My Motors got me to Tap into Radiant Energy

[JohnStone]

@ JohnStone

John, here are the rise and fall times but as you know this is a less than perfect system. This measurement was taken with 18V supply with 160 Hz and 50 Duty. The rise and fall did not change with any frequency change up thru 22 KHz and tested with Arduino pulser.
Rise Time 40 us
Fall Time 40 us
These numbers are from scope which may be total rise and not your 10/90 method. Is not 40 microseconds for the whole thing in the ball park?
These push buttons let you dial in exact numbers but may not be the fastest thing to get from start to high speed.
Dana

Thanks Dana!
- Now we have a base showing up that less than perfect setups work around 40µs = 40µs heating time for any transiton H->L & L->H.
- At every transisiton we get delivered a chunk of heat into the FET. This heat needs to be drained. At low current and low frequency our drain povided (heat sink) might be OK. And we know that increasing frequency OR / AND frequency will demand more drain.
- Imagine a bucket (heat capacity) with a hole at bottom side (heat sink). Increasing level of fluid in the bucket will increase in some extent drain capacity because of more pressure (temp. difference) at bottom hole. Now we can imagine that at a certain level of fluid flow (heat) inot the bucket will get the bucket to overflow => FET death.

With this imagination in mind everybody can "feel" how FETs feel and take measures dor safe operation.
JohnS

[prochiro]

Monster V5.01

@ JohnStone

Last night I finished My Fourth Circuit and checked it out. Tonight after work I ran it up to 8 amps using a car bulb and put the scope on it at the base of fets and the rise and fall is <4.000us there. There was absolutely no heat at fets at this point. I must now change to a larger meter and run to store to get more bulbs. I will then do the same with 24 and 36 volts. There are three tangs on each bulb but only one filiment it glowing. How does one hook up the other filament. What I think I must do is then add bulbs and note fet heat, etc. Then add heat sink and keep going. John, your circuits seem as safe as a baby in its bed.
There is one item that maybe some thought be given to. That is how to fasten the large cable to the fets. I have placed the fets in a vice with the board on top and soldered one 12 gauge wire at a time until large enough. I thought of soldering a brass tube to the cable and sliding it over the fet posts with a lite press fit. That way if the fets go, I do not have to redo the whole back end of the board. Do you maybe have any thoughts about if this is an issue or not. I will report as I go thru this weekend.
Dana

[qvision]

Away for a while but i'll be back

Hello builders

I have had an exciting development with my own device (which uses HER rather than let's HER abuse US !).

I have to put some time, money and energy into this since i have been working on this one principle for two years now.

I just wanted to say i will be back, i have bought a second motor of the type i showed you before and it's a very nice motor for £12, but i have to push on with my own generator first.

I wish you all the best and i am subscribed to UFO's threads so i won't miss anything.

Build well and be Healthy and Happy


All the best,

QV.

[Hitby13kw]

Replication of PWM Driver V5.1

@ JohnStone

Last night I finished My Fourth Circuit and checked it out. Tonight after work I ran it up to 8 amps using a car bulb and put the scope on it at the base of fets and the rise and fall is <4.000us there. There was absolutely no heat at fets at this point. I must now change to a larger meter and run to store to get more bulbs. I will then do the same with 24 and 36 volts. There are three tangs on each bulb but only one filiment it glowing. How does one hook up the other filament. What I think I must do is then add bulbs and note fet heat, etc. Then add heat sink and keep going. John, your circuits seem as safe as a baby in its bed.
There is one item that maybe some thought be given to. That is how to fasten the large cable to the fets. I have placed the fets in a vice with the board on top and soldered one 12 gauge wire at a time until large enough. I thought of soldering a brass tube to the cable and sliding it over the fet posts with a lite press fit. That way if the fets go, I do not have to redo the whole back end of the board. Do you maybe have any thoughts about if this is an issue or not. I will report as I go thru this weekend.
Dana

Dana,
Congratulations on your fourth,

just hatched my first last night.


JohnStone & all

Quick note that we have successfully replicated Sir John’s design and have done basic testing with car headlights 12V 5A to ensure both FET’s are working & sharing the load equally. We have put a few pictures of the circuit in dropbox to share our joy.

https://www.dropbox.com/sh/zd28h9335...nster%20Driver

More detailed testing is to be done after heat sinking the FET’s, but enough has been done for tonight, time to sleep.

Hitby13kw

[JohnStone]

Hi builders of monster driver!
Glad you got it running easily :-)

1. Regarding H4 pinout:

Both filaments share one common pin. Thus you can put them in (1)series, (2)one at a time and in (3)parallel.

2. Check for oscillations at drain/GND and gate/GND. View one single pos and neg edge by adjusting timing to max. at your scope.

