Re: Voltage controlled rheostat?

On Jan 2, 11:44 pm, "Bob Monsen" <rcmon...@xxxxxxxxx> wrote:
"JonathanBusby" <jonathanbu...@xxxxxxxxx> wrote in message

news:fe5bb7de-ffff-4dc0-a06e-d40813454e23@xxxxxxxxxxxxxxxxxxxxxxxxxxxxxx



On Jan 2, 1:38 pm, John Fields <jfie...@xxxxxxxxxxxxxxxxxxxxx> wrote:
On Tue, 1 Jan 2008 11:08:36 -0800 (PST),JonathanBusby

<jonathanbu...@xxxxxxxxx> wrote:
On Jan 1, 1:02 pm,JonathanBusby<jonathanbu...@xxxxxxxxx> wrote:
On Dec 31 2007, 3:01 pm, John Fields <jfie...@xxxxxxxxxxxxxxxxxxxxx>
wrote:

On Sat, 29 Dec 2007 20:18:38 -0800 (PST),JonathanBusby

<jonathanbu...@xxxxxxxxx> wrote:
On Dec 29, 5:53 pm, John Fields <jfie...@xxxxxxxxxxxxxxxxxxxxx>
wrote:
Something Like this: (View in Courier)

MAINS>----------------------+
|
+----------+ |
MAINS?>--| | |
| | |
+----| | +---+---+ +---+
| | |---| +|---|+ |
[SENSOR] |CONTROLLER| |SUPPLY | |FAN|
| | |---| -|---|- |
+----| | +---+---+ +---+
| | |
MAINS?>--| | |
+----------+ |
|
MAINS>----------------------+

where the connections to the supply voltage control terminals are
the outputs from the controller?

--
JF

Yes that's basically it. :)

---
OK, what do you want to do now, and how much money do you have
available to throw at getting the problem solved? Material, that
is.

I think my price cap is around a hundred bucks.

As long as I don't have to start wielding a soldering iron
I'll be happy to help you for free, :-)

Nah. But I want you to come over to my house and complete the
construction
of my 20 ft tall Tesla coil -- I have completed the first steps :
ordering the plans. ;) :P

BTW, do you have a schematic or a link to the controller's data
***?

http://www.t-balancer.com/english/produkt_tban_bigng.htm

http://www.t-balancer.com/download/bng.pdf

--
JF

Thanks a lot for all your help!

Jonathan

Oops. Forgot the fan :

http://www.delta.com.tw/product/cp/dcfans/download/pdf/FFB/FFB120x120....

Part # : FFB1224XHE :
http://www.delta.com.tw/product/cp/dcfans/dcfans_prod_sch.asp?act=sch...
( The one at the top )

---
OK, since it's a brushless fan it's not really designed to be
pulse-width modulated and since you've got a $100 cap on the
project, here's what I'd do: (View in Courier)

A1
+-----+
120AC>-+------|~ +24|-----+
| | | |
120AC>-|-+----|~ -24|--+ |
| | +-----+ | |
| | | |
| | +-----------+ |
| | | A2 |
| | | +-----+ |
+-|-|--|~ -12|-----+-----------------+
| | | | |
+-|--|~ +12|--+ |
| +-----+ | |
| | LT |
0-12VDC>---|--[4K7]---|+ \1636 D
| +---|S >--+-[1000]-+---G IRFZ34E
| | +-|- / | |K S
| | | | [10K] [1N4744A] |
+------+-|--+ | | |
| +--------+ +-----+
| | |+
| [10K] [FAN]
| | |
GND>-------+-----------------+--------------+

A1 is a Kepco RKW 24-4.5K, $79, (or an RKW 24-6.5K for $99) from
Kepco at:

http://www.kepcopower.com/specs/rkw-openspecs2.pdf

http://www.kepcopower.com/rkwo-buy.htm

A2 can be pretty much any 12V wall-wart you might have laying around
as long as its output stays under 15V under high line and no load
conditions.

You'll probably need a heat sink for the MOSFET, but I haven't got a
good handle on what current VS voltage looks like for a brushless
motor. However, assuming that the relationship is roughly linear
means that with 24V across the fan pulling 3.25A through it, 14V (at
the low end of its range) will pull:

14V
----- = 0.583 of 3.25A,
24V

So:

0.583 * 3.25A = 1.89A.

With the supply at 24V and the fan at 14V means the MOSFET is
dropping 10 volts with 1.89 amps through it, so it's dissipating:

P = IE = 1.89A * 10V = 18.9W

Quite a bit of power.

From IR's thermal data for the MOSFET we have:

Tj max = 150C

Rtjc = 2.2C/W

From:

Tj - Ta
Pd = --------------------
Rtjc + Rtcs + Rtsa

we rearrange to solve for the thermal resistance of the heatsink
surface to air:

Tj - Ta
Rtsa = --------- - (Rtjc + Rtcs)
Pd

150C - 50C 2.2C 0.5C
= ------------ - ( ------ + ------ )
18.9W W W

= 2.59C/W

Which is a pretty substantial convection-only heatsink.

Here's an example:

http://www.aavidthermalloy.com/cgi-bin/exdisp.pl?Pnum=60050&LengthUni....

