Cap Dump Circuit

[Mario]

Hi Mike,

If you follow the SOA curve of your devices you should not burn them.

I already wrote this, but if you use an opto the best rise/fall times you will be able to see on your Fets is 4-5 microseconds (not nano). It's simply not possible to go faster because it's the limit of your opto (h11d1).
Only using a driver you will be able to switch faster. So I don't know how you're able to see nanosecond switching if you're using an opto.

One reason for toasting fets can also be this:

When you turn the dump OFF very sharply (talking nanoseconds) and you are not dumping down to the battery voltage (even just 20 or 30V above the battery) you will get a negative spike that can be even a few hundred volts in amplitude. It's like your battery is behaving like a coil, you pulse it, and when you let go you get the flyback spike. This is why it's important to watch the dump pulse on a scope and learn all about it while you change parameters and see what happens.

regards,
Mario

[BroMikey]

No Driver Circuits

Here are a few examples of what can be found all over the web that are good for 5-25 watts. Very low powered. Led lights and blinkers.





Also here are some circuits that are all grown up and qualify for handling higher power with longevity in mind.



In the bottom picture this man was able to drive 4 Mosfets using his own driver with the PNP and NPN toggling like found in prepackaged drivers.

My circuit needs to handle 80vdc at 10amps. 800watt power circuits are far more involved if they are to last.

Industrial apps say that any device operating at 80v and above should be designed using 500v ceiling devices and at present all I see experimenters using are the 200v devices.

So what I have decided is that since my circuits keep burning out i will use 12 mosfets and another more advanced circuit for controlling them.

I will show a picture of my home made heat sink with devices soon. The fan and 4 devices are installed already.

Mike

[BroMikey]

Mosfet Driving Types

Hello Group Members

I have been doing some reading on mosfet drivers and considering all of the words and examples each of you has given me.

What I have found is that low power and high power drivers can do many things but because I do not need multiple functions occurring I will stick to the simple non-inverting high power drivers.

Also according to the big semiconductor companies anything operating at 80v and above use Mosfets that are 500v devices not 200volt devices like I have been instructed to use.

Of course these companies want to sell parts and their rating is for industrial uses.

Although I was having the same burnouts at low voltages.



You will note that to properly drive a Mosfet for high power applications and for highly accurate digital processing that pnp and npn transistors are employed.

Many experimenters use a single 555 timer and caps, resistors with low tolerances so at higher powers this is like a gate swinging wildly.


I could measure -150ns rise and 400ns rise with only 1/16" turn on the dial and crash all of my parts in a short time.

I was using an opto to a single transistor and very sloppy work.

The driver can be properly done without a packaged part as well.

Michael

[BroMikey]

Running up Power

Was running up the power and adjusting the frec and duty and noticing how the rise times and fall times would flip over and I would lose one fet each time. This 555 timer to opto popped or I saw a flash and it sounded like a fire cracker went off.

What happened was my heat sink was hot all day because we went from 40 outside to 90 degrees in one day, the fets were not tight on the sink I should have re-tightened them after a short heat up, to much heat-sink grease and heat, makes for loose parts in a jiffy.

Besides cold solder joints and sinking problems my steering wheel (a weak example of a triggering circuit) being a 555 timer to opto.

It seems you all were right that it would fail.

Well you didn't all say that, you just said "I WOULD DO IT THIS WAY" and so on.

Now I can see why.

The 555 timer triggers an opto and this fires a transistor.

I am and was getting negative and positive triggering all in the same breath.

I tee total mess of conflicting triggers.

My scope would say -140ns rise time and 40ns fall and then if I adjusted the 10 turn pots on the fly the pulses would swing wildly.

You know 500ns rise and 200ns fall and then sometimes I could get it down to 10ns rise and 10ns fall but the problem is when you adjust on the fly if a bad spot develops on your wiper of your pot erratic triggering could blow things up and did several times.

I have the drivers and tl595's and am redesigning. The heat sink is done as of today. I built a new sink that resembles the wheat-field sink. Or whatever that name is.

Bee C-in Yons

Mike

[BroMikey]

Scope Work and Minus Signs

If any of you understand the minus sign on oscilloscope screens let me know.

