Re: [HBRobotics] Digest for hbrobotics - 25 messages in 2 topics
1 view
Skip to first unread message
Chris Tacklind
Nov 23, 2009, 7:59:08 PM11/23/09
to hbrob...@googlegroups.com, Nathaniel Lewis
Nathaniel,
The first time I built a system using a Basic Stamp and a home made H
bridge everything worked great for a while. Then I started having the
problem you mention.
Of course, you would never turn the high side on before turning the
coresponding low side off, like I did. Doing this would be a direct
short of the battery to ground. Fortunately, I used VN920 for the
high side and VNP35N07 for the low side. They are virtually
indestructible and have logic level inputs.
After much fuss, I solved the glitch problem by adding a 1 ms delay
before switching the high and low sides. This smallest delay was
needed to ensure the complete shut down of the FETs. I never bothered
to look up what the minimum spec is. They have continued to work for
years.
-Chris Tacklind
On Tue, Nov 17, 2009 at 4:12 PM, <hbrob...@googlegroups.com> wrote:
> Today's Topic Summary
>
> Power Issue [23 updates]
> HBRC Meeting THIS Wednesday Nov. 18: Machine Vision [2 updates]
>
> Topic: Power Issue
>
> "Nathaniel Lewis" <linux.r...@gmail.com> Nov 17 07:31AM -0800
>
> My single board computer on my Stinger robot resets after a full motor
> reverse. What is the best way to solve this?
>
>
>
>
> "rdo...@bizserve.com" <rdo...@bizserve.com> Nov 17 10:33AM -0500
>
> A few things I have seen done on other robots:
>
> - Additional caps/filtering where the power comes
> into your logic board
>
> - Larger capacity battery pack
>
> - Separate logic battery
>
> Robert
>
> ------- Original Message -------
> From : Nathaniel
> Lewis[mailto:linux.r...@gmail.com]
> Sent : 11/17/2009 10:31:27 AM
> To : hbrob...@googlegroups.com
> Cc :
> Subject : RE: [HBRobotics] Power Issue
>
>
> My single board computer on my Stinger robot resets
> after a
> full motor reverse. What is the best way to solve this?
>
>
>
>
>
> Patrick Luke <pel...@gmail.com> Nov 17 07:53AM -0800
>
> ya what he said;
> make certain your not undervolting when the draw from the motors is under
> load near stall
>
> you mentioned only when it goes into reverse, is the caster underneath
> binding up as it trys to flip around? this would cause a need for more power
>
>
>
>
>
> "Wayne C. Gramlich" <wa...@gramlich.net> Nov 17 10:43AM -0800
>
> What you are experiencing is a downward voltage spike on the
> voltage input to your microcontroller. The microcontroller
> is deciding that it is running out of power and resets.
>
> It was a painful lesson for me to learn, but I now always
> use two separate batteries, one for logic power and
> one for motor power. I have gone to the extreme of
> keeping the two completely electrically isolated and using
> opto-isolators to drive my H-bridges, servos, and
> such. I no longer have spurious resets. There are other
> solutions, but opto-isolators are really cheap.
>
> -Wayne
>
> Nathaniel Lewis wrote:
>
>
>
> giuliano carlini <giul...@carlini.com> Nov 17 11:28AM -0800
>
> Add caps. How many and how big depend on whether you go with a split
> or single power supply. Start with a cap across the motor leads.
> Alternatively, caps from each lead to the case. Or all three. Then
> caps from the h-bridge power supplies (both logic and motor) to
> ground. And from h-bridge signal lines to ground. How big, that
> depends on whether you want a split power supply or a single supply.
> For a split supply, .1uF 100V caps are fine for all the caps, and
> cheap, 3 or 4 cents a pop. For a single power supply, I ended up
> needing .1uF 630V caps from motor leads to case. If you use one
> across the leads, it should also be .1uF 630V.
>
> 1. Single power supply. This is what I did first, cause I was a dufus
> and didn't listen to folks who said to use a split power supply ... In
> addition to the large caps across the leads, and every other cap
> listed above, you'll also need big zener diodes, "head to head"
> electrically between the motor leads. I connect the to the h-bridge
> motor outputs. This seemed to squelch the noise, but the big caps and
> diodes are something like 50 cents to $1 each. And I worry that big
> spikes might still get through. And resetting micros isn't the worst
> of it. I fried a $75 bluetooth chip once ... And other less expensive
> stuff till I figured it out ...
