converting meanwell power suply into Lipo charger

[cor]

Neil,
I traced out the entire NES-350-48.
Interestingly, the overtemp protection of the supply is combined with the fan control and the protection is combined with the overvoltage protection (the two series zeners that were mentioned earlier in this thread, one 1N4746 (18V) and a 1N4755 (43V) that are simply wired between the 48V output and the opto U3 that will trigger SCR1 to shutdown the supply until power is removed.
If you need 12V (referenced to the negative 48V output) or want the fan to run continuously then you should connect it directly to the output of the linear regulator RG1 which is mounted on heatsink HS2 next to the big copper coil and the fan connector. It is as simple as it looks to make the fan run permanently: simply connect the fan pin that is closest to the regulator with the regulator pin that is closest to the fan connector. The transistor Q150 that is wired in between the regulator and the fan connector is to switch the fan on/off, so to make a remote control of the fan, you can take the C pin of SHR2 and connect it to the negative 48V (ground) output.

For current control, I think it is better to change resistor R134 than to mess with R127 and R128, although they are in the same circuit.
Reason is that there is only 59 milliVolt across the R127 and R128 and when you add a circuit with wires and potmeter, I expect that you pick up additional noise and can make the feedback instable. If you just add an extra resistor on the R129 position then I do not expect a problem.
If you want to reduce the current, you could also unsolder the shunt J100 and replace with a higher resistance version, but my suggestion would be to unsolder R134 which is located under the transformer, next to two SMD jumpers of 0 Ohms. Its value is 3k3 and it goes from the 2.5V reference to the pin 6 of U100 which must remain above zero or this opamp will drive the U2 opto to reduce the output of the supply.
If you want to place a potmeter to turn the current down then I suggest that you either use the existing R134 and cut the trace between the SMD resistor and C125 and place the pot between these two points or that you place a 3k3 resistor very close to pin 6 of U100, then a capacitor to ground (pin 4 and 5 of U100) and then a potmeter between this point and the 2.5V reference SHR1.
If you want to be able to turn the current down to zero then you can connect a potmeter (for example 5k) with its ends across SHR1 and the slider connected to R134 (3.3kOhm).
If the output voltage regulation is not allowing you to turn it down far enough then you can replace SVR1 (1k) or R124 (2k) with a larger value.

[fechter]

Good work! I wish we could get a schematic of that.

It's funny how they make changes to the circuit but apparently have lots of older version boards or don't want to make a new one, so end up with strange layout features. Changes seem to happen quite often as well.

It would be important to find out if the current limit can be lowered with R134 and maintain stability.

[NeilP]

Hell Cor , that is amazing

I need to order another like this to try out your findings

Having an easy fan on mod and single resistor to do current mod is what we need, just as you seem to have found

Does R134 need to go up or down from 3k3 to reduce current, sorry did not quite get which way it works. May get it when I re read at home, and nog on iphone, but for now, cant get it

One question, not connected with this supply, but in what you are doing in general. How are you tracing tracks etc? By eye? Or do you use meter as well checking points as you go. And as you do this, how are you recording what you find? Big sheet of paper? Software?

[NeilP]

Regarding the overvoltage protection and the two series zeners, does that mean we are good for up to 61v well 55 to 60 to be safe on these supplies without having to mod the zeners?

[cor]

Good work! I wish we could get a schematic of that.

Thanks! I have a hand-drawn pencil and paper one here... Anyone wants to feed it into a computer for an electronic version? Otherwise I can scan it but it might not be too easy readable.
Most components I noted the value, some include the reference number so you can find them easier on the board.
Almost all of the schematic is straightforward what you would expect, except that some choices are made to use specific inputs and not others - the VFB input of the PWM controller is not used, all feedback is tied to the COMP input.
The voltage and current limit are combined (using U100 and two diodes) into one feedback opto (U2) and the over-voltage protection as well as the over-temp protection share the other opto U3.
The only quite tricky part is the use of a single thermal sensor RTH2 to both turn on the fan (when the voltage on the OTP input to SHR2 has risen to 2.5V) as well as detect overheating (when the OTP voltage has risen high enough to turn on opto U3 through R153 and D121, which triggers SCR1 so the voltage on the COMP input is shorted and the controller shuts down. One of the other smart things is that the over voltage/temp protection will cause a hiccup operation, because SCR1 is tied to a supply voltage from the transformer. Once the supply cap C36 that also feeds Q5 has been drawn down far enough for SCR1 to let go, the controller will restart and the protection cannot engage before the transformer has put some charge in the caps again.

