tag:blogger.com,1999:blog-4339813531032979196.post8051011253053203885..comments2024-02-28T07:32:59.864+00:00Comments on HydraRaptor: Lessons from the A3977nopheadhttp://www.blogger.com/profile/12801535866788103677noreply@blogger.comBlogger23125tag:blogger.com,1999:blog-4339813531032979196.post-53540167702520030462012-08-22T12:03:01.712+01:002012-08-22T12:03:01.712+01:00Yes that can happen if there isn't enough load...Yes that can happen if there isn't enough load on the PSU to absorb it.nopheadhttps://www.blogger.com/profile/12801535866788103677noreply@blogger.comtag:blogger.com,1999:blog-4339813531032979196.post-8644111071145495222012-08-22T11:37:29.957+01:002012-08-22T11:37:29.957+01:00Very informative!
But I have to warn using as sup...Very informative!<br /><br />But I have to warn using as supply voltage above 25V. This is because the motor generates back EMF, which is conducted trough the syncronous rectification mosfets of the driver chip onto the supply rail. So it is possible that the supply rail gets above 35V, because the energy of the back EMF has to go someware. I so destroyed two A3977 drivers, using a motor voltage of 30V and so as the motor was stopped quickly the voltage on the supply rail raised up to 42V, destroying the H-bridges of the driver. Although the supply rail had 4700uF capacitance to dampen the voltage spike. A sufficent method to prevent that is using a power zener diode. I'll use it in a improved future designAnonymousnoreply@blogger.comtag:blogger.com,1999:blog-4339813531032979196.post-36701214176023222722011-03-28T14:29:38.392+01:002011-03-28T14:29:38.392+01:00Great articles and I like the oscilloscope views. ...Great articles and I like the oscilloscope views. Your scope is an analog? What were the values you concluded for setting your controller values? I saw your motor tables on another article you did about the A3977 but I am having difficulty converting to PFD, REF, RC1&RC2 for the MakerBot v3.3 steppers. P.S. I did post to another of your articles. I could not find contact information for you even though you are posting in many venues. Regards...BAShttps://www.blogger.com/profile/11712266694307971483noreply@blogger.comtag:blogger.com,1999:blog-4339813531032979196.post-50716419301885627632011-03-07T18:21:44.175+00:002011-03-07T18:21:44.175+00:00Maybe a bad ground or poor decoupling causes noise...Maybe a bad ground or poor decoupling causes noise on the step signal when the current is high.nopheadhttps://www.blogger.com/profile/12801535866788103677noreply@blogger.comtag:blogger.com,1999:blog-4339813531032979196.post-52833616530715233512011-03-07T18:00:08.015+00:002011-03-07T18:00:08.015+00:00Hi! Great post! I got some trouble with this alleg...Hi! Great post! I got some trouble with this allegro IC. Over a certein value of REF voltage the motor start to become random! (turn without step command). What can be?<br />Thank you!Anonymousnoreply@blogger.comtag:blogger.com,1999:blog-4339813531032979196.post-68498049126385057012010-10-11T00:54:15.856+01:002010-10-11T00:54:15.856+01:00Forgot to paste a link to the similar PIC mcu driv...Forgot to paste a link to the similar PIC mcu driver board here, i think more or less close to the other driver linked above. http://www.piclist.com/techref/io/stepper/linistep/index.htm Was reading some tips on increasing its motoring currents.<br />For the a3977 version 3.3: how to calculate the 50k pot adjustment for a motor (got ~3ohm http://www.reprap.org/wiki/StepperMotor#Keling_KL23H51-24-08B)<br />Tyvm again btw and sry for the previous post, i see it has nothing to do with that transistor, just my brain neurons falling between the two webpages.Anonymousnoreply@blogger.comtag:blogger.com,1999:blog-4339813531032979196.post-15698615379860873012010-10-11T00:35:17.473+01:002010-10-11T00:35:17.473+01:00The A3799 will supply an absolute maximum of 2.5A ...The A3799 will supply an absolute maximum of 2.5A but it will need a good heatsink to achieve that. Without a heatsink the maximum is somewhere between 1.5 and 2A.<br /><br />You can't increase that by adding transistors.nopheadhttps://www.blogger.com/profile/12801535866788103677noreply@blogger.comtag:blogger.com,1999:blog-4339813531032979196.post-83582060631742445292010-10-11T00:21:03.887+01:002010-10-11T00:21:03.887+01:00Could this board (i think currently it is 3.3 vers...Could this board (i think currently it is 3.3 version at makerbot) take smth like a 2.2 or 3.0 A source transistor? Or is that too high versus its 1.0 or 1.5A not sure which is atm. Or it will just fry it all over. Btw, tyvm for all the info in your blogs.Anonymousnoreply@blogger.comtag:blogger.com,1999:blog-4339813531032979196.post-51925330126654755452009-08-04T22:51:50.121+01:002009-08-04T22:51:50.121+01:00Fair enough. You have a point.
