Searching \ for '[EE:] Driving an H-bridge' 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/io/motors.htm?key=h%2Dbridge
Search entire site for: 'Driving an H-bridge'.

Exact match. Not showing close matches.
PICList Thread
'[EE:] Driving an H-bridge'
2003\11\19@025137 by Jinx

face picon face
I'm ready to power up an H-bridge made with 4 N-ch FETS
for motor control. Any advantage, recommendation or advice
as to whether the PWM is on the top or the bottom or does it
not matter ?

--
http://www.piclist.com hint: The list server can filter out subtopics
(like ads or off topics) for you. See http://www.piclist.com/#topics

2003\11\19@033855 by Russell McMahon

face
flavicon
face
> I'm ready to power up an H-bridge made with 4 N-ch FETS
> for motor control. Any advantage, recommendation or advice
> as to whether the PWM is on the top or the bottom or does it
> not matter ?


As a rule the low side drivers are liable to be better for PWMing.
This is because the high side is more liable to have limited bandwidth due
to whatever you use for isolation purposes. If you use opto couplers (as
your circuit does) then the high side is ver likely to be appreciably
slower. Whether this matters depends in part on the PWM frequency.

You can check this easily by running the circuit with a light resistive load
and applying PWM to high and low successively and DC drive to the other
half. Look at the FET gate drive waveforms as you increase the PWM
frequency. The 3 transistor FET drive circuit that you are using will
provide pulses down to around 1 uS in length without too much design effort.
The opto driven high side version us not liable to get anywhere near this
but YMMV.

       RM

--
http://www.piclist.com hint: The list server can filter out subtopics
(like ads or off topics) for you. See http://www.piclist.com/#topics

2003\11\19@070029 by Mauricio Jancic

flavicon
face
One thing that you can do, in case that you *accidentaly* turn on two
transistors of the same pair is to put 2 lamps (60W for example) in
series between each other and both lamps in series with the +V going
from the rectifier to the H-bridge.

Save me 2 times....

Best regards

Mauricio D. Jancic
Janso Desarrollos - Microchip Consultant
http://www.janso.com.ar
spam_OUTinfoTakeThisOuTspamjanso.com.ar
(54) - 11 - 4542 - 3519
Lugones 3238 "B" - C1430ECN
Capital Federal
Republica Argentina
(54) - 11 - 4542 - 3519
MSN: .....jansodesarrollosKILLspamspam@spam@hotmail.com

>>{Original Message removed}

2003\11\19@073428 by Olin Lathrop

face picon face
Jinx wrote:
> I'm ready to power up an H-bridge made with 4 N-ch FETS
> for motor control. Any advantage, recommendation or advice
> as to whether the PWM is on the top or the bottom or does it
> not matter ?

It's usually just a little easier and faster to switch the low side FETs.
This also allows you to relax the specs on the high side drivers, since they
will only be switching when the motor voltage is reversed, not each PWM
pulse.


*****************************************************************
Embed Inc, embedded system specialists in Littleton Massachusetts
(978) 742-9014, http://www.embedinc.com

--
http://www.piclist.com hint: The list server can filter out subtopics
(like ads or off topics) for you. See http://www.piclist.com/#topics

2003\11\19@120733 by gtyler

flavicon
face
I have just done 2 projects like this. the first one used a IR2111 to drive
each side, like this the PWM is on both with no switching problem. The other
used 2 HPCL 4503's and a MC33153 to drive each IGBT as it required
isolation, again no problem driving pwm into either or both. I had to drive
both top and bottom Fets with pwm as I was creating a sine wave, and the
load could be capacitive or inductive. If you only put PWM onto 1 fet in
this case you could get distortion.

George
{Original Message removed}

2003\11\19@122144 by Bob Blick
face
flavicon
face
Jinx said:
> I'm ready to power up an H-bridge made with 4 N-ch FETS
> for motor control. Any advantage, recommendation or advice
> as to whether the PWM is on the top or the bottom or does it
> not matter ?

You've left something out.

Four N-MOSFETs does not equal an H-bridge, you need more parts! You will
need to shift levels to drive the MOSFETs.

Cheers,

Bob

--
http://www.piclist.com hint: The list server can filter out subtopics
(like ads or off topics) for you. See http://www.piclist.com/#topics

2003\11\19@125922 by Edward Gisske

flavicon
face
Bob is right...

you will need to find (or generate) a higher voltage than the pic supplies
to turn on the upper FET's in the bridge. I usually couple with an
opto-isolator to the upper gates. It is cheap and clean.