It would be important to find out if the current limit can be lowered with R134 and maintain stability.

I see no problem with that, because R134 is essentially a current source (or voltage divider if you will) to lift the reference voltage from the shunt resistor by 59mV to keep it above negative output reference, which is the level at the positive opamp U100 input pin 5. As soon as current gets too large, the shunt drops more than 59mV and since it is in the negative output, the upstream side goes more negative than 59mV so the +59mV voltage shift by R127 & R128 no longer is positive and the negative input pin 6 being lower than pin 5 means that the output becomes positive which fires the feedback opto U2 through D122. Reducing the current through R134 will reduce the lift by R127 and R128, so the voltage goes negative at a lower current through the shunt. In theory you can turn the current limit down to zero. I need to fix this NES before I can try that experiment.
To avoid blowing it right away, I want to test as much of it first, so I will feed voltage to the output of the secondary rectifiers to see if the circuits after them work as expected (2.5V references and such) then check the PWM controller by supplying power via Q5 (which is an emitter follower with a zener on its base)
then measure all semiconductors for dead shorts and finally apply AC and see what happens...

[cor]

Regarding the overvoltage protection and the two series zeners, does that mean we are good for up to 61v well 55 to 60 to be safe on these supplies without having to mod the zeners?

I think that is also what the datasheet says.

[NeilP]

I will be the first to admit that I would not really understand exactly what I am doing, and so far only have a basic grasp of the software, but if I can be of help, i can at least try and convert your paper and pencil in to a diagram on the Circuit schematic software that fechter introduced me to when I was trying to get my CC/CV Board working, assuming that is suitable for the job? Richard. Can I use that ? is it the right sort of package for this job?

[cor]

Hell Cor , that is amazing

When I was in highschool, I started fixing TVs as hobby with a screwdriver, a soldering iron and a green LED.
From the sale of the first fixed TV I bought my first analog multimeter.

Does R134 need to go up or down from 3k3 to reduce current

It needs to go up in value. The amount of current through this resistor determines how much current the supply can deliver.
This resistor is connected to a 2.5V reference supply, so you could make a mod that allows the supply to be turned down to zero current by keeping R134 at 3k3 but connecting it to the wiper of a potmeter that sits across the 2.5V reference supply, so that you can choose the reference voltage anywhere between 2.5V and zero which would make the output current of the supply also go down to zero.

One question, not connected with this supply, but in what you are doing in general. How are you tracing tracks etc? By eye? Or do you use meter as well checking points as you go. And as you do this, how are you recording what you find? Big sheet of paper? Software?

Since this is only a single layer PCB, I am tracing the tracks by eye, only when it gets tricky under SMD components or to verify the value of a component do I use a meter. I use pencil and paper. If I have time I may fire up that old evaluation copy of a schematic drawing program, but it will take me a long time to get used to it again and enter it. I'd appreciate if someone has a bit of time, or I may publish my sketches...

[NeilP]

Well, scan in highest res and PM them to me and i'll attempt it
Not promising anything but want to contribute if I can

[neptronix]

Hell Cor , that is amazing

When I was in highschool, I started fixing TVs as hobby with a screwdriver, a soldering iron and a green LED.
From the sale of the first fixed TV I bought my first analog multimeter..

That's pretty heroic. We have some pretty amazing people around here.

