Siglent SAT7.820.687 PSU fail

[gkmaia]

i got my hands on a Lecroy/Siglent scope.

The PSU has failed at the beginning of the inverter block.

Anyone out there with some siglent knowledge who could help me identify these two components, they seem to be fuses to the block, but not sure why two in paralel.

[tautech]

Never seen any Siglent scope PSU carnage, pop a pic if the front panel up so we can identify the model.

Edit
And a full image of the PSU if you don't mind.

With SMPS's the controller type often dictates the schematic and often the manufacturers suggested typical usage is followed very closely.
Grab its datasheet and there should be clues to the component values used.

[gkmaia]

Here are the photos.

siglent sent me a revised board photo and they said L5 was a 1ohm resistor.

The resistor I pulled out of my faulty board tested at 5.5kohm

That is weird because that resistor is the bridge for the positive rail that comes from the SMPS mains and feeds the inverter and the only thing that could have fused it would be a fault with the inverter transformer drawing too much current or a shorted bypass cap...

So my suspicion is the resistor got overloaded, fused, damaged the board and damaged got it conductivity reduced. Does that make sense?

If that is the case, the inverter block is drawing too much current and will fry any other resistor I put in there unless I raise it is power ratio, but then that will stress something on the SMPS block...

I am also not sure why placing a resistor there instead of a fuse. If you look at the silkscreen it was supposed to be a fuse there and that would probably have avoided trashing the PCB.

what are your toughts?

[wraper]

siglent sent me a revised board photo and they said L5 was a 1ohm resistor.

Nope, it's a ferrite bead. Person who wrote you most likely did not have schematics and just measured resistance. There should be some fault on the rail which cased high current passing through it.

[shakalnokturn]

It looks like C23 before CCFL output is the cause of your problems.
You'll want to check transformer or primary transistors have not suffered before replacing those "humm" fuses?

Despite what you have been told, I'd expect L5, L6 to be ferrite chokes.
Adding a real fuse would apparently be a good idea though...

[gkmaia]

It looks like C23 before CCFL output is the cause of your problems.
You'll want to check transformer or primary transistors have not suffered before replacing those "humm" fuses?

Despite what you have been told, I'd expect L5, L6 to be ferrite chokes.
Adding a real fuse would apparently be a good idea though...

C23 was just stained by that glue they put on the connector. It is physically ok. I will check those.

siglent sent me a revised board photo and they said L5 was a 1ohm resistor.

Nope, it's a ferrite bead. Person who wrote you most likely did not have schematics and just measured resistance. There should be some fault on the rail which cased high current passing through it.

The problem now is... what spec I would use for that choke? I got to figure that out.

Or close the rail with an amp meter and find out how much current it is drawing and if it is stable get a fuse or a choke that fits the spec.

Would that be a good approach?

[wraper]

The problem now is... what spec I would use for that choke? I got to figure that out.

Or close the rail with an amp meter and find out how much current it is drawing and if it is stable get a fuse or a choke that fits the spec.

Would that be a good approach?

It's not critical part. You should select one which can hold enough current. As pads are gone, IMHO drill a hole to other side and use through hole ferrite bead. If (missing) terminal on left side goes directly to transformer, chose part with impedance on lower side. Also check if D2 and Q1 are not shorted.

[tautech]

Ah, so it's just on the LV supply side of the LCD backlight supply. 

Was no LCD backlight the only fault ?

Any ferrites on the supply are just to reduce any rail switching noise from entering that LV rail and possibly affecting the scope.
I'd want to check the backlight primary side switching elements and diodes are OK so that DC rail isn't overloaded again.

BTW, yours are the older 5" display models of the current SDS1000CFL series so must be approaching 10 years old.
https://siglentna.com/digital-oscilloscopes/sds1000cfl-series-digital-storage-oscilloscopes/

[gkmaia]

AWeSOME!

You guys are super helpful!

 

[gkmaia]

Ah, so it's just on the LV supply side of the LCD backlight supply. 

Was no LCD backlight the only fault ?

Any ferrites on the supply are just to reduce and rail switching noise from entering that LV rail and possibly affecting the scope.
I'd want to check the backlight primary side switching elements and diodes are OK so that DC rail isn't overloaded again.

BTW, yours are the older 5" display models of the current SDS1000CFL series so must be approaching 10 years old.
https://siglentna.com/digital-oscilloscopes/sds1000cfl-series-digital-storage-oscilloscopes/

Yes just the backlight. I torched the LCD and could see image. So pixels, drivers, processor, etc... everything works. Just backlight is faulty.

[gkmaia]

Did a bunch of troubleshooting also a little schematic.

Current through L5 is at 2 AMPS, no wonder why the poor thing got fried...

Basically Q2 will fuse as it is getting extremely hot in seconds. I did test it on a BK curve tracer and it is all good up to 100v.

With the transformer off circuit nothing happens. Put the transformer back (leaving pins 7 and 8 disconected) Q2 goes nuts. C7 is fine, but there is no oscillation on the transformer primary or feedback. Neither anything out of the secondary.

To me it all points to a problem with feedback windings 2-3-5. But I am not 100% sure. I wont be hard fix the primary and I can do it but just want to make sure I am not missing anything.