Finally, here's a little simulation so you can see how the thing
works:

Version 4
*** 1 880 832
WIRE -224 112 -944 112
WIRE -512 192 -832 192
WIRE -464 240 -608 240
WIRE -336 240 -384 240
WIRE -512 272 -512 192
WIRE -224 272 -224 112
WIRE -608 320 -608 240
WIRE -544 320 -608 320
WIRE -336 352 -336 240
WIRE -336 352 -400 352
WIRE -272 352 -336 352
WIRE -544 384 -704 384
WIRE -512 464 -512 432
WIRE -464 464 -464 432
WIRE -464 464 -512 464
WIRE -224 464 -224 368
WIRE -944 512 -944 112
WIRE -832 512 -832 192
WIRE -704 512 -704 384
WIRE -608 512 -608 320
WIRE -944 640 -944 592
WIRE -832 640 -832 592
WIRE -832 640 -944 640
WIRE -704 640 -704 592
WIRE -704 640 -832 640
WIRE -608 640 -608 592
WIRE -608 640 -704 640
WIRE -512 640 -512 464
WIRE -512 640 -608 640
WIRE -224 640 -224 544
WIRE -224 640 -512 640
WIRE -944 704 -944 640
FLAG -944 704 0
SYMBOL voltage -944 496 R0
WINDOW 123 0 0 Left 0
WINDOW 39 0 0 Left 0
SYMATTR InstName V1
SYMATTR Value 24
SYMBOL voltage -704 496 R0
WINDOW 123 0 0 Left 0
WINDOW 39 0 0 Left 0
WINDOW 3 24 104 Invisible 0
SYMATTR Value PULSE(0 12 0 .1)
SYMATTR InstName V2
SYMBOL res -240 448 R0
SYMATTR InstName R5
SYMATTR Value 8
SYMBOL nmos -272 272 R0
SYMATTR InstName M1
SYMATTR Value IRF1312S
SYMBOL Opamps\\LT1636 -512 352 R0
SYMATTR InstName U1
SYMBOL res -624 496 R0
SYMATTR InstName R1
SYMATTR Value 10k
SYMBOL res -480 256 R270
WINDOW 0 69 58 VTop 0
WINDOW 3 69 56 VBottom 0
SYMATTR InstName R2
SYMATTR Value 15k
SYMBOL voltage -832 496 R0
WINDOW 123 0 0 Left 0
WINDOW 39 0 0 Left 0
SYMATTR InstName V3
SYMATTR Value 36
TEXT -856 672 Left 0 !.tran .2

--
JF

Man! I don't know what to say or how to thank you! You went through
all this effort ( actually it probably took you 5 minutes ) and what
did I do? :) I thank you very much. You're the type of person that
still keeps me on Internet forums -- most have degenerated into
cesspits of off-topic messages, spam, trolls, idiots, flamewares, and
of course assholes. :) Speaking of such things, I'm currently the
target of a flamewar in another newsgroup. Anyway. I tested the
circuit ( not that I thought it wouldn't work ;) and it works
perfectly. There's only one minor and really insignificant detail :
when the control voltage is varied there seems to be around a 30ms lag
before the output responds. I don't know if this is just a property of
the circuit or the simulator. It doesn't matter anyway. I had to
replace the V2 voltage source with a an arbitrary behavioral voltage
source for testing. I'm running ltspice under Wine on top of Kanotix
Linux in LiveCD mode out of a RAM disk. heh. Anyway, here's a link to
a screen capture of the simulator with your circuit running :
http://i209.photobucket.com/albums/bb286/wqo2/snapshot3.png.

Thanks for all your help.

Regards,

Jonathan

Here is a variation that does not require the additional 12V wart. It is
somewhat more complicated, needing 4 transistors and a few more resistors
instead of the opamp. It uses a p-channel mosfet instead of the n-channel.
FWIW. John's heat sink calculations are all the same, since the dissipation
is identical for the pass element. I also think that a diode parallel to the
fan would be a good idea, to prevent a spike from taking out the mosfet.

Thanks to Genome for original circuit.

Version 4
*** 1 880 832
WIRE -688 32 -944 32
WIRE -480 32 -688 32
WIRE -224 32 -480 32
WIRE -688 48 -688 32
WIRE -480 48 -480 32
WIRE -688 160 -688 128
WIRE -480 160 -480 128
WIRE -576 208 -624 208
WIRE -544 208 -576 208
WIRE -576 272 -576 208
WIRE -480 272 -480 256
WIRE -480 272 -576 272
WIRE -944 288 -944 32
WIRE -224 288 -224 32
WIRE -688 304 -688 256
WIRE -272 304 -688 304
WIRE -688 416 -688 304
WIRE -480 416 -480 272
WIRE -832 464 -864 464
WIRE -368 464 -416 464
WIRE -320 464 -368 464
WIRE -224 464 -224 384
WIRE -224 464 -240 464
WIRE -864 512 -864 464
WIRE -688 528 -688 512
WIRE -592 528 -688 528
WIRE -480 528 -480 512
WIRE -480 528 -592 528
WIRE -224 528 -224 464
WIRE -592 544 -592 528
WIRE -368 544 -368 464
WIRE -944 640 -944 368
WIRE -864 640 -864 592
WIRE -864 640 -944 640
WIRE -592 640 -592 624
WIRE -592 640 -864 640
WIRE -368 ...

read more »

You guys are great! Thanks a lot! -- works perfectly in the simulator,
as usual. :) Again I would also like to thank John Fields, Bob
Monsen, G. Duchene, and everyone else who contributed in this thread.

Thanks again,

Jonathan Busby
.