3. I am not pleased with 4µs I will check for improvements if my pulser is ready. But suffitient for now! Please note: it is ten times faster than quick and dirty dirve = 10 times less heat!
Please note that high voltage spikes occuring along coils will increase switching time (more heat!) as there is a backlash from drain to gate. Please observe again if adding a coil.

4. Please measure voltage of spikes (drain / GND) if coil added - just for documentation and alert for all builders.

5. Connecting FETS can be done by crimped ring terminals.. You can solder FET pins to far end of ring in order to not loose soldering heat to the cable. Thus soldering is quick and safe.
Else use a screw.
Please add a strain relief (i.e. those copper pads on PCB) in order to not bend and break FET pins.
In Germany we like to use those UK sockets and plugs for solar and battery connections. Thy are high current rated. In GE there is no chance to mix them up with dangerous mains :-)

Thanks for your activity :-) I am a bit restricted now but tinkering heavily on Arduino. Stay tuned. It will be another building stone to share!
@hit13KW: Your setup corresponds exactly on how I started it. I like it. Not professional at all but functional and quick.

Question: Still not shure if pulsers need to synchronize to commutator switching. Else we could suffer on frequency beats between commutator frequency and pulse frequency.
What effects will happen if switching off performs just before commutator off state?
What will be the suggested pulse position and pulse length?

JohnS

[Cornboy 555]

Greatfull.

Hi builders of monster driver!
Glad you got it running easily :-)

1. Regarding H4 pinout:

Both filaments share one common pin. Thus you can put them in (1)series, (2)one at a time and in (3)parallel.

2. Check for oscillations at drain/GND and gate/GND. View one single pos and neg edge by adjusting timing to max. at your scope.

3. I am not pleased with 4µs I will check for improvements if my pulser is ready. But suffitient for now! Please note: it is ten times faster than quick and dirty dirve = 10 times less heat!
Please note that high voltage spikes occuring along coils will increase switching time (more heat!) as there is a backlash from drain to gate. Please observe again if adding a coil.

4. Please measure voltage of spikes (drain / GND) if coil added - just for documentation and alert for all builders.

Thanks for your activity :-) I am a bit restricted now but tinkering heavily on Arduino. Stay tuned. It will be another building stone to share!
@hit13KW: Your setup corresponds exactly on how I started it. I like it. Not professional at all but functional and quick.

Question: Still not shure if pulsers need to synchronize to commutator switching. Else we could suffer on frequency beats between commutator frequency and pulse frequency.
What effects will happen if switching off performs just before commutator off state?
What will be the suggested pulse position and pulse length?

JohnS

God i'm Glad you are here John Stone!

Warm Regards Cornboy.

[machinealive]

Hey UFO and everyone,

Haven't been able to do anything but catch up with work, I haven't been this busy for a long time, but I am still "chipping away at the stone".
I got my new parts, and rebuilt controller to JS's specs. I tell ya, that opto was one pain in the ass. The output pins are reversed on mine, pin 3 is across from 1, I don't know if any one else found this, Dana, JS.

Anyway, I have the first controller now, plus I built one with 3 drivers on one board , They each work fine seperately, but for the life of me, I can not figure out how to get three to get 3 drivers to switch at once. tried 18v controll pulse. 15v, 5ampn psu across drivers, but no, JS, I hope you or someone can shed light. It's simple but I am blind. I want to switch 4 sets of three drivers, one for each gate. Or is this arduino's job, still learning that one.

I will post some pics or a vid tonight or tomorrow, of controller, scope shots.
JS, I used metal film resistors, tried 3 and 13 ohms, almost no noise on gate, havn'e loaded yet, If you remember the last scope shots I sent you, this is a drastic improvment.

Regards everyone

[JohnStone]

Hey UFO and everyone,

Haven't been able to do anything but catch up with work, I haven't been this busy for a long time, but I am still "chipping away at the stone".
I got my new parts, and rebuilt controller to JS's specs. I tell ya, that opto was one pain in the ass. The output pins are reversed on mine, pin 3 is across from 1, I don't know if any one else found this, Dana, JS.

Anyway, I have the first controller now, plus I built one with 3 drivers on one board , They each work fine seperately, but for the life of me, I can not figure out how to get three to get 3 drivers to switch at once. tried 18v controll pulse. 15v, 5ampn psu across drivers, but no, JS, I hope you or someone can shed light. It's simple but I am blind. I want to switch 4 sets of three drivers, one for each gate. Or is this arduino's job, still learning that one.

I will post some pics or a vid tonight or tomorrow, of controller, scope shots.
JS, I used metal film resistors, tried 3 and 13 ohms, almost no noise on gate, havn'e loaded yet, If you remember the last scope shots I sent you, this is a drastic improvment.

