HP printer supply - repair or scrap?

[brainwash]

I bought a new HP WiFi printer for dirt cheap, something like 10$ including shipping, without the printer head.
http://reviews.cnet.com/multifunction-devices/hp-photosmart-7510-e/4505-3181_7-35042430.html

The power supply is [probably] DOA, a new one costs something like 30$ so I'm wondering if it's any chance to repair it. I've also found out the hard way that the polarized 120uF/400V cap does not have a bleeder resistor.
The outputs are 32V@1A and 12V@0.25A, probably for WiFi standby and soft-on.

From my testing the parts seem fine, nothing burnt, transistors + power mosfet seem ok, transformers seem fine. There's nothing going on on the low-voltage side of the optocoupler so the control part must be broken. Did a quick reflow of the parts but soldering seemed fine anyway.
The controller is this one, just located the datasheet moments ago:
https://www.fairchildsemi.com/ds/FA/FAN6747.pdf

Not sure what to do, on one hand I have access to some nice mechanical parts, stepper motors, color screen, wifi module, etc. On the other hand it's a pretty good printer that could replace my crappy Lexmark WiFi printer, but it still requires a PSU and complete print head.

_{About my Lexmark: the cartridges have an RFID chip on each 4 of them, the printer counts the number of dots dispensed. When that counter reaches 4 million or so the cartridge is marked as 'depleted', even though it's full. If one cartridge is 'empty' the printer does not print. Fuck lexmark.}

I am attaching some PCB pictures, maybe someone has an idea, but I'll study the controller datasheet seems the PSU looks just like the application note.

[Rasz]

holy batman so many logos on the cover
it is even NOM NOM certified!

have you checked pin7? supply voltaeg

[retiredcaps]

have you checked pin7? supply voltaeg

Yes, check pin 7.  It needs to be 16.5V to start.  If the voltage is oscillating, check the likely 10uF capacitor connected to it.  This 10uF capacitor might have high ESR if the adapter is run 24x7?  If you don't have an ESR meter, just replace it to rule it out.

[SeanB]

I will add as well check the small SMT resistors that provide the start up supply. R3, R6 and D1 provide that start up supply. they do go open circuit.

WTF did you do to the second image? It is mirrored!

[mcinque]

Have you checked this component (probably a diode)? It seems blown to me...  Or it's just an artifact/flux?

[retiredcaps]

WTF did you do to the second image? It is mirrored!

I like it.  I have spatial mapping challenges and at first it threw me off, but I'm going to start doing stuff like the above to aid in my own troubleshooting.

Some boards I work on have no pcb designations on the back side which makes it difficult.

[retiredcaps]

Have you checked this component (probably a diode)? It seems blown to me...  Or it's just an artifact/flux?

It is obvious Italians have a better eye care program than Canadians. 

[brainwash]

Hi all, thanks for the feedback.
I haven't had a chance last night to take a look at it, but I will today.
I remember there was some voltage (>50V) on the HV pin, other than that there were some low-level signals (1-5V) going to some of the pins of the controller but nothing else, so at least the startup pin is not properly powered.
I think it's just flux on that part, I checked almost all junctions.

[mcinque]

It is obvious Italians have a better eye care program than Canadians. 

I think it's just flux on that part, I checked almost all junctions.

Try to scrape it to verify it's flux and not a epoxy blow

[brainwash]

No, it was just flux, I've cleaned it up after taking the photos.

I replaced a double diode in the meantime, though I suspect it had nothing to do with it. I'm measuring 5V on the optocoupler receiver side (HV) for about one second after which it switches off. This means there is some signal from the LV side but it could be the OVC protection.

I cannot copy from the pdf (stupid protection) but the overcurrent protection acts like this:
- once the current sense is over threshold it triggers a 15ms counter and checks if Vdd<10V
- if Vdd<10V -> latches off for SCP (shortcircuit protection I assume)
- if Vdd>10V -> counts further to 220ms and latches off for OCP
- latch is reset when Vdd is removed

I'm measuring 11-12V on Vdd so I assume it's not triggering SCP so it might mean a short on the LV side. However, I've connected the ammeter leads directly to the secondary and there is no little to no current reading (the few mA might be capacitively coupled from somewhere else). It reads ~1.6 ohms, so it might be shorted but it's not disconnected.

[mcinque]

I cannot copy from the pdf (stupid protection)

Once the pdf is on the screen hit PRINTSCR key on the keyboard (near F12 key), then run MSPAINT (or any image editing software) and finally paste the clipboard using CTRL+V. There is no image or copy protection that can avoid this.

[SeanB]

PDF protection is only a bit flip. Ghostscript does not care about output device, it will happily print any thing it can display.

[brainwash]

I meant text cannot be copied, I haven't seen that protection in ages and was not prepared for it.
Copying an image does not help since I still have to save it. Alt+printscr is better, it only copies the active window. Actually since win7 there is a snipping tool included with the OS.

The side rant is that you cannot paste images into a website without going through a lot of hoops. At work we switched to Confluence instead of Sharepoint and people have actually started to write documentation for their projects since it's so much easier to do. Pasting a screenshot into the browser creates a file on the server and embeds it in the current page. Pasting a file uploads it to the server and creates a link, drag and drop also works.

I have to say that the software written by these guys (Atlassian I think) is the first one which is actually decent - anyone that used IBM/Telelogic, Doors  and other strange stuff to meet Spice standards knows what I'm referring to - the traceability from requirements to product.
When you do a checkin with the PR/defect number in the commit text it automatically links the code in the repository with the defect/requirements management tool.

At a previous job I remember having to fill an excel, send it to some people to 'sign' it, receive it back, attach it to some request emplate which was sent to the build manager so that he manually builds your stuff by triggering a make file. Oh, and those values in excel had to be filled again into some browser window. And if you somehow crossed a magical number of changed lines in a file, you had to fill in missing unit tests from all the people which added stuff before you, so naturally code changes tended to be very terse.

Sorry for the long rant, it's just that sometimes things exist only so that people who know nothing can have a job. And it also forces the creative people into a mindless pattern.

[Rasz]

http://www.datasheet.net/document/654919/

register, click create snippet