Need Help with a Circuit

[K1JJ]

Hola,

I wonder if someone with a software circuit simulation program can take a look at this four-component current limiter circuit.

This circuit is rated at 70ma, 15V and I need about ~3A at 24V rating.

I would like to use a IRFP260N N channel MOSFET (first choice) or a TIP-120 Darlington NPN (second choice) as  Q1, that I have available here.

Hopefully, the circuit will let the load conduct current until it hits ~3A and then fold back the voltage.

I need to know specific values and how to get a MOSFET working in place of the power transistor.  I will build it as soon as I can get a plan.

Here's the CURRENT LIMITER circuit:  (top schematic, simplest version)

http://www.zen22142.zen.co.uk/Circuits/Power/add-on.html


BTW, just to be sure, this circuit draws current thru Q1's E-C  ONLY when the load does and when R2 = D1/D2, Q1 will begin to turn off, correct?  (No wasteful shunting)


Thanks.

Tom, K1JJ

[WA1GFZ]

Tom,
You need to put this on a good sized heat sink because they generate a fair amount of heat in limit mode. Also they sometimes oscillate while limiting as they heat up. You can use a FET but you have to add more series diodes to generate 4 or more gate volts to turn the FET on. I would put a small cap gate source to keep it from taking off.
Say you want to limit at 3 amps then you need enough heat sink to dissipate drain voltage times limit current worse case.
OR just go to International Rectifier and buy an IPS driver chip with everything in one TO220 package. (IPS511 for example rated for 5 amps with Overcurrent and over temperature shut down) also has a pin that goes low if there is a fault)  I just turned my boss on to a IPS521 10A.that he used to control driving lights on his van. It takes a 5 volt digital level signal to turn it on and off. I've beat the crap out of them and never blew one up. Too bad they are too slow for Class E drivers. I think they also make a low side switch. They are cheap. Great for controlling relays off digital control.
Other guys make drivers like these but I have no history with them. gfz

[K1JJ]

Frank,

Those TO-220  IPS511 look interesting.

I looked up International Rec and Mouser - can't find anything for 3A readily available. Do you have a listing of all the IPS 500 series devices?  Also, they show various devices with inputs from 1, up to 5 leads. What would be the best config for a simple "always on state"  3A current limiter? (no logic needed)

Tnx.

T

[WA1GFZ]

The thing you have will work but you need a good heat sink. I prefer FETs.
I'm not sure if the IPS switches go at 3 amps but why not use the 511 with a series fuse. They are tuff parts and easy to use. I know there are a bunch of them on the IR site. I was just looking at the specs last week. I'm going to use one to control the TR relays in my HPSDR set up so I can get it on the air. Sorry don't have any spare parts. I have run the 521s in parallel to double the current. I did a lot of testing because we were going to use them in a design at one time.
Or just use a transformer / rectifier that won't put out more than 3 amps.
A bigger transformer with a series resistor on the primary works.

[K1JJ]

The current limiter is for a set of stepper motors controlled by the computer using PWM.

The start up current is tremendous, like 25A and then settles down to 300ma at full slewing speed.

The motors will start OK at a 3A limit. I don't want a fuse or anything that will interrupt it cuz it will lose pulse and positioning counts.

I'll wait to see if anyone posts an IRFP260 Mosfet version of the circuit here and go from there.  It's interesting that you thought it would oscillate when heating up. That would probably screw up the PWM wave train and create skipped pulses.  I'll drag the scope out there when it's working to see. 

BTW, this project is for the homebrew automated tracking and GoTo system for the big telescope.

T

[ka3zlr]

Ya know Tom when I built my CNC Milling Machine.. I went to these guys here...

Adaptation wub be FB in your case...Plug an Play man...Lotsa Links out there for Stepper Motor Drives...

http://www.hobbycnc.com/

[K1JJ]

Hi Jack -

OK on the link.  Nice hobbyist website.

The driver boards, software  and computer controls are already finished and working FB. I just need a simple current limiter for the  24V power supply output.

The start up and ending current surge is about 25A. That's how steppers are, I'm afraid. I'm using 24V on 4.8V steppers with a 5A regulated supply. The PWM limits the total power to a safe level. This produces MUCH more torque than running them at normal voltages. No problems doing it.

Frank: I'm going out for a while now, but when I get back I'm going try JSing together an IRFP260N Mosfet limiter.  In that circuit, you said to try 4 diodes in series. Is the 1K resistor OK? Also, where should that Mosfet idle... just enuff to be turned on when the load is drawing no current, like at 20ma or so?  Or does it idle at full current all the time and shunt some off to the load?

