Need help with current limiting. / General Science and Electronics / Forums

Forums

4hv.org :: Forums :: General Science and Electronics

« Previous topic | Next topic »

Need help with current limiting.

previous 1 2 3 4 next last

Move Thread

LAN_403

Patrick

Wed Dec 09 2015, 06:32AM

Registered Member #2431 Joined: Tue Oct 13 2009, 09:47PM
Location: Chico, CA. USA
Posts: 5639

Ive been laying out the circuit board for etching. the IC is in the SOIC-16 package. So ill solder the IC on the trace side, and the other components on the top side. I'm going to use 2 of the TO-247 transistor I have and for now, the transformer will be external, just held in place with short but heavy wire by its primary. The transformer is experimental and may need to be re-wound until I'm happy with it.

Patrick

Sat Dec 12 2015, 06:29AM

Registered Member #2431 Joined: Tue Oct 13 2009, 09:47PM
Location: Chico, CA. USA
Posts: 5639

Progress.

You can see a large blank space at the bottom, that's for hot gluing the current transformer, which will no doubt be re-wound several times.

Some of the traces are pretty bad, so ill etch a few more boards, but I'm looking into ExpressPCB. In any case, ive got colored wire and all the components. so tomorrow ill be populating the board and getting my 13.8 battery replacer and ATX power supplies ready. The SOIC-16 etched section came out perfect.

Electra

Sat Dec 12 2015, 11:55PM

Registered Member #816 Joined: Sun Jun 03 2007, 07:29PM
Location:
Posts: 156

I don't think you'll need to rewind the current transformer that many times, the secondary voltage is developed across a load (burden) resister and its that resistance and turns ratio that determine the secondary voltage. Or even use a combination of fixed and variable resistors if you need adjustment.
I sec (A) = I pri x(Np/Ns)

You can buy ready made ones to save fiddly winding in standard ratios 1:200 etc. Which look epoxied more like a sensor, Note watch out some are only for line frequency 50/60hz.

Patrick

Sun Dec 13 2015, 07:21PM

Registered Member #2431 Joined: Tue Oct 13 2009, 09:47PM
Location: Chico, CA. USA
Posts: 5639

Electra wrote ...

You can buy ready made ones to save fiddly winding in standard ratios 1:200 etc. Which look epoxied more like a sensor, Note watch out some are only for line frequency 50/60hz.

Yes I looked at these too, but I need to learn the lesson of making them. I'm thinking of using 1 turn of each primary, through the ferrite core. then 10 or so secondary turns I guess? then use a voltage divider with minor pot for minor tuning.

Patrick

Wed Dec 16 2015, 12:26AM

Registered Member #2431 Joined: Tue Oct 13 2009, 09:47PM
Location: Chico, CA. USA
Posts: 5639

EDIT : its been a long night. Ill need others like Goodchild to chime in again. I'm unsure how to replace the sense resistor, with the current transformer. I don't know how to put the signal into the IC.

-I already have the signal diode bridge and burden resistor on the board, I just don't know where to send that signal.
-I need to set the oscillator frequency to 75 kHz, don't know how. (the Rt and Ct pins I think)
-I need the current transformer input resolved.
-I think I have the voltage regulation solved with a voltage divider.
- I'm not sure of the ramp pins use or purpose. there was a unitrode/TI pdf that explained this but I cant find it.

I'm trying to figure out my RC time constant for 13us (75kHz) using a 100pF film cap. [100pF and 133kohms I think.)

It does say so in this pin list.

3D printed mount for the core.

Ok here we go :

I'm hoping I counter-wound the CT right.

Ive outlined the mystery components in red.

MEhhehsfdg ee eeee . . . eyes are blurry, time for bed soon. .ew.d.dc. mehhh

Electra

Thu Dec 17 2015, 12:40AM

Registered Member #816 Joined: Sun Jun 03 2007, 07:29PM
Location:
Posts: 156

What's going on there is, their using the current feedback from the sensor resistor or transformer as the ramp input. A part of the ramp voltage from across timing capacitor is ac coupled through the 10nf cap ( probably for compensation ) and reduced through the potential divider formed by the 8.2k and the 1k resistors. The two ramps effectively add at the ramp input pin.

