Canopus ADVC 500 Repair

[Raff]

G'day Everyone,

I bought myself a pro level video digitiser from EBay, sold as working. Its a Canopus ADVC 500.

When I get it and plug the supplied 12 v 1.5 amp plug pack into it, nothing.

Further investigation revels a short inside.

Long long story short.

Inside the 12 volts is boosted to 18 volts and -ve 18 volts by 2 LM2596-adj style switching regulators. There are other smaller regulators scattered about on the board as well but they seem to work fine.

Ive found that the reason for the positive and negative rails is that there is a smattering of op amps for  various things like balanced audio inputs and outputs and some video amps as well.

Some of the op amps are Burr Brown OPA137 and they are powered in a rather odd way.

The positive 18 volt rail is straight from the output cap on the switching regulator, but the negative rail goes through a surface mounted SOT-89 package with "FJ" on it. There are 3 of these odd devices.

Ive found that the short on the 12 volt input goes away if all 3 of those "FJ" devices are removed, but heres the odd thing, if I power up the device, then slide one of those "FJ" things onto its mounting pads everything is ok, it works, but if I power up with the "FJ" device (any of the 3) connected, I get a short.

Just to define a short in the sense of this device, Im using a regulated lab type power supply, the short is when the device draws 3amps plus, its supposed to only draw around half that and i'm only powering the mainboard no daughter boards or front panel LEDS etc

When I do the slide the little "FJ" device onto its solder pads trick, the current draw at 12 volts is around 400mA.

I tried to measure the current drawn from the negative rail by bridging the pads with my meter and it blew the internal 630mA fuse in my meter, so I replaced the fuse and tried the 10A scale, nothing, even though bridging the contacts produces a fat spark. I can only conclude that the current spike is very short.

I measured the "FJ" things (all 3) on my meters, my Jaycar Semiconductor tester announced that they were 3 terminal bi coloured LEDS, no don't think that's right, my $20 chinese Mega328 based tester said back to back diodes on pins 2 and 3. And the diode function on my Fluke 77 says diodes all round, but with a forward drop of 0.6 volts, and a reverse drop of 1.4 to 1.8 volts depending on which pins Im across.

All rather odd.

I initially thought maybe the "FJ" things might be little regulators, but that doesn't make sense to me, the 18 volt rails are already 18 volts and I cant imagine why youd want a spit rail op amp with different voltages.

I then thought maybe a zener diode package, but that doesn't really make sense either.

I've checked the op amps, and the caps on the "output" side of the "FJ" things, I don't think any are shorted, all 3 places where these "FJ" devices were exhibit the exact same weird behaviour, that is sliding the device onto its pads after power up seems to work, but not if the device is connected before power up.

Its doing my head in. 

I hope someone has been down this path before and has a tip or 2.

Thanks for reading my rant,

Raff

[tautech]

A photo will help.

[Raff]

Here's a commented photo.

I hope its clear enough and makes sense.



Raff

[tautech]

I'm thinking V regs too.
Clue being the reversed biased diodes for protection.
Are there any other marking/codes on these SOT-89's?

Partially scrubbed?

[Anks]

If they are voltage regs I would suspect something like these
http://docs-asia.electrocomponents.com/webdocs/0994/0900766b80994b06.pdf

But I've seen back to back diodes look similar but that would be pointless.

[tautech]

Google search "marking code FJ package SOT-89-3" lead me to Richtek RT9261B that is described as a VFM Step-Up DC/DC Converter.
http://www.s-manuals.com/pdf/datasheet/r/t/rt9261b_richtek.pdf
Check the Typical Application for confirmation that its correct.

[Raff]

Thanks for the suggestions guys / gals.

Its an odd way to wire up a circuit for sure.

I'm thinking I might order a handful of those regulators and tack one in and see what happens.

Although just why you'd want a regulator after the 18 volt switching regulator is beyond me, but stranger things have happened before. 

As pointed out, the reverse diode across it and the pinout matches.

Ill keep everyone posted.

Raff

[Raff]

Just put those little SOT89 packages under my microscope, one has "FJ" on it for sure, possibly had something between the F and the J, the other 2 are harder to make out, but may have had an "AY" on them. Maybe "A3" its very faint.

Either way, they all measure similarly on my meter, but the reverse voltage drop of the diodes in the "FJ" one is .2 volts less than the other 2.

Odd indeed.

Raff

[Raff]

Happy update, its working.

Those FJ devices are indeed little linear regulators, 2 are -12 volts and the other -5 volts.

I decided to try and test them off board, so I made up a sort of test jig, they tested fine, so that got me thinking..

I tried winding up the input voltage slowly on my lab supply, at around 2 odd volts in the current draw shot up and the device wouldn't power up.

Then I thought about my sliding the little regulators onto their pads trick, and thought maybe there is a huge but short lived current draw on power up for this thing.

I slid each regulator onto its pads whilst the ADVC500 was powered up and the current draw only rose to around 600mA, hmm.

I then decided to tack them in place once again.

I broke out my ancient and vastly over engineered made in Australia, Scientific Instruments, 10 amp power supply and cautiously hooked it up monitoring the current drawn by the ADVC500.

Imagine my surprise when the ADVC500 powered up no problem!   

I hastily reassembled the whole unit and tried again, yes, powers up fine on the 10 amp power supply, however total current draw is 1.4 A all up.

I didn't even really see any huge spike in current draw on power up, maybe 2 amps at best, odd.

I can only imagine that the supplied PSU brick that came with the ADVC500 is stuffed, and my lab PSU which is an old Jaycar dual floating and tracking one, up to 40 volts and 3.5 amps each side is rather sensitive to current spikes.

For those who may came across this device, the 2 "upper" regulators are -12 volts, the "lower one" -5 volts in spite of them all being marked "FJ" from what I can see.

Thanks for the help guys, it got me thinking and now the ADVC500 works

Raff

[Anks]

the FJ regs come in all sorts of voltages look at the data sheet I posted there is a English version I have seen before but that was the one I got up fist

[Raff]

I looked at the data sheet you posted and from that concluded that perhaps they are Toshiba parts.

Typical damn tiny surface mounted stuff, old blokes like me cant even see the part properly, let alone what might be printed on it