snow white

Decided to revisit the 580EX-II repair after the vintage flash project.

 

 https://growmane.blogspot.com/2023/08/after-1-tube-replacement-in-2020.html

 

 

The pink line continued to be 330V while the discharge pad continued to be 5V.  Past lion verified there was nothing awry in the capacitor enclosure, but the mane cap only gets power from the pink wire & the pink wire was hot.

 It would probably still work if powered by the external battery. 

 

Powered it up while disassembled.  Found 330V on the pink wire & the mane cap, as it should be.  At this point, it was pointing to another intermittently failing wire.

 

 Also noted a nasty mechanism which is supposed to release the horizontal locking pin.  If it breaks or is fouled, there's no way to access the capacitor.  All the articulated joints have detents, so the locking pin is unnecessary.

It uses the locking pin to know if the flash is pointing straight ahead & should be zoomed in, but it uses a proper angle encoder to detect the vertical angle. 

If the angle sensors are disabled, it'll stay permanently zoomed.  You'd have to manually enter a manual focal length to change the flash angle.  There could be a bodge switch which manually exited zoom mode.

Inspected the high voltage boards without finding anything awry.

There's an intriguing clear packaged ASIC which doesn't do anything optical.  It faces blank plastic.  It could contain optically erasable PROM on part of it.

 At this point, discovered the 5V in the capacitor discharge hole wasn't the capacitor but the angle encoder.  The discharge hole was a different hole & the cap was probably installed wrong.

Helas, with it reassembled, it would show a full charge but still wouldn't fire.  The tube showed 330V across the terminals but the trigger wasn't firing.

Page 29 of 580exii.pdf says what electrical events are supposed to happen during flash firing.  The schematic is on page 87 & layout of the trigger circuit is on page 93.  The trick is probing when it's assembled requires soldering to 1 test point at a time.

This painstaking process showed the SCR gate was getting 1 second of 5V 1us pulses. The SCR was creating a 50V drop on C21's output.  This created -20V pulses on the transformer output. C21 output should be 330V divided by R28 & R41.  It holds up well during the pulse train.

The datasheet doesn't have any waveforms or notes about a pulse train.  It's not what the 244T was using & symptomatic of a dead GPIO.  This would require a donor part.
 

The internet raised an intriguing idea of the IGBT being dead.  The GPIO could keep firing if the tube doesn't start conducting.  It's an RJP4301.  The other part which commonly dies is a CR3AS-8ME thyristor.  Sadly, none of this info comes in any goog search.  It all has to be provoked.  The IGBT would have to fail open for it not to flash.

Found shorts in the thyrister between G, anode, cathode.  A & K are connected through the big inductor.  Don't think the G should be shorted.  The IGBT seemed nominal.  As usual, replacement parts have a $10 minimum order.  $20 could buy a new flash.

 The IGBT datasheet had the 1st schematic of a useful flash circuit, should lions decide to completely replace the 244T circuit with a modern TTL circuit.  It looked a lot simpler than what was done in 1983.  Still, the extra inductor & passives might be to achieve faster recycling.

It looks like all the SCRs need diodes on their gates to avoid the inevitable shorting between G & anode.

Burned $60 on 1 replacement IGBT, the minimum SCR order + the $40 of stuff required to cover the shipping.  Of course, if it ends up working it's going to need a bag of new batteries.  It was close to being cheaper to pay for shipping, but the 3 week lead time from China meant all the unnecessary crap which might be needed.

 The fact is it would never be used even if it worked.  Lions have a strange fascination with camera lighting systems.  It led to LED panels for video.  Lighting for still photos has been more difficult because it requires flashes.  RGB lights are the big thing with gootubers, but prisms split those into 3 bands instead of a rainbow.

---------------------------------------------------------------------------------------------------------------

Replaced the IGBT & the thyrister, after a long battle to remove the old ones.  It still didn't work.  Also burned up a ground line.  It seems some battery lines stay hot when the power switch is off.  The microcontroller really seemed to be dead.  The thyrister shorting might have been a symptom or a cause of the GPIO death.

Thus began future contemplation for the remaneing parts.  A lot of money was spent on replacement tubes & batteries.  The easiest outcome would be a studio strobe with TTL.

Even if it worked, it would have been limited by the lack of a laser aiming tool.  The converted 244T alone has hardly been used.