This mode of current control rely's on having an inductor on the output of the transformer for it to work.
You could replace the sense resistor with the burden resistor I think?

An app note that gives a little more insight into the theory, though mainly in respect to buck or boost converters. :-

Edit,
looks like the 120pf cap plays more of a part in the voltage ramp feed, its reactance is lower than the 8.2k resistor.
Anyway hope I've got the general idea right.

Patrick

Thu Dec 17 2015, 03:13AM

Registered Member #2431 Joined: Tue Oct 13 2009, 09:47PM
Location: Chico, CA. USA
Posts: 5639

it keeps sparking when I put power in. I may have killed the IC.

Electra . . . if its that red bit is so complicated I may never get it to work, but ill keep trying.

I don't know what to change from their 1.5 Mhz example, to my 75kHz design. I'm going to read your PDF and re-read the TI/Unitrode PDF's.

EDIT: no matter what I do the IC power pins say 0 ohms, so the traces or other components must have a short.
EDIT: Spark problem solved, a wee bit of wire insulation was missing, in a very small spot.
EDIT: the refercne voltage of 5.1 V is working.
EDIT: the ramp is working, but its at 12kHz. ill need to get it to 75kHz

EDIT: ive got the ramp working at 1.5V and 14uS. ( 70kHz, close enough ). I don't see anything on the gate drive pins though.

pdf. . . .

. . .

EDIT: ill need a snubber for the turn on spike of the CT's leading edge.

I'm operating under the assumption that this pic shows these two comparators working as such; below 1 volt the PWM is unmodified, between 1 and 1.4 volts the PWM is clipped back, as soon as 1.4 volts is hit, the transistors are off entirely.

(or that 0.4 V difference is hysteresis) Sigh this is hard. . .

Electra

Thu Dec 17 2015, 09:40PM

Registered Member #816 Joined: Sun Jun 03 2007, 07:29PM
Location:
Posts: 156

Ok maybe simpler to get it working in voltage mode control first, if you connect pins 6+7 together and to top end of cap CT, (hope that works with this ic). The current transformers burden resistor now just goes to pin 9 via a 1k, and if you include the cap. Like you have it in your app note fig 11.

The two comparators, I'm also a little puzzled by the second one. As the first one at 1v, is pretty straightforward it terminates that cycle regardless of where the PWM is if the output from the current sensor exceeds the 1v threshold. I would have thought that would have done the trick and prevented the current increasing further because now both mosfets ore off.

Now how does it manage to get to 1.4v? delay in the current sense path (your c to filter out the leading edge) or maybe something really bad has happened transformer saturated, and the current rises so fast in one half cycle, that the switching devices cannot turn off fast enough to prevent the increase.
When this does happen though, looks from the block diagram it restarts the error amp what's the reason for this, I'm not too sure possibly it restarts the PWM from zero again.

So under 1v unmodified as you say, Cut back at 1v, don't know if it means if anything happens in between but at 1.4v it shuts down the error amp.

Patrick

Fri Dec 18 2015, 02:23AM

Registered Member #2431 Joined: Tue Oct 13 2009, 09:47PM
Location: Chico, CA. USA
Posts: 5639

Id like to put in bogus signals, then measure the PWM response. I don't know how the RAMP pin figures in. I start with the input between ILM/SD and PGND to see what happens I guess.

Electra

Fri Dec 18 2015, 07:09PM

Registered Member #816 Joined: Sun Jun 03 2007, 07:29PM
Location:
Posts: 156

Don't know what you call bogus signals, If you mean want to run it open loop there's a test circuit for that in the data sheet, you have to set up the op-amp part, simple dc feedback just treat it as any ordinary op-amp ( pins 1,2,3). Turn a pot and see the pwm waveform change.

PWM works by comparing sawtooth wave (that's your ramp input pin) and a control voltage (the output from the op-anp part) using a comparator. At the start of each cycle a latch is set turning the output high, soon as the ramp exceeds the dc level of the control voltage the comparator changes state re-setting the latch, the outputs go low. So increasing the control voltage increases the duty cycle.

previous 1 2 3 4 next last

Moderator(s): Chris Russell, Noelle, Alex, Tesladownunder, Dave Marshall, Dave Billington, Bjørn, Steve Conner, Wolfram, Kizmo, Mads Barnkob