I am testing the rise time on the switching of a mosfet IRFP250 and fall time and the scale is NANO or ns this is nano-seconds.

As I adjust my switch driver I can control time and duty where sometimes I read a minus or this sign (-) in front of the number like say -140ns.

My question is what does the minus sign show? Then I will adjust and the value will move as expected to say 50ns.

I realize that this is a highly technical question that all you theoretical Physicists won't be able to comprehend, so I am looking for a hands on guys who can build something and is able to operate an oscilloscope.

Thanks in advance.

Michael (I will continue to ask this question)

[BroMikey]

Cap Circuit figures

Okay I am learning to look at the HP54100 scope and I was seeing the 55 timer at 500ms(miliseconds)

Now I see the switch but why do I get -.

Most of the adjustments put me from 50ns to 400ns to 10us

Sometimes I get rise times in the negative.

Like (-140ns) rise time and 50ns fall time

I am not sure what this means.

My battery is set and on a pallet, lead bus bars and 150amp wire 30ft long from the basement to the 2nd floor.

The dump works great.

If any of you can answer that question I would be open to hear it. I am learning about scopes.

Michael

[BroMikey]

Capacitor discharging operation

Hi Mike,

Interesting video at

Capacitor Discharging RC source - YouTube

Regards,

VIDBID

Thanks VIDBID

I also need to add one more thing about my specific circuit. These machines are called "ENERGY PUMPS" I noted that if I put an amp meter between the charging batteries I can see amp surges this way.

The thing is that when experimenters make pulsating circuits the fear is that they will lock up their devices (In this case Mosfet's) and burn them out, so as I adjust my pulse circuit to a shorter time down to zero I was afraid that they might not oscillate, right?

Well as it turns out with my diagram even when I turn the time for a pulse to nothing the LED still flickers and the fets still open and the circuit still surges a 20-30amp pulse or some say a 100-300 amp pulse measured in a very short time.

I know this, when I dump that bank with a single discharge it acts similar to a welder at 50-70vdc. Screw driver tips vaporize in a blink.

I have been burning my circuit in and have reached 5-600 watts with little heat. I don't know how to relate the power and potential effects of a large discharge at 500 watts.

One discharge behind the other? Faster and Faster with an ever increasing flow of surging amperage til the source is still fluctuating, but higher.

For example if I run low power and a slow pulse time the amps will charge the cap bank coming up to say 1 amp and then back to zero. However as I increase the speed of the dump, the RC charge curve will require a higher cap bank voltage to deliver the same voltage dump to the batteries.

As I go faster and more voltage, the power levels climb and so far the highest that I have sustained for a single 120vac wall plug without flickering the lights is 900watts. The circuits are not getting hot at all.

I like the extra power that general I won't be using all of the time. The critical factor is the RC charge curve for filling a large bank of caps without throwing the mains on your house when power levels are elevated.

200-300 watts is nothing to a 1500watt wall plug so there will be no problem at those levels even with a messed up design.

Try getting up near 1000 watts sometime of pulsating power and you will see my dilemma.

At first try I found that even a 1 amp input of unregulated power to a discharging cap bank could nearly pin the needle to 15 amp surge demands between dumps for charge up.

Now I will say that the SG Oscillator is a great tool and can improve the efficiency of dumping operations. Many of the guys here have pointed out that when dumping caps it is better to turn the supply OFF for a short duration so as not to interrupt dumping functions.

The Sg oscillator does that and the inventor made these to work together in his patents.

The SG Oscillator can be a solid state circuit as mine has always been and it performs this operation nicely. The SG Oscillator has several modes that it can operate in and when it is connected to a cap discharging bank will be far less disruptive to the dump circuits than wall power.

This compatibility of Oscillator connected to cap dumping circuits was designed by the inventor over years with exhaustive testing.

I can say now that the SG Oscillator is a great tool and as it supplies the cap bank with voltage for this type of radiant pulsating DC, charges caps, and will condition caps to except a charge faster and this increases efficiencies of the design.

I designed my block transformer supply and resistor to have an alternative means of powering my dumps off of a Genset running in the yard when the power fails.

I spent one full week without power a couple of years ago and I promised myself I would be ready this next time the high winds came.