KhiralyFair enough. You have a point.<br /><br />KhiralyAnonymousnoreply@blogger.comtag:blogger.com,1999:blog-4339813531032979196.post-21309469651803395682009-08-04T15:53:08.629+01:002009-08-04T15:53:08.629+01:00If I was trying to design the cheapest system poss...If I was trying to design the cheapest system possible I would use unipolar motors, with four cheap FETs per motor and a single MCU fast enough to do the chopping.<br /><br />Maybe that would work out at about $3 per drive. The Allegro solution would be more like $6 per drive, but less effort and better performance. It also has overload protection, etc.<br /><br />As soon as you go to bipolar you need at least 12 transistors per drive and have to worry about shoot through protection (both transistors must never be on at the same time), so a single chip solution becomes much more attractive.<br /><br />But using two big chips and a micro to do the same job as single chip that is cheaper doesn't make any sense to me unless you really need 3A instead of 2.5A.nopheadhttps://www.blogger.com/profile/12801535866788103677noreply@blogger.comtag:blogger.com,1999:blog-4339813531032979196.post-24825867812972842722009-08-04T15:28:49.718+01:002009-08-04T15:28:49.718+01:00> But you still need two H-bridges per motor, s...> But you still need two H-bridges per motor, so <br />> why not use a chip that has two H-bridges and <br />> the chopper built in?<br /><br />You know when you have a hammer, you see the world as nails;)<br /><br />No, I dont think an arduino is enough, but there is dsPic at microchip, or browse what atmel has oriented to signal processing.<br /><br />I have seen servo drive using a microchip, I have an electric bike (brushless DC motor), where a pic is inside with dozens of FET, and I even saw an inverter drive project for AC motors. Maybe there are some common things in them.<br /><br />But having a target circuit designed for the job has its advantage I agree.<br /><br /> KhiralyAnonymousnoreply@blogger.comtag:blogger.com,1999:blog-4339813531032979196.post-38252591371066617822009-08-04T12:26:06.130+01:002009-08-04T12:26:06.130+01:00Yes a micro with a fast ADC could do the chopping ...Yes a micro with a fast ADC could do the chopping itself. I don't think the Arduno would be up to it, something like an ST ARM Cortex would be ideal. But you still need two H-bridges per motor, so why not use a chip that has two H-bridges and the chopper built in?<br /><br />It gets more practical with a unipolar motor as you only need 4 transistors and a micro, but the motor gives less performance.nopheadhttps://www.blogger.com/profile/12801535866788103677noreply@blogger.comtag:blogger.com,1999:blog-4339813531032979196.post-73099662861445873912009-08-04T11:51:25.784+01:002009-08-04T11:51:25.784+01:00> constant off time choppers
If you have a micr...> constant off time choppers<br />If you have a microcontroller it is not necessary to have a constant off time chopper. (or this 20us comes from the LMD18245T limitation? ie, it has a minimum switching time)<br /><br />I admit that having a microcontoller is a way bigger project then simply buy an off the shelf product. I just wanted raise your attention to this possible solution. (I dont want to convince you at all)<br /><br />Best regards, <br /> KhiralyAnonymousnoreply@blogger.comtag:blogger.com,1999:blog-4339813531032979196.post-51612781892629779252009-08-03T23:32:09.514+01:002009-08-03T23:32:09.514+01:00The two LMD18245T plus an Arduino cost a lot more ...The two LMD18245T plus an Arduino cost a lot more than an A3977 that does the same job.<br /><br />They are still constant off time choppers and will have all the same distortions when run under the same conditions. Their advantage is a bit more current and twice the resolution. There is an Allegro single chip with twice the resolution as well.<br /><br />Size is an issue because I am making an extruder controller that mounts on the extruder, plus PCB is costed by the area.nopheadhttps://www.blogger.com/profile/12801535866788103677noreply@blogger.comtag:blogger.com,1999:blog-4339813531032979196.post-66779427077159441582009-08-03T22:32:16.588+01:002009-08-03T22:32:16.588+01:00> Very expensive, $85
