Searching \ for '[EE]:: A FET for the robot builders' in subject line. ()
Make payments with PayPal - it's fast, free and secure! Help us get a faster server
FAQ page: www.piclist.com/techref/index.htm?key=fet+robot+builders
Search entire site for: ': A FET for the robot builders'.

Exact match. Not showing close matches.
PICList Thread
'[EE]:: A FET for the robot builders'
2007\09\25@065038 by Russell McMahon

face
flavicon
face
Just mentioned in another thread

400A 20V FET
1 milliohm Rdson

       http://www.irf.com/product-info/datasheets/data/irf1324s-7ppbf.pdf

Sensibly priced considering (429A, 20V rated)
$US2.41/1000s Digikey.

If you ever need a truly low Rdson low voltage FET switch this may be
useful.
While true Rdson will be higher than spec sheet nominal with junction
heating, it appears it will only increase by about 30% for sensible
junction temperatures.
Should give 10 to 15 Watt losses and under 0.15 V drop at 100 Amps.

One could almost run a starter motor with this FET [ :-) !!!] as long
as you properly deal with the inductive transients OK. Avalanche
rated, but ... .

What DO the mega-battle-bot people use for motor control?



           Russell


2007\09\25@071911 by Hector Martin

flavicon
face
Not exactly a BattleBot, but FIRST robotics robots use the standard
robot controllers (that we are are required to use), which consist of
two H-bridges with three MOSFETs on each leg (12 total). I forget the
part number, but they probably were some bog-standard model.

We weren't very lucky with ours...

http://marcansoft.com/uploads/frc_victor_fire.avi


--
Hector Martin (spam_OUThectorTakeThisOuTspammarcansoft.com)
Public Key: http://www.marcansoft.com/marcan.asc

2007\09\25@085806 by Steve Howes

picon face
> 400A 20V FET

Is that not a man in a tin can with a big switch?

2007\09\25@102437 by Russell McMahon

face
flavicon
face
>> 400A 20V FET

> Is that not a man in a tin can with a big switch?

Pay no attention to the man behind the curtain ... er with the big
switch


2007\09\25@121158 by William \Chops\ Westfield

face picon face

On Sep 25, 2007, at 3:50 AM, Russell McMahon wrote:

> What DO the mega-battle-bot people use for motor control?

Significant numbers of parallel lesser FETs, I think.  Even
the non-mega RC folk will use parallel FETs in their speed
controllers...

BillW

2007\09\25@152242 by Sean Breheny

face picon face
BTW, Thanks Russell for the info.

One very interesting thing to note about very low RdsON FETs is that
they become more difficult to parallel. When a FET is fully on, the
resistance rises with temperature, so that they tend to self-balance.
However, during the time when a FET is half-on, the hotter FETs will
take more of the current.

So, imagine two FETs in parallel, FET A hotter than FET B. If you are
putting 100A through them together, then as you try to turn them on,
FET A will start ramping up earlier, perhaps to almost 100A, whereas
FET B will turn on later in the switching interval, balancing out the
current and drawing more than FET A does.

With very low RdsON, the on/off switching transients, if they are
frequent enough, may actually dissipate more than the I2R losses
during the full-on time.

This means that the paralleled FETs will no longer self-balance, since
the hotter one will have much higher dissipation during the transient,
therefore overall more dissipation than the cooler one. You can still
have thermal runaway under these conditions.

Sean




On 9/25/07, William Chops Westfield <.....westfwKILLspamspam@spam@mac.com> wrote:
>
> On Sep 25, 2007, at 3:50 AM, Russell McMahon wrote:
>
> > What DO the mega-battle-bot people use for motor control?
>
> Significant numbers of parallel lesser FETs, I think.  Even
> the non-mega RC folk will use parallel FETs in their speed
> controllers...
>
> BillW
> -

2007\09\25@191327 by Jake Anderson

flavicon
face
I'm in the process of building a speed controller for our robot. At the
moment pretty much everybody uses IRF1405's in the to-220 package.
The biggest problem with that fet by the look of things is its voltage
rating. Everybody wants more volts. With the A123 cells people are
starting to run 36 volts and nobody really runs less than 24v.
Putting them in series could make for "interesting" failure modes and a
not particularly robust design i thinks.