>
> 2. Split power supply with separate batteries for the micro and the
> motor. Without optos you'll need a common ground and at least the cap
> across the motor leads. I've built bots with just the single cap with
> split power supplies, but you might as well add them all if it isn't
> too big a hassle. If you don't add them all, here's my order of
> priority: across the motor leads, across the h-bridge power supplies
> and ground, motor leads and case, h-bridge signal pins and ground.
> Other folks may have a different order of priority. Logic power can be
> a 9V battery or a handful of double or triple A's. Motor power should
> be appropriate to the motor.
>
> 3. Split power supply with opto isolators. This is what I do now. It's
> fairly cheap and simple and maximizes the isolation. No need for a
> common ground, so the two power circuits are totally separate. Make
> sure the optos are fast enough for you PWM. Some Darlington optos are
> too slow for really fast PWM. Strictly speaking you don't need caps
> with this approach. But large spikes can radiate noise. May as well
> add at least a few of the caps.
>
>
> giuliano
>
> On Nov 17, 2009, at 7:31 AM, Nathaniel Lewis wrote:
>
>
>
>
> "don clay" <don...@earthlink.net> Nov 17 01:24PM -0700
>
> ya what they said.
>
> Do you have caps (.1) on your motors already? Do you have big enough gauge
> wire
> going to your motors? Is your battery fully charged & powerful enough? Are
> you
> going into full reverse from full forward?
>
> From: Patrick Luke
>
>
>
>
>
> David Wyland <dcwy...@ix.netcom.com> Nov 17 12:37PM -0800
>
> Do not put caps bigger than 0.1 uF across the motor, or you can fry your
> H bridge.
>
> Dave
>
>
> don clay wrote:
>
>
>
> "don clay" <don...@earthlink.net> Nov 17 01:59PM -0700
>
> Oops! That's what I meant .1 uf.
>
> Date sent: Tue, 17 Nov 2009 12:37:53 -0800
> From: David Wyland <dcwy...@ix.netcom.com>
> To: hbrob...@googlegroups.com
> Subject: Re: [HBRobotics] Power Issue
> Send reply to: hbrob...@googlegroups.com
>
>
>
>
> Bob Smith <bsm...@linuxtoys.org> Nov 17 01:26PM -0800
>
> Nathaniel Lewis wrote:
>> My single board computer on my Stinger robot resets after a full motor
>> reverse. What is the best way to solve this?
>
> Opto-isolators. Caps. Split power supplies.
>
> Is there a way you can avoid a "full motor reverse" and
> still meet your system/design requirements? Awhile ago
> we had a discussion about motor ramping. As an experiment
> you might try ramping your PWM pulse width so there is
> never a sudden change. Take, say, a quarter second to
> ramp from 0 to full forward, from full forward to 0, from
> 0 to full reverse, and from full reverse to 0. Can your
> design take a half second to do a "full motor reverse"?
>
> Bob Smith
>
>
>
> Tom Rokicki <rok...@gmail.com> Nov 17 01:45PM -0800
>
> Ramping helps in the normal case, absolutely, but stall
> current can still cause power sag. The cleanest solution
> is of course separate power supplies for logic and motors.
> This alone, even without optoisolators, if you do your
> grounding carefully, will probably solve the majority of your
> problems. Optoisolators are pretty cheap though.
>
> I do love ramping, though, and the effect it can have on the
> apparent "grace" of a robot.
>
>
> --
> Check out Golly at http://golly.sf.net/
>
>
>
> "Alan Marconett" <KM...@SBCGLOBAL.NET> Nov 17 01:50PM -0800
>
> I'm still wondering about the 630V caps!
>
> Alan KM6VV
>
>
>
>
> giuliano carlini <giul...@carlini.com> Nov 17 02:56PM -0800
>
> I tried 100V caps but the micro reset no matter what I did. So I found
> something on the web about using 300V caps and head to head big
> zeners. Difference in price between 300V and 600V caps at digikey was
> nominal, and I wanted the problem FIXED, so I went with the monster
> caps and, I think, 1500 watt zeners ;->
>
> giuliano
>
>
> On Nov 17, 2009, at 1:50 PM, Alan Marconett wrote:
>
>
>
>
> "Alan Marconett" <KM...@SBCGLOBAL.NET> Nov 17 03:02PM -0800
>
> OK, you must be running a really big motor! What voltage/current?
