Prev: ooops...found another bright one .....
Next: Solar powered battery charger for Nickel batteries
From: Tim Williams on 25 Jun 2010 21:54
That circuit doesn't work. Actually, you shouldn't be getting spikes OR squares out of it *anywhere*. The emitter outputs (pins 9 and 10) have no pulldown.

Use a gate driver like this,
http://webpages.charter.net/dawill/tmoranwms/Circuits_2010/12-24_Converter.png
or use TL598 (598 is to 494 what SG3525 is to 3524). Check pinout, pins 8-12 are different.

Better still, use a half bridge gate driver, e.g., IR2101 series. Nix the full bridge, it's only good for increased losses. Replace C7 with 1uF to +30V and 1uF to 0V.

Incidentially, T1 will happily saturate if the voltage is too high for the frequency and primary turns. Saturation means dead FETs. This is very easy to calculate from core geometry and material properties (all flybacks use mix 77 ferrite, as far as I know... Bmax = 0.4T). Further reading:
http://webpages.charter.net/dawill/tmoranwms/Elec_Magnetics.html

Tim

--
Deep Friar: a very philosophical monk.
Website: http://webpages.charter.net/dawill/tmoranwms

"Scott" <smbaker(a)gmail.com> wrote in message news:7d05afef-7f77-4e8f-a7a3-25e16ed9177f(a)h10g2000prh.googlegroups.com...
> I've attempted to build the circuit depicted here:
>
> http://www.instructables.com/files/orig/FVE/I3SC/FR123O01/FVEI3SCFR123O01..png
>
> It's a flyback driver. I understand the basic idea behind the circuit,
> with the TL$94 operating in a push-pull configuration driving the
> mosfets. I have the dead-time adjusted to a value that approximates a
> 50% duty cycle square wave. Pin #9 of the TL494 outputs the square
> wave, and pin #10 outputs the inverse of the square wave. That's a bit
> of a simplification, but I think it's close enough for discussion.
>
> The problem is when I attach my scope to pins #9 and #10. Pin #10
> looks fine, but pin #9's square wave is truncated. The leading edge is
> there, but it immediately drops back to ground. Thus what I'm seeing
> if momentary spikes instead of the "on" portion of the square wave.
> The scope looks sort of like this:
>
> * * * *
> * * * *
> * * * *
> * * * *
> **********************
>
> When it should look like this:
>
> **** ***** ******
> * * * * *
> * * * * *
> * * * . * *
> ***** ****** *
>
> I hope the crappy ascii art comes out somewhat meaningful.
>
> If I disconnect pin #9 from its mosfet, then the output looks fine.
> Reconnect it and the square wave will appear truncated again.
> Strangely enough, pin #10 looks fine. If I switch the TL494 over to
> parallel mode instead of push-pull mode, it'll also work fine.
>
> My only guess is that what's happening is that perhaps the initial
> Gate-Source current of the mosfet is so high that it's causing the
> output transistor in the TL494 to shut down? I've tried substituting a
> larger resistor for the 10ohm R5. I have to go to at least a few
> hundred ohms before the effect will clear up, and this seems so far
> outside the design of the circuit.
>
> I'm using IRF810 as the mosfets.
>
> This happens regardless of whether or not the flyback is connected (so
> it's definitely not inductive kickback from the flyback), and whether
> or not the +30V supply is connected.
>
> Any thoughts?
>
> Thanks,
> Scott
From: legg on 25 Jun 2010 23:09
On Fri, 25 Jun 2010 14:59:33 -0700 (PDT), Scott <smbaker(a)gmail.com>
wrote:

>I've attempted to build the circuit depicted here:
>
>http://www.instructables.com/files/orig/FVE/I3SC/FR123O01/FVEI3SCFR123O01.png
>

Looks POW-ful, to me.

Please review the 494 spec.

....and stop calling this rubbish a full-bridge flyback.

RL
From: Scott on 26 Jun 2010 00:01
> That circuit doesn't work.  Actually, you shouldn't be getting spikes OR squares out of it *anywhere*.  The emitter outputs (pins 9 and 10) have no pulldown.

