Tuning cheap 42V charger

[WheezyRider]

A while back I bought a couple of cheap 42V chargers as spares, so I could have one at work, at home etc. I got two for just under £10 each. They are quite a lot smaller than the SANS charger and looked good:



However, when I checked the output, I was disappointed to see that it was nearly 43V!




Which is quite worrying and I am glad I checked. I wondered how many other people had just plugged their battery in and used this type of charger. Especially as I've found that a lot of basic multimeters are not very accurate in the 200V range. A 1% error at 42V is typical for many, some may be 2% out, and if 42V is critical for reasonable battery life, it's easy to end up damaging things with these cheap PSUs, which are obviously not tuned in the factory to give 42V or less.

I opened up one of the PSUs, hoping to find an adjustment pot. However, there wasn't. So these chargers have sat on my shelf for a couple of years.

Recently I was forced to look into ways of tuning a switched mode PSU, as I bought an expensive charger, which I wanted to run at 41V.

Mikes eBikes 3A charger

I picked up one of these recently and thought I'd give it a once over. It looked quite compact, came with a fan, already had an XT60 connector and I could do with something that charges a bit faster than my SANS chargers. It uses a PSU PCB company that also makes the PSU boards for the Vruzend...

www.pedelecs.co.uk

Here I was able to find the voltage regulator and voltage divider quite easily as there were signs that someone had adjusted this manually to give 42V. But what about this untuned charger?

From the circuit board I was able to find the current and voltage regulator circuits:







There are two TL431 regulators, I found the one in the red circle was the voltage control. The yellow circle is the opto isolator to the high voltage circuit. The blue circle is the top side of the voltage divider and the purple circle is the bottom side of the voltage divider.

I desoldered the 120k resistor in the top side of the voltage divider and soldered in a 4.7k pot and 100k and 15k resistors in series:



This gave 43.2V Max and 41.5V Min. So a good start. I replaced this with a 100k, 10k and a 10k pot and could get 39.7V min and 43.2 Max. From this I was able to use 100k, 10K, 3.3k, 680R in series as fixed resistors to give me 41V. I then drilled holes in the PCB so the resistors could be mounted on top as there was a lot of space on the top side of the board:





I tidied up the length of the resistor leads and added shrinkwrap. Then I put it all back together and was glad it all still worked

In terms of operation, it is now giving 41V out at 1.8A. It draws 87W and outputs 76W, which is actually better than the SANS.

I'm very glad I did this. I now have a spare charger and have saved more electrical waste from landfill - and my investment was not wasted

I could add a switch with 2 resistor networks, one for 41V and one for 42V, but for now this is good enough.

The real take home from this is that it's really important to check the voltage from your charger with a decent multimeter, to ensure it is giving out no more than 42V!

 

[cyclebuddy]

I think you're overthinking this. I routinely change the Sans output caps to Panasonic 63v Low ESR caps which are specifically designed for SM PS purpose. The cheap capacitors Sans fit are 50v rated and - lets be honest - total crap: In this application, they're hammered hard and will easily fail... sooner or later.

As you already know, low ESR implies low series resistance, and that means a charger designed for 42v then fitted with Low ESR caps will then output a slightly higher voltage. My measurements are 43v+.

In my experience, in e-bike charging applications, that marginally higher voltage really doesn't matter. The BMS copes just as well.

 

[WheezyRider]

I think you're overthinking this. I routinely change the Sans output caps to Panasonic 63v Low ESR caps which are specifically designed for SM PS purpose. The cheap capacitors Sans fit are 50v rated and - lets be honest - total crap: In this application, they're hammered hard and will easily fail... sooner or later.

As you already know, low ESR implies low series resistance, and that means a charger designed for 42v then fitted with Low ESR caps will then output a slightly higher voltage. My measurements are 43v+.

In my experience, in e-bike charging applications, that marginally higher voltage really doesn't matter. The BMS copes just as well.

I think that's a good thing to do to prolong the life of the charger.

However, if the low ESR capacitors result in a higher output voltage, then this will increase the voltage going to the PSU regulator, which will then tell the HV side to lower the input voltage to compensate, bringing the output voltage back to the preset voltage. So I'm not sure why that change would result in you seeing a higher output voltage? How accurate is your multimeter?

Do BMS come with high voltage cut off as default? Probably not, except high end stuff? So if 43V does go into the BMS, it will probably trigger the balancer to bleed off the excess maybe, but these are only designed to handle mA. Maybe ok for a short time, but not good if someone leaves the battery connected to the charger for a long time.

Have you ever checked what voltage is going to the battery from the BMS when your charger reaches 43V+?

Maybe there are enough fail safes, but I'd rather not take my chances. I also prefer to charge to 41V to extend battery life.

 

[cyclebuddy]

How accurate is your multimeter?

I'm a qualified - albeit retired - electrician. My 3 multimeters are all Fluke and all recently calibrated.

I do understand your concerns. But in reality it's fine. 2 of my e-bike batteries are 2016/7, and still give a measured capacity (on a measured charge/discharge cycle) equal today to that as they were when new. The third is 2019, used less, but still measures as new.

They are all Samsung cells, but cheapies - not the premium sort.

 

[WheezyRider]

I'm a qualified - albeit retired - electrician. My 3 multimeters are all Fluke and all recently calibrated.

I do understand your concerns. But in reality it's fine. 2 of my e-bike batteries are 2016/7, and still give a measured capacity (on a measured charge/discharge cycle) equal today to that as they were when new. The third is 2019, used less, but still measures as new.

They are all Samsung cells, but cheapies - not the premium sort.

That's good

I've been charging to 41V for quite some time now and all my packs are working as well as they did when new and I hope they will keep going for a long time. They are all "cheapies" and one pack is supposed to be LG celled, but is probably fake.

The thought of having to fork out early for a new battery pack, or worse a battery fire because I've used dodgy cheap components does worry me though.

I am curious about the voltage increase due to the caps changed not being regulated down. Can you explain why the capacitor change would result in a higher output voltage? Is it due to some kind of ripple that confuses the regulator?

I'm not an electronics engineer, just someone who dabbles.

 

[cyclebuddy]

As I understand it, regular capacitors present a higher resistance and low ESR caps present lower resistance. When measured, the 42v Sans charger reads 42v before the change, and 43+ a bit open voltage when the those output caps are changed.

Ideally, yes, charging to 41v would IMHO be a better idea for battery pack longevity. Nealh's own basic 2-wire chargers include a pot to alter output, but that isn't an option for me. My own 2 Sans chargers designed for Phylion batteries are 5-wire connected and have a circuit mod to include an onboard chip in place of that pot to monitor potential issues (overcharge, overheat etc).

I didn't really look into it further than that. I simply changed the caps to more reliable, higher rated, higher quality ones designed specifically for power supplies, and those 3 chargers have all been working fine for years now.

I have noticed though that all my chargers initially cut-out after an empty to full charge at 41v. You have to disconnect and then re-connect the chargers to then top-off/balance the charge properly at 41.7v. That's with 2 5-wire Phylion batteries, a 3-wire TranzX one, and two 2-wire Fiido batteries (the later an unmodified charger).

I think you can worry too much about these things. I only changed the Phylion charger caps because they were obviously crap, and if they were to fail they'd cause significant damage elsewhere rendering the charger useless. Being as they are tailored for a Phylion battery with 5-wire connection and integrated "policing" chip, a £10 42v charger replacement wouldn't be an option (they'd be £65-£100 to replace).