ZDP-189
New Member
This was a totally out-there DIY camera hack.
Tonight I completed a long shelved project to build a valuable lens buy cannibalising junk. I successfully converted my Sigma 15-30mm f/3.5-4.5 EX DG lens from SA mount to Canon EF.
Quite some time ago, I pickled up a second hand Sigma SD-9 DSLR. It was a lovingly kept kit with all the gubbins, including a fantastic lens, the Sigma 15-30mm f/3.5-4.5 EX DG. I experimented with the camera, did a lot of fun and test shooting with the Foveon X3 sensor, but at the end of the day, the camera body was slow, the sensor insensitive, it took only the very oldest memory cards, and had the effective resolution of a 3.4MP camera. So I modified it, converting it to IR and then discovered that the sensor was too insensitive to IR to shoot handheld because even with the ICF off, it wasn't very IR sensitive (either due to the body or the lens coating). So I converted it back and shot with it some more, then abruptly one day I turned it on and the sensor had gone all rainbow colours. It was effectively dead. I was left with a bunch of junk: a severely boogered camera, a gear bag, a camera strap and a lens in an obscure mount.
A little research showed that the Sigma SA mount is based on the Canon EF mount. It uses the same protocols, and the contact block is laid out the same way, but the mount bayonet is based on a simplified Pentax PK mount. I believe the reason is because they make third party lenses for Canon, Nikon, Sony and Pentax with registers from 44mm to 46.5 mm and throat diameters from 37mm to 54mm, so Sigma picked a bayonet specification that would work with them all, and a control standard that matched the biggest third-party lens market. What that meant was I could convert the SA lens to EF, which is my primary system.
It's simple in theory. I figured out that lens will fit in the EF mount throat and it won't project backwards and hit the mirror. The control system is EF so I should retain full functionality. Within Sigma's range it's the ideal lens, lacking stabilisation and being very wide, not particularly fast and with a focussing range from under a foot to past infinity (for IR photography), so it's a little more tolerant to register inaccuracy. I happily sauntered off to the store and bought a junk EF (not EF-S) lens to provide the bayonet.
However it's not quite so simple in practice. You hear all over the net how easy it is to do, but nobody shows what you have to actually do to get it to work. Having taken the bayonet off the lens and cut the connector ribbon, I had to strip the EF bayonet and knock out the lens / glass window on the back. I bored out the plastic throat seal till it would match the 33.5mm throat of the rear element housing of the lens. I used a lathe rather than a file to achieve concentricity. I might have gotten away with filing it out, or just leaving it off, but it would have left a ugly gaping hole showing the PCB and wires and will let in lots of dust which can get back out and onto the sensor.
I then measured the bayonet from the base of the plate to the back of the flange of the SA mount including shims. I had to use a vernier height gauge and hang the contacts of the EF mount off the measuring table as they wouldn't allow it to lie flat.
This is an absolutely critical measurement and must be reproduced to very tight tolerances on the EF mount. If the mount was too thin, I'd have added shims, but it was too thick, so I had to turn it down on the lathe to within about 0.05mm (a twentieth of a millimetre). At that thickness, how hard the screws are tightened makes more difference, so I figured I was in the ballpark.
After that, most of the screw holes matched up, but one was out of position, so I drilled out and countersunk an extra hole on the EF bayonet in the correct place. Without this, the lens could break off the bayonet. I decided epoxy won't be best, because it will make the lens unservicable had I made a wiring mistake, and because the thickness of epoxy would throw the register out.
The next thing I did was fully desolder both ends of the data connection - the back of EF data connector and the lens PCB that had been connected to the lens. I used a shot length of IDE ribbon cable (seven wires of) to connect the two together. All I did then was screw it back together and test it.
Guess what? It was totally functional. It auto-focuses and confirms focus. Early testing shows it close focuses and infinity focuses at wide aperture and loots sharp. Maybe not sharp like my 17-40mm f/4L, but pretty sharp. Indeed, focus seems to be faster and more positive than I remember it on my Sigma body and I've been shooting it full frame on the 5D Mark II. The Sigma only managed a 1.7x crop factor; it's almost like a four thirds sensor. With a full frame sensor, the lens is just about as wide as you can get a in rectilinear zoom lens for EF (yes, I know about the new Sigma 12-24mm F4.5-5.6 EX DG ASP HSM, but it's $1,150 and this was made from junk).
I'll post test pictures in the morning when the light's better.
Disclaimer: I don't recommend people try this themselves because it's not your typical easy home project involving a stanley knife and gaffer tape. You might ruin a good lens and camera body and you conceivably might even hurt yourself. BTW, it would surely invalidate your lens warranty too.
