Use for the Tesla Switch

[Matthew Jones]

AHHH!!! Relief we can take the "Have You Seen" posters down now.

LOL

Matt

[nvisser]

Come on Leroy. You can do better than that. After you went missing for a few months you must have something to tell us about the TS.
Welcome back

[ldissing]

...

Now that we have MosFet switches which are very
close to the 'ideal' it is possible to construct pulse
width modulated 'boost' circuits which are extremely
efficient.

With modern switching circuits it is possible to duplicate
the 'capacitor switching charge pump' function by means
of inductive charge/discharge in a very small package
with better than 90% efficiency.

The basic 'desulfator' circuit is a simple example.

With tiny 'point of load' boosters to create the voltage
needed for driving the MosFet switches, the switching
circuit could be driven from a solar panel of about 2 Volts
at considerable amperage. Or any convenient voltage
and current.

The output could be a stable DC of virtually any voltage
within reason. Or pulses if desired.

For charging a battery bank of lead-acid batteries, pulses
are the most effective.

SeaMonkey,

I have read your posts but have seen no actual circuits. You come to this group with some good references, so I am not bashing you. If you would be so kind as to post a circuit that incorporates your ideas, experience, etc. it would be easy to validate/verify your comments.

Leroy

[SeaMonkey]

SeaMonkey,
Now i see You understand my point of view.


That said, how about considering a most efficent TS build:
A. Power source: Photovolatic or mimic it with a current limiting PSU
B. Receiving load: A battery (I have NiCd 56Ah, and PbS to spare)
C. Mode: parallel receive charge, serial deliver to load.

Is this in line?


1. I would suggest a "crippled" variant of my design, where we "shunt short" all high side switches (effectively leaving them out of the design) so we deal with low side only (most simple config).
2. We can't avoid floating middle switch, this could impose a bit of a problem?
3. We need investigate how we supply the gate of the middle floating one with power the best way?
4. I have a $7 AVR that's able to be proggramed with C code (ATmega48 on a "07301" form JYETech) that has 3x8 I/O pins and runs off 5V@12Mhz, has "razor sharp" leading edges (CMOS Shmitt) and could drive few IRS21851 (go on, look it up here) that would drive few IRF3205 or IRFZ48 (doubled or quad) to "kick" hard? I already "Qued" some LEDS with it

So either we go up in the sky or down to earth in flames!
B)

Stevan C.

The "Charge Pump" circuits (TS) are an old technology
which became popular long before the modern
'switching' technology became a reality.

Now that we have MosFet switches which are very
close to the 'ideal' it is possible to construct pulse
width modulated 'boost' circuits which are extremely
efficient.

With modern switching circuits it is possible to duplicate
the 'capacitor switching charge pump' function by means
of inductive charge/discharge in a very small package
with better than 90% efficiency.

The basic 'desulfator' circuit is a simple example.

With tiny 'point of load' boosters to create the voltage
needed for driving the MosFet switches, the switching
circuit could be driven from a solar panel of about 2 Volts
at considerable amperage. Or any convenient voltage
and current.

The output could be a stable DC of virtually any voltage
within reason. Or pulses if desired.

For charging a battery bank of lead-acid batteries, pulses
are the most effective.

[Matthew Jones]

About 40 years ago I cut the top off a to-3 style transistor and it worked as a solar cell. I put it in clear epoxy and still have it and just ran across it a few days ago.
Since I only recently have been reading this thread and probably won't be able to get through all 99 pages may I ask if anyone has tried a mechanical setup like Patrick Kelly shows in his energy book which Electrodyne corp. used? The schematic for this was recently changed. I'm trying to think of ideas for how one could build the mechanical switch as I believe there may be an unseen effect that results from building it that way versus using solid state switching.

Ya I have built 6 mechanicals. about 7 mechanical / solid state. Another 8 solid state. I have tried just about every configuration of them that is outand I have come up with a few to top it off.
Mechanical has got some advantages. Mostly they keep the potential up. But you can do it with mechanical relays just as well and with a lot less work.
Look in my signature for the ISCC it worked pretty well.
The biggest problem is finding something that will hold up over a period of time. You start pulling large load, and thats what you need to do to get a positive effect, the start to run rough and eventually fail from the points.