>
> Alan KM6VV
>
>
>
>
> giuliano carlini <giul...@carlini.com> Nov 17 03:20PM -0800
>
> Just a tiny solarbotics gm8 at 9V into a SN754410 and no load so maybe
> 100ma.
>
> I was putting it through a torture test. 1/2 second forward, 1/2
> second reverse, repeat. No caps, 5 seconds or less to reset. I got a
> few minutes with 100V caps sprinkled all over. 18 hours with the 600's
> and zeners. Might have gone longer, but I needed my notebook to go to
> work.
>
> giuliano
>
> On Nov 17, 2009, at 3:02 PM, Alan Marconett wrote:
>
>
>
>
> "Alan Marconett" <KM...@SBCGLOBAL.NET> Nov 17 03:29PM -0800
>
> That's not all that big. I wouldn't expect the voltage rating matter, as
> long as it was 25V or more (if that). Same goes for the current rating of
> the Zener.
>
> Alan KM6VV
>
>
>
>
> "Nathaniel Lewis" <linux.r...@gmail.com> Nov 17 03:45PM -0800
>
> I don't have any caps on the motors at all, but there is 3000uF across the
> SBC power lines and 1000uF across the battery. I'm guessing I should put
> more on the battery. I would like to use two batteries, but I am using an
> expensive Lithium Ion pack for RC Helicopters and I don't want to get
> another one if there are other options. Another thing I realized was that I
> should use a PID algorithm to smooth the motor operation.
>
> Nathaniel
>
> -----Original Message-----
> From: don clay [mailto:don...@earthlink.net]
> Sent: Tuesday, November 17, 2009 12:24 PM
> To: hbrob...@googlegroups.com
> Subject: Re: [HBRobotics] Power Issue
>
> ya what they said.
>
> Do you have caps (.1) on your motors already? Do you have big enough gauge
> wire
> going to your motors? Is your battery fully charged & powerful enough? Are
> you
> going into full reverse from full forward?
>
> From: Patrick Luke
>
>
>>
>>
>
>
> --
>
> 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=.
>
>
>
>
>
> "Nathaniel Lewis" <linux.r...@gmail.com> Nov 17 03:48PM -0800
>
> Why does the voltage matter so much? I thought that if you had 2x your
> maximum voltage on you caps you be fine as long as you had enough
> microfarads.
> Nathaniel
>
> -----Original Message-----
> From: giuliano carlini [mailto:giul...@carlini.com]
> Sent: Tuesday, November 17, 2009 3:20 PM
> To: hbrob...@googlegroups.com
> Subject: Re: [HBRobotics] Power Issue
>
> Just a tiny solarbotics gm8 at 9V into a SN754410 and no load so maybe
> 100ma.
>
> I was putting it through a torture test. 1/2 second forward, 1/2
> second reverse, repeat. No caps, 5 seconds or less to reset. I got a
> few minutes with 100V caps sprinkled all over. 18 hours with the 600's
> and zeners. Might have gone longer, but I needed my notebook to go to
> work.
>
> giuliano
>
> On Nov 17, 2009, at 3:02 PM, Alan Marconett wrote:
>
>> 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=
>>
>>
>
> --
>
> 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=.
>
>
>
>
>
> e...@okerson.com Nov 17 05:48PM -0600
>
> Back EMF from the motors changing directions could easily go over 100v.
>
> Ed Okerson
>
>
>
>
> "Wayne C. Gramlich" <wa...@gramlich.net> Nov 17 03:59PM -0800
>
>
> Nathaniel Lewis wrote:
>> another one if there are other options. Another thing I realized was that
>> I
>> should use a PID algorithm to smooth the motor operation.
>
> The PID algorithm does not necessarily smooth out motor operation.
> If your PID algorithm is trying to get you towards +X and you feed
> it a new position target of -X, you will get the same motor reverse
> situation.
>
> Adding ramping, where you restrict how fast you can change the
> direction and with of your pulses will help a lot.
>
> Is there some reason why the other battery pack needs to be Lithium
> Ion? The battery pack that feeds your processor can be simple
> alkaline batteries and it will run for a long time before you
> drain the batteries.
>
> -Wayne
>
>
>
> giuliano carlini <giul...@carlini.com> Nov 17 04:00PM -0800
>
> I didn't expect it. But it's easy to replicate.
>
> Arudino, SN754410NE, GM8, 8x.1uF 100V caps, 9V battery and snaps, a
> breadboard, and if desired some leds and resistors to provide visual
> feedback for the signal pins.
>
> Connect the battery to both breadboard power rails.