T

[KA1ZGC]

I just turned my boss on to a IPS521 10A.that he used to control driving lights on his van. It takes a 5 volt digital level signal to turn it on and off. I've beat the crap out of them and never blew one up. Too bad they are too slow for Class E drivers. I think they also make a low side switch. They are cheap. Great for controlling relays off digital control.

By "low side switch" do you mean opposite polarity, like -5 volt switching?

I've been meaning to wire some of the digital outputs on the NHRC-5 controlling my repeater to the relays on Tim's wire array so he could switch patterns from the trailer. Sounds like these might do the trick.

There is one caveat, however: the phase-reverse relay snaps, pops, and bangs like a mother when there's a thunderstorm within 20 miles, and I'm a bit leary of lashing something between that and my repeater. I'm guessing these devices probably wouldn't care much for it, either!

[WA1GFZ]

Tom,
You are going to need more than 4 diodes. The threshold voltage for a FET is up around 4 vots so you will need more like 6 of them maybe a few more to saturate it. The 1 K resistor is fine. The 10 ohm resistor is too high. At 3 amps you will have 30 volts across it. At 1 ohm you will need a 10 watt with a peak power at 10 watts at the resistor. Put a .001 or .01 across gate- source to keep it from oscillating. Don't use too large of a power supply. To set it up you need enough diodes to keep the fet on at rated current so if you need 4 volts to turn the FET on and 3 volts drop across sense resistor you will need 7 volts of diode. A zener will actually be more stable than diodes over temperature. 

[K1JJ]

OK, on all.

Yes, I meant to say 4 volts, not 4 diodes, but will play with that value. I have some zeners.

When you say "rated current"  do you mean the FET will have to IDLE at 3A all the time if we are doing a 3A current limiting scheme?

Take a look at the schematic and figure if it idles at 20ma like a series pass or 3A as a shunt.  The author does not say.

If it needs to idle at 3A, that's 60W of heat for nothing and I may reconsider the design.

How about the IPS511.... does that idle at full current?

T

[WA1GFZ]

When the output current is low the device is saturated. As you increase current past the regulation point the device becomes resistive meaning the device comes out of saturation and becomes more resistive. It is common to use low voltage steppers at higher voltage to motivate them. Just make sure you don't drive them hard enough to get them hot.
The 511 is not a current limiter it shuts down for a short time until the junction cools down. Another limiter you could do is a small 24 volt supply with a big cap after the bridge. This way you get high peak current from the cap but not for a long period of time because the transformer will not supply it. BTW add a cap Drain to ground and maybe one on the source if it still oscillates. fc

[K1JJ]

OK, that's good and makes sense.

So, the MosFet will dissipate little power until the load draws more than 3A.  That's what I was hoping for from the beginning but wanted to be sure.

Yes, the steppers stay cold now even testing with long 60 minute slews.

I'm using a regulated 24V 5A supply. No need for a big cap.

 I really don't want those big 25A spikes cuz it can cause havoc with the computer and associated driver circuitry.  I think just this limiter will be OK-fine.

Thanks for the help.  I'll build it up tonight and see.

T

[WA1GFZ]

One thing about diodes you can vary the trip point to suit your needs.
GL

[K1JJ]

Frank,

It's working great!

Have it JS wired on the breadboard with about 0.5 ohms of emitter and about 5 volts on the gate. At 2A it just starts to fold over. Some more diodes and I'll be all set.

I put the MosFet on a 4" aluminum slab and it doesn't show any signs of heat when full limit.

I'll take a look for oscillations once it is on the circuit board later.

T

[WA1GFZ]

Great, There will be about 75 watts of heat on a dead short.

[K1JJ]

Hola Frank,

Well, here it is.  Just a little Mr. Wizard type project, but it does the job.

I ran it at the required 3A and find the MosFet and 0.5 ohm resistor drop the output about 2 volts, as you would expect.  But when I put a 7A load on, it does the job and the output voltage drops in half, which is great.

Looking at the scope, there are no oscillations, even down to the 20mv level.

Also, the current limiting is instantaneous. I can key the big load on and off and see a squarewave.

I have enuff extra diodes in there to tap any limiting current I need.

It's a cool little current limiter - shuda made a few years ago for other projects.

Thanks for the help, OM!