Another thing to be aware of: The FETS will not turn off immediately, so put
a little dwell between turning the motor off in one direction and on in the
other. A pic can reverse the outputs in well under a microsecond. It is
unlikely that the FETs will turn off that fast, unless you have a really
stiff driver circuit, so you end up with a short circuit (maybe two) across
the motor supply. If you don't pay attention to this you will have either
big current spikes on the motor line or the sweet smell of combusting FET.

Also, the motor will act like a generator when coasting. Depending on how
you do the driving, you can end up with some weird bias on the FET switches
until the motor stops. Clamp each FET with a diode S-D to deal with this
issue. When you reverse a motor running full speed, you will get a current
spike of about  twice the locked rotor current. This can be a bunch of
current! Size the H-Bridge FETs accordingly. You can figure out the locked
rotor current by measuring the resistance of the motor when at rest and
dividing that number into the motor voltage.

I dislike H-Bridges, and only use them when I have to. The problem with
driving them with a PIC is dealing with the power-up transient before the
PIC initializes the ports. Be clever and take into account the power-up
state of the PIC output lines when you design the circuit.

If I am not dealing with a gazillion motor reversals, I reverse with a DPDT
relay and use one FET for PWM. It is a lot safer and often cheaper than a
full FET H-bridge.

In short....H-Bridge design is a non-trivial art...

Edward Gisske, P.E.
Gisske Engineering
608-523-1900
gisskespamKILLspamoffex.com


{Original Message removed}

2003\11\19@192848 by Jinx

face picon face
> Four N-MOSFETs does not equal an H-bridge, you need
> more parts! You will need to shift levels to drive the MOSFETs

I have more parts ! I have lotsa lotsa parts !

http://home.clear.net.nz/pages/joecolquitt/4nchfeth.html

Many thanks to Russell for his technical input

Still to find proper testing time to give it thorough work-out so
there may need to be the odd tweak. For example, diodes
bigger than PBYR745 if necessary

--
http://www.piclist.com hint: The list server can filter out subtopics
(like ads or off topics) for you. See http://www.piclist.com/#topics

2003\11\19@201118 by Tom

flavicon
face
Hi Jinx

Russell has probably mentioned this stuff already but in case he didn't...

1. If your circuit ever sees a stalled motor, then there will be a contest
of sorts.  Who will give out first?  The power supply? The fets? The
wiring? Just wondering...

2. You are being quite conservative in your gate drive setup. For example,
your bottom fets have their excitation clamped at 7.5 volts.  Is this
enough to guarantee good turn on at all loads and temperatures?

3. Finally, dynamic effects.  This is something to test carefully with a
scope to ensure you are getting good turn on at the intended operating
frequency.  Look closely for inductive spikes (you look to have taken great
care to snub these) and of course, verify operating temperature.  Anything
getting hot? Cook dinner on it?

Good luck.
Tom

--
http://www.piclist.com hint: The list server can filter out subtopics
(like ads or off topics) for you. See http://www.piclist.com/#topics

2003\11\19@214205 by Jinx

face picon face
> 1. If your circuit ever sees a stalled motor, then there will be a
>  contest of sorts.  Who will give out first?  The power supply ?
> The fets ? The wiring ?

The only universal stall detector I could think of is a motion
sensor. If there's PWM and the motor ain't going nowhere.....

Because this could be used with any number of motors, I'm not
sure if a current or voltage sensor is going to work without any
tinkering for the individual motor

> 2. You are being quite conservative in your gate drive setup.
> For example, your bottom fets have their excitation clamped
> at 7.5 volts.  Is this enough to guarantee good turn on at all
> loads and temperatures?

http://www.irf.com/product-info/datasheets/data/irl3402.pdf

According to the datasheet, output is virtually the same for
comparable Vds voltage at 25C and 150C (Vds = 2.25V - 7V).
Vds = 5V seems to be a good compromise between output
and the 7V maximum

> 3. Finally, dynamic effects

That's probably where the fun will start

--
http://www.piclist.com hint: The list server can filter out subtopics
(like ads or off topics) for you. See http://www.piclist.com/#topics

2003\11\20@005050 by Bob Blick

face
flavicon
face
On 20 Nov 2003 at 13:21, Jinx wrote:
> I have more parts ! I have lotsa lotsa parts !
> http://home.clear.net.nz/pages/joecolquitt/4nchfeth.html

Here's a problem. You have a 5.1 volt zener gate-source on the upper
fets. But you are driving the gates with almost 20 volts and the drain is
at 12 volts(and maybe less). So the zener diode is really upset when
you turn on the gate, but the source can't possibly be 14.9 volts.