[cor]

Well, scan in highest res and PM them to me and i'll attempt it
Not promising anything but want to contribute if I can

Hi Neil, that would be great.
I will make a scan at work, here I only have a camera and I doubt that the result is very readable:

[NeilP]

Hey, that will get me started...at least if nothing else, working on getting my head around the software again..I only had some work stuff i really should be doing today..so that can wait

I have done a perspective crop using photo shop to 'square it up' and printing it out now

Even if I just try placing an few components and finding the ones that are not in the built in library of Express SCH

[NeilP]

Did you maybe take that at a higher res on the camera and reduce for the forum? If so could you maybe e-mail me the full res copy?..if not no worries, i'll wait for the scan

Trying to zoom in and the res goes a bit

Before you do the scan, could you maybe re mark with Heavy Dot wires lines that are actually connected as opposed to just cross over. I can probably guess some of them..but not all.
For instance U100 Pins 3 and 4 have cap across..but would not then be connected common to pins 1 and 2...I think....

[NeilP]

Got this far so far. need clarification in the red areas...
Component after R1 and before C1
RTH1? relay?

other side of transformer


Have added the ExpressSCH file in case anyone else wants to work on a bit more

[Alan B]

Just in case folks don't know Neil is using free software from Expresspcb.com to draw the schematic. It takes very little time to learn this simple schematic drawing package. It is a great way to share schematics.

[NeilP]

Yes, sorry I did not make that very clear. I have included the link and again here ExpressSCH

It is a bit longwinded to do though. Lots of the components like inductors and transformer coils have to be all drawn manually. I have d/loaded various other library components, but still not enough to do all the work.

That little bit above so far took probably 4 hrs . Ok I am working from a print out of Cor's orignal, so that is making things a bit harder

[Alan B]

One thing you can do is use rectangles for transformers and coils. No real need to draw them the old fashioned way. Much faster to do.

[NeilP]

True, but it makes it easier to see what it all is when you come to look at it ..or someone else looks at it later..Plus once you save them as Custom components you can use them again later

[cor]

Got this far so far. need clarification in the red areas...

Great progress! Don't be afraid to go ahead and draw what you think it shows, I will review and correct where it differs with my sketch.

Component after R1 and before C1

That ZNR1 is not a usual zener but a surge protector which could be thought of as two zeners in series, one reversed, so that the voltage limit in each direction is the same (the voltage here is AC). Most likely it will protect (conduct) at a voltage around 600V. It serves to remove high voltage spikes to protect the bridge rectifier which might otherwise break down from too high reverse voltage. The symbol for it is a simple Z

RTH1? relay?

RTH1 is a temperature dependent resistor, most likely a NTC (Negative Temp Coefficient) which gets lower resistance when it warms up. It has a "precharge" function, to reduce inrush current when you first plug in the supply and the big 680uF 200V capacitors are empty, so the grid will try to fill them instantly. This resistor is initially a high value, so the current is relatively low and does not blow the fuse. The current through this resistor warms it up, dropping the resistance while the capacitors are charging so the voltage difference between grid and capacitors reduces and the power consumption in the resistor reduces. The relay will short-circuit this resistor as soon as the power supply starts working, so there is no unnecessary loss in the input conditioning circuit once the "precharge" has completed.

other side of transformer

I will try to scan that clearly. In essence 3 times a winding with a single diode and a snubber across it. A snubber is a series circuit of a resistor and a capacitor to dampen transients from the diode switching off. The topmost winding actually comes out the top of the transformer and this powers the fan circuit which is regulated to 12V. The bottom secondary winding has triple connection pins for higher current capacity, otherwise all 3 windings would have been able to come out on the transformer pins just like the primary side, because it has 6+8 pins. (2 pins not populated on primary side)

The only drawing error I see is the connection of the line that runs down from the diode but instead it should be connected to the lower contact of the topmost primary winding. At the bottom of the transformer there is an R across the diode, but this should be an R and C in series.

Do you have a datasheet with drawing of TL3845P? You may be able to download a symbol for the TL384x from Texas' website. Great work so far!

[NeilP]

RTH1 is a temperature dependent resistor, most likely a NTC (Negative Temp Coefficient) which gets lower resistance when it warms up. It has a "precharge" function, to reduce inrush current when you first plug in the supply and the big 680uF 200V capacitors are empty, so the grid will try to fill them instantly. This resistor is initially a high value, so the current is relatively low and does not blow the fuse. The current through this resistor warms it up, dropping the resistance while the capacitors are charging so the voltage difference between grid and capacitors reduces and the power consumption in the resistor reduces. The relay will short-circuit this resistor as soon as the power supply starts working, so there is no unnecessary loss in the input conditioning circuit once the "precharge" has completed.

