Cap Dump Circuit

[BroMikey]

Heat reduction method

PS update

Been finishing up now, 12 fets are running also changed the 30 amp meter out to read pulses because it was heating the shunt. I have been burying the needle so got a 200 amp meter in it's place. The pulses are as high as 50 amps but I didn't know that til I switched meters.

I will be showing the added coils in video after I am sure of success but so far with 12 brand new FET's running for the first time today giving 40-50 amp pulses, these suppression coils help to balance out electrical pressures thus helping to even out heating issues.

Of course next on the list is to finish the new control board for further reduction of heating due to incorrectly steering the FET's.

Mike

[BroMikey]

Coils of Emitter and Colloector

I ran all day and adjusted my timing circuit around to see if there is a change.

The added coils do lower heat on the sink is noticeable. Coils used this way are sometimes referred to as "ARC SUPPRESSION" coils.

I am seeing comments from others who compare my cap bank to theirs.

Let me stress that the way my bank is made is way more powerful than the average bank of capacitors. I know I have many cap banks.

For instance I have 2 banks that I compared one against the other.

Both banks were charged to 50vdc or 65,000uf. The physical size comparison is bank two is the size of a softball and bank one is size of a beach ball.

Same voltage and same microfarad size. Bank one seems to be 5X more powerful. Don't ask me why but I think rating vary with physical size.

At any rate, the pop coming from the physically smaller bank is far far weaker.

This is puzzling and unless an experimenter has build a bank of this using 250vdc caps they have nothing to compare it with.

It is 4-5 times more powerful than my 65,000 uF bank.

Now having laid the ground work for this next statement, I conclude by saying that a larger cap bank requires more Mosfet's to handle the power, so when the average JOE EXPERIMENTER flies by this forum and sees my uF value he is thinking how he has done that many times no problem.

Also I reread the Patent on the John Bedini energy pump and it say a cap bank of 132,000uF is chosen to operate the unit. Therefore I recommend that experimenters not automatically assume that the uF value on their caps are equivalent to all other capacitor Banks. That would be a gross misconception.

If we all look back at various web data surrounding John Bedini's capacitor discharge circuits we find that for charging a 12vdc battery a 75-80vdc cap is selected. The next most common capacitor voltage available for a 36vdc or 24vdc would be 200-vdc or 250vdc following that pattern.

However what I see is experimenter using the same 80-100vdc capacitors to dump 70vdc. Using low voltage caps to dump 75-85vdc throws the rated values off by double and triple what the cap bank should be.

Like I said I know I have multiple cap banks and they all have a job.

Mike

[BroMikey]

Coil Suppression

Here is my latest addition to my dump. It ran much cooler I must say.

I still need to get away from the slow switching speeds but for now it

is best to lower heat other ways first.

BroMikey's Science Projects

Thanks Farmhand

Mike

Here is the link to the video I just did tonight.

CapDump 12devices 4coils Bromikey Alum Batteries - YouTube

Today I did 60 amp pulses per second charging a 24vdc battery set.

This video shows 12 fets and 4 balancing coils of 12 turns each, 1" dia when wound on 1/2" cpvc hot water pipe. The energy stored in both coils (Or all 4 coils 2 for each bank) creates pressure on both sides of the waterfall so avalanche occurs in more of an evenly distributed non sparetic rather uniform action.

I looked at the Bedini energy pump circuit and saw John was using a high wattage very low ohm (.005) resistance and thought to myself only a coil could do that. It seems I was right after using 8awg (45 amp) wire to make the coils and testing I find that this solves most if not all of my mosfet heating issues without the use of higher accuracy driver.

I have noticed posts that show experimenter doing everything right using high quality driving circuits and the heating issue will still be a problem.

Use coils for low induction short term energy storing and balancing.

They work beauty like.

Mike

[BroMikey]

Thinking it through.

Really that was a stupid question on my part to ask Farmhand if the 400hz 10amp pulse was as fast of a charge as straight DC.

When I thought it through better I realized that the current is off 90 percent of the time so it can't really charge to fast.

Great post Farmhand you got me thinking beyond my 3 pulses per second.

The only thing I don't get is how do you get a 10 amp pulse from your caps?

But I will think about it more and come back.