Ok, everything you've said makes sense, I like the looks of this
IR2101 driver, and might order a few and give it a shot. I've also
been looking at other drivers (TC4420, TC4429, MIC442x equivalents,
etc).

However, I'm confused about why the circuit works at all. I understand
the lack of a pull-down and how that should be causing the gate to be
stuck high and the mosfet to be stuck on. It does actually mostly
work, as does this schematic:

http://www.hackerbotlabs.com/wp-content/uploads/2009/02/arc-speaker.png

Albeit this last one works rather poorly, due to overheating mosfets
(most of the comments speculate that adding a mosfet driver would
solve that problem). The problem with these instructables websites is
that the original design never seems to get updated; somewhere down in
the ~ 1100 comments someone who knows what they're doing has probably
submitted a corrected design. At least I have learned a bit about the
TL494, which was part of my goal with trying to build the project.

Thanks again for all the help!


From: Tim Williams on 26 Jun 2010 09:47
"Scott" <smbaker(a)gmail.com> wrote in message news:52e87f5f-0912-42a3-a4eb-d5e7a29809ec(a)h10g2000prh.googlegroups.com...
> http://www.hackerbotlabs.com/wp-content/uploads/2009/02/arc-speaker.png
>
> Albeit this last one works rather poorly, due to overheating mosfets
> (most of the comments speculate that adding a mosfet driver would
> solve that problem).

No surprise there. And it needs a snubber. That primary is not well coupled, most of the energy delivered will go into the transistor as avalanche. It's a wonder it survives at all.

Tim

--
Deep Friar: a very philosophical monk.
Website: http://webpages.charter.net/dawill/tmoranwms
From: Jamie on 26 Jun 2010 11:50
Scott wrote:
> I've attempted to build the circuit depicted here:
>
> http://www.instructables.com/files/orig/FVE/I3SC/FR123O01/FVEI3SCFR123O01.png
>
> It's a flyback driver. I understand the basic idea behind the circuit,
> with the TL$94 operating in a push-pull configuration driving the
> mosfets. I have the dead-time adjusted to a value that approximates a
> 50% duty cycle square wave. Pin #9 of the TL494 outputs the square
> wave, and pin #10 outputs the inverse of the square wave. That's a bit
> of a simplification, but I think it's close enough for discussion.
>
> The problem is when I attach my scope to pins #9 and #10. Pin #10
> looks fine, but pin #9's square wave is truncated. The leading edge is
> there, but it immediately drops back to ground. Thus what I'm seeing
> if momentary spikes instead of the "on" portion of the square wave.
> The scope looks sort of like this:
>
> * * * *
> * * * *
> * * * *
> * * * *
> **********************
>
> When it should look like this:
>
> **** ***** ******
> * * * * *
> * * * * *
> * * * . * *
> ***** ****** *
>
> I hope the crappy ascii art comes out somewhat meaningful.
>
> If I disconnect pin #9 from its mosfet, then the output looks fine.
> Reconnect it and the square wave will appear truncated again.
> Strangely enough, pin #10 looks fine. If I switch the TL494 over to
> parallel mode instead of push-pull mode, it'll also work fine.
>
> My only guess is that what's happening is that perhaps the initial
> Gate-Source current of the mosfet is so high that it's causing the
> output transistor in the TL494 to shut down? I've tried substituting a
> larger resistor for the 10ohm R5. I have to go to at least a few
> hundred ohms before the effect will clear up, and this seems so far
> outside the design of the circuit.
>
> I'm using IRF810 as the mosfets.
>
> This happens regardless of whether or not the flyback is connected (so
> it's definitely not inductive kickback from the flyback), and whether
> or not the +30V supply is connected.
>
> Any thoughts?
>
> Thanks,
> Scott
The schematic shows them as JFET's not mosfets. if that makes any
difference.. Could be a miss use of symbols. there is a difference..



First  |  Prev  |  Next  |  Last
Pages: 1 2 3 4 5
Prev: ooops...found another bright one .....
Next: Solar powered battery charger for Nickel batteries