Would be great to know what are the factory readings at J2.

[tautech]

CFL service manual, P 48
http://siglentna.com/wp-content/uploads/dlm_uploads/2017/10/SDS1000CFL_ServiceManual_en.pdf

Those are 7" display but it shouldn't be much different for the older 5" display models.

[tautech]

R7 = 01E = 1M
http://kiloohm.info/eia96-resistor/01E

R47 should be 470
Check the epoxy dipped cap C7 on the primary side. It sets the RC frequency of the push-pull and if it's drifted (seen this in a Tek) then it won't self oscillate.

[gkmaia]

Thanks tautech.

Updated version of the schematics. Tested R42, 43 & 8. C7 is a 334 and reads 333nf so well in range.

I have replaced Q1 & Q2 with TIP50 3A and the LCD now works but till the TIP50s get hot, then it shuts off.

From what I understood... the main board lowers J2-2 and that makes Q1 conduct.

C7 sets the frequency for Q2 and Q3 to conduct and generate feedback for the primary.

The secondary now outputs power to the LCD at a rate of 44Khz and Vpp 4kv. Then it starts dropping till around 2.5kv and that is when the LCD shuts off.

As the frequency remain stable I am ruling out C7 as an issue.

The pattern is the same if the LCD lamp is disconnected.

Also when the LCD lamp shuts off Q1 gate is still LOW so I can discard the main board detected something wrong and shut off the circuit.

The problem is definitively on the inverter block.

The whole circuit seems to be working. So I guess everything points out to my transformer.

Probably something wrong with my secondary... does it seem reasonable?

[tautech]

Thanks tautech.

Updated version of the schematics. Tested R42, 43 & 8. C7 is a 334 and reads 333nf so well in range.

Values that aren't right and/or don't look right to me:
R42 = 470 ....from your image.
L5 and R6 

To tidy up reassign the transistor numbering to match the PCB.
Shame you have the PCB to paw over and we don't. 

Now I dunno if one of those scopes has LED backlight (the other one) and this one uses CCFL but either way, if the Xformer is toast a LED backlight conversion might be the better way to go.
This might be worth a watch:
https://youtu.be/smrMoIpvYTk

[gkmaia]

this inverter circuit seems to be pretty standard as well as the transformer. I may find one on RS components... will have a look.

[wraper]

this inverter circuit seems to be pretty standard as well as the transformer. I may find one on RS components... will have a look.

Why would you bother? It's one of last parts you should expect failing. Check resistance to know if it's not open. Check inductance for short in winding.

[gkmaia]

this inverter circuit seems to be pretty standard as well as the transformer. I may find one on RS components... will have a look.

Why would you bother? It's one of last parts you should expect failing. Check resistance to know if it's not open. Check inductance for short in winding.

I think I did run over every component on this block I cannot see anything else to blame.
Resistance is and insulation is easy to test. I did test resistance but not insulation yet.
How would you suggest test inductance (for shorts) without a datasheet?

[tautech]

Have you checked to see if the inverter runs without the backlight connected ?
I know you mentioned something earlier that it wasn't oscillating but was it under load or not ?

[wraper]

How would you suggest test inductance (for shorts) without a datasheet?

If there is short, measured inductance will be suspiciously low as well as Q value.

[gkmaia]

Have you checked to see if the inverter runs without the backlight connected ?
I know you mentioned something earlier that it wasn't oscillating but was it under load or not ?

it runs just fine without the backlight. Outputs the same voltage and fails the same way. By dropping voltage slowly and heating Q2/3.

Then only difference is that instead of failing in 30 sec with backlight, it may fail with 50 seconds.

Do you see anything to blame on the secondary?

O maybe my TIP50 I used to replace are not ideal for the circuit?

[wraper]

Dunno why you thought TIP50 is good replacement for 2SD1802. 1/5 of rated current, 8x of rated voltage. Much lower gain, especially at higher current.

[gkmaia]

Dunno why you thought TIP50 is good replacement for 2SD1802. 1/5 of rated current, 8x of rated voltage. Much lower gain, especially at higher current.

That was the only thing I had at hand yesterday night to give it a go. So I went for it. And yes, it is not ideal... but that is what I had at the time.

So I did some digging on some old boards and found a D1585 that is much closer to 1802 gain and still has good margin and well... it is working!!!

Temperature stable... has been running for a while.

I guess my mistake was to test the hot Q2 while I was high yesterday night and I may surely have hallucinated looking at curve tracer and seeing a healthy transistor there on that nice and shining green crt!

[gkmaia]

Have you checked to see if the inverter runs without the backlight connected ?
I know you mentioned something earlier that it wasn't oscillating but was it under load or not ?

Thanks the help mate! It is running with no problem for more than an hour!

[tautech]

Have you checked to see if the inverter runs without the backlight connected ?
I know you mentioned something earlier that it wasn't oscillating but was it under load or not ?

Thanks the help mate! It is running with no problem for more than an hour!

Cool, well done !
As we do these we all gain a better understanding and greater knowledge.

Now onto the next one ! 

Thanks to wraper too for guidance and sanity checks.