Russell McMahon wrote:
{Quote hidden}

2007\09\26@011742 by Paul Anderson

face picon face
On 9/25/07, Jake Anderson <jakespamKILLspamvapourforge.com> wrote:
> Everybody wants more volts. With the A123 cells people are
> starting to run 36 volts and nobody really runs less than 24v.
> Putting them in series could make for "interesting" failure modes and a
> not particularly robust design i thinks.
>
I'm not really that up on motors, but it seems to me, per ohm's law,
if you have a lower voltage and greater current carrying capacity, you
should be able to even it out.  Barring other losses, the likes of
which I don't recall at the moment.

--
Paul Anderson
VE6HOP
.....wackyvorlonKILLspamspam.....gmail.com
http://www.oldschoolhacker.com
"May the electromotive force be with you."

2007\09\26@071936 by Russell McMahon

face
flavicon
face
>> Everybody wants more volts. With the A123 cells people are
>> starting to run 36 volts and nobody really runs less than 24v.
>> Putting them in series could make for "interesting" failure modes
>> and a
>> not particularly robust design i thinks.

> I'm not really that up on motors, but it seems to me, per ohm's law,
> if you have a lower voltage and greater current carrying capacity,
> you
> should be able to even it out.  Barring other losses, the likes of
> which I don't recall at the moment.

For the motor as a whole power scales (to an adequate approximation)
linearly with voltage and current - double the voltage and halve the
current and you have the same power, as per Ohm's law, as you say.

BUT resistive loss in a given conductor is proportional to the square
of the current. Halve the current and you get quarter of the previous
loss. This means that if you use the same wiring and double the motor
voltage you reduce your wiring losses by a factor of 4! OR you can use
wiring 1/4 the area and get the same losses.

When you are using tens of amps wiring losses, connector resistance
losses and even switching device lead losses can start to become
significant. Consider a 12V, 50A system. Nominal power is 12 x 50  =
600 Watts. Effective load resistance is V/I = 12/50 = 0.24 ohm.
Obviously the odd tenth of an ohm in the wiring is unnacceptable. Even
0.01 ohm of wiring will dissipate I^2R = 2500 x 0.01 = 25 Watts or
about 5% of the system power. If you double the voltage and halve the
current then the same wiring will give losses of about 6 Watts or
about 1% of the system power. It would be nice to not have to reduce
total wiring losses to 0.01 ohm, so lets try a 48V system and ... :-).


       Russell
.

2007\09\26@075934 by Gerhard Fiedler

picon face
Russell McMahon wrote:

> When you are using tens of amps wiring losses, connector resistance
> losses and even switching device lead losses can start to become
> significant. Consider a 12V, 50A system. Nominal power is 12 x 50  = 600
> Watts. Effective load resistance is V/I = 12/50 = 0.24 ohm. Obviously
> the odd tenth of an ohm in the wiring is unnacceptable. Even 0.01 ohm of
> wiring will dissipate I^2R = 2500 x 0.01 = 25 Watts or about 5% of the
> system power. If you double the voltage and halve the current then the
> same wiring will give losses of about 6 Watts or about 1% of the system
> power. It would be nice to not have to reduce total wiring losses to
> 0.01 ohm, so lets try a 48V system and ... :-).

Is there any chance that a step-up switcher that transforms 12V/50A into
24V/25A or a higher voltage costs less than 10 USD? Efficiency wouldn't be
terribly important.

Gerhard

2007\09\26@081440 by Jake Anderson

flavicon
face
Russell McMahon wrote:
{Quote hidden}

48v has been done. I believe if you go over 50V nominal (ie not a float
charge voltage) you need to get approval from your event organiser.
I wonder how feasable it would be to make a series connection using
those FETs.
Hmmm or if there is some way of making your speed controller such that
you connect it up to say 16 volts worth of battery, then series your
speed controllers.
Smoking a section of ESC then just means reduced speed, thats not such a
bad idea perhaps.

The downside to those fets is probably the thermal conductivity, yes it
can pump out 25W of heat but you need some way of getting that heat out
whilst keeping the die just this side of plasma.