> Connect arduino power input to breadboard.
> Place the hbridge between the two halves of the breadboard.
> Connect h-bridge pins 8,9 to power, and 4,5,12,13 to ground.
> Place caps from pins 1, 2, 7, 8, 9 to ground.
> Solder caps from each GM8 motor lead to the case, and between the leads.
> Solder 20ga wire from each motor lead to a 4 pin header. Pull the two
> middle pins.
> Push the header into the breadboard for h-bridge pins 3,6
> Wire up the LEDs and resistors if you want to
> Connect arduino 11 to h-bridge 1, arduino 12 to h-bridge 2, and
> arduino 13 to h-bridge 7.
>
> Write and run a simple arduino program that write 5v to digital 11 and
> alternates between writing 5v to digital 12 and 0 to digital 13 for
> half a second; and then writing 0v to digital 12 and 5v to digital 13.
> Add some printlns in setup and at each transition, including the
> current value of millis so you can detect a reset.
>
> Might have got something wrong in the details, I'm typing this in off
> the top of my head, but it's something close to this.
>
> If anyone does this, let me know it works for you ...
>
> giuliano
>
> On Nov 17, 2009, at 3:29 PM, Alan Marconett wrote:
>
>
>
>
> Michael Prados <mpr...@gmail.com> Nov 17 04:04PM -0800
>
> lots of good suggestions, here's a couple others-
>
> 1. Hopefully they are there somewhere anyway, but there should be
> flyback diodes to limit the inductive spikes! In some cases, these
> are already built into the H-Bridge. Spikes from switching inductive
> loads without flyback diodes are the number one factor I hear in
> stories of microcontrollers resetting in mechatronic systems.
>
> http://en.wikipedia.org/wiki/Flyback_diode
>
> 2. DC-DC converters are increasingly cheap and compact, and can go a
> long way to smooth out noise in a wide frequency range. Here's one
> I've been using with an AVR (we used a similar one on Swarm):
>
> http://search.digikey.com/scripts/DkSearch/dksus.dll?Detail&name=PT5101C-ND
>
> -mike
>
>
> --
> sent from my computing machine
>
>
>
> "Alan Marconett" <KM...@SBCGLOBAL.NET> Nov 17 04:10PM -0800
>
> Oh, you're perfectly fine; the cap might be a little bigger. How big are
> 630V caps? The 25v or 50v ceramic .1uF caps are easy to tuck onto the rear
> end of a motor.
>
> Alan KM6VV
>
>
>
>
> giuliano carlini <giul...@carlini.com> Nov 17 04:10PM -0800
>
> A 9V, handful of AAs, or a 7.2 or 9.6 NiCd will do just fine to power
> the micro. Only the motors need the current capacity the LiPo's can
> supply. But, even for the micro use some sort of rechargeable. No
> reason to fill landfills with dead batteries.
>
> Caps across the batteries aren't enough. You need them close to the
> motors and h-bridge.
>
> PIDs are good for not overshooting your target speed or oscillating
> wildly about it. They won't do a slow start. If anything, they tend to
> do a hard start and then slow as they get closer to the target. You
> want to send the output of the PID into a limiter. The limiter will
> set a max limit on the speed change over short bits of time. It should
> let you change speed rapidly, but not full forward to full reverse.
>
> giuliano
>
> On Nov 17, 2009, at 3:45 PM, Nathaniel Lewis wrote:
>
>
>
>
> Topic: HBRC Meeting THIS Wednesday Nov. 18: Machine Vision
>
> Camp Peavy <cpe...@yahoo.com> Nov 17 10:26AM -0800
>
>> I’m looking forward to this presentation. I am working on a vision system
>> and maybe this might help me add functionality and increased performance!
>
> Mike Thompson rocks!
>
>
>
>
> "Alan Marconett" <KM...@SBCGLOBAL.NET> Nov 17 10:35AM -0800
>
> Me too! I've got a CMU3 on request for Christmas!
>
>
>
> Alan KM6VV
>
>
>
> _____
>
> From: Camp Peavy [mailto:cpe...@yahoo.com]
> Sent: Tuesday, November 17, 2009 10:27 AM
> To: hbrob...@googlegroups.com
> Subject: RE: [HBRobotics] HBRC Meeting THIS Wednesday Nov. 18: Machine
> Vision
>
>
>
>
>> I'm looking forward to this presentation. I am working on a vision system
> and maybe this might help me add functionality and increased performance!
