high current 0 to 40 volt pwm and v-reg power supply / General Science and Electronics / Forums

Forums

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

« Previous topic | Next topic »

high current 0 to 40 volt pwm and v-reg power supply

previous 1 2 3 4 next

Move Thread

LAN_403

Sulaiman

Sat May 19 2012, 04:55AM

Registered Member #162 Joined: Mon Feb 13 2006, 10:25AM
Location: United Kingdom
Posts: 3140

At the power levels that you are considering neither the pfc
(power factor correction - converts the ac input to dc at a higher voltage than the ac peak with a good power factor)
nor the pwm
(pulse width modulation - converts the dc to a (variable) regulated low voltage dc)
sections are 'basic'.
The design/test effort is such that I can see often the two are designed by different people/teams.
There are ics that control both the pfc and pwm function
For a good start go here

brandon3055

Sat May 19 2012, 08:43AM

Registered Member #4548 Joined: Mon Apr 23 2012, 03:52AM
Location: tasmania
Posts: 271

thanks Sulaiman i think i might try using one of my many tl494's becaus i don't want to go as far as bying specilized parts for this project but i'm not 100% shore how to set it up i'm thiking of using one errorr amp to control the voltage and the other to control the current aswell as the dead time control (duty cycle) to control the current and please correct me if i am wrogn the only thing that will limit my maximum current will be the mosfets (i have a fuw irfp250's aswell as a lot of irfp460's) and of cors my transformer

B.T.W any sugestions on what the circuit should look like would be very helpfull becaus i dont fully understand how to set up the tl494

Sulaiman

Sat May 19 2012, 10:12AM

Registered Member #162 Joined: Mon Feb 13 2006, 10:25AM
Location: United Kingdom
Posts: 3140

If you go to the URL pointed to by me and type tl494 into the search box you will find all kinds of
Technical Documents, example circuits, application notes, a forum ...........

P.S. for all the 'ZVS' addicts here, the UC2871 etc. ICs and notes are interesting
(for ZVS up to 40 V dc and 200 kHz, ideal for audio-modulated zvs)

I am currently 'playing' with the UC2825, a typical pwm ic that has
high speed,high current mosfet drive capability and is quite cheap/easily available.
(switches at small sstc high frequencies)

P.P.S. when looking to buy or for information, the UC-series of IC numbering means that
UC1xxx is -55C to +125C, the same as
UC2xxx is -40C to +85C, the same as
UC3xxx is 0C to +70C with slightly 'looser' specifications.

brandon3055

Sat May 19 2012, 12:58PM

Registered Member #4548 Joined: Mon Apr 23 2012, 03:52AM
Location: tasmania
Posts: 271

I did go there and serch for tl494 but i didn't find any useful circuit exemples I will go and have another look

edit: i have just done more reading then i ever wanted to do in my lifetime (about two pages) and this is the circuit i have cume up with
]expresspcb_1.zip[/file]

please note this is just a very basic idea of what the finished circuit will look like it is missing things like voltage and current displays the entire power input section and the OCD wich i will work on if and when i get it up and running. it is also probobly gowing to take a lot of trial and error toget the voltage controle working properly.
(any help with that or if there are any other problems with the circuit would be most aprciated)

Sulaiman

Sun May 20 2012, 12:24PM

Registered Member #162 Joined: Mon Feb 13 2006, 10:25AM
Location: United Kingdom
Posts: 3140

Had a quick look at your circuit diagram;
1) You seem to be feeding back the Input/+40V supply for the error amplifier whereas it's normal to sample the output/load voltage.
2) The output configuration looks wrong to me.

Since your supply voltage will probably be too high to directly connect across the TL494, you will not be able to copy most of the simpler circuits on the 'net.
Using 12V for the TL494 as in your circuit diagram is good.

To enable simple voltage feedback the -ve of the output needs to be connected to the -ve of your supplies, this means a switch (transistor) between the +40V supply and (diode to 0v and output inductor) load,
the easiest is pnp bipolar or p-channel mosfet but they are more expensive than the npn/n-channel versions so a little extra work is required.
OR
you can drive n-channel mosfets as in your diagram with the load +ve connected to +40 +ve BUT you will now need opto-isolated voltage feedback or equivalent.

brandon3055

Sun May 20 2012, 01:47PM

Registered Member #4548 Joined: Mon Apr 23 2012, 03:52AM
Location: tasmania
Posts: 271

Thanks i didn't notice that problem i guess I have a bit of redesigning to do

Edit: would it be possible to use an opamp or comparitor atached to the 40v rail with a floating negative (it think that's what its called) to feed the error amplifier through an optoisolator I will update my schematic tomorrow afternoon to give you a better idea of what i mean but ignoring all that do you think the this setup can actually work?

edit: i think i have fixed the problem please tell me what you think (some of the resister valuse may need tweeking)
]tl494_revision1.zip[/file]

Neet Studio

Mon May 21 2012, 05:38PM

Registered Member #4037 Joined: Fri Jul 29 2011, 03:13PM
Location:
Posts: 86

LM317 circuit:
The feedback path of the LM317 is not closed at the output, so voltage regulation under variable load would be poor as it is relying on the transistors behaving as emitter followers. Expect temperature drift to your output voltage as the transistors heats up.

Comment to your MOT/relay design:
It is okay if you are using fixed voltage output and manually select the relay for your output.

