Steadicam Prototype v3 Test
I think I’m making some progress! Here is some video I shot of the steadicam prototype arm curling some weights. There are still some obvious problems, but at least it’s encouraging.
3A style topstage
Starting to think about the other parts of the steadicam, working on a 3A style top stage, here’s what I have so far. I worked off some photos I found online, so the dimensions are a total guess. I found a rack and pinion gearset on mcmaster carr, the rest should be pretty straighforward. Going to glue the rack in the top plate, and add some screws to add friction to the sliding parts, as well as some hard stops to make sure the camera doesn’t fall off.
Changes changes
Did some thinking, and I’ve made several changes to the cam/ride mechanism which should help with the issues I’ve been having.
1) made the bottom of the lift assembly thicker so it will have less chance to twist on it’s bearings. I may at some point change this further to grip the post tighter- right now it’s just a set screw, but it might need to be a clamp instead.
2) the ride mechanism gets new larger bearings and shafts- going from 1/4″ to 3/8″ should help with the twisting.
3) I moved the lift adjust screw to the side so the pivot for the ride mech. can be closer to the post, meaning a shorter moment on the twisting action
4) see #2- did the same thing on the bone side of the ride mechanism.
5) changed the cam, now it’s a single piece with a 3/8″ post sticking out.
Hopefully these changes will help.
Design + Development = wow, I respect engineers even more now.
I added a second cam and connecting rod to my arm yesterday. It definitely solved the twisting problem. In the long run, I’d rather just have a single cam, because adding the mechanism to adjust the offset from the cam will be twice as difficult with 2 cams. For now, twin cams, fixed connecting rod length.
The most obvious problem right now is that as the cam switches from pushing UPwards on the connecting rod, to pushing DOWNwards on the connecting rod, the whole assembly clicks and jumps a bit.
I’m also getting way too much friction in all the little rods and bearing surfaces, part from the design, and part from my lack of holding tight tolerances in the machining process. This friction is so great that the arm is sluggish, and I doubt would do much stabilizing in use.
Prototypes are frustrating
I finally completed a prototype of my Steadicam G-series style arm. Yippie!
The bad news is that there are some pretty severe design flaws in my plans. Not sure how the G-series arms cope with this issue- that the cam at the top is pushing only on one side of the top of the spring attachment mechanism… makes the whole thing torqued sideways, as you can see in the top view. Going to try to add a matching cam and rod to the other side to compensate. I’m starting to see how things get over-engineered.
Shock mount for boom pole
So I finally made the first usefull thing with my new machining skills and tools. My good friend Aaron sent me a boom pole that he got free for some reason, and finally got around to making an adjustable adaptor to mount my DIY shock-mount on the pole. The shock mount is 4″ ABS pipe, cut off in the middle, and some elastic shock cord from an old pair of pants.
The top of the boom pole has a 5/8″ threaded rod poking out of it, so I machined up some 1/2″ aluminum, added a thingy to adjust the angle of the shock mount, and voilla! I actually made something usefull! Probably not going to get it anodized, since it’s not exactly fine craftsmanship. It works though.
DIY HMI pt 3
After even more research, I’ve decided to retrofit one of my Lowel Omni fixtures with a 70W HQI bulb.
~$86 for the bulb - Iwasaki EYE Color Arc MTD70/D #56227 which is4000 lumens, 6000hrs, 6500K 92CRI. the bulb is 120mm wide, so it’ll have to go at a 45 degree angle in the square fixture. The specs say it has an up to 15min restrike time, but I’m betting it’ll be more like 5mins, like the other Iwasaki HQI bulb that I have in my soft box. Here are the specs for the bulb- http://www.eyelighting.com/tb/MH/ColorArc/EQS-N-52-78-56227.pdf
~$98.43 for the ballast- 120v-277v input - electronic ballast for 70W bulb
~$5 for a lamp holder for this bulb
I ordered 1 bulb, 1 balast, and one lamp holder from Steve from Dyna-Brite Lighting in Tuscon (Phone: 1-520-882-4404). I’ll have to do a bit of metal work to get everything inside the fixture, and then will need to make some special cables to connect the light to the ballast, and ballast to the wall.
