Should I worry about shoothrough? / High Voltage / Forums

Forums

4hv.org :: Forums :: High Voltage

« Previous topic | Next topic »

Should I worry about shoothrough?

previous 1 2 3 4 next

Move Thread

LAN_403

Ash Small

Mon Apr 25 2011, 08:04PM

Registered Member #3414 Joined: Sun Nov 14 2010, 05:05PM
Location: UK
Posts: 4245

Steve McConner wrote ...

Ah, no, I meant that the tunnel diode was used as part of the circuit that sensed motor current and limited it. I just pointed it out to give a feel for the general archaic weirdness of the circuit.

I wasn't recommending it as some sort of band-aid for your circuit, if you can even find a tunnel diode nowadays.

Yep, the highest value current limiting diode I found was around 6 mA.

Either I'll use a current limiting resistor, or I might 'improvise' a current limiting diode using a partially switched on power BJT (resistor on the base).

(probably makes more sense than using an IGBT with a zener on the gate)

ScotchTapeLord

Mon Apr 25 2011, 08:05PM

Registered Member #1875 Joined: Sun Dec 21 2008, 06:36PM
Location:
Posts: 635

It seems the subject has already been cleared, but yes, the current through the primary will cancel out and will effectively not be loaded down by the core and secondary, leaving only, as Steve said, the DC resistance, which will mean that while it is a short amount of time, it will be a significant amount of energy, just like shoot-through, as Steve also pointed out regarding an earlier post.

If you are going to replace R1 and R4 with the primary then that is correct and I misunderstood what you wanted to do with the circuit. I thought you were going to build it as it was and use the outputs going to the GDT primary, in which case it would still work but the resistors would not be serving any purpose but to cause additional heating.

I still stand behind my argument that your circuit may not be as close to 50% as you want. This is important in a center tapped design since you cannot add an AC coupling/DC block capacitor to the primary. There are consequences if it is too drastic.

Hate to say it, but if ICs added complexity, they wouldn't be in use. It'll save you a lot of trouble if you just learn how to use the tl494. You can get away with very few external components after you understand the chip. If you want higher power handling and want to keep it simple, I think you can use a PNP transistor as a buffer for each output.

Ash Small

Mon Apr 25 2011, 08:26PM

Registered Member #3414 Joined: Sun Nov 14 2010, 05:05PM
Location: UK
Posts: 4245

I'm taking all this on board Scotch Tape, as I will probably move on to IC's later.

I want to start with 'first principles' first though,

I would also advise other 'newbies' to take your advice, and use IC's. I'm not saying what I want to do is 'better'.

I just want to use individual components rather than IC's, to start with, anyway.

Steve Conner

Mon Apr 25 2011, 08:32PM

Registered Member #30 Joined: Fri Feb 03 2006, 10:52AM
Location: Glasgow, Scotland
Posts: 6706

Well, nowadays I think discretes are harder to understand than ICs, as well as less commonly used. Maybe applying IC building blocks is the new "first principles".

But if it helps, I remember building a very similar power flip-flop circuit as a teenager. I used the astable's collector resistors to drive the bases of two PNP power transistors (2N4403s?) which I then used to drive a centre tapped power transformer, whose CT was grounded. It made a perfectly good power oscillator/SMPS type thing, though rather inefficient. So I'm sure you could get it to work too, if you wanted.

The Mazzilli/ZVS circuit is just the two-transistor astable, but using diodes for feedback instead of capacitors. So in the absence of any RC time constants, the only frequency determining element is the tank circuit, or the magnetic core in the ferroresonant square-wave version. (Which I have seen used in one hobbyist project, as a "master oscillator" for gate drive of a bigger inverter.)

Ash Small

Tue Apr 26 2011, 05:21PM

Registered Member #3414 Joined: Sun Nov 14 2010, 05:05PM
Location: UK
Posts: 4245

ScotchTapeLord wrote ...

I still stand behind my argument that your circuit may not be as close to 50% as you want. This is important in a center tapped design since you cannot add an AC coupling/DC block capacitor to the primary. There are consequences if it is too drastic.

I didn't directly address this point yesterday, but as I see it, if the balance is a problemI should only need EITHER a trim pot in parallel with R2 OR R3, or a trimming capacitor in parallel with C1 OR C2. either way, it is only one extra component and I have dozens lying around. I see this as part of the FUN of building this circuit.

ScotchTapeLord wrote ...

Hate to say it, but if ICs added complexity, they wouldn't be in use. It'll save you a lot of trouble if you just learn how to use the tl494. You can get away with very few external components after you understand the chip. If you want higher power handling and want to keep it simple, I think you can use a PNP transistor as a buffer for each output.

If I use a TL494, I'll still need a couple of PNP power transistors (which I don't currently have), as it is I'm planning to use a couple of Motorola 2N3055H's (a matched pair) that I have lying around (I have several of these, all from two batches, so matching them shouldn't be a problem).

