diy solar

diy solar

How about a solar generator repair?

The board isn’t very clearly defined.
That would take all the fun out of it. Just please be careful ;-)
Who knows, if you can get them cheap enough on eBay perhaps you'll soon have a nice sideline of refurbs?

The charging board, and the main board are all one...
You can probably recognize the BMS portion of the PCB by the cell wires and the thick(er) wires for the battery +/- to the BMS (or possibly thicker traces on the PCB).

I will post some pictures of it all tonight.
Sweet!
 
I can help if you need it. (EE circuit designer). Those look like 4 8-pin fets which typically have DDDGSSSS pinout. I can annotate the connections on that picture or you can take a higher resolution one which would make it easier to see the connections. Based on the low res one the solder globs are shorting pins which are already supposed to be shorted anyway...
 
Ok, here is the dc side of tw board I see a lot of labeled ports i guess I can hook the pack back up and test voltages etc
 

Attachments

  • 20210424_205728.jpg
    20210424_205728.jpg
    653 KB · Views: 24
  • 20210424_205645.jpg
    20210424_205645.jpg
    123.7 KB · Views: 23
  • 20210424_205625.jpg
    20210424_205625.jpg
    108.3 KB · Views: 22
  • 20210424_205547.jpg
    20210424_205547.jpg
    679.5 KB · Views: 22
  • 20210424_205533.jpg
    20210424_205533.jpg
    98.2 KB · Views: 22
  • 20210424_205516.jpg
    20210424_205516.jpg
    131.4 KB · Views: 20
  • 20210424_205441.jpg
    20210424_205441.jpg
    203.5 KB · Views: 23
I see red blade fuse F1 in the first picture under the white RTV by the two caps.
Can you take picture of the whole Top and Bottom side of the board?
 
The blade fuse is good, battery voltage measured on both sides.

I will take a wide pic of the entire board. I didn’t figure the high voltage side was important yet...
 
Yes, the battery sends 10.9V to the blade fuse and the BMS sense leads all test correct. 3v 7v 10.9v
 
Put alcohol on the board (wet it), then try to power it up. (work fast)

Look for any place where the alcohol evaporates VERY quickly. Chip, around a resistor etc.

You can sometimes find where the power stops flowing through the circuit that way.
It will often terminate in a high resistance and generate heat at a Mosfet or some transistor....or a IC chip.
 
Here are some suggestions:

- the zeroth thing: if there are any digital circuits like a microcontroller or other processor, you can probably kiss the chances of fixing this goodbye as even if you replaced the part you wouldn't be able to program it or even acquire the replacement firmware. (This is assuming that part of the design is at fault). I didn't see any though.
- the first thing is to power it on and carefully probe every voltage which you know should be some value (ex: all the ones with silkscreen labels like +5, -5, +12.6, etc.) and see if they are the voltage they purport to be. You can use an obvious ground plane for the negative terminal, which you can see is a region that a bit "taller" underneath the soldermask. Anything that says "VDD / GND" you should probe the voltage of those and record it; it's a DC power supply rail.
- based on the gigantic transformer, it's likely this is an "isolated" regulator / MPPT / whatever the hell it is. So everything on the top half of the board relative to the transformer has one ground and some supplies, and everything on the bottom half has another ground and its own supplies. So, stick to one side or the other when you're probing around. You can confirm this as there is often an obvious "split" visible on the board to denote that nothing should cross that boundary in the design, except the transformer.
- look for any swollen or burned-looking electrolytic capacitors, the "cylinder" ones. Those have a much shorter lifespan than the SMT components like resistors and ceramic capacitors. One strategy if you really discover nothing of interest is to just start replacing things. I would start with power FETs and electrolytic capacitors in that case
- check any fuses, as someone mentioned above
- look closely at every smt component for cracked solder joints or cracked ceramic capacitors; for example using a microscope or magnifying glass
- look for any integrated circuits with melted-looking plastic cases or discolored areas of the board, as things tend to get hot before they burn out
- the transformer (big thing with yellow tape) is probably fine. With the board off, you should read something small like 0.1 ohms between the primary leads and between the secondary leads (and any tertiary coil, if there is one), and you should read infinite / megaohms from primary-to-secondary. It's rare that these fail though.

Knowing nothing about this design, the vertical board with the small toroidal transformer or inductor looks like a high voltage power supply; they're often on separate boards that get built and tested individually then inserted into lower-voltage boards. That's totally a guess though. The huge transformer is almost certainly part of a regulator on the power-path, meaning if this is an MPPT or something, it's part of the buck regulator, along with a hefty set of mosfets for chopping the current. The "large" stuff at the bottom of photo *625 looks like an output filter, possibly smoothing the output of a switching regulator. It might be an L-C filter with the red thing being a big (tantalum?) capacitor and the vertically-mounted toroidal coil, also wrapped in yellow tape, being an inductor in that filter. If this is the case, the coils are probably fine and the only thing to check would be that the capacitors are not blown.

If you have the part number for this product you can search for a wiring diagram and/or ask the manufacturer. Old electronics they would give out the wiring diagrams since they warranted servicing (ex: AM radios, appliances). But for this type of design they might not give that out as it's proprietary information. If you have the brand and model number of what we're looking at, please share it.

To me, the board looks like it's in pretty good shape. So my money is on a blown capacitor or FET.

Edited several times as I thought of more things.
 
