how to turn a voltage regulator into a current regulator

i'm curious if there is a simple way to turn a voltage regulator into a current regulator.

take an IC that has a voltage feedback sensitivity that uses a resistor network to jump the output to ground through the feedback pin. let's take the MAX1674 with its 1.3V .03nA reference as an example.

MrAl has done it with the ZLT+ circuit, but i haven't really looked at it well to figure out how it ticks (assuming i can) and it's a slightly different set-up. i seem to much prefer the higher, and much more accessible, resistance feedbacks of the maxim chips.

btw, i found this on the net and thought it was fairly good but could be made simpler with an IC like the MAX603.

Hi papasan,

Zetex makes a current sensor ZXCT1010
(and others) that measures current
and turns it into a voltage. By
selecting the right resistor value,
you can get any output voltage you want.
This means, at least in theory, you can
use it with any voltage regulator that
has a feedback voltage sense input.

There are a number of people playing around
with this idea, but i mayself havent actually
had a chance to test this ic out yet.
There may be some stability issues if you try
to use this chip with a switcher that
operates at a frequency over 200kHz.
It's worth a try i guess, and it's not
an expensive ic.

Good luck with your LED circuits,
Al

i take it, from your message, that their is no simple work-around, like a cap/res network or something. alright, i'll look into the zetex chip. thanks.

Hi there again papasan,

Well, there is a 'current mirror' circuit,
which is made from two transistors and
maybe 3 or 4 small resistors. You
might find one in an ic package too
already made, but the Zetex chip looks
easier to deal with in general and it's
also very small in size.

Good luck with your LED circuits,
Al

if i got this right and i wanted to get a 1.3V output at .3A then Rsense=4.33 and Rout=333.33. and if i wanted to adjust the output to .21A out then Rsense=6.19 and Rout=476.2. makes it harder to change the output on the fly (via pot or even a switch) which is my main reason for going with the maxim chips. and i get into very odd resistance values that may be hard to find.

perhaps someone with real experience on this chip would care to comment?

Hello again papasan,

Actually, you use a set value for the
sense resistor (0.1 ohms) and you adjust
the output resistor to any value you need.
It should be very simple to use a pot
for the output resistor as long as you
limit it's low end value. This way
you can adjust the output with the pot.

The simple rule is the output current
is 1/1000 of the load current when using
a 0.1 ohm resistor. This makes it easier.

Say you need 0.3 amp output and you need
a sense voltage of 1.25 volts.
0.3 divided by 1000 equals 0.0003.
1.25 divided by 0.0003 equals 4167 ohms.
So you use a value close to 4167 for the
output resistor.

Slightly simpler is to divide 1.25 by
the output current:
1.25/0.3=4.167
and then take this value in k ohms,
so this would be 4.167k for the resistor.

If you use a 10k pot then connect it in
series with a 4k resistor and you should
get a range of adjustment. The 4k fixed
resistor is needed so you dont accidentally
adjust it to a much higher current then
you want, so keep in mind that if the
resistance went down to only 2k you would
get twice the output current, or 0.6 amps.

The resistor values arent really that
critical, and you can usually assume if you
replace a certain value with a higher or lower
value that the output current change will
be according to about the same ratio.
For example,
if you use a resistor 10% higher in value,
you will get about a 10% decrease in output
current.

Good luck with your LED circuits,
Al

Good luck with it,
Al

I've got two prototypes using the Zetex IC in the feedback loop to try to convert a voltage mode regulator to a current mode regulator.

Neither circuit behaves well. The first prototype is the LM2621 and zetex and I started a separate topic just for it. I abandoned it since I could not control the stability and the inductor current ramped excessively to high values under certain condtions (>4A). Efficiency when I did get it running was just so-so.

I just completed a MAXIM1524 with external FET prototype that also uses the Zetex ZHCT1009 to make this a current mode regulator. It too has similar loop stability and high peak inductor current with >6A of peak inductor current (yikes). When I got this to regulate I measured about 71% efficiency. The loop stability has to be solved for both cases and I think once that is accomplished the basic conversion could be applied to most voltage -> current mode converter conversions.

The ZXCT1009 GBW is affected by how large Vsense is. The larger Vsense the higher the BW is on the IC. This should be taken into consideration when looking at cycle by cycle loop control and ensure the ZXCT1009 has sufficient BW to not incur too excessive phase delay.

I'm using (5W design) a 0.15 ohm sense resistor (0.1V Vsense) and a 615 ohm resistor biased with one diode drop on the output.

My equations for the IC per the data sheet is:

Vout = 0.01 * Vsense * Rout.

.615 = 0.01 * .1 * 615 ohms.

With 0.1V Vsense and a 0.15 ohm sense resistor it should regulate around 0.1V / 0.15 ohm = 666mA.

Vfb = Vdiode + Vout = .6 + .615 ~= 1.23V

-WayneY

Hello again,

I looked up the current mirror chip on
the Zetex site. Over $4.00 USD; another
good reason to use the Zetex high side
current monitor chips instead

--Al