Right now I am collecting parts to build 1 relay that has all the mechanical contacts needed to switch. But tungsten is not cheap.And I would like the contacts to actually be able to hold up over a period of time with a large load.

If you post your idea's I'll let you know what I have tried or if I have any advice.

Matt

[ewizard]

Hi SeaMonkey,
Do You know about the way to turn a To-3 packed BJT into a tiny solar cell?
Is this possible for a MOSFET too?

Wonder why I ask?

Stevan C.

About 40 years ago I cut the top off a to-3 style transistor and it worked as a solar cell. I put it in clear epoxy and still have it and just ran across it a few days ago.

Since I only recently have been reading this thread and probably won't be able to get through all 99 pages may I ask if anyone has tried a mechanical setup like Patrick Kelly shows in his energy book which Electrodyne corp. used? The schematic for this was recently changed. I'm trying to think of ideas for how one could build the mechanical switch as I believe there may be an unseen effect that results from building it that way versus using solid state switching.

[StevanC]

SeaMonkey,
Now i see You understand my point of view.


That said, how about considering a most efficent TS build:
A. Power source: Photovolatic or mimic it with a current limiting PSU
B. Receiving load: A battery (I have NiCd 56Ah, and PbS to spare)
C. Mode: parallel receive charge, serial deliver to load.

Is this in line?


1. I would suggest a "crippled" variant of my design, where we "shunt short" all high side switches (effectively leaving them out of the design) so we deal with low side only (most simple config).
2. We can't avoid floating middle switch, this could impose a bit of a problem?
3. We need investigate how we supply the gate of the middle floating one with power the best way?
4. I have a $7 AVR that's able to be proggramed with C code (ATmega48 on a "07301" form JYETech) that has 3x8 I/O pins and runs off 5V@12Mhz, has "razor sharp" leading edges (CMOS Shmidt) and could drive few IRS21851 (go on, look it up here) that would drive few IRF3205 or IRFZ48 (doubled or quad) to "kick" hard? I already "Qued" some LEDS with it

So either we go up in the sky or down to earth in flames!
B)

Stevan C.

[StevanC]

Vissie,
My device still runs, and shows no fatigue yet, how about MOSFET ones?

Yes I know TS canbe made to run on MOSFET, I plan to make one too, but I just fail to see real advantages for doing so. Rest assured as soon as I investigate I will come back with a nice beffy report

^cold virtuall beverage for all on the thread

Stevan C.

[SeaMonkey]

SeaMonkey,
I hope I didnt offend You? I only wanted to point it out, that we had a fair share of do it with "MOSFET" newcomers that later only found out that the circuit for some strange reasons just fries them, while achieving incredible performance with BJT.
Despite BJT being regarded as devices of past, we see here, in action, they outperform whats touted as the leading edge of the latest and best we can do.

Then there are quite few assumptions we all accept "per se" out of the textbooks that are just plain wrong:
A. The complexity of IC's serves their reliability (<- wrong - no complexity serves reliability - maybe SEVERS their reliability. They are complex because they where faulty and are patched to work- that's not reliability and only high revisions, distilled off their patchy complexity are made reliable)
B. The small cost of the R&D divided per unit makes the sheer R&D price affordable (<- this applies only to mammoth series, how about initial R&D with dozens of devices? I find this faulty too. There are now some methods of cost effective initial R&D, but we all pay for it and did pay a lot for it in the past (and in vain too))
C. The MOSFET are key superior to BJT <- how come then BJT are still produced and sold? how come we didn't switch to FET altogether? On Space faring? Go figure...
D. We still are honestly clueless what makes BJT/FET trully "CUE" - we have teories and models only - but no clue! no law is there yet AFAIK
E. We still can't tell amperes from potential (NOT VOLTAGE!) and are basically clueless to electricity as well, to be able to understand under point "D."

So I truly welcome Your honest contribution to this thread, while I wished to spare You some of the already chewed up stuff we had in the past (not necessarily on this thread).

Okay?


Stevan C.

No Stevan, no offense has been taken. You raise several
very good points.

The Bipolar Transistor is inexpensive and suitable for many
applications. When one learns how to 'squeeze' maximum
performance from them they can still be an excellent choice.

For high current DC switching applications the MosFet is
capable of performance and efficiency far beyond what
the bipolar transistor is capable of due to its very low
'ON' resistance and its ability to switch very rapidly.