I mean you can figure that the pulse is 10 amps 400 times per second but you can't see it on a meter.

Doing the calc's by math and not meters I am getting a 400,000 amp pulse

Not that the big one is better. Got to try it both ways now.

Mike

[BroMikey]

10 amp pulse@400hz

I heard the new bedini chargers for the small number ten something uses this idea to charge batteries of all kinds using 8 pulses per second. He said that on a video.

I heard Peter L talk about how a rest interval of some many microseconds was all that was needed to stop the super heating condition that conventional charge exhibit.

I know that heating of plates damage batteries and anytime an unattended charging unit is left a few seconds to long damage will occur.

I have some li-ion batteries in a cordless gun that I fried in 1 hour. The time before the pack charged fine but the next time it burned up because I guess I set it to high?

With so many functions to zap a battery up fast without heating a sensor of some kind would keep the battery safe. Maybe one cell can be fitted with a laser temp probe for giant batteries.

This way a tiny computer control could have a better feedback. Then a 400hz 10amp pulse could drive up the battery to 13vdc and drop back power levels to 3 pulses per second @30 amps instead of 100X more power of forced charging.

You see it is still a forced charge at 400X 10 amp pulses per second. I don't know how to explain this but it just seems like it to me that a battery would be charge in a few hours.

Mike

[BroMikey]

400hz Charging

Hey Farmhand

I should of asked you if charging a battery using 400hz is about as fast as a straight DC charger?

Also I think the caps are not being used so this would mean a direct flow from source to charge battery except for some ripple.

I did try that on my dump because I could go up to 20 or 30hz but the input will just lay right over, no fluctuation a steady flow of current but when I did it I wasn't sure what to expect of what was going on in the scope shot but the output side would lay over then pulse then lay over then pulse like a heart beat.

The current looked steady on but I know it was not. Like I said the output acted strange. The Fets cooled right off as they were not working very hard and the caps only pulses once every couple seconds. This would mean that the battery was being forced up by faster pulses and more current and I will assume much more heating of the battery plates along with a speedy charge.

Sometimes at the beginning of a charge cycle it might be nice to drive up the battery faster without the use of Direct Current because that temptation will arise when time runs out.

That 10 pulse is alot. 10 amp pulse? 400 times per second?

Mike

[BroMikey]

Coils and shunt resistors

Farmhand wrote:

I'm now using a microwave oven transformer primary

A coil with less resistance but a with the appropriate inductance is better

Second shot shows the circuit working at 400 Hz and pulsing the battery with

10 Amp rectangles of current for 260 uS, it's 400 Hz 10% duty, That really

pumped up the battery.



At 400hz I don't think you need many caps do you? I mean at 400hz is almost a straight shot from the supply but it does solve some problems.

I thought of my 4 strand twisted 14awg coil I have. I could use that in generator mode or just the plain coil with no circuits.

Got me think

Mike

[BroMikey]

Fast turn off

yeah you got it, kinda, the coils inductance slows the flow of current from the
supply to the dump caps, I'm using smoothing caps because I 'm using a
transformer, I can't setup a large setup like yours but I can set up a smaller version.

So with the right coil between the supply and the dump caps and the right
"on" time and "off" time we should be able to get down to almost zero current
at switch off due to the coil storing the energy and causing the dump cap to
drop voltage without drawing directly from the supply during the dump time,
the dump caps fill when the switch is off because the coils "momentum"
causes it to discharge into the dump caps due to the switch turning off.

The intention isn't to heat up the coil or anything and a coil with less
resistance but a with the appropriate inductance is better.

I'm now using a microwave oven transformer primary, and with 30,000 uF I
can get the cap to fully discharge in around 70 mS or less which leads to an
almost no current situation at switch off. it takes about 350 mS to recharge
the cap and it goes to a slightly higher voltage due to the storage of energy
in the big coil. The peak current is the same if the voltage is the same and
the resistance of the current path on discharge is the same. This is a way to
isolate the supply from the load during the dump without using a second
switch "except for the diode".

I'll include some shots of the action of the voltage on the caps in that mode.
as well as a shot of the applied voltage and resulting current when the
discharge is short and the capacitor does not get fully discharged.

The coil and dump cap is a "resonant charging circuit" but it is very low "Q"
being so low a frequency. I can do from about 1 per second to about three
per second and keep a voltage rise on the dump caps.