The Td(off) of these should be long enough to give me a reasonable dead time. (although they don't give a figure in the datasheet, I'm hoping I can measure it using my 'scope)

The only other components I'll need are, as I said before, two resistors and two capacitors (maybe a trimmer or trim pot as well), I'm sure I'll need at least as many extra components if I used the TL494.

I will need some form of current limiting, but a light bulb would do the job and won't short out if it blows. This seems to be the main advantage over using a power BJT in the linear region (if that is the correct terminology), but even so, I could use a 'few' of the 2N3055H's I have lying around in parallel, to keep the current through each one low enough that the risk of blowing one is negligible, just put a trim pot on each base (in parallel with another resistor if needed, for finer control, etc) and adjust until I get sufficient current. This should be more efficient than using a light bulb or current limiting resistor.

I have all these lying around in abundance. All I need to buy are the four IGBT's (Â£6 each) for the bridge, and the zeners to protect the gates.

(I think I also have sufficient big stud diodes and electrolytics for the mains rectifier/smoothing circuit, but I'll have check)

As always with these things, it's the construction of heatsinks and enclusures that is the most time consuming part.

I still think that two resistors, two capacitors and a light bulb (and maybe a trimmer) is simpler than a TL494 and associated essential components though. although obviously whatever I use to regulate the current will waste power, (unless I only run it in the dark and use the current limiting bulb for illumination

)

James

Tue Apr 26 2011, 05:30PM

Registered Member #3610 Joined: Thu Jan 13 2011, 03:29AM
Location: Seattle, WA
Posts: 506

All you need to make a TL494 run open loop is a capacitor and a couple of resistors. Another option is a gated oscillator, I've been using a 74HC14 with a bridge made of TIP120 and TIP125 transistors to drive a GDT. This way you have only a single resistor and capacitor in the oscillator so your duty cycle is nearly perfect with no drift.

74HC14, TL494 and similar parts are under 50 cents each.

Ash Small

Tue Apr 26 2011, 06:42PM

Registered Member #3414 Joined: Sun Nov 14 2010, 05:05PM
Location: UK
Posts: 4245

James wrote ...

All you need to make a TL494 run open loop is a capacitor and a couple of resistors. Another option is a gated oscillator, I've been using a 74HC14 with a bridge made of TIP120 and TIP125 transistors to drive a GDT. This way you have only a single resistor and capacitor in the oscillator so your duty cycle is nearly perfect with no drift.

74HC14, TL494 and similar parts are under 50 cents each.

Well I can certainly see the sense in having only one resistor and capacitor. All the circits I looked at had a minimum of three resistors and at least one capacitor.

Do you have a link to the circuit you describe?

ScotchTapeLord

Tue Apr 26 2011, 06:49PM

Registered Member #1875 Joined: Sun Dec 21 2008, 06:36PM
Location:
Posts: 635

Alright, I think we've advertised ICs enough. I'll even withhold my feelings toward the 2n3055...
I think it's good to make such oscillators for proof of concept. If you don't mind making adjustments and high quiescent operating current then the approach is fine.

The main reason I think a lot of people are talking about the tl494 is because it is relevant to the topic of eliminating shoot-through (without shorting the supply out).

Ash Small

Tue Apr 26 2011, 07:37PM

Registered Member #3414 Joined: Sun Nov 14 2010, 05:05PM
Location: UK
Posts: 4245

ScotchTapeLord wrote ...

Alright, I think we've advertised ICs enough. I'll even withhold my feelings toward the 2n3055...

Again, using 2N3055's is part of the 'fun'. I don't know if I'll stick with them, but if I can get it to work I will at least have proved the skeptics wrong.....personally I think the trick is in the heatsinks.....watch this space

ScotchTapeLord wrote ...

I think it's good to make such oscillators for proof of concept. If you don't mind making adjustments and high quiescent operating current then the approach is fine.

It's educational as much as anything, I can get the 'scope probes everywhere, which you can't do with an IC. The current concerns me a bit too, but I can always use it for heating and lighting in the winter

ScotchTapeLord wrote ...

The main reason I think a lot of people are talking about the tl494 is because it is relevant to the topic of eliminating shoot-through (without shorting the supply out).

I appreciate that, and I'd give that same advice myself. I think part of this is about building 'the circuit I never built as a kid in the seventies'.

I'll report progress here, but making the heatsinks is delaying things at the moment.

James

Wed Apr 27 2011, 12:23AM

Registered Member #3610 Joined: Thu Jan 13 2011, 03:29AM
Location: Seattle, WA
Posts: 506

Ash Small wrote ...

Ok so maybe you need 3 resistors, I don't remember, either way we're talking pennies here. I've breadboarded TL494s several times using parts I had laying around on the bench.

If you delete the optional protection circuitry, this is very simple

I see he's updated the gate drive from discrete transistors to ICs, but either one works just fine.

If you want to reinvent the wheel and use a discrete multivibrator instead of a 39 cent purpose designed IC go for it, it will probably work to some extent, but it's more effort for less flexibility.

previous 1 2 3 4 next

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