2007\09\26@082340 by Martin Klingensmith

face
flavicon
face
Generally speaking they use something similarly rated to the IRF1405,
paralleled. There was another part that some people were using but I
lost interest in OSMC when they got upset that I foiled some of their
OSMC-selling propaganda (but that's OT).
--
Martin K

William "Chops" Westfield wrote:
> On Sep 25, 2007, at 3:50 AM, Russell McMahon wrote:
>
>> What DO the mega-battle-bot people use for motor control?
>
> Significant numbers of parallel lesser FETs, I think.  Even
> the non-mega RC folk will use parallel FETs in their speed
> controllers...
>
> BillW

2007\09\26@092939 by Russell McMahon

face
flavicon
face
> Is there any chance that a step-up switcher that transforms 12V/50A
> into
> 24V/25A or a higher voltage costs less than 10 USD? Efficiency
> wouldn't be
> terribly important.

That may be pushing the price a bit, but it's close to doable.
The smps circuit is "trivial".
But you need a suitable FET, a suitable Schottky and a suitable coil.

If this were a one off or a vast volume product the coil could be
custom fabricated. Air core would do at the cost of radiation issues.
Digikey or your favourite supplier's catalog will give FET and
Schottky pricing.
I'd guess an all up parts cost of $US20 to $US30 would be more likely
but approaching $10 may be possible with special selection care.

I'm doing the above at 100 Watts plus at 80-90% efficient. Should
scale reasonably well.


       Russell






2007\09\26@092939 by Russell McMahon

face
flavicon
face
> 48v has been done. I believe if you go over 50V nominal (ie not a
> float
> charge voltage) you need to get approval from your event organiser.

ELV (Extra low voltage) is usually 50 VDC or 32 VAC in most
administrations. This is the voltage below which it's allegedly
reasonably safe to work with and there are relaxations on various
safety standards issues and probably also certification of who is
allowed to build things that are publicly accessible.

In practice 50 VDC can be lethal if you allow it to be.

12 VDC also under extreme conditions. (Hint: Placing car battery
connected terminals across your heart using saline wetted pads may
fail the Darwin test)(It was tried on a prisoner volunteer. They
failed to revive him!!! :-(  ).

.


               Russell


2007\09\27@052900 by Howard Winter

face
flavicon
picon face
Russell,

On Thu, 27 Sep 2007 01:30:03 +1200, Russell McMahon wrote:

> > 48v has been done. I believe if you go over 50V nominal (ie not a
> > float
> > charge voltage) you need to get approval from your event organiser.
>
> ELV (Extra low voltage) is usually 50 VDC or 32 VAC in most
> administrations.

That's certainly the case in the UK - I don't know if they chose that because it's the level across a phone line (so you can use telephone cable rather than anything
more protective/expensive) or if the telephone people went to the max they were allowed!

What confuses a lot of newbies is that 240/415V is called "Low Voltage" by people involved in the supply.

> 12 VDC also under extreme conditions. (Hint: Placing car battery
> connected terminals across your heart using saline wetted pads may
> fail the Darwin test)(It was tried on a prisoner volunteer. They
> failed to revive him!!! :-(  ).

Ugh!  Why on Earth would anyone volunteer for this?

Cheers,


Howard Winter
St.Albans, England


2007\09\27@055450 by Jinx

face picon face

> > fail the Darwin test)(It was tried on a prisoner volunteer. They
> > failed to revive him!!! :-(  ).
>
> Ugh!  Why on Earth would anyone volunteer for this?

Well, he did get out sooner.......

Doesn't sound half so risky or scary as drug trials (as long as it's
not me of course)

2007\09\27@055519 by Alan B. Pearce

face picon face
>>(It was tried on a prisoner volunteer. They
>> failed to revive him!!! :-(  ).
>
>Ugh!  Why on Earth would anyone volunteer for this?

He was probably "volunteered" ...

2007\09\27@084443 by Russell McMahon

face
flavicon
face
>>>(It was tried on a prisoner volunteer. They
>>> failed to revive him!!! :-(  ).
>>
>>Ugh!  Why on Earth would anyone volunteer for this?

Prisoners who have long terms have been known to be offered
significant inducements to participate in such trials. It was almost
certainly not thought that death would occur but rather probably that
they MAY cause fibrillation which would need resuscitation.

The point is, few would consider that there was ANY prospect that 12V
could be fatal under any circumstances. It's useful to be aware that
under extremely unusual circumstances it can be. This may make one
just a bit warier of eg 20 to 50 VDC supplies which are usually
treated with impunity.

