Fechter's Mini Meanwell Limiter Board

[fechter]

I'm need to revise the 3 pot version as well. It does need some serious corrections. I'm just trying to get the mini limiter going first.

The Ver3 layout also has an extra set of holes for the shunt, so double current will be supported by adding a second shunt. If not using the second shunt (even if you are I guess), the extra holes could be used for wires going to an analog meter. With a single shunt, 10 amps will be 50 mV across the shunt. I've seen some nice analog meters with 50 mV scales on Allelectronics and some other surplus places.

[NeilP]

I'm need to revise the 3 pot version as well. It does need some serious corrections. I'm just trying to get the mini limiter going first.

Ah that is OK then, as long as it is not being side lined

[teklektik]

Maybe this is just a personal thing, but that function is more useful than the current limiting. Current limiting can be done on the board, much simpler with the pot/resistor mod. To my mind, if someone is going to the trouble,effort and expense of building the board it has to offer more than you can get with 20 minutes work and less than a couple of dollars worth of components ..I a pot and some wire to do resistor mod (S or NES series) ...

Yep - it's a personal thing Moreover, the new HVC functionality makes an equivalent boardless direct MW hack less attractive.

I see the HVC can do that , but it requires an external input..cant that be incorporated in the board??

No - a primary use of the HVC is to easily hook the MW up to CellLogs or similar that do cell-level rather than pack-level monitoring. As one example using the new v.3 board: if the pack is hooked for cell-level monitoring and low rate balance charging in parallel with MW bulk charging, then the first cell that hits CellLog HVC terminates the MW bulk charge and the pack finishes and floats on the balance chargers. Here the extra functionality of the 3-pot board is unnecessary and is accomplished by other means.

This brings us to a fundamental difference in the 1 and 3 pot solutions: the 1 pot solution with HVC is more like a building block for DIY (cell-level) charging solutions with other unspecified external elements while the 3 pot board provides more of a complete pack-level charge solution in its own right.

So - different strokes for....

[NeilP]

I think maybe I did nto explain myself very well. I see your points and understand...

the 3 pot version does HVC monitoring too, which I do use on occassion when I have apcked connected up to me tppacks HVC/LVC parallel boards. I think more what I am tryign to say is..why re design and use the single pot board, when the three pot already does it. This time spent revising the single pot board and addding extra features to it may have been better spent revising the 3 pot board rather than biinging up the spec of the single pot baord to that approcahing the 3 pot version

I suppose the advantage of the single pot board is that it is cheaper because of less components, but the difference in cost had to be marginal

[Degull]

Fetcher's V3 Mini Meanwell Limiter boards have been order and should be ready next week. This is a very limited run and the price is going to be around $10 per board. The price might be considered a little high for such a tiny board but this is because we decided to order a small quantity. It's hard to judge the demand and neither of us would like to be stuck holding a bunch of unsold boards.
I will start a new thread once we receive the boards. If anyone would like to pre-order now just send me a PM and I will start a list.

[fechter]

On to the 3 pot version...
Below is the latest revision that should solve issues with the older design. The HVC interface works differently though, as this has a different intended function. In the new design, when HVC tiggers, it will cut off the charge current and latch in the float mode. The older design lowered the current but not to zero and would not latch off.

The way the circuit is intended to work is similar to the mini limiter in the way it adjusts the current limit. In addition, when the charge current drops below a certain level set by the EOC current pot, the output voltage of the Meanwell will drop to a 'float' voltage, set by the float pot. Since this stops the charge current, the circuit will latch off until the pack voltage gets drained below the float setting. In practice, you'd set the CV setting on the Meanwell to your maximum charging voltage (say 4.2v/cell for example). On end of charge, the float setting would be 4.0-4.1v/cell and the charge current would be zero (or slightly draining) until the cells reach the float setting. The pack then needs to be discharged somewhat below the float setting before the current can go back up.

I'm not sure how useful the HVC input is. It would save some stuff if it was left out.
I'm also not sure if it needs a manual restart. Normally it should not need it if the voltages are dialed in right, but could be used to 'top off' a nearly full pack.

3 Pot Meanwell Charge Controller 4.1.jpg (45.61 KiB) Viewed 543 times

[teklektik]

I'm not sure how useful the HVC input is. It would save some stuff if it was left out.
I'm also not sure if it needs a manual restart. Normally it should not need it if the voltages are dialed in right, but could be used to 'top off' a nearly full pack.

As with the mini-limiter, the HVC input does offer some options from external cell-level monitoring - so seems valuable and can be left unpopulated if not desired.

I think the only question is the desired behavior of that external input. Just playing Devil's Advocate here - if the external input did not latch, then when the external alarm was removed, charging would resume until the internal cutoffs bought charging to conclusion normally or until the alarm re-asserted. This might cycle in and out of alarm a while until all the cells became charged enough to trip the internal cutoff. If this doesn't seem like objectionable behavior then removing the HVC latch could be good. If so:

(I'm kind of rushing this out, but I think this is correct ):
It seems the latching effect as it relates to the HVC input is tied in part to the response time of the MW in comparison to the voltage drop at the collector of Q. If a bit of damping is introduced with a small RC hung from U2b to Gnd (across R11+R8), then when HVC goes high, the voltage at the Q collector will fall more slowly presenting a declining voltage to U2 indicating a declining 'apparent' current. The apparent current will seem very high initially in the limiter zone, drop into the middle no-limiter range, and finally fall to the float cutoff level. A true pack current higher than cutoff will prematurely terminate this decline with a voltage across R8, allowing normal charging to resume. The RC constant should only need to be large enough to safely allow the MW response to outrun the voltage decline across the no-limiter voltage zone and should not materially affect normal operation.

A side effect of this modification is that a manual reset could be accomplished with a pushbutton from HVC to Gnd. (Another way to look at this is that manual reset takes an additional RC, and by the way, the HVC works a little differently....)

[fechter]

Yes, I think you got it. Previous version was intended to work with a shunt based balancer and HVC was intended to throttle the charge current.

It could be designed to either latch or not latch. I'm just not sure what most people would want.

[teklektik]

It could be designed to either latch or not latch. I'm just not sure what most people would want.

I vote for 'no latch'. However, if the above modification actually works (?), then the latching can be enabled by omitting parts - so it's two mints in one...

[fechter]

The new mini limiter boards arrived.

First thing I noticed was where I messed up and put a part label across two traces creating a short. I personally fixed every board in the batch and tested with an ohmmeter to resolve the problem. I should know better by now. In the layout file it looks like there's a tiny bit of clearance but the actual board came out shorted.

Ver.3 oops.jpg (64.38 KiB) Viewed 473 times

Other than that, they look great. I have one built and bench testing was stellar. Construction is a breeze too with the improved layout. I'll post pics when I get a chance in a new sale thread.

[teklektik]

... bench testing was stellar.

!

[mwkeefer]

Fechter - +2 me for this one, I'll take bare PCB, Kit or Assembled units... many more if testing goes well, I have a stock pyle of meanwells I don't want to manually modify if unneeded... to tie into any of my HVC systems would also be quite nice as a safety feature since I do monitor cells at all times (when bike is in operation and while charging, not in resting state at this time).

-Mike

[fechter]

See the Ver. 3 sale thread:
http://endless-sphere.com/forums/viewtopic.php?f=31&t=36958