Regards everyone

Hi Machine,
1. I am so sorry about your problems with SFH617. I checked all sources and found one minor error (being posted already).

The letters C and E at pins 3/4 are changed in circuit diagram (sorry) but pinning is consistent over all sources. No idea on how pin 1/3 can be changed - or did understand wrong what you expressed?

Additionally: Some data sheets name pin2 as C -> refers to good English (cathode) but confusing with C=collector (pin4)

2. Multiple drivers: It depends! But before I utter some hints I'd like to know what oscillator output you have and waht voltage. With Arduino pin, LM339 or LM324 it will be a problem while 555 should be OK.

3. Please elaborate your switching plan! Why using 3 drivers in parallel? 300A steady current and 600A peak? You will not have 100% duty on your drivers! Do you own a Tesla EV and want to rework it?
Do you want to switch those 4 sets at once or consecutive? Who controls the sequence?
Please specify your desire -> I will post a circuit diagram for your application.
JohnS

[machinealive]

Hey JS,

Wow, I thought I would check again, I was sent 2 different optos, the sfh610a-3, and the sfh615a-3, and they are opposite. I was using the 610, while checking the 615 datasheet, that only took me 3 days to figure out .

Yea john I want to switch big amps but less per fet, less heat. So now I know I need low ohms in series with motor. When motor coils were all parallel, we get highest RPM, but fets fried, Coil in series saved fets, but RPM, went down. I feel i need to work on getting that sereies resistance down, as low as i can, but i need to stop that heat. A fan would not have saved the last fets.

I still have the small 2 fet driver working now, and I plan on seeing how fast we can take that setup, with and withoput a BBQ . The 3 driver setup per gate is gonna be for safety/ruggedness,later.
It's something I'm going to have to see, I just gotta.

The controller is an LM2903, like a 393, i'm using 18v input. An, IPAZ 7555, would drive them? Can the arduino drive 3 drivers at same time, using 3 different output pins? I have and want to use arduino, but am still learnig, I either start doing that or work with motor setup, using controller I have, which is really all I need right now.


I just got the drivers going using the 7555 as predriver, thanks John, I forgot about the 555. Can't wait to see how hot this gets.

[JohnStone]

....
Yea john I want to switch big amps but less per fet, less heat. So now I know I need low ohms in series with motor. When motor coils were all parallel, we get highest RPM, but fets fried, Coil in series saved fets, but RPM, went down. I feel i need to work on getting that sereies resistance down, as low as i can, but i need to stop that heat. A fan would not have saved the last fets.

I still have the small 2 fet driver working now, and I plan on seeing how fast we can take that setup, with and withoput a BBQ . The 3 driver setup per gate is gonna be for safety/ruggedness,later.
It's something I'm going to have to see, I just gotta.

The controller is an LM2903, like a 393, i'm using 18v input. An, IPAZ 7555, would drive them? Can the arduino drive 3 drivers at same time, using 3 different output pins? I have and want to use arduino, but am still learnig, I either start doing that or work with motor setup, using controller I have, which is really all I need right now.

Machine,
here you have a circuit for 3 drivers:

Very probably it will do fine but data shee guarantees 6mA only .

I am sure you will not kill those Infineon FETs but others (300V + higher RDSon)

Ardunino can drive 40mA per pin but not all at same time. It is no good idea to load those pins seriously because it can disturb anaolg functons. Prepare to use a 555 as driver or a transistor.

You may try to drive 4 FETs at driver 5.1. We used a very strong driver. Add another 100nF at driver pin 5/8. Measure how fast FETs are if dirving 4 individuals. Post switching time for documentation - please.

Data from IPW60R041 (worst case @125 °C and high amps):
@100A one pair will have 4V drain/GND -> 400W loss at 100% duty cycle. With 4 FETs you suffer 200W loss once again at 100%.......
So do not operate with tiny heat sinks!
Now imagine conditions if FETs RDSon factor 10!!!!

You did not answer if you intend to drive all 4 sets with same signal or do you intend to use a sequence.

You can use Nico's generator with manual control by just adding 4 switches and some resistors to Arduino. Works fine on my bench. Soon I will have it with PC control. Waiting for Nico to repair program damage caused by my redneck programmimg skills.
BTW: Nico's generator operates one single pin only with PWM. This is fixed property by hardware of Arduino.
I hope those facts help!
JohnS

[JohnStone]

Nico GEN with PC control

Hi friends,
now I managed to modify Nico's GEN program in order to get it controlled by PC. Please accept this is a sample only and is still faulty in some extent.



