Switch mode supply mod

[Arlo1]

I blowed mine up at 10 amps how do you get 18?

 

[Hyena]

24 or 48v ?
I parallel charge 6S at 25.1v

 

[Arlo1]

24 or 48v ?
I parallel charge 6S at 25.1v

2 48v 350w meanwells in series!

 

[Hyena]

10 amps through a 48v supply (set to 50v no doubt) is 500w which is well over their 350w rating. 450w is the most I'd run, 400w for safety.
I'm sure I told you this about 6-10 months ago :lol:

 

[Arlo1]

10 amps through a 48v supply (set to 50v no doubt) is 500w which is well over their 350w rating. 450w is the most I'd run, 400w for safety.
I'm sure I told you this about 6-10 months ago :lol:

Haha yeh and since i listened to your advice i have not blown any up!

 

[myzter]

I just got a reply from ebay seller : kitsparts888
regarding the listing with "Current Limit Control VR"

limit control VR at the RHS of the connector

So I emailed the seller back because looking at the pictures of the PS they sell..
I think I can see an adjustment dial/screw on both sides of the input/output terminals.


by Hyena :

Anyone played with higher current units ?
I just came across 500w ones on ebay that claim "Current Limit Control VR"
http://cgi.ebay.com.au/500W-24V-20A-Swi ... 3a637f3558
I get 18 amps out of my 24v 350w unit reliably so I wonder if there's too much different in these. Probably just an additional fet, but having an adjustable current limit out of the box could make them worth the extra.

 

[NeilP]

One of the other posters here has asked the same question about the higher current units...but not the little 500W units...talking the big ones here...3kW or 1.5kW in a single supply and they are load balancing so you can stick two or three in parallel for real fast 5C charging

http://www.meanwell.com/search/spv-1500/spv-1500-spec.pdf
http://www.meanwell.com/search/rsp-3000/default.htm

think I did see something about not being able to use the Voltage trimmer though if load sharing is enabled..maybe they mean voltage trimmer is not available on all but one of the supplies...and one of them controls all the voltages...that would make sense.

No idea what these units cost though...Not planning on looking it up yet though...but the extra cost is probably easily offset by the fact you have one unit to wire up rather than two or three.


Here is my unit finally finished

I got my two S-350-48s wired up.
Mods done are:
Fan mod Q5 resisor C-E bridge with switch
CC/Cv board switchable in/out for series or parallel ops
SVR1 replaced on both supplies
SVR2 added on both supplies
These were mounted externally ...see the front panel picture...the letter box like slot.
Actually used two pots for SVR1 on each supply and two pots for SVR2 on each supply

All pots were 16 turn cermet 5k trimmers

They are connected up to the boards via a 4 pole change over switch...so I have two voltage limits and two current limits...higher volt limit...lower current limit

The second 4 pole switch changes between series and parallel. It also switches the negative feed so, in parallel the negative feed is direct from the PSU, and in series it is through the Goodrum Fechter CC/CV board
Small switch in lower middle, is double pole and switches the CC/CV board voltage feed and voltage sense line in and out. That was an after thought fit ...as the CC/CV board was pulling the voltage of one board down when the units were in parallel

CC/CV board VR 1 2 3 just below letter box

HVC connection and Float enable jumper also bought out to front panel

Small jumper loop wires at the bottom are monitoring points. I can put in ammeter in series their to monitor current from each supply when they are running in parallel...Seeing he current I can adjust the external voltage trimmers to get even load sharing.

Swiches along the top of the board...two over ride switches for the meanwell fans ( Switch across collector -Emitter of Q5...a better mod than the thermistor mod as ..according to Commanda.

Middle switch directly drives the middle fan

Good point of doing it this way, the thermistor "stage 2" protection still works as designed. If it does overheat (because the fan failed), it drives the Dead Time Control on the TL494 to cut the current by reducing the pwm.

The people that mess with the thermistor lose this.

 

[hydro-one]

Nice !! Best meanwell ive seen!! Glad to see others are enjoying the meanwells. I currently have 6 running for people and they are pretty good. i also have a pile of blown ones for spares :lol:

 

[NeilP]

Are you in the UK or USA? I need a fan for my Right hand one...it only starts with a flick of a pen and is not running at full power.

I have had an offer from someone else here in the UK that may have a spare fan, but if not, could you have a look for me? 60mm square 15mm deep 12V 0.13A. Want one exactly the same as the current one, so it does not affect the performance of one when they are in series

 

[hydro-one]

canada and unfortunatly the fans seem to be cheapo as i dont even have any spares. i would invest in two better quality fans
as i think the unit will blow if the fan dies. (because of thermister mod, changes the thermal protection circuitry?)

mike

 

[commanda]

canada and unfortunatly the fans seem to be cheapo as i dont even have any spares. i would invest in two better quality fans
as i think the unit will blow if the fan dies. (because of thermister mod, changes the thermal protection circuitry?)

mike

Mike,

If you leave the thermistor circuit alone, and permanently enable the fan with the Q5 mod (switch in parallel or wire link Collector-Emitter), the over-temp protection should still work. This works by driving the DTC (dead time control) of the TL494, which decreases the pwm.

Also, things tend to be more reliable when operated conservatively. I run my S-350-48 at 60 volts and 5 amps (300 watts).

And yes, replacing the fan with a better quality ball-bearing unit is a sensible upgrade.

