[Gumstix-users] Power Supply Filtering

 [Gumstix-users] Power Supply Filtering From: Grahame Jordan - 2006-06-14 06:27:33 ```Hi, This is an electronics question. I hope someone can help. I have a problem with my motor controller. http://www.pololu.com/products/pololu/0205/ I am not actually using a motor but a resisitive load. When the PWM really starts working I get an increadable amount of noise on my power lines such that it locks up the gumstix. More than 4V peak to peak 3.686400 MHz / 375 = 9830.4 Hz - Maximum is 10Khz for VHN3SP30 Which just happens to the the frequency of the noise. The higher the current I draw the more noise. This is obvious. I can aleviate the problem greatly by putting a big electrolytic capacitor across power to ground 4700uF say. Now the noise is ~ 1V - 0.6V peak to peak and the gumstix runs OK. Is there a better way of filtering this noise Thanks -- Grahame Jordan Glass Expansion 15 Batman St West Melbourne VIC 3003 PH: +61 3 9320 1111 FAX: +61 3 9320 1112 ```

 [Gumstix-users] Power Supply Filtering From: Grahame Jordan - 2006-06-14 06:27:33 ```Hi, This is an electronics question. I hope someone can help. I have a problem with my motor controller. http://www.pololu.com/products/pololu/0205/ I am not actually using a motor but a resisitive load. When the PWM really starts working I get an increadable amount of noise on my power lines such that it locks up the gumstix. More than 4V peak to peak 3.686400 MHz / 375 = 9830.4 Hz - Maximum is 10Khz for VHN3SP30 Which just happens to the the frequency of the noise. The higher the current I draw the more noise. This is obvious. I can aleviate the problem greatly by putting a big electrolytic capacitor across power to ground 4700uF say. Now the noise is ~ 1V - 0.6V peak to peak and the gumstix runs OK. Is there a better way of filtering this noise Thanks -- Grahame Jordan Glass Expansion 15 Batman St West Melbourne VIC 3003 PH: +61 3 9320 1111 FAX: +61 3 9320 1112 ```
 Re: [Gumstix-users] Power Supply Filtering From: Dave Hylands - 2006-06-14 07:27:41 ```Hi Grahame, > This is an electronics question. I hope someone can help. > > I have a problem with my motor controller. > http://www.pololu.com/products/pololu/0205/ > > I am not actually using a motor but a resisitive load. When the PWM really > starts working I get an increadable amount of noise on my power lines such > that it locks up the gumstix. More than 4V peak to peak > > 3.686400 MHz / 375 = 9830.4 Hz - Maximum is 10Khz for VHN3SP30 > > Which just happens to the the frequency of the noise. > The higher the current I draw the more noise. This is obvious. > > I can aleviate the problem greatly by putting a big electrolytic capacitor > across power to ground 4700uF say. Now the noise is ~ 1V - 0.6V peak to peak > and the gumstix runs OK. > > Is there a better way of filtering this noise There are several techniques that all help to reduce the amount of noise. 1 - Regardless of whether you're using a single battery or multiple batteries, you should run a ground wire and power wire from the battery to the motor + and - on the H-Bridge. Run a separate ground and power wire from the battery to the gumstix. This is referred to as a star configuration. 2 - If you have two batteries, make sure that the grounds are connected. 3 - Twist the power and ground wires together. 4 - Use optical isolation. This requires a second battery for sure. In this case you don't want a common ground. You need to be sure that the optoisolators will switch fast enough for whatever PWM signal you're sending. Obviously, you power source needs to be able to deliver the appropriate amount of current. What type of power source are you using? If batteries, what type of batteries? The cheapy carbon-zinc batteries are absolutely terrible. They have a very high internal resistance and will experience big voltage sags when trying to draw larger amounnts of current from them. NiMH or NiCAD are much better for this (LiPo or Li-Ion are also pretty good). -- Dave Hylands Vancouver, BC, Canada http://www.DaveHylands.com/ ```
 Re: [Gumstix-users] Power Supply Filtering From: Grahame Jordan - 2006-06-15 00:49:55 ```Hi Dave, I am pulling 6A on the single model and up to 19A on the dual model (heating & cooling) These are only resistive loads, Peltier and/or Resistor. The motor driver is perfect for controlling temperature, etc. I am using a laptop style power supply on the single model and a laboratory PS on the dual model. I will look into splitting the grounds and see if it helps. My 5V line is supplied by a ADP3050AR-5 which being a chopper would expect that it should remove all noise, but that may be on the ground. It has a filter on the output to filter out the 200KHz switching frequency, maybe I can add a filter to get rid of the 9.8KHz. Thanks Grahame Jordan On Wed, 14 Jun 2006 17:27, Dave Hylands wrote: > Hi Grahame, > > > This is an electronics question. I hope someone can help. > > > > I have a problem with my motor controller. > > http://www.pololu.com/products/pololu/0205/ > > > > I am not actually using a motor but a resisitive load. When the PWM > > really starts working I get an increadable amount of noise on my power > > lines such that it locks up the gumstix. More than 4V peak to peak > > > > 3.686400 MHz / 375 = 9830.4 Hz - Maximum is 10Khz for VHN3SP30 > > > > Which just happens to the the frequency of the noise. > > The higher the current I draw the more noise. This is obvious. > > > > I can aleviate the problem greatly by putting a big electrolytic > > capacitor across power to ground 4700uF say. Now the noise is ~ 1V - 0.6V > > peak to peak and the gumstix runs OK. > > > > Is there a better way of filtering this noise > > There are several techniques that all help to reduce the amount of noise. > > 1 - Regardless of whether you're using a single battery or multiple > batteries, you should run a ground wire and power wire from the > battery to the motor + and - on the H-Bridge. Run a separate ground > and power wire from the battery to the gumstix. This is referred to as > a star configuration. > > 2 - If you have two batteries, make sure that the grounds are connected. > > 3 - Twist the power and ground wires together. > > 4 - Use optical isolation. This requires a second battery for sure. In > this case you don't want a common ground. You need to be sure that the > optoisolators will switch fast enough for whatever PWM signal you're > sending. > > Obviously, you power source needs to be able to deliver the > appropriate amount of current. What type of power source are you > using? > > If batteries, what type of batteries? The cheapy carbon-zinc batteries > are absolutely terrible. They have a very high internal resistance and > will experience big voltage sags when trying to draw larger amounnts > of current from them. NiMH or NiCAD are much better for this (LiPo or > Li-Ion are also pretty good). ```
 Re: [Gumstix-users] Power Supply Filtering From: Leon Heller - 2006-06-14 08:27:03 ```----- Original Message ----- From: "Grahame Jordan" To: Sent: Wednesday, June 14, 2006 7:27 AM Subject: [Gumstix-users] Power Supply Filtering > Hi, > > This is an electronics question. I hope someone can help. > > I have a problem with my motor controller. > http://www.pololu.com/products/pololu/0205/ > > I am not actually using a motor but a resisitive load. When the PWM > really > starts working I get an increadable amount of noise on my power lines such > that it locks up the gumstix. More than 4V peak to peak > > 3.686400 MHz / 375 = 9830.4 Hz - Maximum is 10Khz for VHN3SP30 > > Which just happens to the the frequency of the noise. > The higher the current I draw the more noise. This is obvious. > > I can aleviate the problem greatly by putting a big electrolytic capacitor > across power to ground 4700uF say. Now the noise is ~ 1V - 0.6V peak to > peak > and the gumstix runs OK. > > Is there a better way of filtering this noise That's still a lot of noise. I'd use a proper low-pass filter with an inductor and two capacitors. A transient suppressor (AVX TransGuard) would help, also. Leon ```
 Re: [Gumstix-users] Power Supply Filtering From: Adam McLeod - 2006-06-14 20:46:35 ```It sounds like your power supply can't source as much as your motor controller is trying to sink. Can you increase the resistive load to see if the noise decreases? On 6/14/06, Grahame Jordan wrote: > I am not actually using a motor but a resisitive load. When the PWM really > starts working I get an increadable amount of noise on my power lines such > that it locks up the gumstix. More than 4V peak to peak ```
 Re: [Gumstix-users] Power Supply Filtering From: Grahame Jordan - 2006-06-15 01:00:17 ```Hi Adam, My leads from the PS were getting a bit warm at 16A :) so I increased the gauge of wire. It was quite suprising that I was getting 2A+ drop on that wire, so that wasn't helping. However I think that the noise is inherent with the switching of high currents. My PS is a 20A supply and the problem occurs at 4-6A and only gets worse as the current increases. The large capacitors help a lot but it still leaves a hairy (0.6 - 1V p-p) 12V. Thanks Grahame On Thu, 15 Jun 2006 06:46, Adam McLeod wrote: > It sounds like your power supply can't source as much as your motor > controller is trying to sink. Can you increase the resistive load to > see if the noise decreases? > > On 6/14/06, Grahame Jordan wrote: > > I am not actually using a motor but a resisitive load. When the PWM > > really starts working I get an increadable amount of noise on my power > > lines such that it locks up the gumstix. More than 4V peak to peak > > _______________________________________________ > gumstix-users mailing list > gumstix-users@... > https://lists.sourceforge.net/lists/listinfo/gumstix-users -- Grahame Jordan Glass Expansion 15 Batman St West Melbourne VIC 3003 PH: +61 3 9320 1111 FAX: +61 3 9320 1112 ```