Download

• Download, start it - it will unpack automatically.
• The program file for Arduino xxx.ino shall be uploaded into Arduino nano, uno ... via Arduino programming environment.
• Set the virtual com port to COM9 as the PC programm uses it as default. Port number can be changed but I did not try it yet.
• PWM-Out at Ardunino is pin 9
• No furhter HW at Arduino required but those 4 buttons provided by Nico are supported as before.
• The folder contains PC-Program. The file with red "P" icon is the corresponding Application program.

Though faulty this program can be used for basic control of the generator. The reason I post it is to show a sample on what we can do on PC level. It might be necesary later on in order to perform sophisticated control.

Note: This is just a sample for those knowing Arduino. Please do not suspect to get a basic introduction into Arduinno and programming environment.

I suggest to not use your precious PC if you operate severe and noisy (electrically) motor equipment. Small setups will not disturb your PC. It is recommended to have an old lap top PC connected to it's genuine PSU while having no other connection to any GND lead except via USB cable to your setup.
If you connect ethernet make sure you have no cable screen connected to your lap top PC. Prefer to use WLAN!

BTW: No, I am no PC programmer at all. I used a graphic programming environment (just learn it!) dedicated to universal control applications. ProfiLab The program above is being composed like below.

You see above about 50% of the program above. Top circuit -> duty control, center -> serial receive and ascii processing, bottom -> button processing including serial transmission. The serial monitor is being hidden in a macro. Dashboard controls and gauges are inherent part of the components applied.
The program supports writing data to CVS (i.e. for MS Excel or OpenOffice) files in order to valuate data taken later on (i.e. torque and power, COP).

JS

[machinealive]

hi johnstone,

Sorry I didn't post sooner but everyone at my house has the flu but me. John, first I will be using the 3 drivers to pulse motor with same signal. Once I get that under control, I hope to pulse each gate in sequence. Each gate with it's own driver cluster, I am using 3 cause thats all that fit on board.

I had the set up running tonight with the two fet driver. I was using 24v, with the extra coil in series, with the motor, worked ok but fets warmed up quick. I am using cheaper fets, until I get control, 600v 50amp continous, but larger Rson.

I have to build a heavy duty rheostat, or something. Any suggestions anyone, I would love to hear.

I'm gonna send you some scope shots, js.

[JohnStone]

hi johnstone,

Sorry I didn't post sooner but everyone at my house has the flu but me. John, first I will be using the 3 drivers to pulse motor with same signal. Once I get that under control, I hope to pulse each gate in sequence. Each gate with it's own driver cluster, I am using 3 cause thats all that fit on board.

I had the set up running tonight with the two fet driver. I was using 24v, with the extra coil in series, with the motor, worked ok but fets warmed up quick. I am using cheaper fets, until I get control, 600v 50amp continous, but larger Rson.

I have to build a heavy duty rheostat, or something. Any suggestions anyone, I would love to hear.

I'm gonna send you some scope shots, js.

Hi Machine,
you need to consider reality:

- We need FETs rated for high voltage in order to stand those voltage spikes. We are not alowed to kill them.
- At standard FET technology HV property comes along high RDSon up to 0.5...1Ohm. (worst case derating factor 2 to be considered) Therefore I choose Infineon coolmos. Their technology offers 0.041 Ohm at high voltage (temp and current derating => 0.08Ohm per FET). Thus you get @ 100A 100% duty a dissipation of about 200...400W with coolmos instead of 1 KW - that is the good news. The bad one is you need to drain up to 400W anyway. For future units we will use superiour SiC technology from CREE. (57$ for one piece just now - but they will come down soon!)
- Please prepare to attatch two good heat sinks with fan to those two FETs and mount heatsinks stableon to PCB. You'll find heat sinks in scrap PC PSUs. Maybe you can reuse case, fan, bunch of cables and heatsinks and mount your driver into the case.

- Please do not go for rheostat! If inevitable use bunch of H4 bulbs. Calculate them to not light at current expected. They will light if overcurrent.
Preferrable is to first reduce voltage to 12V, get FET driving stable and then go for 24V. Repeat procedure. for 36, 48 V ........
JS

[Donald Haas]

MS and LES

Hi All,

I follow at a distance because I'm a slow learner on purpose.

As a learning exersize I made up the Mad Scientist LM393 circuit and added Les' complimentary transistor pair to it and etched it.

It's running and the scope shots look like it would be OK to feed the monster.

"Feed the monster and loose the Lady"

Sounds like the hook for a hit tune.

bro d

[JohnStone]

Hi All,

I follow at a distance because I'm a slow learner on purpose.

As a learning exersize I made up the Mad Scientist LM393 circuit and added Les' complimentary transistor pair to it and etched it.

It's running and the scope shots look like it would be OK to feed the monster.

"Feed the monster and loose the Lady"

Sounds like the hook for a hit tune.

bro d

HI,
that is good news! Proceed and share your success or problems!
JS