Recording RAW to SSD

Started by PeteTomov, June 05, 2013, 01:32:21 AM

Previous topic - Next topic

0 Members and 1 Guest are viewing this topic.

PeteTomov

A buddy and I had a crack at making a SATA to CF adapter over the weekend. The technology is nothing special and adapters to use CF cards as SATA drives already exist, but until now I guess  nobody's had reason to want one going the other way.

For the purposes of the test we hard wired our adapter to a Lexar USB 3.0 reader, since this was a DYI job and we couldn't make the adapter small enough to fit into a CF slot, and I didn't feel like hard wiring it to my 5d  ::). The important part is that the OCZ Agility SSD we had for the test benchmarked at exactly the same speed as when it was connected via the SATA cable(around 200MB/s write speed).

While this is as far as the actual test went, yesterday I broke out my mad design skills (not really) and designed a box that screws into the bottom of the camera like a battery grip and holds an SSD and a big enough battery to keep it going for at least 2 hours(around 6000mAh). Then I got my buddy to design the actual adapter so that it fits into a CF card body(that's sorta what he does for a living) that connects to the SSD box via a SATA cable and we've got a complete solution that record as fast as the 5d can write to a cheap-ish(cheaper than CF X1000 cards)memory.

We can't actually manufacture this our self so we contacted some manufacturers to get prices and I was just wondering if you guys are generally interested in that sort of a device. We're probably going to have to order at least 200, so I was thinking that if there's interest maybe I can start a kickstarter project. Obviously, the more we order the cheaper they are going to be so I was even thinking that if we can get a lot of orders we can donate some of the money to Magic Lantern guy to continue doing their magic  :).

Anyway, I just wanted to get your thoughts on the subject. I will have some renders of the design in a day or two, and I'll also update you on potential price as soon as we get one.

Audionut

I'm sure you will get much more interest if you can post some video of the design in action, showing things like recording speed.

yobarry

Quote from: Audionut on June 05, 2013, 01:47:50 AM
I'm sure you will get much more interest if you can post some video of the design in action, showing things like recording speed.

I agree, if you can show examples of what's capable and get yourself a nice battery grip mimicked design, you might find yourself  with a sweet kickstarter!

PeteTomov

Quote from: Audionut on June 05, 2013, 01:47:50 AM
I'm sure you will get much more interest if you can post some video of the design in action, showing things like recording speed.

That is a really good point.

If anyone else wants to see anything in particular, keeping in mind what we have for now(read the first post), post what you want see and I'll get to shooting some videos it in the next few days.

hirethestache

I too would like to see proof-of-concept. Im working with a colleague that is an engineer at Makerbot on an interface, and it'd be great to see results.
@HireTheStache
www.HireTheStache.com
C100, 5D3, 5D2, 6D

mogs

This CF extender may be helpful in your research.

http://store.calexium.com/en/316-cf-extender-290-mm-with-flat-ribbon.html

Using a CF reader as a lab rat is an excellent idea.

Roman

So just to be clear though, it hasnt been tested connected to an actual camera?

As that's half the battle by the sounds, making sure it's all a format etc that the camera is willing or capable of interacting with.

I wouldnt be getting carried away with getting 100s manufactured etc at this stage, if it's not been proven to work with a camera yet.

Cityeater

Good to see some progress in this area, thanks for your hard work. My concern is that for those of us that work with camera rigs the extended battery grip wouldn't be ideal with the equipment we've got and might make some gear (for instance follow focus) incompatible due to the new lens height.
For me, I like everything you've suggested except if I could choose I would prefer to have a simple external box that I can mount. I need a little more flexibility and I find the battery grip inconvenient.
Looks like a good project.

On a side note could I suggest a mounting option on the bottom (if you go that way) so its not a single screw but also has a registration pin like a normal film cam for base plates and so forth.

EDIT
It hasn't been attached to a camera yet? I think you'll find that's the tough bit, its more than just a physical interfacing problem. Looks like a good project though, best of luck.

heavygrafix

i also thought a solution as battery grip isn´t optimal, so i designed some SSD housing that is directly connected to the Rig Rails. A Battery Pack (10Ah), that is powering both, camera and SSD (or some other stuff on the Rig) is (in my mind) the best solution.

if you did not saw my rendering yet:


5D Mark III, SP 24-70 2.8 Di VC, KB 1000x 64Gb, KB 1050x 128GB

PeteTomov

Quote from: mogs on June 05, 2013, 03:25:37 AM
This CF extender may be helpful in your research.

http://store.calexium.com/en/316-cf-extender-290-mm-with-flat-ribbon.html

Using a CF reader as a lab rat is an excellent idea.