>
>
>
> Mike Thompson rocks!
>
>
>
> --
>
> 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=.
>
>
>
>
> --
>
> 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=.
>
The first time I built a system using a Basic Stamp and a home made H
bridge everything worked great for a while. Then I started having the
problem you mention.
Of course, you would never turn the high side on before turning the
coresponding low side off, like I did. Doing this would be a direct
short of the battery to ground. Fortunately, I used VN920 for the
high side and VNP35N07 for the low side. They are virtually
indestructible and have logic level inputs.
After much fuss, I solved the glitch problem by adding a 1 ms delay
before switching the high and low sides. This smallest delay was
needed to ensure the complete shut down of the FETs. I never bothered
to look up what the minimum spec is. They have continued to work for
years.
-Chris Tacklind
On Tue, Nov 17, 2009 at 4:12 PM, <hbrob...@googlegroups.com> wrote:
> Today's Topic Summary
>
> Power Issue [23 updates]
> HBRC Meeting THIS Wednesday Nov. 18: Machine Vision [2 updates]
>
> Topic: Power Issue
>
> "Nathaniel Lewis" <linux.r...@gmail.com> Nov 17 07:31AM -0800
>
> My single board computer on my Stinger robot resets after a full motor
> reverse. What is the best way to solve this?
>
>
>
>
> "rdo...@bizserve.com" <rdo...@bizserve.com> Nov 17 10:33AM -0500
>
> A few things I have seen done on other robots:
>
> - Additional caps/filtering where the power comes
> into your logic board
>
> - Larger capacity battery pack
>
> - Separate logic battery
>
> Robert
>
> ------- Original Message -------
> From : Nathaniel
> Lewis[mailto:linux.r...@gmail.com]
> Sent : 11/17/2009 10:31:27 AM
> To : hbrob...@googlegroups.com
> Cc :
> Subject : RE: [HBRobotics] Power Issue
>
>
> My single board computer on my Stinger robot resets
> after a
> full motor reverse. What is the best way to solve this?
>
>
>
>
>
> Patrick Luke <pel...@gmail.com> Nov 17 07:53AM -0800
>
> ya what he said;
> make certain your not undervolting when the draw from the motors is under
> load near stall
>
> you mentioned only when it goes into reverse, is the caster underneath
> binding up as it trys to flip around? this would cause a need for more power
>
>
>
>
>
> "Wayne C. Gramlich" <wa...@gramlich.net> Nov 17 10:43AM -0800
>
> What you are experiencing is a downward voltage spike on the
> voltage input to your microcontroller. The microcontroller
> is deciding that it is running out of power and resets.
>
> It was a painful lesson for me to learn, but I now always
> use two separate batteries, one for logic power and
> one for motor power. I have gone to the extreme of
> keeping the two completely electrically isolated and using
> opto-isolators to drive my H-bridges, servos, and
> such. I no longer have spurious resets. There are other
> solutions, but opto-isolators are really cheap.
>
> -Wayne
>
> Nathaniel Lewis wrote:
>
>
>
> giuliano carlini <giul...@carlini.com> Nov 17 11:28AM -0800
>
> Add caps. How many and how big depend on whether you go with a split
> or single power supply. Start with a cap across the motor leads.
> Alternatively, caps from each lead to the case. Or all three. Then
> caps from the h-bridge power supplies (both logic and motor) to
> ground. And from h-bridge signal lines to ground. How big, that
> depends on whether you want a split power supply or a single supply.
> For a split supply, .1uF 100V caps are fine for all the caps, and
> cheap, 3 or 4 cents a pop. For a single power supply, I ended up
> needing .1uF 630V caps from motor leads to case. If you use one
> across the leads, it should also be .1uF 630V.
>
> 1. Single power supply. This is what I did first, cause I was a dufus
> and didn't listen to folks who said to use a split power supply ... In
> addition to the large caps across the leads, and every other cap
> listed above, you'll also need big zener diodes, "head to head"
> electrically between the motor leads. I connect the to the h-bridge
> motor outputs. This seemed to squelch the noise, but the big caps and
> diodes are something like 50 cents to $1 each. And I worry that big
> spikes might still get through. And resetting micros isn't the worst
> of it. I fried a $75 bluetooth chip once ... And other less expensive
> stuff till I figured it out ...