Every year the power goes out here even if it is only 8 hrs.

Some are using Oscillators or inverters as they are called without much of a coil of wire to power a cap dump.

In my case I am building an inverter Oscillator combination switchable from spike mode to Genmode to rene mode.

It is coming.

If you look at my last video I show this hardware right at the very beginning.

It is another circuit just like this cap dump circuit, exactly. The only

difference is the rate of oscillations is 8000-17,000HZ instead of 3-5hz.

Mike

[vidbid]

Hello group

Here is a video of my progress and how I can run over 500 watts of surging power into batteries.

https://www.youtube.com/edit?video_id=tAflPiIM1gc

Mike A special thanks to all my helpers.

Hi Mike,

Interesting video at

Capacitor Discharging RC source - YouTube

Regards,

VIDBID

[BroMikey]

RC source and Cap Dump Ciruit

Hello group

Here is a video of my progress and how I can run over 500 watts of surging power into batteries.

https://www.youtube.com/edit?video_id=tAflPiIM1gc

Mike A special thanks to all my helpers.

[BroMikey]

Cold solder joints

Hello Group Members



I would like to add that a cold solder joint ignited and I got the entire cap bank dump right across that gap. Needless to say my Fet went up in smoke. But I have plenty of Fets and this is a small price to pay to learn.

I think someone said that each fet is good for 750 watts and will run cool so 6 fets should be good for 4000 watts Plus.

If anyone has noticed the "radiant" chargers are being sold using a conventional 24vdc wall adapter connected to a cap dump. This simple setup will also employ some form of resistor to limit cap charging.

Like AC and others have stated turning off the supply for a few micro-seconds is another option generally acquainted with high quality computerized industrial equipment.

Mike

[BroMikey]

Dump Down Crash

Hello Mike


I should mention that if your using PWM at a relatively low frequency with big current then On resistance is huge in my opinion. You can always bump the driver to compensate for a larger gate capacitance but R-DS is a nightmare with low voltage/high current. As well it may sound counter-intuitive but in one project I found it was easier and more efficient to boost the voltage to max, switch it then buck it back down rather than take a major hit with I2R losses when switching.

AC

Hello Ac and group members

Here is what happened to my 555 timer to opto to 6 mosfet's.

Each mosfet has around 500ohms on the gate and a 10k divider arrangement and what would happen is each time the dump came I could hear a "TICK" that would get louder as the power increased.

My dump down voltages was out of control and as soon as I added atleast 100ohms inline between the opto and the entire set of six gates I regained control of "dump down to" voltage.

As it turns out the entire circuit setup is fully functional without blowing an occasional fet due to out of control crash landings at the bottom of the dump.

The "Tick" Sound is gone and the circuit can handle alot of power with some heat evolution.

IT IS RUNNING NOW AND ALL NIGHT WITH NO PROBLEMS

I am very happy.

I am dealing with RC time constants for charging the bank of caps for the least amount of loss. For instance I use a set of toroidals (ONE is a variable) rated at 2.5kVA to HUGE bridge to 5000UF 200v caps then to an RC charge resistor that is HUGE also.

It gets the job done for now and either way some heat will be lost with any large power circuit. Fans must be used when switching large amounts of power no matter how you cut the cake.

For now as in the past I use heater element wire from a clothes dryer cutting to length for proper resistance so it is only gently warm so a fan is added.

This slows the charge time of the cap bank so hardware and energy loses are minimized.


Dumping down to the battery bank voltage causes an ever increasing "TICK" noise that will smash your junctions.

Case dismissed.

Thank you all for keeping me focused while I experiment. It is your advice that gave me something to go by when I am in the middle of testing a circuit.

Michael

[BroMikey]

Deeply Descriptive

Hello Mike


The 7667 is nothing special, a cheap 1-2 amp generic mosfet driver I believe. The key here is to use a driver so the rise/fall is super fast while the 555 output is slow and adding an optocoupler is like watching paint dry.

I use a mosfet driver so the on/off transition or slope is very fast regardless of the transition from the source switching it such as the 555. Then I use the driver to switch another set of mosfets to source/sink the gates of the mosfets actually performing the switching operations. Think of the driver and source/sink mosfets as a very large mosfet driver in itself because it is and it's cheap.