For high current AC switching applications one must
not forget that the 'body diode' is an integral part of
the MosFet and therefore, it is often necessary to
use two MosFets to avoid 'problems.'

When a MosFet 'fries' it is generally due to either an
avalanche, or excessive body diode current at polarity
reversal on the Drain. When using a MosFet certain
precautions must be incorporated into the circuit design
to prevent 'overstressing' it - once that is done it performs
very reliably while dissipating little heat.

Whether a bipolar transistor or a MosFet is best for any
given application is often a matter of personal preference.
Some choose a device that they're more familiar with or
more comfortable with. Since the MosFet is 'relatively'
new on the scene, many experimenters have yet to
learn how to use them or to experience first hand their
superior characteristics.

Why is the bipolar transistor still manufactured and
utilized? Because it is still a good choice for many
applications and functions - particularly 'linear' apps.
Even for certain switching applications it is capable
of adequate performance, particularly at the higher
voltages.

But, for power handling and switching requirements
which demand the utmost efficiency and flexibility,
the MosFet is becoming the device of choice.

The MosFet has one unique property - once it is
'gated on' it will conduct current flow in either
direction equally well (effectively shunting the
body diode).


As with all things in life - knowledge comes to the
rescue.

Regarding 'chips.' Chips are good!

I'm with you on the 'complexity' issue. Complex is
not necessarily the best way to go - unless there
are absolutely no 'simpler' alternatives. I really like
'simple.'

[Matthew Jones]

Vissie's right, but we have not heard anything more from Dave since, On the actual device so it might not have ran very long or he put it in the closet.



Matt

[nvisser]

Hi Stevan
Please refer to post 314 on this thread.
It is about a Tesla switch that runs on mosfets
Vissie

[StevanC]

SeaMonkey,
I hope I didnt offend You? I only wanted to point it out, that we had a fair share of do it with "MOSFET" newcommers that later only found out that the circuit for some strange reasons just fries them, while acheaving incredible performance with BJT.
Despite BJT being regarded as devices of past, we see here, in action, they outperform whats touted as the bleading edge of the latest and best we can do.

Then there are quite few assumptions we all accept "per se" out of the textbooks that are just plain wrong:
A. The complexity of IC's serves their reliability (<- wrong - no complexity serves reliability - maybe SEVERES thyer reliaility. They are complex because they where faulty and are patched to work- that's not reliability and only high revisions, distilled off theyr patchy complexity are made reliable)
B. The small cost of the R&D divided per unit makes the sheer R&D price affordable (<- this applies only to mamoth series, how about initial R&D with dozens of devices? I find this faulty too. There are now some metods of cost effective initial R&D, but we all pay for it and did pay a lot for it in the past (and in vain too))
C. The MOSFET are key superior to BJT <- how come then BJT are still produced and sold? how come we didn't switch to FET alltogether? On Space faring? Go figure...
D. We still are honestly clueless what makes BJT/FET trully "CUE" - we have teories and models only - but no clue! no law is there yet AFAIK
E. We still can't tell amperes from potential (NOT VOLTAGE!) and are basically clueless to electricity as well, to be able to understand under point "D."

So I truly welcome Your honest contribution to this thread, while I wished to spare You some of the already cheved up stuff we had in the past (not necesserly on this thread).

Okay?


Stevan C.

[StevanC]

a long, long trip before ... ;-)

Stevan C.,

Yes, there have been recent advances in the art of
bipolar transistor manufacture (with the advent of the
Low Vce or BISS switching transistors) which has revived
their popularity.

@others:
BISS=Breaktrough In Small Signal : Vcesat~0.25V, Ice~2A, high gain (hFE) etc.

I would disagree here, I meant properly driving any transistor, let alone a especially good one: If properly designed, a BJT can offer competing performance if not even surpassing a MOSFET, reliability is one for example:
1. If a design is "cost efficient (=cheap) the MOSFETS just "fall like flies" never mind they are touted "more reliable",
2. If a design takes "all measures" in account: we deal with a device comparable to a BJT driven one - both are remarkably reliable, but the later significantly less complex?

The transistor you mention is designed for low distortion
linear applications but its construct (perforated emitter)
is the new design. It could function well in switching
applications when properly driven.

It does (I did my 101 of MJL21194, and so some readers of this thread)

Yes, it is possible to make the SSG circuit with MosFet
devices. Yes, it will have better efficiency than most
bipolar transistor versions.

