I need a charger schematic YLT6722000

[Alexbp]

before repairing the battery 49v after 5 minutes of charging 54v means that the battery is charging, now 54v

[Alexbp]

Before that, everything was fine for a year and a half.

[Alexbp]

after 2 months a 49v battery connected to a 2.3v motorcycle

[Alexbp]

battery

[eccerr0r]

16S7P Li-ion... at 54V it's almost dead, so it'd suck up a lot of current.

Don't know, would have to reverse engineer that circuit in depth to figure out what's wrong and why the fr207 is heating up.

[Alexbp]

This is understandable, but the current consumption is 2.10A

[Alexbp]

it turns out if the battery is completely dead this charger is very weak

[Alexbp]

how it is possible, the developers should know about it

[eccerr0r]

Again the MUR1060FCT is the diode dealing with the 2A/130+ watts to the battery pack. The FR207 is only powering the electronics to control the charge and theoretically should be far less than 1A.

[Alexbp]

I understand that, but why does the FR207 burn out?

[eccerr0r]

That would require a more careful reverse engineer of your circuit...Don't know. Has it always been like this?

[Alexbp]

ordered mur1060fct and fr307 order will arrive within the next week

[Alexbp]

I will try to replace it, I see no other way, although it is unlikely to do any good

[Alexbp]

If I understand correctly, it is necessary to change the scheme, but it would entail a lot of costs

[Alexbp]

perhaps there is a better solution to this problem

[CapLeaker]

Originally Posted by Alexbp

without load all cold voltage at the output 67.4v, when you connect the battery is strongly heated FR207, the output is 54v

Sounds about right if the battery is totally flat

[redwire]

FR207 diode datasheet calls them (FR107,207,307) "fast recovery" but not really. It's a bit slow, rated 500nsec recovery. Higher voltage diodes are also slower. FR207 is 1,000V which is not needed at all. 1A should be enought for the charger control and LM358.
We don't know the charger's frequency, but I find a slow rectifier with no snubber will run hot and even fail, despite the load being small.

I would upgrade to a faster recovery diode with lower voltage, under 400V. Like UF4004 (50nsec) or BYV38-200 (30nsec) etc.
A mistake I have made is thinking the diode failure is due to current, so I upsize the diode but it's worse because bigger diodes are typically slower. FR307 is still 500nsec. slow.

The charger might have a design problem with spikes/ringing coming out of the transformer, if there is no snubber network across the secondary-side rectifier diodes. Or the primary snubber is inadequate.

[CapLeaker]

Originally Posted by Alexbp

I understand that, but why does the FR207 burn out?

Put a higher power rated and faster diode in place of the fr207. Theoretically you could put an ultra fast recovery diode there and heat sink it. These fr whatever diodes are not the best diodes to start with and I usually replace these with something higher rated. Then I never had anymore troubles with it.

[eccerr0r]

Worth trying a faster diode, but still does not explain why the FR207 though always under load, only heats up when the main diode is loaded? Control logic should be about the same load regardless of main output load.

These chinese mains PSUs tend to be lower frequency due to size of the switching FET needed...

[redwire]

The charger is pretty much at the max. for low cost architecture using a flyback converter for 135W. At this power level, the design has to be well done but this one is basic and low cost.

Under heavy load charging a battery, the transformer gets max duty cycle so both secondary windings will see the ringing and spikes at their worst.

The low voltage rail with D3 FR207 would get hot if the diode is slow and trying to switch off when the waveform is ringing, the diode will turn on and off many times during one pulse. When the diode is stuck on, slowwwwly turning off it will allow a -ve spike to flow through the 100uF 35V cap. LM358 can run to 36V but not handle -ve spikes. There's no other reason I think for it to heat up, how could it see a few A otherwise.

Main rectifier MUR1060 has R31//R32 4R7's with C7, as a snubber but R31 looks like dark splotch. If R31 is burnt, then this confirms the PSU has some design issues causing extra stress to the rectifiers. You would have to poke around with a scope to see if my guess is good.

R21 (27?) and R23 by U3 LM358 also look cooked, but maybe it's flux?

It could also be an the primary RCD snubber, as I see the usual 100V film cap there and they usually undersize the big resistor or use a crappy FR107 there.