Amanda

 

[NeilP]

I was looking at getting new fans anyway, just thoughtn`i would see if Icould find any originals first
You are correct about the thermistor mod loosing further overheat protection...but I did not do that mod, I fitted a bypass switch across the transistor Q5

 

[auraslip]

I just got two of these genuine s-150-27 meanwells off ebay for $20 each.
Rated at 5.6a - very small, but no fan. These were gonna be my "at home chargers" but they may be small enough to carry around in the battery box. Probably need fans for that though!


click for bigger view.

I'm going to use fetchers limiting board with these, unless anyone can see a way I could limit the current an easier/cheaper way. I'm also going to be using isolated cell-logs and relays to break the 110v line with the HVC alarm. This is a short term solution until I can afford a new bms for my 16s ping pack.

 

[auraslip]

So in doing research on limiting current as well as using a cell-log's alarm port for HVC, I had an idea.
Basically, the lower resistance of R33 the lower the current. So if you jumped R33 via the alarm port of a cell-log the current should drop to nothing right? It'd be super easy to implement. Two wires go from R33 to the cell-log alarm port.
I know the alarm port isn't isolated though. That might make things a little bit more complex. Unless the power supply is isolated.....

 

[olaf-lampe]

So in doing research on limiting current as well as using a cell-log's alarm port for HVC, I had an idea.
Basically, the lower resistance of R33 the lower the current. So if you jumped R33 via the alarm port of a cell-log the current should drop to nothing right? It'd be super easy to implement. Two wires go from R33 to the cell-log alarm port.
I know the alarm port isn't isolated though. That might make things a little bit more complex. Unless the power supply is isolated.....

That wouldn't reduce the current and the Cell-logs open collector output may not switch completely to ground, if the voltage comming from R33 is too low. The relais version is much better IMHO.
-Olaf

 

[NeilP]

That wouldn't reduce the current and the Cell-logs open collector output may not switch completely to ground, if the voltage comming from R33 is too low. The relais version is much better IMHO.
-Olaf

Well, reducing the resistance of R33 WILL reduce the current, that is why the R33 R37 SVR2 mod works Well it is not actually a mod..it is part of the original design) ...so putting a direct short across R33 would knock the current right back to nothing. I know not if it would work with a cell log...as Iknow nothing of them...but if as you say there is a realy version, then that would be the better option, and rather than shorting across R33, dropping the value to a low amount to reduce the current to a trickle current would be better I think

 

[olaf-lampe]

I stand corrected; with a cell-log it would be an ON/OFF behaviour rather than a regulation.
-Olaf

 

[fechter]

That wouldn't reduce the current and the Cell-logs open collector output may not switch completely to ground, if the voltage comming from R33 is too low. The relais version is much better IMHO.
-Olaf

Well, reducing the resistance of R33 WILL reduce the current, that is why the R33 R37 SVR2 mod works Well it is not actually a mod..it is part of the original design) ...so putting a direct short across R33 would knock the current right back to nothing. I know not if it would work with a cell log...as Iknow nothing of them...but if as you say there is a realy version, then that would be the better option, and rather than shorting across R33, dropping the value to a low amount to reduce the current to a trickle current would be better I think

An alternate way to do this would be to use the CellLog to trigger a relay or optocoupler and use the output to pull up on the voltage sensing line. This will drop the output voltage of the supply. As soon as the supply voltage drops below the pack voltage, the charge current will be zero (or slightly negative). If you just lower the charge current, it might continue to charge the cells, just at a lower rate.

 

[wineboyrider]

So in doing research on limiting current as well as using a cell-log's alarm port for HVC, I had an idea.
Basically, the lower resistance of R33 the lower the current. So if you jumped R33 via the alarm port of a cell-log the current should drop to nothing right? It'd be super easy to implement. Two wires go from R33 to the cell-log alarm port.
I know the alarm port isn't isolated though. That might make things a little bit more complex. Unless the power supply is isolated.....

Great idea! Did you get those meanwells from allelectronic.com ?
Keep us posted if the celllog idea works. Would definitely merit a new post if it does

 

[auraslip]

 

[auraslip]

An alternate way to do this would be to use the CellLog to trigger a relay or optocoupler and use the output to pull up on the voltage sensing line. This will drop the output voltage of the supply. As soon as the supply voltage drops below the pack voltage, the charge current will be zero (or slightly negative). If you just lower the charge current, it might continue to charge the cells, just at a lower rate.

Ok, that should be easy to implement. Just break the connection between svr1 positive and negative.

Great idea! Did you get those meanwells from allelectronic.com ?

I got them off ebay... they still have 3 of them left.

 

[auraslip]

So for all of you wondering: the current limiting and voltage adjustment section of the s-150 is near identical to the s-350. I removed r33 and installed a 1k pot in parallel with the other 4k resistor on the current sensing line.

One thing I noticed is that the resistance of the pot cannot be too low or too high. It will only power up if it's with in the correct range. If it goes out of this range while operating, the voltage will drop to 14v even after the r33 pot returns to the correct value. This makes me wonder if my original idea of using a celllog and relay to jump r33 is viable. Since the voltage drops and latches well below the voltage of the battery pack, their should be no current flowing to the battery. (what about to the PSU?) I'll have to experiment with this and something similar with svr1.


I got it charging at 5.5amps - the rated current is 5.6amps - I wonder how much over I can go? Would adding fans help in any situation?

 

[NeilP]

Good to know

Is that a $0.99 ebay DVM I see there? I had three...think I will order some more

 

[auraslip]

Yup. It actually is better than the $20 one from wal-mart. The one from wally world doesn't measure resistance correctly, or at all.

 

[Arlo1]

Yup. It actually is better than the $20 one from wal-mart. The one from wally world doesn't measure resistance correctly, or at all.

Which one is which?