Also if you put the capacitance and inductance in this calculator it will tell
you the resonant frequency of the two together which will tell how much
inductance is required for a given capacitance to get a certain frequency.

Resonant frequency calculator
Resonant Frequency Calculator

The first shot shows the supply capacitor voltage in blue and the dump cap
voltage in yellow, as the mosfet turns on the dump cap discharges but no
current flows directly from the supply to the load due to the coil causing a
delay and storing energy, which is released when the mosfet turns off.

Second shot shows the applied voltage and resultant current when the
mosfet is switched on and off before the cap can discharge, this requires a
fast turn off to do without excessive heating of the switch.
This shot shows the circuit working at 400 Hz and pulsing the battery with 10 Amp
rectangles of current for 260 uS, it's 400 Hz 10% duty, That really pumped up the battery.

Cheers

Hey Farmhand

Good run down on coil energy storing and inductance figures.

As you know I have been running coils in this location but no diode. Why didn't I think of that?

My coils are made of nickle chromium wire used in heating elements on clothes dryers. No real inductance, yet inductance should be part of this operation.

Glad you keep pointing these things out plus showing how this looks.

I am an oscilloscope quadriplegic. I can't show myself what's up yet. I have hooked a 1X probe across the collector and emitter on the MOSFET and I can get all kinds of things.

I can get the sawtooth dump.

I can dial it in to see nS and uS pulsing.

I will get better but in mean time YOU are there with the shots, good man.

I am thinking that my inductance coil is going to have to be 12awg wire and about 500 feet. Ball park. "Probably"? I should put taps along the coil to change input speeds or whatever you call that.

Yeah I am thinking a big coil now for sure .

Diode too.

Then this "CSR" current sensing resistor or some say "Load resistor or shunt resistor" I think this should be a coil also? If a coil is used a higher wattage low resistance value can be reached cheaply and energy is not wasted but stored.

Farmhand wrote:
This shot shows the circuit working at 400 Hz and pulsing the battery with 10 Amp rectangles of current for 260 uS, it's 400 Hz 10% duty, That really pumped up the battery.


Now you have my attention. I will be back to comment on this portion of your ability to pulse a battery with the drivers you show in the photograph.

I am getting excited now more than ever as I build my driver board.

It is so simple to make a board but knowing how to use these boards with diodes hooked to CSR and coil inductance's is way beyond.

I will have to repeat these ideas many times to the group as I remember them for myself.

You are leading the way Farmhand, great work.

Mike

[Farmhand]

yeah you got it, kinda, the coils inductance slows the flow of current from the
supply to the dump caps, I'm using smoothing caps because I 'm using a
transformer, I can't setup a large setup like yours but I can set up a smaller version.

So with the right coil between the supply and the dump caps and the right
"on" time and "off" time we should be able to get down to almost zero current
at switch off due to the coil storing the energy and causing the dump cap to
drop voltage without drawing directly from the supply during the dump time,
the dump caps fill when the switch is off because the coils "momentum"
causes it to discharge into the dump caps due to the switch turning off.

The intention isn't to heat up the coil or anything and a coil with less
resistance but a with the appropriate inductance is better.

I'm now using a microwave oven transformer primary, and with 30,000 uF I
can get the cap to fully discharge in around 70 mS or less which leads to an
almost no current situation at switch off. it takes about 350 mS to recharge
the cap and it goes to a slightly higher voltage due to the storage of energy
in the big coil. The peak current is the same if the voltage is the same and
the resistance of the current path on discharge is the same. This is a way to
isolate the supply from the load during the dump without using a second
switch "except for the diode".

I'll include some shots of the action of the voltage on the caps in that mode.
as well as a shot of the applied voltage and resulting current when the
discharge is short and the capacitor does not get fully discharged.

The coil and dump cap is a "resonant charging circuit" but it is very low "Q"
being so low a frequency. I can do from about 1 per second to about three
per second and keep a voltage rise on the dump caps.

Also if you put the capacitance and inductance in this calculator it will tell
you the resonant frequency of the two together which will tell how much
inductance is required for a given capacitance to get a certain frequency.