If it works, then I’ll have an openface/spot/flood adjustable Lowell Omni that puts out ~ 300W worth of light (in Tungsten terms) and pulls 70W from the wall. Not to shabby for a hair/rim light.
DIY HMI continues
Started doing more research on a small HMI I can use for a hair light. Looks like there are several 200W HMIs on the market- by Arri, Joker, Frezzi to name a few. The Frezzi is $3500 for a hot-restrike unit that includes pelican case, soft box, filters, barn doors, and ballast. Similar unit from Arri that also includes a DC ballast costs $5200. Wow. This is for a single 200W HMI (puts out about 600W worth of light compared to a tungsten) in a fresnel housing, with accessories- but wow! That’s not cheap.
Turns out they all seem to use the same bulb- Phillips 200W MSR HR HMI bulb. Bulbs cost around $120 from Barndoor Lighting. The bulb has a GZY 9.5 base, and is 80mm (3.2″) tall.
Further research lead me to find only 1 compatible ballast for this bulb- made in Germany. Schiederwerk EVG 2-20 ballast roughly - $380, then you also need an ignitor - HZG 8-25 (Hot Restrike!), part number 32 058 1000. Ignitor - $95. The US distributor is Jeffrey Van Etten @ Professional Lamps jfsales@plilamps.com
So we’re at $600 for bulb, ballast, and ignitor. You still need a fixture, a box for the ballast (it’s a circuit board) and cabling.
If you can get away without hot-restrike, then things get a lot cheaper. Coollights has CDM (similar tech to HMI) bulbs 150W in an Arri knockoff shell with a nice ballast and cable for $460 or so. Takes 5 minutes to warm up, but is cheap, bright, and apparently well made.
Food for thought. Don’t know if a doubling in price is worth it for hot-restrike (one could get a similar arri knock off for $270, and replace the guts with an HMI and build a ballast box- but I bet you’re still near $1000 or more when done- but you’d have a pretty sweet 200W HMI for $1000, which isn’t all that bad.
FEA follys
Turns out the FEA I’ve been doing with Solidworks 2007 was based on some mis-information. I was using Cosmos Express (the simple/beginner’s FEA tool in SW2007) and once I realized that I needed some sort of saftey factor on my parts, suddenly I couldn’t make anything strong enough to hold the 300lbs that the spring would be pulling inside each arm segment. I decided to look at some different materials than 6061-T6, and that’s when I found that SW’s material library had the yield strength of 6061 @ 8000psi. Wikipedia and ALCOA list it as a yeild strength of “over 35,000PSI” (http://en.wikipedia.org/wiki/6061_aluminum) which is quite a big difference! Now that I made that change in SW, suddenly all my parts are more than strong enough.
I also re-designed my spring holding apparatus- I’m running 5 steel springs of a smaller diameter to get enough range of motion, and hold enough weight. The upside is that I can also run 4 if I find that it’s way too strong for my camera, or I can run just 1 for very light weights.
Canon 5DMkII makes me want to accept 30P
Seems like Canon will not be coming out with any sort of firmware upgrade to get 24P out of the 5DMkII. It also seems like RED is pushing back it’s promised deadlines (which they also promised not to stick to, so it’s hard to blame them).
This makes me really want to think more seriously about the 5DmkII for video. There are now workarounds for all the main shooting issues (manual f-stop - twist the lens, or use manual nikons, manual ISO with * button to lock it down, manual shutter speed - http://www.vimeo.com/2425650). They might not be pretty, but they work. The only real stumbling block is the 30P shooting mode. If I can just convince myself that the 30P isn’t so bad, then I’d jump on the bandwagon. The upsides are many- most importantly, big sensor, and great low-light performance. Now instead of obsessively reading Reduser.net (which replaced my obsessive reading of DVXuser.net) I can read http://www.5dfix.com/ and other 5D sites.
For run-n-gun documentary work, it looks just amazing- you could get away with just a few lights (or none at all with some natural light), a small audio kit, and you’re good to go. You could fly a very light stabilizer rig to help with the handling issues- Here’s some footage from switzerland- http://www.bildblog.ch/canon-5d-mk2-expedition/ they are just using a hand-held glidecam, it’s not perfect, but it’s not too bad for being so damn portable.
Come on canon, or hackers! make my day.
keep looking »