If the relays are switched on/off on the fly, then the output might go out of regulation on break and might goes overshoot when the next relay makes contact as the feedback circuit has limited bandwidth. Also you'll need a bit of hysteresis if the relay might chatter is your output sits at the threshold of the next bank. The first silly idea pops into my head of using a LM3914 bar graph IC in dot mode as a range select.

Can't comment on newer designs as they are in ExpressPCB formats. I am not going install yet another package just for that. A pdf or png (or even a link to your site) would be easier.

I am not a big fan for opto-isolator especially when it is not crossing isolation boundaries. Doing analog feedback with an optoisolator might look deceptively simple (but isn't) especially for newbies. Optoisolator have low bandwidth and it will add propagation delay in your feedback path. That means you'll need to do some work in feedback analysis/design to avoid stability issues.

brandon3055

Mon May 21 2012, 09:51PM

Registered Member #4548 Joined: Mon Apr 23 2012, 03:52AM
Location: tasmania
Posts: 271

Sorry I should of mentioned those zip files do contain the express PCB design but they also contain an image of the circuit (it don't have a web page)

Neet Studio

Mon May 21 2012, 11:36PM

Registered Member #4037 Joined: Fri Jul 29 2011, 03:13PM
Location:
Posts: 86

Ah... Didn't know about the .EMF file as it is not a common one. I actually worked with the format.

I am assuming that you are trying to build a buck (step down) regulator.

Starting point:
TI's webpage for the part would be a good one:

There are 5 application notes right there, starting off with the first one:
Designing Switching Voltage Regulators With the TL494 (Rev. E)

If you are bored with reading theory etc, skip to Design Example. That should at least give you some idea what your circuit should look like.
You'll need to scale that up a bit with proper gate driver chips. Transistor voltage follower drivers with NPN/PNP can probably drive 200-300mA and drive a couple of nF gate capacitance. You'll need more than that for a switcher the size you are building.

--------------------------------------- ---------------------------
Feedback path:
Your LM339 is comparing input & output voltage. Do you intend to use the raw input as a reference source at all?
Depending on the load and AC line voltage, an unregulated 40V source can change a lot! You are also going to have AC ripple at your output!

You already have a reference source at pin 14 inside the TL494 which is spec'ed for 2mV (typ) over the entire input range. With a 12 regulated power, that reference source is only going to be more stable.

You have the Error Amplifier at pin 1, 2 and the output at pin 3 specifically designed for voltage feedback! No need for the LM339 etc at all. Also LM339 gives a logic level not an analog signal and the optoisolator. :P

TL494 MOSFET driver: OMG!
- You seems to have the 2 opposite phase output connected as "wired or"! (Might work??)
- you are going to have a *lot* of hard time driving with 5X N-MOSFET without a (or even a few) proper MOSFET driver chip.
- why need 5 MOSFET in parallel? I made a 30A synchronous buck converter with a H bridge with only 2 TO-262 (SMT version for TO220) in parallel for the prototype. I ran 50-60A (constant power ftw!) charging a supercap from 0V with nothing other than the PCB for cooling and it wasn't even warm.
- gate series resistance is too large (see below)
- First off, you are driving that with a 15K passive pull down. There is no way you could discharge all that gate charge fast enough for anything.
Home work: Calculate the RC time constant as a sanity check (R=15K and C = 5x gate capacitance of your MOSFET) and that should give you an idea how bad it is!

- Your load with see close to 40V when MOSFET is on and whatever when it is off. Not sure if this is what you want. If it is like a heater or light bulb that you actually wants a PWM waveform, ditch the output caps. As is, your capacitor will blow up as it won't be able to handle the ripple current (charging to 40V and discharging to load) all by itself. If not, as a buck converter as you are missing a L and at least a diode. See

What you have right now would be chopping the DC voltage down. The only energy when the MOSFET are off are stored in the 2000uF cap which isn't very good for high current. What you want is to store the energy in an inductor - the higher current you have, the more it would store as E = 1/2 I^2*L (until it saturates)

At above 10A, you might want to look into a synchronous rectifier for the above topology for higher efficiency if you use low voltage at high current a lot. But hey that's for the next rev.

Funny thing about switch mode power supply. They are not too forgiving for tinkering. You'll need to have something more or less right to not blow stuff up. But you already know that...

brandon3055

Tue May 22 2012, 12:24AM

Registered Member #4548 Joined: Mon Apr 23 2012, 03:52AM
Location: tasmania
Posts: 271

Ok so just to make thighs clear its actually meant to be a more amadvanced version of the 555 PWM below the only problem with a PWM is it more of a current controller rather then a voltage controller because even on low duty cicle the cap would just charge to full voltage and then drop drastically when a load is aplied so the idea of this circuit was just to let the capacitor charge until it reaches a preset voltage (set by the competitor) and then switch it off until the voltage drops slightly at which point it would would switch back on and this would happen at a faily high speed depending on how much current was being drawn (the PWM would be more of a current limiter then anything else) and the inductor is gust to obsorb some of the current spike
Also the reason I don't just conect the error amp straight to the output is because as sulaiman pointed out (unless I use p-chanel mosfets) the positive output will be fixed and the negative will change

I will also point out that this is just my latest idea and I will still consider new ideas if anyone has any as long as they are not to complex

edit:here is that simple pwm (i have used this circuit to controle loades uo to 50A at 40v with 4 irfp250's but as i said abuve it is no good for a voltage regulator)

edit:if o get time i will breadboard the circuit tomorow just to see what hapens and inthe proscess maby i will fined a way to make it work

previous 1 2 3 4 next

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