Resonant frequency calculator
Resonant Frequency Calculator

The first shot shows the supply capacitor voltage in blue and the dump cap
voltage in yellow, as the mosfet turns on the dump cap discharges but no
current flows directly from the supply to the load due to the coil causing a
delay and storing energy, which is released when the mosfet turns off.

Second shot shows the applied voltage and resultant current when the
mosfet is switched on and off before the cap can discharge, this requires a
fast turn off to do without excessive heating of the switch.
This shot shows the circuit working at 400 Hz and pulsing the battery with 10 Amp
rectangles of current for 260 uS, it's 400 Hz 10% duty, That really pumped up the battery.

Cheers

[BroMikey]

Nice Test Dude

I've been busy making some shots and pictures for you, I'll go back and read
the posts I missed soon.

Here below is a circuit I tried tonight, I used 26 volts dumped in a 250 mS
dump time with 500 mS recharge for the dump caps. The 45,000 uF caps had
not fully discharged in 250 mS, There was still about 2 volts and some current
when the mosfet turned off.

I got about 13 Amps peak current on dumping that declines. The current
sense resistor is 0.8 Ohms and it got hot but the mosfet stayed cool to the
touch. Maybe if you put a 0.5 Ohm resistor in series with the mosfets Sources
it might help.

Also

Shot 1) is the wave form of the applied voltage in blue and current yellow. Note
the voltage scale and the yellow voltage is divided by 0.8 to get amps through the battery.

Shot 2) is the voltage on the supply caps and the dump caps. Note the dump
caps drop voltage and this reduces the current a lot before switch off.

Shot 3) is my gate wave form taken on the gate while working.

Shot 4) is the fall time of the gate voltage. This allows the turn off with
current still flowing without excessive heating, because the cap was not fully discharged.

This indicates 2 Ohms resistance in the discharge path, if I took out the 0.8 Ohms
current sense resistor the peak current would be more like 21 Amperes. I think

Cheers

P.S. If I put my ear near the wire to or from the battery or near the mosfet I
can hear an audible tick....tick....tick in the mosfet and wires, it sounds like a
spark plug sparking, but it is a good sound, the sound of shock and awe,
If I was to scope different parts of the wires I would see ringing at HF probably.

In cicuits that switch randomly like my solar circuit the clicking sound becomes a jittering buzz,
like a code or logic language.

..

2nd P.S We can see by the initial voltage and the initial current on discharge
that 26 volts x 13 amperes = 338 Watts applied initially to the battery, but
declining, the circuit is drawing a fluctuating power of between 25 and 65 Watts from the wall socket.

Actually, maybe the battery voltage needs to be subtracted from the 26 volts cap voltage for some calculations.

...

And the coil I used with 2.7 Ohms resistance got real hot so need more
inductance and less resistance in the coil, maybe a MOT primary would work.

Pretty tricky Farmhand

I am learning something new putting in coils of high wattage in the place you show. Now I get it so the heat goes to the coil and not the MOSFET.

Nice waveforms, lots of good work, what a man. So this way a coil could save the FETS. Plus a coil done right won't burn the energy just slow if and store it. Humm....? I don't know.

You have got the wheels turning. I like the looks of your driver board.

My input is 1000 watts. I charge my cap bank to 100vdc and 3 pulses per second empties the bank beginning at 65-70vdc and down to 36vdc so I am operating so fast it is like an always "ON" condition. Very little rest interval. 3 pulses dumping with each pulse taking 300mS to discharge so 50mS charge up time.

Me bee smokin dem fets

Waveform 1-4 is very interesting. I am glad to see these because I am not sure how to do it over here.

Mike

[Farmhand]

I've been busy making some shots and pictures for you, I'll go back and read
the posts I missed soon.

Here below is a circuit I tried tonight, I used 26 volts dumped in a 250 mS
dump time with 500 mS recharge for the dump caps. The 45,000 uF caps had
not fully discharged in 250 mS, There was still about 2 volts and some current
when the mosfet turned off.

I got about 13 Amps peak current on dumping that declines. The current
sense resistor is 0.8 Ohms and it got hot but the mosfet stayed cool to the
touch. Maybe if you put a 0.5 Ohm resistor in series with the mosfets Sources
it might help.

Also

Shot 1) is the wave form of the applied voltage in blue and current yellow. Note
the voltage scale and the yellow voltage is divided by 0.8 to get amps through the battery.