50 V in a telephone exchange will cause annoying minor shocks in ones
hands if you brush against terminal pairs carrying it. 75VAC ? ringing
has quite a kick. Inductive kickback from a homing step by step
selector (those were the days) is extremely painful.



           Russell

2007\09\27@093407 by Peter Todd

picon face
-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

On Thu, Sep 27, 2007 at 01:30:03AM +1200, Russell McMahon wrote:
{Quote hidden}

Or the guy who was learning about the dangers of electricity in his Navy
electronics class. Namely how the relatively high resistance of your
skin prevents current flow, but that your body has a fairly low internal
resistance. He thought this was interesting, and decided to measure his
own internal resistance by grabbing an old style 9-volt powered analog
multimeter and jabbing both leads into his thumbs. Unfortunately, he
found out that his internal resistance was rather low, terminally so.


Hmm... Peircings and electrocution? Especially fresh ones...

- --
http://petertodd.org
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.6 (GNU/Linux)

iD8DBQFG+4Uz3bMhDbI9xWQRAvFeAKCjeX8kjU8Ve4SoUg6k6WxzD8d1ZQCfSQPN
EI4TO6OV4Qb4sZoMc9DKlwg=
=Bwla
-----END PGP SIGNATURE-----

2007\09\27@093714 by Gerhard Fiedler

picon face
Russell McMahon wrote:

>> Is there any chance that a step-up switcher that transforms 12V/50A into
>> 24V/25A or a higher voltage costs less than 10 USD? Efficiency wouldn't
>> be terribly important.
>
> That may be pushing the price a bit, but it's close to doable. The smps
> circuit is "trivial". But you need a suitable FET, a suitable Schottky
> and a suitable coil.
>
> If this were a one off or a vast volume product [...]

<g> It would be in that "sweet" spot in between...

> [...] the coil could be custom fabricated. Air core would do at the cost
> of radiation issues.

I need to look into what I can do with a custom made air core.

> I'd guess an all up parts cost of $US20 to $US30 would be more likely

That's what I also figured.

Thanks,
Gerhard

2007\09\27@100812 by Russell McMahon

face
flavicon
face
MC34063 + FET + Schottky + Coil + very few passives + something to
mount it on + not much else.

Although, at the 500 Watt power level, what are you wanting to do that
won't justify > $US10?



           Russell

{Quote hidden}

2007\09\27@160253 by David VanHorn

picon face
>
> Or the guy who was learning about the dangers of electricity in his Navy
> electronics class. Namely how the relatively high resistance of your
> skin prevents current flow, but that your body has a fairly low internal
> resistance. He thought this was interesting, and decided to measure his
> own internal resistance by grabbing an old style 9-volt powered analog
> multimeter and jabbing both leads into his thumbs. Unfortunately, he
> found out that his internal resistance was rather low, terminally so.
>

Sounds like an urban myth to me.

2007\09\27@172546 by Paul Anderson

face picon face
On 9/27/07, David VanHorn <EraseMEmicrobrixspam_OUTspamTakeThisOuTgmail.com> wrote:
> >
>
> Sounds like an urban myth to me.
>
Same here, but it looks like snopes hasn't addressed this one yet.

--
Paul Anderson
VE6HOP
wackyvorlonspamspam_OUTgmail.com
http://www.oldschoolhacker.com
"May the electromotive force be with you."

2007\09\27@173244 by Gerhard Fiedler

picon face
Russell McMahon wrote:

> MC34063 + FET + Schottky + Coil + very few passives + something to
> mount it on + not much else.
>
> Although, at the 500 Watt power level, what are you wanting to do that
> won't justify > $US10?

A motor. The problem is the oomph per price; currently we're using a 36 Nm
peak motor (40 A peak, 100 rpm no-load max) that costs ~25 USD. We're
working mostly on the far end of the typical motor curve, which is not
good. But I haven't yet found another one with a higher gearing in the same
price range. So either expanding my options into the 24 V range could help
with finding a better suited motor, or boosting the 12 V supply and drive
it current-controlled to compensate for the voltage losses in the MOSFETs
and connections would also improve things.

Gerhard

2007\09\27@174017 by Gerhard Fiedler

picon face
Russell McMahon wrote:

> The point is, few would consider that there was ANY prospect that 12V
> could be fatal under any circumstances.