>
> 2. Split power supply with separate batteries for the micro and the
> motor. Without optos you'll need a common ground and at least the cap
> across the motor leads. I've built bots with just the single cap with
> split power supplies, but you might as well add them all if it isn't
> too big a hassle. If you don't add them all, here's my order of
> priority: across the motor leads, across the h-bridge power supplies
> and ground, motor leads and case, h-bridge signal pins and ground.
> Other folks may have a different order of priority. Logic power can be
> a 9V battery or a handful of double or triple A's. Motor power should
> be appropriate to the motor.
>
> 3. Split power supply with opto isolators. This is what I do now. It's
> fairly cheap and simple and maximizes the isolation. No need for a
> common ground, so the two power circuits are totally separate. Make
> sure the optos are fast enough for you PWM. Some Darlington optos are
> too slow for really fast PWM. Strictly speaking you don't need caps
> with this approach. But large spikes can radiate noise. May as well
> add at least a few of the caps.
>
>
> giuliano
>
> On Nov 17, 2009, at 7:31 AM, Nathaniel Lewis wrote:
>
>
>
>
> "don clay" <don...@earthlink.net> Nov 17 01:24PM -0700
>
> ya what they said.
>
> Do you have caps (.1) on your motors already? Do you have big enough gauge
> wire
> going to your motors? Is your battery fully charged & powerful enough? Are
> you
> going into full reverse from full forward?
>
> From: Patrick Luke
>
>
>
>
>
> David Wyland <dcwy...@ix.netcom.com> Nov 17 12:37PM -0800
>
> Do not put caps bigger than 0.1 uF across the motor, or you can fry your
> H bridge.
>
> Dave
>
>
> don clay wrote:
>
>
>
> "don clay" <don...@earthlink.net> Nov 17 01:59PM -0700
>
> Oops! That's what I meant .1 uf.
>
> Date sent: Tue, 17 Nov 2009 12:37:53 -0800
> From: David Wyland <dcwy...@ix.netcom.com>
> To: hbrob...@googlegroups.com
> Subject: Re: [HBRobotics] Power Issue
> Send reply to: hbrob...@googlegroups.com
>
>
>
>
> Bob Smith <bsm...@linuxtoys.org> Nov 17 01:26PM -0800
>
> Nathaniel Lewis wrote:
>> My single board computer on my Stinger robot resets after a full motor
>> reverse. What is the best way to solve this?
>
> Opto-isolators. Caps. Split power supplies.
>
> Is there a way you can avoid a "full motor reverse" and
> still meet your system/design requirements? Awhile ago
> we had a discussion about motor ramping. As an experiment
> you might try ramping your PWM pulse width so there is
> never a sudden change. Take, say, a quarter second to
> ramp from 0 to full forward, from full forward to 0, from
> 0 to full reverse, and from full reverse to 0. Can your
> design take a half second to do a "full motor reverse"?
>
> Bob Smith
>
>
>
> Tom Rokicki <rok...@gmail.com> Nov 17 01:45PM -0800
>
> Ramping helps in the normal case, absolutely, but stall
> current can still cause power sag. The cleanest solution
> is of course separate power supplies for logic and motors.
> This alone, even without optoisolators, if you do your
> grounding carefully, will probably solve the majority of your
> problems. Optoisolators are pretty cheap though.
>
> I do love ramping, though, and the effect it can have on the
> apparent "grace" of a robot.
>
>
> --
> Check out Golly at http://golly.sf.net/
>
>
>
> "Alan Marconett" <KM...@SBCGLOBAL.NET> Nov 17 01:50PM -0800
>
> I'm still wondering about the 630V caps!
>
> Alan KM6VV
>
>
>
>
> giuliano carlini <giul...@carlini.com> Nov 17 02:56PM -0800
>
> I tried 100V caps but the micro reset no matter what I did. So I found
> something on the web about using 300V caps and head to head big
> zeners. Difference in price between 300V and 600V caps at digikey was
> nominal, and I wanted the problem FIXED, so I went with the monster
> caps and, I think, 1500 watt zeners ;->
>
> giuliano
>
>
> On Nov 17, 2009, at 1:50 PM, Alan Marconett wrote:
>
>
>
>
> "Alan Marconett" <KM...@SBCGLOBAL.NET> Nov 17 03:02PM -0800
>
> OK, you must be running a really big motor! What voltage/current?
>
> Alan KM6VV
>
>
>
>
> giuliano carlini <giul...@carlini.com> Nov 17 03:20PM -0800
>
> Just a tiny solarbotics gm8 at 9V into a SN754410 and no load so maybe
> 100ma.
