Free Day at Sparkfun

[Tim Craig]

Caught this on the PARTS list.  On Jan 7, 2010 the first 1000 orders at Sparkfun get $100 off their order.

 

http://www.sparkfun.com/commerce/news.php?id=305

 

Tim

I would rather have a mind opened by wonder than one closed by belief.

 

[Bob Smith]

I would appreciate any comments on the dual H-bridge circuit
available at http://www.linuxtoys.org/robots/dual_H_sch.png

The connector goes to a BaseBoard FPGA card. The FPGA does
all of the control modes, PWM, and anti shoot-through timers.
When the FPGA is not not connected all inputs to the TC426s
are high putting both motors into high-side braking.


The goal was to keep it simple, have medium current capability
and to use only parts available at my local electronics store
(Anchor Electronics on Walsh in Santa Clara).

All comments are welcome.

thanks
Bob Smith

[Wayne C. Gramlich]

Bob:

Since the diodes are 3A Schottky diodes, your H-Bridge is
limited to 3 Amps per bridge. I suspect that an L298 hooked
up in 3.5 mode will be cheaper. I'll point you at my design:

<http://gramlich.net/projects/rb2/midimotor2/rev_c/index.html>

Also, with the L298, you do not have to worry about shoot-through
timing issues. Lastly, I no longer install C8 and C9 on my
schematic; I use a couple of ferrite beads instead.

-Wayne

> --
>
> You received this message because you are subscribed to the Google Groups "HomeBrew Robotics Club" group.
> To post to this group, send email to hbrob...@googlegroups.com.
> To unsubscribe from this group, send email to hbrobotics+...@googlegroups.com.
> For more options, visit this group at http://groups.google.com/group/hbrobotics?hl=en.
>
>
>

[Bob Smith]

Wayne C. Gramlich wrote:
> Since the diodes are 3A Schottky diodes, your H-Bridge is
> limited to 3 Amps per bridge.

Thanks very much for looking at the design.

The 1N5822 has an average current of 3 amps. Its
_peak_ current (8.3 ms single half sine wave) is
80 amps, so I think these should be fine as back
EMF diodes. The MOSFETs are rated as 10 amps and
this is the limit for the H-Bridge as a whole.

thanks
Bob Smith



> Bob Smith wrote:
>> http://www.linuxtoys.org/robots/dual_H_sch.png

[Alan Marconett]

Looks OK to me Bob. But I don't know; can you do that in Linux?

I didn't realize Anchor was still around! So few are. Did you make up a
PCB?

Alan KM6VV

[Bob Smith]

Alan Marconett wrote:
> Looks OK to me Bob. But I don't know; can you do that in Linux?

The control modes, PWM counters, and a watchdog timer
are in the FPGA. The Linux device driver will have a
nice "user interface" for all of the messy registers
down in the FPGA.

> I didn't realize Anchor was still around!

In fact, they recently expanded their store and have
expanded their inventory to match.

> Did you make up a PCB?

I'm only making one of these for now. The Verilog
for the controller is done and I'm building a demo/dev
board for the guy writing the Linux device driver for it.


BTW_1: the watchdog timer is a safety feature. If you
don't write to its register every XXX milliseconds, the
watchdog shuts down all PWM output.

BTW_2: the Linux driver will try to arrange the PWM
pulses so that they overlap as little as possible. While
kind of a detail, it does mean the battery might last a
little longer since the I-squared losses will be lower
in the battery cables.

thanks!
Bob

[Wayne C. Gramlich]

Bob:

I'm not enough of a Power FET person to evaluate your
design at any level of specificity. My recollection is
that high-power MOSFET H-Bridges exclusively use N channel
devices because the most P channel devices have a higher
RDS(on). The problem with an all N-channel H bridges is
you have to get your on voltage higher than V+. In a
medium current situation, the mixed N/P channel design
should work fine, tho'.

If you are pumping 10 Amps through motor and turn off
the FET, you are going to get 10 Amps right back at you.
Whether or not it will decay in 8.3 mS is not obvious
to me; it really depends on the moment of interia of
the motor system. If the motor has a big fly wheel,
it may take a while to get the current back down.
The 1N5822 may take a certain amount of abuse and then
decide to go on strike.

I recommend that you put as large a capacitor across
the V+/V- as possible to deal with surge current. 10uF
seems a little small to me.

Just my $.02,

-Wayne

[Alan Marconett]

I don't remember the details, but for Linux, EMC & EMC2, CNC controller
programs; I think special builds are/were used to allow the pulse generation
without interruptions. The code is available, it might offer some
additional hints...

Alan KM6VV

> -----Original Message-----
> From: Bob Smith [mailto:bsm...@linuxtoys.org]
>

> <snip>