T

[WA1GFZ]

Tom,
A square wave means the circuit is oscillating. This is OK if your stepper driver can handle it. The losses will be lower in this mode because you are in a PWM sort of mode. Try reducing the gate source cap value if it is a problem for the stepper. You are setting up a time constant between the 1 K pullup on the gate a with the gate source cap.  You might try moving the cap from the source to the other side of the .5 ohm sense resistor to give it some NFB.
This could also cause another problem though with the inductance of the resistors making a HF oscillator.
Here is a thought. Put the right value cap on the output so it will support a stepper pulse and let the limiter stay in pwm mode. The cap charges between pulses.

[Mike/W8BAC]

Tom,

Aren't you a graduate of Cloutier University of Science and Art? I don't see a single clip lead in your work. I'm afraid I'm going to have to contact the facility in Townsend regarding this. They won't be happy!

Mike

[K1JJ]

Mike,

heheheh  -  Guilty as charged. Though the test breadboard used about 12 clip leads, so I'm not a total loss... :-)


Frank:  No problemo. The squarewave I mentioned is good - there is no oscillation. I used it to describe the instant response of the circuit when I rapidly keyed the load on and off. The DC waveform on the scope would step up and down so quickly it looked like a squarewave. Instant response to overloads.

I'm gonna try the board with the steppers later today and see if there are any problems in the real whirl.   There always are. I'll keep your stability ideas in mind.

Thanks again for the advice.

T

[K1JJ]

Frank,

The current limiter worked out perfectly with the steppers.  Very smooth operation.

I was able to limit the current to only 1.5A at startup and shut-off. Thats about 20 times less current spikes than without the limiter.

The steppers start ramping up at about 12V (1.5A) and peak at about 28V (200ma) at full speed. This is good cuz a stepper's torque gets lower and lower as it speeds up, unlike a conventional motor.  A higher voltage adds torque. At the full 28V I was able to get it to spin faster than ever before without stalling.

I added some 2" steel flywheels and this made it smoother and even faster. 

During slow tracking, I set it to run at about 13V which is best for the least vibration and smooth steps.

The heatsink now gets slightly warm when tracking at 13 volts. Glad I used a big 1/4" slab.

Before, I had to use two separate supplies, one for tracking and one for slewing - and had relays, etc.  Now just the one fixed 28V supply does the whole trick.

 I should post the schematic now that it's finalized.   I've always hesitated to make a limiter cuz they usually required lots of transistors and parts. This one is just four major parts. (not counting the string of diodes for adjustment) So easy to build.

Happy camper.

Later -

T

[WBear2GCR]

beware of the slab getting warm... it has high thermal mass and poor radiation of heat, so if you run the thingie a lot and it is a warm day, it will get really hot much faster than you expect... you can find a nice finned heatsink that is probably suitable out of an old found on the street or in ur junk pile computer monitor... usually... the fins do a much better job of transfering the heat to the air than does a chunk... fyi...

              _-_-bear

[WA1GFZ]

Tom make sure you have catch diodes across each winding of the stepper to reduce turn off transients. When switching the low side the anode goes on the driver and cathode goes to positive common of the stepper.

[K1JJ]

Bear,

I plan to mount all the electronics on one plexiglass board and have a small fan blow air across.  The 1KW class E rig I built a few years back taught me the lessons of heat sinking... :-)


Frank:  Here's the actual scope stepper driver board schematic below. I built it homebrew about 12 years ago and still use it. Just went through a complete telescope rebuild and want to get a PCB for it instead of the p-p wiring I have now.

With the free software it uses, I can take a 200 step/revolution stepper and break each step down into 40 precision micro-steps. I also set each step to move identical to each other (via software) and also normalized the currents in each winding. This produces a smooth stepper action, just like a DC servo motor.  Each step equates to about 0.2 arc seconds of movement in the sky, once the gearing is figured in, which is too small to be seen by the eye, even thru the eyepiece.

The tracking requires very tiny steps, whereas the slewing to each object requires speed. Dynamic range was increased using the flywheels. The scope will slew about 4 degrees / second - not bad for steppers.

The sofware also has sophisticated scope MOUNT error corrections that will compensate for all the mechanical errors - resulting in accurate goto slews of under 5 arcminutes error across the sky.  I use it for the 22" mirror alt-az siderostat mount system here.

Here's the website with the full details on the system. It took me quite a long time to fully understand and put everything into operation... :-)

http://www.bbastrodesigns.com/StepperDoc.html

Later -

T

[WA1GFZ]

It has snubbers so you are all set. They take the place of catch diodes and might be quieter and more efficient. I just threw away a couple big steppers I should have dropped them off. About the size of a hardball.