Tonight I completed a long shelved project to build a valuable lens buy cannibalising junk. I successfully converted my Sigma 15-30mm f/3.5-4.5 EX DG lens from SA mount to Canon EF.
Quite some time ago, I pickled up a second hand Sigma SD-9 DSLR. It was a lovingly kept kit with all the gubbins, including a fantastic lens, the Sigma 15-30mm f/3.5-4.5 EX DG. I experimented with the camera, did a lot of fun and test shooting with the Foveon X3 sensor, but at the end of the day, the camera body was slow, the sensor insensitive, it took only the very oldest memory cards, and had the effective resolution of a 3.4MP camera. So I modified it, converting it to IR and then discovered that the sensor was too insensitive to IR to shoot handheld because even with the ICF off, it wasn't very IR sensitive (either due to the body or the lens coating). So I converted it back and shot with it some more, then abruptly one day I turned it on and the sensor had gone all rainbow colours. It was effectively dead. I was left with a bunch of junk: a severely boogered camera, a gear bag, a camera strap and a lens in an obscure mount.
A little research showed that the Sigma SA mount is based on the Canon EF mount. It uses the same protocols, and the contact block is laid out the same way, but the mount bayonet is based on a simplified Pentax PK mount. I believe the reason is because they make third party lenses for Canon, Nikon, Sony and Pentax with registers from 44mm to 46.5 mm and throat diameters from 37mm to 54mm, so Sigma picked a bayonet specification that would work with them all, and a control standard that matched the biggest third-party lens market. What that meant was I could convert the SA lens to EF, which is my primary system.
It's simple in theory. I figured out that lens will fit in the EF mount throat and it won't project backwards and hit the mirror. The control system is EF so I should retain full functionality. Within Sigma's range it's the ideal lens, lacking stabilisation and being very wide, not particularly fast and with a focussing range from under a foot to past infinity (for IR photography), so it's a little more tolerant to register inaccuracy. I happily sauntered off to the store and bought a junk EF (not EF-S) lens to provide the bayonet.
However it's not quite so simple in practice. You hear all over the net how easy it is to do, but nobody shows what you have to actually do to get it to work. Having taken the bayonet off the lens and cut the connector ribbon, I had to strip the EF bayonet and knock out the lens / glass window on the back. I bored out the plastic throat seal till it would match the 33.5mm throat of the rear element housing of the lens. I used a lathe rather than a file to achieve concentricity. I might have gotten away with filing it out, or just leaving it off, but it would have left a ugly gaping hole showing the PCB and wires and will let in lots of dust which can get back out and onto the sensor.
I then measured the bayonet from the base of the plate to the back of the flange of the SA mount including shims. I had to use a vernier height gauge and hang the contacts of the EF mount off the measuring table as they wouldn't allow it to lie flat.
This is an absolutely critical measurement and must be reproduced to very tight tolerances on the EF mount. If the mount was too thin, I'd have added shims, but it was too thick, so I had to turn it down on the lathe to within about 0.05mm (a twentieth of a millimetre). At that thickness, how hard the screws are tightened makes more difference, so I figured I was in the ballpark.
After that, most of the screw holes matched up, but one was out of position, so I drilled out and countersunk an extra hole on the EF bayonet in the correct place. Without this, the lens could break off the bayonet. I decided epoxy won't be best, because it will make the lens unservicable had I made a wiring mistake, and because the thickness of epoxy would throw the register out.
The next thing I did was fully desolder both ends of the data connection - the back of EF data connector and the lens PCB that had been connected to the lens. I used a shot length of IDE ribbon cable (seven wires of) to connect the two together. All I did then was screw it back together and test it.
Guess what? It was totally functional. It auto-focuses and confirms focus. Early testing shows it close focuses and infinity focuses at wide aperture and loots sharp. Maybe not sharp like my 17-40mm f/4L, but pretty sharp. Indeed, focus seems to be faster and more positive than I remember it on my Sigma body and I've been shooting it full frame on the 5D Mark II. The Sigma only managed a 1.7x crop factor; it's almost like a four thirds sensor. With a full frame sensor, the lens is just about as wide as you can get a in rectilinear zoom lens for EF (yes, I know about the new Sigma 12-24mm F4.5-5.6 EX DG ASP HSM, but it's $1,150 and this was made from junk).
I'll post test pictures in the morning when the light's better.
Disclaimer: I don't recommend people try this themselves because it's not your typical easy home project involving a stanley knife and gaffer tape. You might ruin a good lens and camera body and you conceivably might even hurt yourself. BTW, it would surely invalidate your lens warranty too.