 Re: [Gumstix-users] Power Supply Filtering From: Dave Hylands - 2006-06-15 01:17:37 ```Hi Grahame, > My PS is a 20A supply and the problem occurs at 4-6A and only gets worse as > the current increases. The large capacitors help a lot but it still leaves a > hairy (0.6 - 1V p-p) 12V. So it's possible that even though it's a 20A supply, it may not be able to deal with the instantaneous inrush of 4-6A, especially with a resistive load. Inductive loads are much more forgiving this way (I'm not an analog guy - so might be conjecture on my part). I've heard of people adding inductance when driving motors with really low inductance to reduce the current draw from the PS. It has something to do with recirculating currents (I think). With a pure resistive load, the PS has to be able to deliver the huge load. When a motor has inductance it acts like a small generator during the portion that it's not being driven, so you don't need as much current from the PS. This is the source of some of my information about PWM. It's very motor centric, but probably useful nonetheless. http://www.4qdtec.com/pwm-01.html -- Dave Hylands Vancouver, BC, Canada http://www.DaveHylands.com/ ```
 Re: [Gumstix-users] Power Supply Filtering From: Adam McLeod - 2006-06-15 04:09:36 ```I agree. Given that everything shares a ground, I find it unlikely that EMI would create a disturbance that the filter caps on the Gumstix couldn't take care of. An inductor in series with the load will act as a momentary current source if the power supply current drops, doing essentially the same thing as a capacitor in parallel with the load (apologies if I'm stating the obvious). Since both components impede current/voltage from dropping quickly, either or any combination of them are make a form of low-pass filter. An inductor/cap combination as Leon reccomends, or just an inductor might give better performance than one massive cap. Adam On 6/14/06, Dave Hylands wrote: > So it's possible that even though it's a 20A supply, it may not be > able to deal with the instantaneous inrush of 4-6A, especially with a > resistive load. Inductive loads are much more forgiving this way (I'm > not an analog guy - so might be conjecture on my part). ```
 Re: [Gumstix-users] Power Supply Filtering From: ken staton - 2006-06-17 04:13:28 ```On 6/13/06, Grahame Jordan wrote: > Hi, > > This is an electronics question. I hope someone can help. > > I have a problem with my motor controller. > http://www.pololu.com/products/pololu/0205/ > > I am not actually using a motor but a resisitive load. When the PWM really > starts working I get an increadable amount of noise on my power lines such > that it locks up the gumstix. More than 4V peak to peak > > 3.686400 MHz / 375 = 9830.4 Hz - Maximum is 10Khz for VHN3SP30 > > Which just happens to the the frequency of the noise. > The higher the current I draw the more noise. This is obvious. > > I can aleviate the problem greatly by putting a big electrolytic capacitor > across power to ground 4700uF say. Now the noise is ~ 1V - 0.6V peak to peak > and the gumstix runs OK. > > Is there a better way of filtering this noise This topic keeps coming up. I've drawn some pictures to illustrate the problem: http://pubs.gumstix.com/documents/app_note_2.pdf Regards, Ken ```
 Re: [Gumstix-users] Power Supply Filtering From: Lawrence Harris - 2006-06-17 05:30:14 Attachments: Message as HTML ```One solution if the power spikes are short is to put smaller bypass capacitors as close to the load as possible. These provide the current when it's needed most and then 'recharge' from the main supply between spikes. You might experiment with 0.1 to 1 uF capacitors at the gumstix power connector and at the power connector to your load. Small capacitors liberally used at all the power/ground 'load' points allow a great deal of energy to be stored without the need for large bulky capacitors but work best if the load spikes are very short like the switching load of electronic circuits. Lawrence ken staton wrote: > On 6/13/06, Grahame Jordan wrote: > >> Hi, >> >> This is an electronics question. I hope someone can help. >> >> I have a problem with my motor controller. >> http://www.pololu.com/products/pololu/0205/ >> >> I am not actually using a motor but a resisitive load. When the PWM really >> starts working I get an increadable amount of noise on my power lines such >> that it locks up the gumstix. More than 4V peak to peak >> >> 3.686400 MHz / 375 = 9830.4 Hz - Maximum is 10Khz for VHN3SP30 >> >> Which just happens to the the frequency of the noise. >> The higher the current I draw the more noise. This is obvious. >> >> I can aleviate the problem greatly by putting a big electrolytic capacitor >> across power to ground 4700uF say. Now the noise is ~ 1V - 0.6V peak to peak >> and the gumstix runs OK. >> >> Is there a better way of filtering this noise >> > > > This topic keeps coming up. I've drawn some pictures to illustrate the problem: > http://pubs.gumstix.com/documents/app_note_2.pdf > > Regards, > Ken > > > _______________________________________________ > gumstix-users mailing list > gumstix-users@... > https://lists.sourceforge.net/lists/listinfo/gumstix-users > ```