Last edited by a moderator:
Drinking game: For each wrong guess, we have to drink a shot. It won't actually help @Supervstech, but it'll make it more fun. No guessing while driving!

Photographic quality is outstanding! Perpendicular to the board so we can see the writing on the PCB is great, angle shots we can look for discoloration/bulging.

Some of the photos are a bit of a puzzle ... what would help me is a wide-angle "reference shot" of everything with "markers" on stickies. Every other photo could then be referenced back to that marker. Ideally, that photo keeps getting updated as you test through the circuit with voltages/test-results. For example, color over traces with green for 12V, blue for 5V, Green dots on things you've tested (I typically use the windows built-in "snipping tool" as it's fast/easy at that sort of stuff).

It would also help if the photos had shorter more meaningful names, otherwise when I ask if the white connector in photo 20210424_205533 is of the balance leads most people won't bother to look as too time-consuming to correlate. (update: also helps me from making mistakes, just had to update one of the photo names in this post).

What's the battery's cell configuration (e.g., xSyP)?

Is CH2 FAN in 20210424_205547 a photo blemish, or is the connector actually partially melted? If so, that might be a trouble area.

At the top of 20210424_205547, water damage or did a protective layer on the PCB melt off? What component is on the other side?
Might be worthwhile to test it.

To me, the board looks like it's in pretty good shape. So my money is on a blown capacitor or FET.
Nice list! Using the mosfet testing instructions in the video from post #17 those should be easy to test.
As it makes sense for the display to be DC powered and it doesn't light up, I suspect a mosfet in the BMS keeping DC power from flowing to the rest.

Put alcohol on the board (wet it), then try to power it up. (work fast)
So that's how they tested that before IR Imagers!
 
Last edited:
Can you put a working bms on it and see what happens? Then you've isolated the problem.
 
Drinking game: For each wrong guess, we have to drink a shot.
Lol, I would be dead already, I think. Fortunately there's no way to be wrong when you're throwing out suggestions! Here are some other notes that may or may not be useful. More just observations since circuit boards are fun to look at and we might all learn something from it:

- The "split" I mentioned between the two isolated sides is where the two physical slots and the Q.C. sticker are in picture *441. The transformer straddles this.

- J2 also appears to straddle the split and is likely an opto-isolator providing a signal across the isolated halves of the design. Is it burned-looking or is that my imagination?

- D17/D18 atop picture *441 (and probably 2 more to the left under the foam) are back-to-back diodes so the 4 of them together is probably a bridge rectifier. This is smaller and probably to supply the circuitry on one of the the vertical boards.

- D1-D4 are another bridge rectifier in picture *625, on the output side of the transformer. These are pretty big so this is probably the higher-current / lower-voltage side of the design and is rectifying the chopped signal coming out of the secondary of the transformer.

- The 4 FETs from the first photo (Q130, etc) seem like non-switching power path FETs for control circuitry (because they're smaller and not heatsinked). Since DC input is in the bottom left of photo *441, you should probably start by tracing the voltage through those FETs. Or, with it off, trace the connectivity first and/or use mosfet testing techniques like @svetz mentioned. I think their source is the left 3 header pins of the row of headers that ends with "GND" at the bottom of photo *441, and their drain is the shared "island" in the middle where all 16 pins of the 4 devices are connected. Their gate is the singleton pin of each package.

- Not sure if there is more stuff to see under the various pieces of foam, but it's worth looking under there. If the foam is for cushioning to some structure, maybe something got crushed underneath it from excessive force.
 
Last edited by a moderator:
typically have DDDGSSSS pinout
This was a mistake, should have said "SSSGDDDD". The reason source/drain matters is because discrete MOSFETs have a diode included from source to drain. Integrated FETs only have source/bulk and drain/bulk diodes unless you instantiate one explicitly between source and drain.
 
Here is an overall top and bottom shot.

The fan connector has silicone on it just like all connectors did.

There are no signs of failure or heat.
 

Attachments

  • 20210424_214312.jpg
    20210424_214312.jpg
    631 KB · Views: 13
  • 20210424_214300.jpg
    20210424_214300.jpg
    768.9 KB · Views: 13
Put alcohol on the board (wet it), then try to power it up. (work fast)

Look for any place where the alcohol evaporates VERY quickly. Chip, around a resistor etc.

You can sometimes find where the power stops flowing through the circuit that way.
It will often terminate in a high resistance and generate heat at a Mosfet or some transistor....or a IC chip.
I have an IR camera... I can take a look at it... but nothing gets hot I can see...
the battery was in when I got it. And nothing powers up at all... the ONLY indicator I have of function is the LED strip blinks when plugged in...
 
Also...
This unit does not have a display of any kind.

the only indicators of function would be the led strip lighting up.
 
Ah, I didn't read the first post very carefully. I guess we're looking at a BMS / MPPT / Inverter all in one! I could go back and update my lists but I think they're sort of general observations anyway and still mostly valid.
 
Ah, I didn't read the first post very carefully. I guess we're looking at a BMS / MPPT / Inverter all in one! I could go back and update my lists but I think they're sort of general observations anyway and still mostly valid.
Yup.
Mppt, BMS,inverter, and usb buck converter all integrated. Very annoying.
 
As for the foam pads, they are VERY well attached. Only the one pad covering the BMS labels can move at all.
I could soak it in wd40 and see if they loosen up.
I don’t have a heat gun... I guess I need to get one.
 
Back
Top