You can put the zener before the npn-pnp buffer, but really you should
rethink the gate supply, and have separate supplies for each upper fet,
and the supply should be referenced to the source, not ground.

I heard IRF2112 mentioned earlier, I use them, too, and approve. Duty
cycle can't be 100%, but if you're driving with a PIC you can make sure
100% doesn't happen.

Cheers,

Bob

--
http://www.piclist.com hint: The PICList is archived three different
ways.  See http://www.piclist.com/#archives for details.

2003\11\20@013326 by Russell McMahon

face
flavicon
face
> On 20 Nov 2003 at 13:21, Jinx wrote:
> > I have more parts ! I have lotsa lotsa parts !
> > http://home.clear.net.nz/pages/joecolquitt/4nchfeth.html

> Here's a problem. You have a 5.1 volt zener gate-source on the upper
> fets. But you are driving the gates with almost 20 volts and the drain is
> at 12 volts(and maybe less). So the zener diode is really upset when
> you turn on the gate, but the source can't possibly be 14.9 volts.
>
> You can put the zener before the npn-pnp buffer, but really you should
> rethink the gate supply, and have separate supplies for each upper fet,
> and the supply should be referenced to the source, not ground.
>

I think we've got version creep (or uncreep) here.
The "gate zener lower than drive supply" problem  was identified some while
ago and Jinx adjusted things accordingly.
Its done gone and got back the way it were before.

The ideal is to have the feed resistor from the 555 pump (here = 10R)(which
is too low) such that it passes the average current OK. Switching peaks are
provided from the driver collector caps (here 100 uF). The gate zener needs
to be at least as large as Vdrive - V_high_supply. If its lower than this
then the gate zener will conduct throughout the on cycle (or until the pump
reservoir cap is exhausted).

I'm not sure exactly what you mean by "referencing the supply to the source.
(I understand the words but there are several different possible meanings
here.) The 555 supply must pump from ground of course. The here 20v clamp
zeners could instead be 5v1 or so referenced to the 12v supply. This gives
you a better clamp relative to supply when supply varies slightly. This has
some merit if the 12v supply varies fairly widely as the FETS have a rather
low max gate voltage which must not be exceeded.

> I heard IRF2112 mentioned earlier, I use them, too, and approve. Duty
> cycle can't be 100%, but if you're driving with a PIC you can make sure
> 100% doesn't happen.

It's a nice IC but not really needed. You could use the circuit as is to
self pump the voltage multiplier which is presently driven by the 555. The
disadvantage is, as you say, that you can't stop the circuit and still get
drive. The 555 is an easy solution when such low voltages are used.
When you get to 100+ volts then special pumps start to look more attractive.


       Russell McMahon

--
http://www.piclist.com hint: The PICList is archived three different
ways.  See http://www.piclist.com/#archives for details.

2003\11\20@041512 by Alan B. Pearce

face picon face
>you will need to find (or generate) a higher voltage than
>the pic supplies to turn on the upper FET's in the bridge.
>I usually couple with an opto-isolator to the upper gates.
>It is cheap and clean.

And probably the easiest way to do it is a chip such as the Intersil
HIP4080, or some of the other HIP408x family. This chip will look after the
high side drive, and has programmable overlap control to stop both devices
on the one totem pole from conducting together. Reasonable current drive for
making the FETs switch fast - provided you use modern FETs with minimized
gate capacitance. If trying to use older FETs with high gate capacitance it
is difficult to get the dead time long enough.

--
http://www.piclist.com hint: The PICList is archived three different
ways.  See http://www.piclist.com/#archives for details.

2003\11\20@115855 by Bob Blick

face
flavicon
face
Russell McMahon said:
>> IR2112, blah blah...

> It's a nice IC but not really needed. You could use the circuit as is to
> self pump the voltage multiplier which is presently driven by the 555.
> The disadvantage is, as you say, that you can't stop the circuit and
> still get drive. The 555 is an easy solution when such low voltages are
> used. When you get to 100+ volts then special pumps start to look more
> attractive.