Shot 2) is the voltage on the supply caps and the dump caps. Note the dump
caps drop voltage and this reduces the current a lot before switch off.

Shot 3) is my gate wave form taken on the gate while working.

Shot 4) is the fall time of the gate voltage. This allows the turn off with
current still flowing without excessive heating, because the cap was not fully discharged.

This indicates 2 Ohms resistance in the discharge path, if I took out the 0.8 Ohms
current sense resistor the peak current would be more like 21 Amperes. I think

Cheers

P.S. If I put my ear near the wire to or from the battery or near the mosfet I
can hear an audible tick....tick....tick in the mosfet and wires, it sounds like a
spark plug sparking, but it is a good sound, the sound of shock and awe,
If I was to scope different parts of the wires I would see ringing at HF probably.

In cicuits that switch randomly like my solar circuit the clicking sound becomes a jittering buzz,
like a code or logic language.

..

2nd P.S We can see by the initial voltage and the initial current on discharge
that 26 volts x 13 amperes = 338 Watts applied initially to the battery, but
declining, the circuit is drawing a fluctuating power of between 25 and 65 Watts from the wall socket.

Actually, maybe the battery voltage needs to be subtracted from the 26 volts cap voltage for some calculations.

...

And the coil I used with 2.7 Ohms resistance got real hot so need more
inductance and less resistance in the coil, maybe a MOT primary would work.

[BroMikey]

No the rating for the IRFP460 is 20 amperes for continuous source to
drain current and 80 amperes for pulsed source to drain current, however we
can't expect to reach those values without at least passive cooling with heat
sinks or possibly with active cooling (fans). If not going over 100 volts applied
why the need for a 500 volt part ?


I found that once I could reliably switch the mosfets all the way on and all
the way off quickly then I could make the mosfets handle a lot more power
with less heating.

Have you got a gate drive wave form to look at ?


Cheers

Hello Farmhand

Thank you for reaffirming what I need to hear that like you said above, MOSFET HANDLING A LOT MORE POWER.

No I don't have the wave shape just yet it is coming. I am out in the burning sun working on AC units ETC.....................................

Last Night I did some calculations for a switchable 1,2 3, and 4 pulses per second. I am going with a 100uf cap and resistance is 4000, 6000, 8000 ohms if I remember right but been dragging my feet making sure I do things right cause i never tried this one before.

To answer your question as to why I went from a 200v device up to a 500 volt device is in the industrial applications handbook by I think ONSEMI.

The handbook says 200v and 250volt devices are good up to a max of 70 volts and that these 200v and 250volt devices are generally operated best in the 50 volt range.

I was looking around to try and understand why my 200volt fets were blowing at 70-80vdc and the handbook also said that for industrial designs 80vdc and above were more stable and performed longer using the next step up in device voltages being a 500vdc ceiling.

Also it was brought to my attention that a cap dump to battery often produces a reflected voltage surge of up to 3X the dumped voltage and did explain why unexpectedly even running at low power level these lower voltage parts failed.

Also I am using to large coils of wire in the form of 2 separate toroid's, one a variac and the other a 3kva size step down. As the caps are filled and dumped the coils surge right along with the process. Inductive surging can cause voltage fluctuation.

So with all of this in mind I decided to go with a higher voltage rating of 500vdc. It was a great move because now i can go for weeks or months running a 300-500 watt input where before those input levels would make short work of my devices.

I have come along way in the last month learning by burning FETS out.

The only way I could burn out this last set of 6 FET's and I was trying to. Was to max out my supply at 100vdc using a surging amp draw from 7-12 amps. I left it in a 100 degree room for 2 days waiting for the moment of truth BOOM!!!! a straight shot to the supply at 5amps.

Oh yeah one more thing I need to say about device voltages and current ratings. A 500vdc max ceiling generally is operated at around 100-150vdc but will not be able to run the full amp rating. The full amp rating is often shown at a 10volt or 30-50vdc range so these ratings change with the weather. So unless you specify exact application and show the actual device operating I wouldn't even guess at how many amps a 20 amp device might be good for. Well if I guessed I would say more like stay within a temp range measuring while it runs and turn it down. Only then will you know for each application.