The general rule is to hammer the gate hard and fast which can get a little tricky. For instance switching big current means big losses (I2R) so the on/off transition must be fast and the On resistance(R-DS) of the mosfets as low as possible.

I should mention that if your using PWM at a relatively low frequency with big current then On resistance is huge in my opinion. You can always bump the driver to compensate for a larger gate capacitance but R-DS is a nightmare with low voltage/high current. As well it may sound counter-intuitive but in one project I found it was easier and more efficient to boost the voltage to max, switch it then buck it back down rather than take a major hit with I2R losses when switching.

AC

Hello AC

Looks like you got your feet wet a few time doing the tronix. Great pointers.

I am looking at quad drivers now, all of the guys got me thinking.

In your post you talked about gate capacitance and fast switching in a way that made me think the max ratings are there because some switching times are very slow for specific design, but where the MAX speed can be used with multiple fets the capacitance might get shared so more fets can operate safely.

Thanks AC.

Mike

Mike

[Allcanadian]

Hello Mike

I have never heard of this so yeah. Is this the only part number you have? Any Idea how many 2500pf gate Mosfet's one of those darlin's can handle?

Ill go look see at the ICL 7667 dual mosfet driver ASAP.

The 7667 is nothing special, a cheap 1-2 amp generic mosfet driver I believe. The key here is to use a driver so the rise/fall is super fast while the 555 output is slow and adding an optocoupler is like watching paint dry.

I use a mosfet driver so the on/off transition or slope is very fast regardless of the transition from the source switching it such as the 555. Then I use the driver to switch another set of mosfets to source/sink the gates of the mosfets actually performing the switching operations. Think of the driver and source/sink mosfets as a very large mosfet driver in itself because it is and it's cheap.

The general rule is to hammer the gate hard and fast which can get a little tricky. For instance switching big current means big losses (I2R) so the on/off transition must be fast and the On resistance(R-DS) of the mosfets as low as possible.

I should mention that if your using PWM at a relatively low frequency with big current then On resistance is huge in my opinion. You can always bump the driver to compensate for a larger gate capacitance but R-DS is a nightmare with low voltage/high current. As well it may sound counter-intuitive but in one project I found it was easier and more efficient to boost the voltage to max, switch it then buck it back down rather than take a major hit with I2R losses when switching.

AC

[BroMikey]

Driver testing

Actually, I operate at 85VDC and >30,000uF but my pulse is only 2.5-3ms. For what you do some cooling will be necessary. Your duty cycle seems quite high. Is this something you figured out experimentally?
I was going to build cap pulser controlled by microchip with 6 pairs of Fet's but got busy with my other projects a.t.m. Besides, I got rid off all batteries except two and I have couple other ways to charge them using solar power.
I sent you PM.

Regards
V

Yes i learned much from experimenting and I had some tc4420's coming but may try others if I was sure.

Great to see your work in electro-Med. The people need you with all of the planned sickness against the populous.

Looks like you have some slippery handed folks running around outside HUH?

Either way your work is an added help and great for learning by.

Mike

[BroMikey]

Very Well Then

@BroMikey

There is an easier option I think, I was using a ICL 7667 dual mosfet driver and only using one channel (A) to drive both a P ch mosfet and an N ch mosfet as a half bridge with a common gate. It's basically a flip flop however the half bridge current can be very large so it can drive a huge number of mosfet gates.

Basically the mosfet driver doesn't allow a slow transition, when it crosses the turn on threshold it switches at max speed as well as the turn off threshold. So why buy a bunch of mosfet drivers when we can use one to make a half bridge switch just as fast and handle big amps to drive other switching mosfet gate capacitance(s). It's stupid simple and can be upgraded to handle 100+ amps to drive the switching mosfets... if that's your thing, lol.

I do this because circuit boards suck and so does soldering. So I want the least number of components to get the job done.

AC

Hello AC

I have never heard of this so yeah. Is this the only part number you have? Any Idea how many 2500pf gate Mosfet's one of those darlin's can handle?

Ill go look see at the ICL 7667 dual mosfet driver ASAP.

Thanks for your comment.

All comments excepted.

Mike