Ok, thanks

Looking at the switch part of RTH1, is there an activation coil and connector drawn in, just below it.Looks like but cant be sure from the drawing

The only drawing error I see is the connection of the line that runs down from the diode but instead it should be connected to the lower contact of the topmost primary winding. At the bottom of the transformer there is an R across the diode, but this should be an R and C in series.

Umm can see where you mean but cant see how it should be, for that to work as I see it drawn, that puts a dead short across the capacitor that comes off the bottom coil of the topmost coil. If that cross is not a connection, then it seems to be drawn showing the diode connecting from top of top primary coil to top of second primary coil, next to a capacitor, resistor and capacitor in series.
it is often difficult to see where two wires just cross or are an actual connection


is going to be difficult to do any more for the next few weeks...the bloody girlfriend has broken her hand, falling off her bike..so is at home all the time. I was lucky to do what I did yesterday morning. She came home from a friend at midday..and all work stpped...just the incessant whittering that women see to have to constantly engage in. Have been on the PC since 8 this morning, it is now 1015, and I still have not done more than look at the Schematic to write this reply

[cor]

The RTH1 is a separate component. The switch across it is part of a relay with a dotted line to a coil that connects to the negative output of the bridge rectifier and the other side of the coil goes through a 510 Ohm resistor to the emitter of Q5 which delivers the power supply for U1 once the unit is running.
My convention for connecting wires is a plan cross is not connected unless a dot is placed on the cross.
T-junctions by definition connect, with or without dot.
Keep asking questions where things are unclear. I will scan the sketch later today and send it to you.
Sorry to hear about the GF. I am sure she did not plan to break her hand or fall off the bike in the first place and she probably feels bad enough about it in addition to the pain from her hand and other places that may be chafed/bruised. So be kind to her and she might even recover faster
I am in no hurry with the schematic, though I might in the mean time sketch the mods to run the fan and to control the current so others could try it before this work is done.

[NeilP]

Yes probably better that I wait for the scan

Or is that Capacitor, resistor capacitor string from top of prmary top coil to top of priary second coil..and no connection to the bottom of the top most coil at teh bottom..The only connection to the bottom of the topmost coil disappearing off downwards to the Pair of FETS?
Ah yes..guess that is it..two FETS back to back?

Still does not explain where the connection from the top of the second prmary coil connects to

So mayb the only connection from the bottom of the first prmary coil is to the Drain of the RH FET?

OK< thinki get the Relay bit..was thinking the RHT1 and relay unit were one combine package

[NeilP]

Just to make sure we are talking about the same diode

The only drawing error I see is the connection of the line that runs down from the diode but instead it should be connected to the lower contact of the topmost primary winding.

Primary side.jpg (10.48 KiB) Viewed 455 times

Bearing in mind a cross is not connected, that diode there is not connected to the topmost primary winding


it is definitely connected to the top of the second winding in the picture. and not the lower contact of the topmost. I shall await the scan

[cor]

The bottom of the top primary winding is connected to the R and to the C in addition to the Drain of the FETs. The FETs have all contacts connected (they run in parallel) though the gates are connected by a pair of 5.1 Ohm resistors to the driver.
The top primary winding has a snubber (R and C in series) across it for when the FETs switch off you get a large voltage spike that must be dampened.
The top and middle primary windings are connected by a capacitor and the middle winding is attached to the neg supply and its top contact connects only to the cap and a diode to the pos supply. So the only function of this winding is to give a hard limit how high the flyback voltage of the transformer can go. Since the duty cycle of the controller is at max 50% you would not expect the flyback voltage to be higher than the forward switching voltage, but the middle winding actually clamps it, so that all the voltages on the transformer are limited and stresses on the connected FETs and diodes reduced.

[cor]

Just to make sure we are talking about the same diode
Bearing in mind a cross is not connected, that diode there is not connected to the topmost primary winding
it is definitely connected to the top of the second winding in the picture. and not the lower contact of the topmost. I shall await the scan

Correct, the diode goes from the top of the middle winding to the top of the topmost winding (pos supply).