I just ordered one of these, thanks for the link. The only problem is that we think it'll lower the transfer rate significantly due to the high frequency which is why we think the actual adapter should be built into a CF card body. But at the very least we can use it to do a proof of concept test with a camera.

...and sorry for getting the wrong forum, I was looking for a Hardware section but it was about 1am here and I must've missed it

Quote from: Roman on June 05, 2013, 03:48:03 AM
So just to be clear though, it hasnt been tested connected to an actual camera?

As that's half the battle by the sounds, making sure it's all a format etc that the camera is willing or capable of interacting with.

I wouldnt be getting carried away with getting 100s manufactured etc at this stage, if it's not been proven to work with a camera yet.

Not really, the SSD acts 100% like a CF card on the other side of the adapter. Also, as I mentioned earlier, the technology is nothing special and has been done before in other forms.

We will get it inside a camera somehow to prove the concept but it probably won't run at full SSD speed.

The thing about the manufacturing is the in order for adapter to work 100% at those speeds the components have to be very precise and that requires machines that we don't have. And ordering one to be manufactured cost pretty much the same as ordering 50.


Quote from: Cityeater on June 05, 2013, 09:11:45 AM
Good to see some progress in this area, thanks for your hard work. My concern is that for those of us that work with camera rigs the extended battery grip wouldn't be ideal with the equipment we've got and might make some gear (for instance follow focus) incompatible due to the new lens height.
For me, I like everything you've suggested except if I could choose I would prefer to have a simple external box that I can mount. I need a little more flexibility and I find the battery grip inconvenient.
Looks like a good project.

On a side note could I suggest a mounting option on the bottom (if you go that way) so its not a single screw but also has a registration pin like a normal film cam for base plates and so forth.

EDIT
It hasn't been attached to a camera yet? I think you'll find that's the tough bit, its more than just a physical interfacing problem. Looks like a good project though, best of luck.

The bottom mount is just so that you can easily use the box without a rig. You can put anywhere you like. The only limitation is that the cable has to be fairly short or it won't work properly.

Quote from: heavygrafix on June 05, 2013, 10:10:31 AM
i also thought a solution as battery grip isn´t optimal, so i designed some SSD housing that is directly connected to the Rig Rails. A Battery Pack (10Ah), that is powering both, camera and SSD (or some other stuff on the Rig) is (in my mind) the best solution.

if you did not saw my rendering yet:


I hadn't seen your design. Looks pretty cool.
We're going for a slightly different approach with a smaller all-in-one box with removable batteries.


PeteTomov

I just scrolled through the forum and noticed how many similar projects there are...should have probably done that before we started work on it  ;D

heavygrafix

Quote from: PeteTomov on June 05, 2013, 11:06:24 AM
I hadn't seen your design. Looks pretty cool.
We're going for a slightly different approach with a smaller all-in-one box with removable batteries

so would you use 1,8" SSDs or 2.5" Ones?

i think 1.8" would be better with my design. but 2.5" are cheaper in fact of Gb/$
5D Mark III, SP 24-70 2.8 Di VC, KB 1000x 64Gb, KB 1050x 128GB

heavygrafix

i just got on funny idea. don't take it serious..... What if we can record directly to DDR3 RAM.. just kidding.


I real like what´s going on these days here. And if somebody needs some rendering, i can do this 4 you awesome guys. And at least because so many people working at similar projects, it would be fine to connect and swop some ideas and experiences. Together it would be awesome to get the kickstarted project done.
5D Mark III, SP 24-70 2.8 Di VC, KB 1000x 64Gb, KB 1050x 128GB

PeteTomov

Quote from: heavygrafix on June 05, 2013, 11:15:52 AM
so would you use 1,8" SSDs or 2.5" Ones?

i think 1.8" would be better with my design. but 2.5" are cheaper in fact of Gb/$
We are actually hoping to be able to record higher the full hd resolution in crop mode with the speed we got (though we still don't know how fast the 5d can write), so 2.5" seems more appropriate considering file sizes.

Quote from: heavygrafix on June 05, 2013, 11:24:32 AM
i just got on funny idea. don't take it serious..... What if we can record directly to DDR3 RAM.. just kidding.


I real like what´s going on these days here. And if somebody needs some rendering, i can do this 4 you awesome guys. And at least because so many people working at similar projects, it would be fine to connect and swop some ideas and experiences. Together it would be awesome to get the kickstarted project done.