Yepp, but the LMD18245T ...> Very expensive, $85<br /><br />Yepp, but the LMD18245T part (DMOS full bridge) costs alone 15.47$ and it requires 2/motor. <br />However this is responsible for the nice curve....<br />Maybe its not "that" expensive this way.<br /><br />http://parts.digikey.com/1/parts/577573-ic-h-bridge-3a-55v-220-lmd18201t-nopb.html<br /><br />So Im thinking replacing this expensive part, and it would suddenly turn into a real option.<br /><br />>and very big<br />Cant be an issue... If it really that big, we can use smd parts.<br /><br />KhiralyAnonymousnoreply@blogger.comtag:blogger.com,1999:blog-4339813531032979196.post-91705538738800335892009-08-03T13:22:05.092+01:002009-08-03T13:22:05.092+01:00Very expensive, $85, and very big.Very expensive, $85, and very big.nopheadhttps://www.blogger.com/profile/12801535866788103677noreply@blogger.comtag:blogger.com,1999:blog-4339813531032979196.post-26603066191793496242009-08-03T13:03:04.591+01:002009-08-03T13:03:04.591+01:00nophead: Have you evaluated the AVR based open sou...nophead: Have you evaluated the AVR based open source stepper motor controller?:<br />http://www.avrstmd.com/<br /><br />It seems they have microstepping, with nice waveform and 3A/coil. Maybe Im mistaken, Im just curious...<br /><br />KhiralyAnonymousnoreply@blogger.comtag:blogger.com,1999:blog-4339813531032979196.post-45050870157655433642009-08-03T04:58:37.875+01:002009-08-03T04:58:37.875+01:00I just finished reading all of your posts from the...I just finished reading all of your posts from the very beginning (well, skipping the posts I deemed boring) and I must say, it's easy to see that you're very dedicated to getting things working and improving them, so good job. It's people like you that make the world go round.Joeyhttps://www.blogger.com/profile/06084982197916412528noreply@blogger.comtag:blogger.com,1999:blog-4339813531032979196.post-16491601354193543562009-08-02T13:50:17.631+01:002009-08-02T13:50:17.631+01:00Also it should be possible to calculate the compon...Also it should be possible to calculate the component values from the motor resistance, which you can measure with a multimeter. <br /><br />I have been doing the sums but can't make the figures match the measured values at the moment. Maybe I have got something wrong somewhere, or perhaps the eddy current losses account for the discrepancy. Still I might be able to publish a worst case formula.nopheadhttps://www.blogger.com/profile/12801535866788103677noreply@blogger.comtag:blogger.com,1999:blog-4339813531032979196.post-36936621230725706542009-08-02T13:40:45.627+01:002009-08-02T13:40:45.627+01:00Well hopefully a bit more chance having read this,...Well hopefully a bit more chance having read this, but yes it is not as straight forward as I would like. However, you never get worse performance than a non microstepping drive would achieve.<br /><br />It should be possible to do a software implementation of micro-stepping that self tunes and / or tells you when it can't acheive the required current waveform.<br /><br />Some of the problems come from it being a constant off time chopper. Constant frequency would probably be less dependent on the motor.nopheadhttps://www.blogger.com/profile/12801535866788103677noreply@blogger.comtag:blogger.com,1999:blog-4339813531032979196.post-27526173444387844282009-08-02T06:35:22.350+01:002009-08-02T06:35:22.350+01:00Whew! What chance has your average ten thumbed ama...Whew! What chance has your average ten thumbed amateur got to achieve high performance using those sorts of chips using surplus steppers without an oscilloscope and a clone of your brain? :-(Forrest Higgshttps://www.blogger.com/profile/17208965471464716174noreply@blogger.comtag:blogger.com,1999:blog-4339813531032979196.post-87546300831622734022009-07-30T08:58:54.641+01:002009-07-30T08:58:54.641+01:00Hi Phil,
My blogging has got behind real time by...Hi Phil,<br /> My blogging has got behind real time by about a month, so I am doing my best to catch up.<br /><br />I have just made the jump to SMT at home. I find soldering is no problem with paste and a small oven or a hot plate with PID control. I will be blogging details soon. You can buy suitable ovens in places like ALDI cheaper than a decent soldering iron.<br /><br />Also MakerBot have now started selling assembled boards, so they are no longer limited to easy to assemble chips.<br /><br />Any "details" that I can clarify?nopheadhttps://www.blogger.com/profile/12801535866788103677noreply@blogger.comtag:blogger.com,1999:blog-4339813531032979196.post-40275685739151584582009-07-30T08:38:23.751+01:002009-07-30T08:38:23.751+01:00Wow, once again very impressive. I thought you mus...Wow, once again very impressive. I thought you must have been on holiday as you havent posted for a while - silly me to doubt that you would be beavering away in the background!<br /><br />I have to say I wish I understood all of the details but it is a real insight into the way we all SHOULD be working.<br /><br />Do you think this kind of chip will make its way into a new stepper motor driver or has it already been discounted? Ive avoided SMC as my soldering experience is limited but I can see the time when I have to make the jump!Anonymoushttps://www.blogger.com/profile/04321669798401086867noreply@blogger.com