DIY analog constant-current discharger?

[meteor]

Assuming I've got current measurement and data capture devices like a PC-interface DMM and/or an eMeter/Whattmeter, is there a "simple" way to make a basic, variable, analog constant-current discharger?

I assume that the components would be just a stable resistance (one that doesn't change much with heat), a bunch of MOSFETs (running in a linear mode), and some sort of rudimentary feedback circuit?


I'm imagining an ultra-simple device with nothing more than a pot to adjust the current drawn from the cell/pack.

The critical issue is that it maintains a constant-current load.

I guess a pot to adjust a LVC would be useful as well.


FYI, let's assume that I have a heavy-duty DC power supply which I could use to calibrate the unit, without resorting to draining batteries to get the appropriate amperage and cut-off.



Ideas?

[LT]

A couple of links you muight find interesting

http://www.rcgroups.com/forums/showthread.php?t=379167
http://www.elecdesign.com/Articles/Index.cfm?AD=1&ArticleID=4056

And for a couple of articles in German look at
http://www.elektromodellflug.de/technikausw.htm look at "entlader" and "stromsenke" dowards end of page.

[meteor]

A couple of links you muight find interesting

http://www.elecdesign.com/Articles/Index.cfm?AD=1&ArticleID=4056




Thanks!


The one based on the Maxim components looks like the closest to my concept.


But, having the four different returns unnecessarily complicates the circuit.

Could you, or someone else, show how to eliminate the "low-power" returns without effecting the circuit operation?


Also, I assume that you can cascade more IRF540's (or similar) to increase the current dissipation capacity? (does each IRF540 need a distinct 1k resistor on the gate?)



Thanks!

[z-matrix]

Thanks!


The one based on the Maxim components looks like the closest to my concept.


But, having the four different returns unnecessarily complicates the circuit.

Could you, or someone else, show how to eliminate the "low-power" returns without effecting the circuit operation?


Also, I assume that you can cascade more IRF540's (or similar) to increase the current dissipation capacity? (does each IRF540 need a distinct 1k resistor on the gate?)



Thanks!

hi,

well,

unfortunatly, you can not parallel mosfets operating in the linear region this simple, each mosfet needs its own source resistance, gate resistance, and driver electronics,

use irfz44 if you want more power, and put it on 0.5kg aluminium heatsink, it will handle 44w os power dissipation in a to-220 package!
(this mosfet can be used safely in the operating voltage range of 45V, the IRF540 type is 100V one, with higher gate source resistance)

[meteor]

unfortunatly, you can not parallel mosfets operating in the linear region this simple, each mosfet needs its own source resistance, gate resistance, and driver electronics


Really?


I can't say about the gate driver side, but I know that most older dischargers used paralled MOSFETS, like the IRFZ40.


The source and drain of five discharge MOSFETS were simply a parallel circuit in a Victor Engineering cycler, with one load shunt.

They were all attached to a massive heatsink, but they were electrically isolated from each other on the heatsink.


I can supply photos of the implementation if you like...

[Mr.RC-CAM]

Could you, or someone else, show how to eliminate the "low-power" returns without effecting the circuit operation?
It appears to me that if you do not want the max current protection of the tapped load outputs, just use the 10A tap and replace RA/RB/RC with a single 100 ohm resistor.

[z-matrix]

Really?


I can't say about the gate driver side, but I know that most older dischargers used paralled MOSFETS, like the IRFZ40.


The source and drain of five discharge MOSFETS were simply a parallel circuit in a Victor Engineering cycler, with one load shunt.

They were all attached to a massive heatsink, but they were electrically isolated from each other on the heatsink.


I can supply photos of the implementation if you like...

you may parallel mosfets if you use them only for switching like you mentioned, so now, the load resistors will dissipate the poer, not the mosfets, and thus, mosfets will have there drain source resistances in parallel when turned on.
There are many problems you have to think about if you manage to use mosfets in parallel and operating them in the linear region, like:

if mosfet gets warmer, internal resistance gets larger, so=> more heat is generated
if mosfet gets warmer, the gate source threshold voltage decreases so=> current will increase at same gate source voltage

they do not like to work in parallel this easy, as can be seen.

[simingx]

I have a circuit based on a Maxim appnote here:
http://www.rcgroups.com/forums/showthread.php?t=200602

Does up to 15A per MOSFET (don't even TRY going above this, even though the MOSFET specs may state higher!!).

Note the MASSIVE heatsink in the pictures.