I actually considered a CF card raid so the whole thing can be smaller but I also wanted something cheaper.

Also, thanks for the offer. I consider myself fairly decent 3ds Max 'artist' but I realize I'm wrong and take up on that :)

Michael Zöller

Pete, can you please post some pictures and benchmarks? Thanks!
neoluxx.de
EOS 5D Mark II | EOS 600D | EF 24-70mm f/2.8 | Tascam DR-40

heavygrafix

5D Mark III, SP 24-70 2.8 Di VC, KB 1000x 64Gb, KB 1050x 128GB

PeteTomov

The prototype is a my buddy's place where we worked on it. We'll get together over the weekend and take some pictures, maybe some videos and run some benchmarks. We should also have a response on price by then so we might think about setting up the kickstarter as well.

The thing about this project is that while we would love to share out project with world, probably get a small office and hire someone to "handle things", maybe get a couple free adapters for ourselves  in the process, we're not actually expecting to make any serious money of it. We'll keep the price as low as possible and share any extra money we get with the Magic Lantern team who I imagine do this for the for the same reasons we do.

During the week we both have other jobs,well he actually does electronics design which is why we're able to do this in the first place and I myself am trying to start a video game company(it's a long story), so expect the more substantial updates during the weekends.

Steven

So, have you been able to actually record from the camera to the attached SSD yet?

AnotherDave

I'm rather excited to see this thing!  I'd make a video of yourselves using it so people could see it working.

5DanielMIII

 
Quote from: PeteTomov on June 06, 2013, 03:04:38 AM

with the Magic Lantern team who I imagine do this for the for the same reasons we do.

As in sharing info for free so more people can contribute to develop? Like they (ML-team) do;-)
I am thankful that Jesus and MagicLantern exist, because they both change lives!

PeteTomov

Just a quick update since I've been really busy and haven't been able to check forum as much as I'd like.

We finalized the design and got a response from the manufacturer. Initial manufacturing of the first 200 units is going to be a bit more expensive than we anticipated. With the additional cost of batteries,handling, shipping and extra units to cover warranty(we managed to negotiate warranty with the manufacturer but we can't keep people waiting while they manufacture some more) we're looking at a $50 000 kickstarter. This means the price of a shipped unit(removable rechargeable batteries included) will be $250. While obviously larger quantities will bring the per unit price down I was original hoping to keep the price at around $150.

I'll set up the kickstarter as soon we have some material to show but in the meantime what do you, guys, think of the price? I guess it's not too expensive for a 5d accessory but I was really hoping to keep it even more affordable. 

Expect pictures, final specs and stuff soon.

PeteTomov

Quote from: 5DanielMIII on June 10, 2013, 10:08:33 PM

As in sharing info for free so more people can contribute to develop? Like they (ML-team) do;-)

As in giving people affordable (we obviously can't make hardware free but we are looking to make it as cheap as possible) tools to help them make their work better.  ;)

haemma

Sounds good! Hopefully the kickstarter project works well! I definetly will support this.

5D3shooter

I'm very interested also, I'm so sick of CF cards.

Just so you know, if you do kickstart this, you will definitely receive more than $50,000...probably $250,000
A LOT of people want this. 

squig

Quote from: PeteTomov on June 13, 2013, 11:17:15 AM

We finalized the design and got a response from the manufacturer. Initial manufacturing of the first 200 units is going to be a bit more expensive than we anticipated. With the additional cost of batteries,handling, shipping and extra units to cover warranty(we managed to negotiate warranty with the manufacturer but we can't keep people waiting while they manufacture some more) we're looking at a $50 000 kickstarter. This means the price of a shipped unit(removable rechargeable batteries included) will be $250. While obviously larger quantities will bring the per unit price down I was original hoping to keep the price at around $150.


$250 sounds reasonable and is about what I was expecting. Keep in mind that a Ninja costs $700 and produces an inferior recorded image. My advice is to run the specs and design by the peeps on the forum just to make sure you're on the right track. Personally I think a 2x LPE6 hot swappable battery mount is a good way to go. Some people have been asking for an under the camera mount but this is a bad place to put it, it raises the centre of gravity between the camera and rig/stabiliser and puts the camera too high over rods for a mattebox. The best option is to just put some 1/4" threads on it so people can mount it wherever they want.

EDIT: 6000mAh for 2 hours? Really? The Ninja can run longer with less power.