You have no idea how eagerly we wait someone come a prove that claim?

By use of the new Low Vce transistors and 'signal
conditioning' chips it is possible to improve the efficiency too.

Also, using Schottky Diodes will increase efficiency.

I just wouldn't bet on that (but could live with it if proven)...

Your comments regarding the IR characteristics of
silicon semiconductors is interesting. They are good
IR emitters and are, conversely, very sensitive to it.

I was surprised on the benchmark...
We are ever more ripped off of devices You can really experiment and find something useful...
I wonder how a BJT would react if lit upon by an other BJT that is switching?

The point you make about 'near saturation' operation
is a good one - it is best to go into 'deep' saturation
in order to minimize conduction losses which cause
heat. The trade-off is that 'turn-off' from deep
saturation is slowed due to 'charge storage' in the base
region. Provisions must be made to eliminate charge
storage at turn-off in order to speed up the transition.

Fortunately, MosFets do not have that problem and
are able to switch with better efficiency.

Only thing I found MOSFETs are superior to BJT is their swift turn-off time - One can drain a FET in nanoseconds.
But then again, their drive energy can't be the part of the useful current being switched, which means they are bond to always have a loss - the switching costs current and one has to supply it extra to the power handled.

You are correct. The bipolar transistor is far from being 'dead.'
In certain applications it is still a very good choice - especially
when it is of the type that has very low saturation losses and
switching losses.

Or a designer can take the properties a BJT inherently has as his advantages.

Especially the GaAs transistors of the past have not seen their "last dance" yet

I always admire the inventors of the middle ages:
They had so rough tools, so little resources and so good solutions that endured and proved reliable despite spartan in design.
They are in the first place my models i try hard to be worthy of... the ancient windmills, watermills, sailing ships...

Best regards,
Stevan C.

[SeaMonkey]

Stevan C.,

Yes, there have been recent advances in the art of
bipolar transistor manufacture (with the advent of the
Low Vce or BISS switching transistors) which has revived
their popularity.

The transistor you mention is designed for low distortion
linear applications but its construct (perforated emitter)
is the new design. It could function well in switching
applications when properly driven.

Yes, it is possible to make the SSG circuit with MosFet
devices. Yes, it will have better efficiency than most
bipolar transistor versions.

By use of the new Low Vce transistors and 'signal
conditioning' chips it is possible to improve the efficiency too.

Also, using Schottky Diodes will increase efficiency.

Your comments regarding the IR characteristics of
silicon semiconductors is interesting. They are good
IR emitters and are, conversely, very sensitive to it.

The point you make about 'near saturation' operation
is a good one - it is best to go into 'deep' saturation
in order to minimize conduction losses which cause
heat. The trade-off is that 'turn-off' from deep
saturation is slowed due to 'charge storage' in the base
region. Provisions must be made to eliminate charge
storage at turn-off in order to speed up the transition.

Fortunately, MosFets do not have that problem and
are able to switch with better efficiency.

You are correct. The bipolar transistor is far from being 'dead.'
In certain applications it is still a very good choice - especially
when it is of the type that has very low saturation losses and
switching losses.

[StevanC]

WHY Stevan asks

Yes, tell us why you ask...

1. The BJT is especially susceptible to "far IR" or in other terms high energy heat.

2. A sharply triggered BJT, produces some heat

3. A BJT conducting current near it's VceSAT region produces heat very quickly (and a lot)

4. The BJT is more susceptible (=produces more output) to "far IR" the more cool it is.

5. Sharply triggered BJT if the duty cycle is low enough hardly warms the heat sing - we consider it "cold"

***

A. But we know heat is generated

B. And we know there are switching losses

C. And we know a BJT can recycle it's switching current if the circuit is advantageously designed

D. And we know a BJT can convert some amount of heat ("far IR") to electricity. Just how much, seems to depend of the temperature gradient of the BJT layers.

***

@SeaMonkey:
Did You ever study the MJL21194 generation of BJT? ("post 2000" era of BJT?)

Could You devise a SSG type of circuit based on MOSFET switching devices, with same or better efficiency?

I just try to point some key properties of BJT that have made them recognized as potentially the preferred devices in the future, and that by key semiconductor manufacturers...

Stevan C.
P.S.
sorry for the delay