Surely you jest. The circuit as I saw it had gobs of parts, including some
glue logic and two optoisolators. Optoisolators! This is a 12 volt circuit
that shares a common ground, for Heaven's sake! Using something like the
IR2112 takes away the need for a whole mess of parts, and they are cheap
and easy to get.

For that matter, since it's only 12 volts you could just as easily do the
upper half with P-channel fets, talk about easy and cheap. You could do
the whole thing with a third the parts.

On the other hand, if this is supposed to be a big beefy thing, where's
the current limiting?

I hope I'm not too critical here, but these are not issues to pooh-pooh.
H-bridges are not rocket science but you need to put the right parts where
it matters.

Cheerful regards,

Bob

--
http://www.piclist.com hint: The PICList is archived three different
ways.  See http://www.piclist.com/#archives for details.

2003\11\20@235537 by Hopkins

flavicon
face
This all sounds great,

Can one of you guys with the experience in the H bridge circuits publish a
design or give a web site that would outline a good building block for the
type of circuit.




---
Outgoing mail is certified Virus Free.
Checked by AVG anti-virus system (http://www.grisoft.com).
Version: 6.0.541 / Virus Database: 335 - Release Date: 14/11/2003

--
http://www.piclist.com hint: The PICList is archived three different
ways.  See http://www.piclist.com/#archives for details.

2003\11\21@034107 by Jinx

face picon face
> And probably the easiest way to do it is a chip such as
> the Intersil HIP4080, or some of the other HIP408x family

I picked up a couple of those from RS the last time you
recommended them. They look very nice and would be
quite easy to control with s/w. As I recall they're like the
H-bridge equivalent of a stepper controller. This particular
H-bridge was already under way so I've not had a chance
to try the HIP4080 yet but it's certainly on the to-do list

--
http://www.piclist.com#nomail Going offline? Don't AutoReply us!
email .....listservKILLspamspam.....mitvma.mit.edu with SET PICList DIGEST in the body

2003\11\21@060131 by Ishaan Dalal

flavicon
face
> > And probably the easiest way to do it is a chip such as
> > the Intersil HIP4080, or some of the other HIP408x family

I don't know; these seemed ideal, I got 3 pieces [HIP4081] from Arrow (who
btw are a *lot* cheaper than Digikey for almost every part, esp. PICs, if
you can wait for their ground shipping (3-10 days)). Spent hours trying to
get them to work. Either they would do nothing, or the chip would go into
some kind of "controlled latchup"; it would start gobbling up about 0.5A,
and start getting warm, but _still_ do nothing. No responses to control
inputs, nothing on the scope or FETs at the outputs.

Maybe I got 3 duds, or I was making some dumb mistake; YMMV.

I found the National LMD18xxx integrated H-bridges to be good if you are
prototyping and your max continuous current is <3A, max pulsed <6A.

What about the FET drivers Microchip makes? Yeah, they're not 1-piece like
the HIP408x, but have people tried them? Experiences?

Cheers,
Ishaan

--
http://www.piclist.com#nomail Going offline? Don't AutoReply us!
email EraseMElistservspam_OUTspamTakeThisOuTmitvma.mit.edu with SET PICList DIGEST in the body

2003\11\21@060738 by Alan B. Pearce

face picon face
>Can one of you guys with the experience in the H bridge circuits
>publish a design or give a web site that would outline a good
>building block for the type of circuit.

I would suggest you look at the Intersil HIP408x family application notes.

See http://www.intersil.com/products/partsearch.asp and enter HIP408 to
search and you will get the whole family of datasheets listed. You need to
follow each device to find all the application notes.

--
http://www.piclist.com#nomail Going offline? Don't AutoReply us!
email listservspamspam_OUTmitvma.mit.edu with SET PICList DIGEST in the body

2003\11\21@062434 by Alan B. Pearce

face picon face
>Can one of you guys with the experience in the H bridge circuits
>publish a design or give a web site that would outline a good
>building block for the type of circuit.

Further to my previous advice, look at the Netional Semiconductor LMD18200,
and the couple of application notes as well. I was not aware of these before
Ishaan posted about them.

http://www.national.com/pf/LM/LMD18200.html and down the bottom of the page
for app notes.

--
http://www.piclist.com#nomail Going offline? Don't AutoReply us!
email @spam@listservKILLspamspammitvma.mit.edu with SET PICList DIGEST in the body

More... (looser matching)
- Last day of these posts
- In 2003 , 2004 only
- Today
- New search...