It's the current, not so much the voltage that is fatal, and where the
current flows. So if you take care that the low voltage causes all the
current it is able to cause, and make sure that current flows where maximum
damage can be expected, you can lower the dangerous voltage quite a bit,
apparently.

Gerhard

2007\09\27@193002 by Russell McMahon

face
flavicon
face
>> Or the guy who was learning about the dangers of electricity in his
>> Navy
>> electronics class. Namely how the relatively high resistance of
>> your
>> skin prevents current flow, but that your body has a fairly low
>> internal
>> resistance. He thought this was interesting, and decided to measure
>> his
>> own internal resistance by grabbing an old style 9-volt powered
>> analog
>> multimeter and jabbing both leads into his thumbs. Unfortunately,
>> he
>> found out that his internal resistance was rather low, terminally
>> so.

> Sounds like an urban myth to me.

It may well be. But I'm reasonably sure that my reference to a
prisoner who died from a 12V DC supply applied to the heart was
reputable. Despite the fact that that came from my memory of a
reference from an unknown paper many years ago and pre-internet, I
have little reason to think that it was made up. But, you need to
trust my long-term memory. [Trusting my short term memory can be less
advisable :-). Once it gets in it tends to stay in at a rather better
than average level. Remembering to notice any given thing is another
matter.]



       Russell




2007\09\27@213546 by Gerhard Fiedler

picon face
Russell McMahon wrote:

>>> Or the guy who was learning about the dangers of electricity in his
>>> Navy electronics class. Namely how the relatively high resistance of
>>> your skin prevents current flow, but that your body has a fairly low
>>> internal resistance. He thought this was interesting, and decided to
>>> measure his own internal resistance by grabbing an old style 9-volt
>>> powered analog multimeter and jabbing both leads into his thumbs.
>>> Unfortunately, he found out that his internal resistance was rather
>>> low, terminally so.
>
>> Sounds like an urban myth to me.
>
> It may well be. But I'm reasonably sure that my reference to a prisoner
> who died from a 12V DC supply applied to the heart was reputable.
> Despite the fact that that came from my memory of a reference from an
> unknown paper many years ago and pre-internet, I have little reason to
> think that it was made up. But, you need to trust my long-term memory.

No need really to trust your memory. E.g.
<http://www.allaboutcircuits.com/vol_1/chpt_3/4.html> provides some data
points. I remembered (long term memory :) that some 20 mA AC are considered
dangerous, and that page confirms that. We apparently can stand higher DC
values, but if the DC pulse comes in the right (or wrong) phase of the
heart cycle, it may approach the danger of AC.

Given the resistance numbers there, it doesn't sound unreasonable that a
12V shock with very good skin contact or even a 9V shock direct into the
blood stream may be deadly; possibly not always, but sometimes.

OTOH, someone jabbing leads into his thumbs, whether connected to a battery
or not, may deserve a Darwin award and all that comes with it :)

Gerhard

2007\09\27@234647 by Jake Anderson

flavicon
face
Gerhard Fiedler wrote:
{Quote hidden}

Is this a geared motor? I am slowly plodding towards making something
with similar specs as a brushless motor.

2007\09\28@074628 by Gerhard Fiedler

picon face
Jake Anderson wrote:

>>> MC34063 + FET + Schottky + Coil + very few passives + something to
>>> mount it on + not much else.
>>>
>>> Although, at the 500 Watt power level, what are you wanting to do that
>>> won't justify > $US10?
>>
>> A motor. The problem is the oomph per price; currently we're using a 36
>> Nm peak motor (40 A peak, 100 rpm no-load max) that costs ~25 USD.
>> We're working mostly on the far end of the typical motor curve, which
>> is not good. But I haven't yet found another one with a higher gearing
>> in the same price range. So either expanding my options into the 24 V
>> range could help with finding a better suited motor, or boosting the 12
>> V supply and drive it current-controlled to compensate for the voltage
>> losses in the MOSFETs and connections would also improve things.
>>
> Is this a geared motor? I am slowly plodding towards making something
> with similar specs as a brushless motor.

Yes, that's brush DC motor with a worm gear; seems to be the least
expensive construction. Or at least the only one I found that seems to be
produced in high enough numbers to be cheap enough :)

One thing I'm missing is reading the speed of the motor before the gear;
that would improve the controller quite a bit.

Gerhard

More... (looser matching)
- Last day of these posts
- In 2007 , 2008 only
- Today
- New search...