simply laid down (next to each other), and many of them are ICs anyway. Despite this, they are a total of 210 and if you have a system with 8 of these cards we are talking of assembling 1680 components just for the I/O section.

Although much much simpler, there will be also the Total Recall+Keyboard+Transport card (this one is also much smaller) and the Motherboard (which is more like a back-plane connector with a few components) to assemble. The Ultimation I/O card will be also a bit simpler (less components) than the VCA I/O card, but similar in terms of complexity.

Now that we are finally able to show you how a final through-hole card is going to be, it's much easier to assess the complexity of the entire DIY project, considering the complete set of cards a system will need and the effort necessary to assembly it.

Knowing this information, we believe that at this stage it's an obligation to ask the entire community how they feel about this. Is this a project you, as future clients, will be willing to handle as a DIY project? Or, would you instead prefer we abandon the DIY goal and do it as a ready-made system (using SMD technology)?

I know this goes out of the initial project motivation, but we believe it's paramount to reach out to you all and hear what you have to say about it before we pursuit any further. This way we can decide knowing that whatever decision we take will meet the overall forum opinion.

For us eventually moving to an SMD route will have a big impact in terms of business model, capital investment and commitment. So it's very important to get this right from start.

Please be as complete as possible in your comments and reasons behind your opinion. Also lets not be shy to rise questions and start a discussion if necessary. Anything that help us steer in the right direction is welcome.

Thank you,

Project Manager

Artur
« Last Edit: February 22, 2016, 02:51:10 PM by Artur D'Assumpção »

sintech

  • Global Moderator
  • Hero Member
  • *****
  • Posts: 839
  • Karma: +61/-0
    • Bristol, UK
My opinion, it’s always going to lean towards the original workflow, ha:

From a commercial stand point, SMD is the way forward. Trough hole is great from a DIY point of view, and that only helps a very small amount of people. (DIY, like DIY computers appeals to one kind of person)

If you consider it as a product, it’s ability to run the original G 4.10 software and act as a modern 1:1  plug and play replacement, will only go in it’s favour... people like CLA, Clearmountain, just want to run in auto pilot., and they have to be your target customers

I’m customer number 1

A lot of new people don’t get the G software, and how it makes records.. it’s about feeling the music, and reacting whilst in the zone, and not tinkering with a mouse!, to looking for a fader in PT, that breaks you out of the zone!

Lot of alternatives solutions, seem to loose features to the original, which is vast.

It’s ability to be open source is a bonus., and new software will be attractive to new users and maybe people that don’t have a computer.

I do wonder how well other solutions are doing commercially? (here’s the competition)

Pelle’s I liked, but it lacked Recall at the time.

The Tangerine system, is VCA controlled via midi, they didn’t put up a convincing video, showing the interface in action… it’s very similar to the a Wizzer system offered in the late 90’s that did banks of 8 VCA’s to midi, scalable. I wonder how many studios bought that?

Dramastic, maybe showed the product doing just recall over three years ago.. and then nothing., I put this down to the magnitude of writing an automation system from scratch.

matt

  • Jr. Member
  • **
  • Posts: 93
  • Karma: +2/-0
    • Reutlingen, Germany
Hi Artur,

although I don't have a 4K (yet), if this all was for my 5K these are my thoughts on this :

I'd compare the time to spend for soldering, testing, debugging (my soldering) assembling and such against the cost of a SMD solution that could be manufactured by a machine in larger numbers more easily (?).
Although possible, I don't want to solder SMD.
So if I want to DIY, SMD is not an option.
But if the number of requested boards could be manufactured maybe in China or such for small(er) money at a good quality (althogh these two thigs might cut each other out), the price per I/O board would drop to a number where it doesn't make sense (to me) to DIY anymore.

What would be helpful in such a comparison are prices for eacch of the options.
Could you help with that ?

So finally, I'd opt for a machine manufactured I/O board (due to its complexity and number of boards needed) and have the rest of the boards as a DIY option.

Another question is how do I repair something if a board brakes ?
With SMD I'd probably have to replace the whole board whereas with the DIY board I could swap an IC easily especially if I chose a socketed IC layout.

So, for me personally, it finally comes down to a price discussion.
And it's hard for me to really value what you guys already put in in terms of time and money for your R&D ......