>
> I was putting it through a torture test. 1/2 second forward, 1/2
> second reverse, repeat. No caps, 5 seconds or less to reset. I got a
> few minutes with 100V caps sprinkled all over. 18 hours with the 600's
> and zeners. Might have gone longer, but I needed my notebook to go to
> work.
>
> giuliano
>
> On Nov 17, 2009, at 3:02 PM, Alan Marconett wrote:
>
>
>
>
> "Alan Marconett" <KM...@SBCGLOBAL.NET> Nov 17 03:29PM -0800
>
> That's not all that big. I wouldn't expect the voltage rating matter, as
> long as it was 25V or more (if that). Same goes for the current rating of
> the Zener.
>
> Alan KM6VV
>
>
>
>
> "Nathaniel Lewis" <linux.r...@gmail.com> Nov 17 03:45PM -0800
>
> I don't have any caps on the motors at all, but there is 3000uF across the
> SBC power lines and 1000uF across the battery. I'm guessing I should put
> more on the battery. I would like to use two batteries, but I am using an
> expensive Lithium Ion pack for RC Helicopters and I don't want to get
> another one if there are other options. Another thing I realized was that I
> should use a PID algorithm to smooth the motor operation.
>
> Nathaniel
>
> -----Original Message-----
> From: don clay [mailto:don...@earthlink.net]
> Sent: Tuesday, November 17, 2009 12:24 PM
> To: hbrob...@googlegroups.com
> Subject: Re: [HBRobotics] Power Issue
>
> ya what they said.
>
> Do you have caps (.1) on your motors already? Do you have big enough gauge
> wire
> going to your motors? Is your battery fully charged & powerful enough? Are
> you
> going into full reverse from full forward?
>
> From: Patrick Luke
>
>
>>
>>
>
>
> --
>
> 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=.
>
>
>
>
>
> "Nathaniel Lewis" <linux.r...@gmail.com> Nov 17 03:48PM -0800
>
> Why does the voltage matter so much? I thought that if you had 2x your
> maximum voltage on you caps you be fine as long as you had enough
> microfarads.
> Nathaniel
>
> -----Original Message-----
> From: giuliano carlini [mailto:giul...@carlini.com]
> Sent: Tuesday, November 17, 2009 3:20 PM
> To: hbrob...@googlegroups.com
> Subject: Re: [HBRobotics] Power Issue
>
> Just a tiny solarbotics gm8 at 9V into a SN754410 and no load so maybe
> 100ma.
>
> I was putting it through a torture test. 1/2 second forward, 1/2
> second reverse, repeat. No caps, 5 seconds or less to reset. I got a
> few minutes with 100V caps sprinkled all over. 18 hours with the 600's
> and zeners. Might have gone longer, but I needed my notebook to go to
> work.
>
> giuliano
>
> On Nov 17, 2009, at 3:02 PM, Alan Marconett wrote:
>
>> 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=
>>
>>
>
> --
>
> 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=.
>
>
>
>
>
> e...@okerson.com Nov 17 05:48PM -0600
>
> Back EMF from the motors changing directions could easily go over 100v.
>
> Ed Okerson
>
>
>
>
> "Wayne C. Gramlich" <wa...@gramlich.net> Nov 17 03:59PM -0800
>
>
> Nathaniel Lewis wrote:
>> another one if there are other options. Another thing I realized was that
>> I
>> should use a PID algorithm to smooth the motor operation.
>
> The PID algorithm does not necessarily smooth out motor operation.
> If your PID algorithm is trying to get you towards +X and you feed
> it a new position target of -X, you will get the same motor reverse
> situation.
>
> Adding ramping, where you restrict how fast you can change the
> direction and with of your pulses will help a lot.
>
> Is there some reason why the other battery pack needs to be Lithium
> Ion? The battery pack that feeds your processor can be simple
> alkaline batteries and it will run for a long time before you
> drain the batteries.
>
> -Wayne
>
>
>
> giuliano carlini <giul...@carlini.com> Nov 17 04:00PM -0800
>
> I didn't expect it. But it's easy to replicate.
>
> Arudino, SN754410NE, GM8, 8x.1uF 100V caps, 9V battery and snaps, a
> breadboard, and if desired some leds and resistors to provide visual
> feedback for the signal pins.
>
> Connect the battery to both breadboard power rails.
> Connect arduino power input to breadboard.
> Place the hbridge between the two halves of the breadboard.
> Connect h-bridge pins 8,9 to power, and 4,5,12,13 to ground.