Ball Park: A device capable of passing 280 watts operating as an oscillator like in a PC power supply will have a rating of 30amps at 500vdc and this should give folks some sort of example for applications that target longevity.


Anyway my FETS were all running far over their max. I pushed them to 380 watts each which is near double what they should be operated at.

Thanks for staying there till I make the grade on this step up in steering gates. It is good to have someone remind me that the new circuit will be better.

And yes a wave form will be posted for your review.

Till then I have a million questions but won't bore everyone just yet.

Mike

[Farmhand]

No the rating for the IRFP460 is 20 amperes for continuous source to
drain current and 80 amperes for pulsed source to drain current, however we
can't expect to reach those values without at least passive cooling with heat
sinks or possibly with active cooling (fans). If not going over 100 volts applied
why the need for a 500 volt part ?

Mosfets can switch DC so that it flows continuously for however long is desired.
They have a continuous current rating.

IRFP460 pdf, IRFP460 description, IRFP460 datasheets, IRFP460 view ::: ALLDATASHEET :::

The current and voltage ratings are only related in that a higher voltage will
cause a higher current through a given resistance. So if you apply 500 volts
to a 5 Ohm load through an IRFP460 then the current rating will be exceeded.
All of them. A higher voltage part will not help deal with more current.

I found that once I could reliably switch the mosfets all the way on and all
the way off quickly then I could make the mosfets handle a lot more power
with less heating.

Have you got a gate drive wave form to look at ?


Cheers

[BroMikey]

160 amp device

I am correcting my last post of 200 amps. The tracker 5 I am looking at is a 160 amp unit for driving raw wattage into batteries like sound amps do to speakers.

This tracker operation is a cool subject.

Look close in the video. John uses 16 amp PNP devices. There are 24 of them for a max amp handling of 380 amps but JOHN never runs them over half of that for MAX and MAX output is rarely ever reached with any good setup.

https://www.youtube.com/watch?v=y93IwhOGWB4

Am I getting through?

24 huge power transistor twice the mass of these FETS.

24 devices at 8 amps each oscillating at whatever frec adds up to 192 amp handling AT 38vdc.

So at 100vdc the same setup might only be good for 3 amps for each Transistor.

Does anyone know that a cap dump is not operating in microseconds>

This beasty is on for 300mS 3 times per second the rise is a few "uS" that is "MICROSECOND" so the dump is always on for the most part.

A dump should be rated the same way continuous power flow ratings are made like a contractor application.

It is off some but mostly always on and dumping. Off time is very few mS "MILISECONDS" during the length of 1000 miliseconds.

Am I reaching anyone?

Mike

[BroMikey]

Engineering Jargon

You see when Farmhand talks about a tiny part switching 160 amps it is a technical marvel but only applies to a micro second rating.

All of the graphs show that a device at 77 degrees can throw alot of amps for say 50-100 micro-seconds. And then if the process is repeated this is called the "repetitive" rating and drops. Then if the "TIME" goes beyond to say double 200 microseconds the amp handling goes way way down.

Switch mode power supplies work well with high frequency pass transformer and oscillations of the FET can do wonders.

In my case as the "TIME" is extended to 300 mili-seconds the Fet can no longer pass it's maximum value. A good way to understand what a cap dump needs in terms of power handling is to do this.

Take a large battery charger and charge a battery with it at a rate of 20 amps but do this through a MOSFET that is fully turned on. After several seconds the device will begin smoking.

I visualize my capacitor wires like channels of a river all coming together into one large canal. Now I want to break up my stream into small branches for redistribution.

The small branches are my FETS processing the flow coming from the canal.

In my case it looks like 12 FETs will be enough to absorb the flow of energy and then pass the power into another location. If my branches are not big enough then the pressure and force increases as the canal is much bigger and is over powering these pathways.

Water pressure and electrical pressures can resistance and damage.

Take John Bedini tesla tracker 5. John uses 24 16 amp rated devices to pass 200 amps. The max rate is 380 amps approx. Plus 200 amps is never reached during the machines operation.

John B Track 5 is using large wire and the way to figure is by common rules. A conductor can only pass so much and if each tiny leg of 24 devices adds up to a 200 amp wire size, you have just figured out how to size things.

I chaet, I watch and follow the best.

Mike

Mike