Regarding the larg(er) base of the 4K / 6K / 8K consoles and that every owner could use this computer, I think it's best to prepare for a larger amount of cards needed to get back some ROI for you developers. If the cards are not too expensive, they're accepted by more people and such more 'motivation' will be send back for further development.

Just my 2ct

Matt
« Last Edit: February 23, 2016, 09:28:08 AM by matt »

marcmozart

  • Sr. Member
  • ****
  • Posts: 483
  • Karma: +29/-0
    • Frankfurt/Germany
    • Mozart & Friends
Hi Artur,

I very much like the work you guys have been doing. My vote is for through-hole, it's be easier to service and repair if anything ever breaks.

Also, I think most people who run an E/G-console have some DIY-skills, otherwise you can't keep them running economically.

The "market" for this product is non-existing. It's something enthusiasts are doing for and with enthusiasts.

Just my 2 cents...
1992 SSL 4048 G-Series
Mix Engineer Blog
http://www.mixedbymarcmozart.com

Artur D'Assumpção

  • Moderator
  • Hero Member
  • *****
  • Posts: 661
  • Karma: +23/-0
  • Sintra, Portugal
Hi guys,

Thanks for all the replies so far. Just to clarify, this discussion is only to access what path should we take regarding the hardware format (throughole-DIY or SMD-non-DIY). Regarding the software our commitment is still the same, we intend to pursuit the same workflow of SSL so the transition to the new system is seamless to the user.

Here are the pros and cons of Throughole-vs-SMD for this project as I see it (in no particular order):

Throughole:

* It's intended only as a DIY solution;
* We won't be able to provide a fully assembled system due to assembling cost - assembling is the user responsibility or a third party that wishes to make some extra €€€;
* Serviceable - easy to troubleshoot and replace any component;
* More real-state (cards are bigger) - PCB price point is higher, but the rack capacity is still the same size, either SMD or Throughole makes no difference;
* Higher price - DIP/through-hole components have a higher price. Due to more PCB real-state,  PCB size is also more expensive. Price ratio can potentially reach 2:1.
* lower packing, handling and exporting fees - in the limit we can provide only PCBs, microcontrollers and a few selected components, while the user purchases the rest of the BOM directly from electronic suppliers (ex: mouser or farnell). Exporting to US is as easy.
* lower investment necessary - We just need to manufacture the PCBs we need and provide the materials necessary (leaving the rest to the user). For SMD we need to rent a pick & place and build a minimum amount of fully assembled cards to make it viable.
* No competition - with this format we will not be competing with the other solutions.


SMD:

* Non-DIY solution - system sold as any other off-the-shelf product;
* lower component count - with SMD we can take advantage of more modern technologies which pack more features in one single IC, reducing the total number of components necessary;
* cheaper technology - taking advantage of newer IC technologies translates to a cheaper cost overall, we believe we can get a ratio of 2:1.
* smaller PCBs necessary - reduces price of PCB
* virtually non-serviceable - service is much more limited in SMD
* higher packing, handling and exporting fees - we need to ship the entire system, fully assembled and tested. Europe is not problem, US might be though.
* much higher investment - this means we need to commit to a minimum amount of fully assembled cards to make it viable.
* competition - we will be competing in the same league as the rest of the solutions, confined to similar business models, market risks, market rules and competing options.

I know much of this is from a business point of view. Although we didn't enter this project to get rich or make money, we also don't want to loose and at least pay for the personal investment, so we have to consider all these angles.

By the way, I am curious, how much would you be willing to pay for a full system? (consider also if it's DIY)

I would love to hear your thoughts... 