> Place caps from pins 1, 2, 7, 8, 9 to ground.
> Solder caps from each GM8 motor lead to the case, and between the leads.
> Solder 20ga wire from each motor lead to a 4 pin header. Pull the two
> middle pins.
> Push the header into the breadboard for h-bridge pins 3,6
> Wire up the LEDs and resistors if you want to
> Connect arduino 11 to h-bridge 1, arduino 12 to h-bridge 2, and
> arduino 13 to h-bridge 7.
>
> Write and run a simple arduino program that write 5v to digital 11 and
> alternates between writing 5v to digital 12 and 0 to digital 13 for
> half a second; and then writing 0v to digital 12 and 5v to digital 13.
> Add some printlns in setup and at each transition, including the
> current value of millis so you can detect a reset.
>
> Might have got something wrong in the details, I'm typing this in off
> the top of my head, but it's something close to this.
>
> If anyone does this, let me know it works for you ...
>
> giuliano
>
> On Nov 17, 2009, at 3:29 PM, Alan Marconett wrote:
>
>
>
>
> Michael Prados <mpr...@gmail.com> Nov 17 04:04PM -0800
>
> lots of good suggestions, here's a couple others-
>
> 1. Hopefully they are there somewhere anyway, but there should be
> flyback diodes to limit the inductive spikes! In some cases, these
> are already built into the H-Bridge. Spikes from switching inductive
> loads without flyback diodes are the number one factor I hear in
> stories of microcontrollers resetting in mechatronic systems.
>
> http://en.wikipedia.org/wiki/Flyback_diode
>
> 2. DC-DC converters are increasingly cheap and compact, and can go a
> long way to smooth out noise in a wide frequency range. Here's one
> I've been using with an AVR (we used a similar one on Swarm):
>
> http://search.digikey.com/scripts/DkSearch/dksus.dll?Detail&name=PT5101C-ND
>
> -mike
>
>
> --
> sent from my computing machine
>
>
>
> "Alan Marconett" <KM...@SBCGLOBAL.NET> Nov 17 04:10PM -0800
>
> Oh, you're perfectly fine; the cap might be a little bigger. How big are
> 630V caps? The 25v or 50v ceramic .1uF caps are easy to tuck onto the rear
> end of a motor.
>
> Alan KM6VV
>
>
>
>
> giuliano carlini <giul...@carlini.com> Nov 17 04:10PM -0800
>
> A 9V, handful of AAs, or a 7.2 or 9.6 NiCd will do just fine to power
> the micro. Only the motors need the current capacity the LiPo's can
> supply. But, even for the micro use some sort of rechargeable. No
> reason to fill landfills with dead batteries.
>
> Caps across the batteries aren't enough. You need them close to the
> motors and h-bridge.
>
> PIDs are good for not overshooting your target speed or oscillating
> wildly about it. They won't do a slow start. If anything, they tend to
> do a hard start and then slow as they get closer to the target. You
> want to send the output of the PID into a limiter. The limiter will
> set a max limit on the speed change over short bits of time. It should
> let you change speed rapidly, but not full forward to full reverse.
>
> giuliano
>
> On Nov 17, 2009, at 3:45 PM, Nathaniel Lewis wrote:
>
>
>
>
> Topic: HBRC Meeting THIS Wednesday Nov. 18: Machine Vision
>
> Camp Peavy <cpe...@yahoo.com> Nov 17 10:26AM -0800
>
>> I’m looking forward to this presentation. I am working on a vision system
>> and maybe this might help me add functionality and increased performance!
>
> Mike Thompson rocks!
>
>
>
>
> "Alan Marconett" <KM...@SBCGLOBAL.NET> Nov 17 10:35AM -0800
>
> Me too! I've got a CMU3 on request for Christmas!
>
>
>
> Alan KM6VV
>
>
>
> _____
>
> From: Camp Peavy [mailto:cpe...@yahoo.com]
> Sent: Tuesday, November 17, 2009 10:27 AM
> To: hbrob...@googlegroups.com
> Subject: RE: [HBRobotics] HBRC Meeting THIS Wednesday Nov. 18: Machine
> Vision
>
>
>
>
>> I'm looking forward to this presentation. I am working on a vision system
> and maybe this might help me add functionality and increased performance!
>
>
>
> Mike Thompson rocks!
>
>
>
> --
>
> 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=.
>
>
>
>
> --
>
> 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=.
>
Reply all
Reply to author
Forward
0 new messages