Cheers,

Artur



« Last Edit: February 23, 2016, 11:32:30 AM by Artur D'Assumpção »

StarF666

  • Jr. Member
  • **
  • Posts: 80
  • Karma: +1/-0
    • Berlin, Germany
Hello guys,

I finally had a talk about this today with my partner and in short, we would go with either solution. In my honest opinion, you already named all the pros and cons of the different aproaches and I'm sure everybody in this thread would be ok with both. SMD and ready to use system will probably target a broader user group. And while I agree with Marc, that most owners have some servicing skills and endless dedication to their desk, not everybody is so much into soldering. In fact, most of us just want to focus on making music. On the other hand, those, who really dug into the giblets of the beast, like the idea of "all access" to even the last bits and pieces although I have to admit, that DIY can be a pita, especially with such a high parts count.
I think you should ask yourself, what you want to get out of your work. If you want to make money, I'd go for all-inclusive SMD. But that would mean running a business and all that. If you just wanted to dedicate something very big to the SSL legacy, go DIY.
One last thing: Maybe there is a third way? You could start with the DIY system. Let us "freaks" be the early adopters. You'd get a lot of feedback and the DIYers would be kind of beta testers. Maybe there is something to tweak in the circuit. It might be possible to change small "bugs" on a DIY board. The software could evolve and once you are satisfied and feel you are ready for running a business off of your design, you could simply make SMD cards with the same footprint/dimensions to stay compatible the DIY system. That might be a little more expensive but not to an extend to make it uneconomical. I doubt, "real estate" is that much of an issue with an SSL.
I want to close my 0.02$ with the answer to your question for the price. I'm not sure about an absolute number. If I'd ask my partner, I think up to 2000€ for our 48 channels would be absolutely ok. But we agreed that we would absolutely go SMD as long it's not more than 1,5x the DIY price. Above that we would think about DIY.

Anyway, I have to tell you once again, thank you for your dedication.

All the best,
Martin

waltzingbear

  • Sr. Member
  • ****
  • Posts: 329
  • Karma: +13/-0
    • Portland, Oregon
    • Waltzing Bear Audio
I have a variety of feelings regarding what direction you might take.

1) thru hole components are going away, period.

2) SMD does not mean it cant be partial DIY. We all need to be able to do SMD, it is a reality.

3) The comments about vending to others is correct, many will not be able/willing to do a DIY project.

4) the producer and /or engineer should be making music, not soldering.

5) The total real cost will be less with SMD built boards than DIY, but the incremental cost can be less with DIY, either SMD or thru hole.

6) A mix of types is not out of order, the major component boards prefab, lower component boards DIY. Not sure if thats effective or not.

7) The numbers will be small. DIY spreads the money risk.

thats all  that pops into my head at the moment.

Thanks for your efforts so far.
Alan

Alan Garren
Waltzing Bear Audio

KoleThomsen

  • Newbie
  • *
  • Posts: 11
  • Karma: +0/-1
Through-hole/DIY here.
If you haven't got the skills to solder the PCBs, you probably haven't got the skills to install the SMD-version either. You could just get your local tech to assemble the DIY-solution for you, and add it to the total cost of the installation.

Great project - keep up the good work!  :)

Artur D'Assumpção

  • Moderator
  • Hero Member
  • *****
  • Posts: 661
  • Karma: +23/-0
  • Sintra, Portugal
Hi guys,

thank you so much for the great inputs! It will help a lot in our decision.

Just to clear out, from our current cost drafts we think we can get an SMD card production costs near half the price of the through-hole price (we haven't really materialized a SMD design we can't make a 100% accurate prediction, but it will be cheaper).  We've been a long way optimizing costs and we already were able to reduce almost 65% of the overall price per through-hole card.

Like Alan said, for us the risk of a through-hole solutions is spread among all clients, while on a SMD solution the initial investment has to be much bigger to really make it worth. - there might be here an option which is to sell on pre-order basis (assuring a minimum number of units and that all funding is guaranteed from start).

SMD DIY is out of the question, you have to know what you're doing and have the proper tools. Many things can and will go wrong with less experienced users. Said that, it doesn't mean we can't offer an SMD system IKEA style. So everything is disassembled (cards are populated) and you just assemble it.

Offering a ready made system also might bring other issues, such as, conforming with country laws, legal aspects associated to warranties, fully testing the systems, etc.. etc... This is all something we can't neglect since it can eat up a lot of the overall price.

Rest assured we will find the best way to make this work and everyone happy! :D

Cheers,

Artur
« Last Edit: February 23, 2016, 09:35:28 PM by Artur D'Assumpção »

PelleG

  • Full Member
  • ***
  • Posts: 103
  • Karma: +8/-1
    • Stockholm, Sweden
    • Automan
Question to all forum members - project direction (participation requested!)
« Reply #9 on: February 23, 2016, 10:38:02 PM »
Maybe I'm not allowed to talk here :-)
Well, being a punk rocker I do it anyway.

I thought about debugging. Through hole and smd. Even though studio owners know their way with an iron and a scope. Debugging computer boards is a different thing.

I also wanna warn you about dil sized micro controllers. I (and a lot of others) had a lot of trouble with the old school packages. They're not made and tested thoroughly for production. Merely prototyping.
Atmega may be different, not sure.
I haven't had a single failure from my smd batches. Several on the DILs. Even the multiplexers DILs are failing frequently for me.

I still had through hole on my backplane on my first batch because I use it for other things as well. It's a good way of breaking out tons of io from usb. I then could easily tweak new kinds of prototypes.

..by the way, have you all seen Jeff Steigers new take on this?
I wonder how many automation nerds we can fit in a boat :-)
« Last Edit: February 23, 2016, 10:40:28 PM by PelleG »

Artur D'Assumpção

  • Moderator
  • Hero Member
  • *****
  • Posts: 661
  • Karma: +23/-0
  • Sintra, Portugal
Hi Pelle,

Of course you're allowed, all input is welcome! ;)

That's a very good thing to know which I wasn't aware of the DIL/DIP packages. So far I haven't had any problems, but I guess for big batches of components the probability of hitting a faulty one is higher, giving your experience. We'll take this in consideration for sure! - Have you found the same problems with components other than micro controllers?

I am curious, what was your threshold for the min amount of boards produced (in the PCB Assembly factory) to make it viable? (rough estimate)

Did you also found out that you could get almost a ration of 2:1 in price vs throughhole?

Yeah I already heard about the  Jeff Steigers, it seems for such a small market everyone is jumping in. Something will crack.

Thanks for your inputs! :D

Cheers,

Artur
« Last Edit: February 24, 2016, 10:29:24 AM by Artur D'Assumpção »

nick_hepfer

  • Newbie
  • *
  • Posts: 1
  • Karma: +1/-0
Hi everybody; Hi Artur,

my vote goes to the through-hole. Mainly because of trouble shooting, changing components, etc.

I also see the benefits of SMD and why it makes more sense in some circumstances.
For selling the whole, assembled system, on the market, this may be the best way.

But keeping the initial thought "A computer from the community - for the community" , maybe the DIY trough-hole version is more appropriate inside this community.

So why not doing both? A DIY for the core of the community and a fully assembled (SMD) for selling to the outside world.

An IKEA style semi DIY/SMD thing doesn't make much sense in my opinion, because owing an old SSL you are either a world class engineer with an assistant, who simply wants (and can afford) to buy a running system; or you've refurbished your console by yourself, which is why a bit of soldering doesn't hurt.

Since the trough-hole version requires less upfront investment and is less "risky", we should definitely keep that option for the community.
A high priced, assembled SMD version could be sold as a business; but you really have to consider the (shrinking) market for those consoles and of course you have to compete with the other manufacturers..

Just my thoughts... :)

Cheers,
Nick

Artur D'Assumpção

  • Moderator
  • Hero Member
  • *****
  • Posts: 661
  • Karma: +23/-0
  • Sintra, Portugal
That's a very good point also! Thanks Nick! :D

Artur D'Assumpção

  • Moderator
  • Hero Member
  • *****
  • Posts: 661
  • Karma: +23/-0
  • Sintra, Portugal
Hi guys,

I am curious about your opinion of a semi-DIY solution? This semi-DIY would be SMD (already assembled) and all the rest would be left for you to do, ex: connector soldering, any left-over through-hole component soldering, connecting cables, mounting the case, etc.

Does this make any sense, versus the full-DIY (throughhole) and the complete assembled system (SMD)? Or do you share Nick's opinion on this?

Cheers,

Artur

StarF666

  • Jr. Member
  • **
  • Posts: 80
  • Karma: +1/-0
    • Berlin, Germany
I think I'm with Nick. Do it SMD or DIY or do it both (and make connectors/cards compatible) but don't mix it up itentionally. I'm still for starting with a DIY solution and continue development with the community and make a commercial solution once you feel your ready.

Martin