Reply by John Larkin April 15, 20132013-04-15
On Tue, 09 Apr 2013 17:37:22 -0500, Tim Wescott <tim@seemywebsite.com>
wrote:

>So, I'm working on a motor drive board for a customer. Due to a desire >for performance improvements within the existing space, we're going from >the traditionally-accepted brute-force power supply to a switcher. > >I have this nice board and software set that's carefully designed to >treat the power supply with kid gloves -- and it works very well with the >power supply that we used for prototyping. > >But for space reasons they went out and found what appears to be the >smallest possible power supply with the apparent specifications needed >(24V and 10A): > >http://www.powerconversion.com/assets/lps200-m_ds_1226129797.pdf > >It's not that this power supply sucks, per se. It's just that things >that work just fine with the original supply (A TDK/Lambda SWS300-24) >make the Emerson supply either exceedingly unhappy or downright dead. >Worse, I'm not sure what's taking out the power supplies. > >So -- > >Does anyone have any mileage with these Emerson supplies? After I told >the customer "why yes, that should do" I began to notice the name >"Emerson" on this group, and the view of their stuff seems to be pretty >dim. I'd like to know if you feel these are OK, or comprehensively bad, >or what. > >Also, I've carefully designed things so that any time the motor is >generating the power goes into a resistor and not into raising the motor >supply rail (and thus backfeeding the supply). But I did notice that >doing so kills one of these Emerson supplies (the TDK supply mostly takes >it in stride, or if you're really egregious about things it turns off >until you cycle power). > >Do you know of anything else that would kill one of these supplies? >While I've eliminated the backfeeding problem, the board still has a >current draw that can vary from a few tens of milliamps to ten amps over >the space of less than a second -- if it would make the supply happier I >can change this so that whenever the motor is turning the board is always >drawing some minimum power. Efficiency will suffer -- but we're not >after efficiency, we're after performance in a small space. > >Also, if anyone knows of a supply that's the same size or not much >bigger, but comes from a manufacturer that's known to make bulletproof >stuff, I'd appreciate it.
I suppose Thompson has dried up on ideas, so consider this: https://dl.dropboxusercontent.com/u/53724080/Circuits/Power/BackFeed.JPG -- John Larkin Highland Technology, Inc jlarkin at highlandtechnology dot com http://www.highlandtechnology.com Precision electronic instrumentation Picosecond-resolution Digital Delay and Pulse generators Custom laser drivers and controllers Photonics and fiberoptic TTL data links VME thermocouple, LVDT, synchro acquisition and simulation
Reply by John Larkin April 14, 20132013-04-14
On Sat, 13 Apr 2013 21:33:03 -0700, josephkk <joseph_barrett@sbcglobal.net>
wrote:

>On Wed, 10 Apr 2013 10:11:36 -0700, John Larkin ><jlarkin@highlandtechnology.com> wrote: > >>On Wed, 10 Apr 2013 08:37:19 -0700, Jim Thompson >><To-Email-Use-The-Envelope-Icon@On-My-Web-Site.com> wrote: >> >>>On Tue, 9 Apr 2013 21:59:16 -0700 (PDT), Bill Sloman >>><bill.sloman@ieee.org> wrote: >>> >>>>On 10 Apr, 13:08, John Larkin >>>><jjlar...@highNOTlandTHIStechnologyPART.com> wrote: >>>>> On Tue, 09 Apr 2013 21:43:44 -0500, Tim Wescott <t...@seemywebsite.please> wrote: >>>[snip]. >>>>> >>>>> >When the motor starts to draw power the motor rail drops: when it drops >>>>> >down to match the power supply voltage things are switched on again and >>>>> >it draws power from the supply. >>>>> >>>>> That sounds a little tricky. You could just hang an active zener across the >>>>> power supply output to clamp at, say, 25 volts. That might keep the ps from >>>>> crowbarring, too. >>>>> >>>>> I'm sure Jim would be delighted to design an accurate, 10 amp active zener >>>>> circuit for you. >>> >>>More demented snarkiness. >>> >>>But rather trivial to design, even with off-the-shelf discrete's... >>>think a TL431 plus a power PNP. >> >>Show us. > >Crikey Larkin, even i can see how to do that. Just how dumb are you? > >?-)
I want to see Thompson do it. He sometimes does hilarious stuff when he goes off-chip. -- John Larkin Highland Technology Inc www.highlandtechnology.com jlarkin at highlandtechnology dot com Precision electronic instrumentation Picosecond-resolution Digital Delay and Pulse generators Custom timing and laser controllers Photonics and fiberoptic TTL data links VME analog, thermocouple, LVDT, synchro, tachometer Multichannel arbitrary waveform generators
Reply by Jim Thompson April 14, 20132013-04-14
On Sat, 13 Apr 2013 21:33:03 -0700, josephkk
<joseph_barrett@sbcglobal.net> wrote:

>On Wed, 10 Apr 2013 10:11:36 -0700, John Larkin ><jlarkin@highlandtechnology.com> wrote: > >>On Wed, 10 Apr 2013 08:37:19 -0700, Jim Thompson >><To-Email-Use-The-Envelope-Icon@On-My-Web-Site.com> wrote: >> >>>On Tue, 9 Apr 2013 21:59:16 -0700 (PDT), Bill Sloman >>><bill.sloman@ieee.org> wrote: >>> >>>>On 10 Apr, 13:08, John Larkin >>>><jjlar...@highNOTlandTHIStechnologyPART.com> wrote: >>>>> On Tue, 09 Apr 2013 21:43:44 -0500, Tim Wescott <t...@seemywebsite.please> wrote: >>>[snip]. >>>>> >>>>> >When the motor starts to draw power the motor rail drops: when it drops >>>>> >down to match the power supply voltage things are switched on again and >>>>> >it draws power from the supply. >>>>> >>>>> That sounds a little tricky. You could just hang an active zener across the >>>>> power supply output to clamp at, say, 25 volts. That might keep the ps from >>>>> crowbarring, too. >>>>> >>>>> I'm sure Jim would be delighted to design an accurate, 10 amp active zener >>>>> circuit for you. >>> >>>More demented snarkiness. >>> >>>But rather trivial to design, even with off-the-shelf discrete's... >>>think a TL431 plus a power PNP. >> >>Show us. > >Crikey Larkin, even i can see how to do that. Just how dumb are you? > >?-)
What can you say... John "snarky demented NOLA white trash" Larkin always responds that way. One of Slowman's "peers" >:-} ...Jim Thompson -- | James E.Thompson | mens | | Analog Innovations | et | | Analog/Mixed-Signal ASIC's and Discrete Systems | manus | | Phoenix, Arizona 85048 Skype: Contacts Only | | | Voice:(480)460-2350 Fax: Available upon request | Brass Rat | | E-mail Icon at http://www.analog-innovations.com | 1962 | I love to cook with wine. Sometimes I even put it in the food.
Reply by josephkk April 14, 20132013-04-14
On Wed, 10 Apr 2013 10:11:36 -0700, John Larkin
<jlarkin@highlandtechnology.com> wrote:

>On Wed, 10 Apr 2013 08:37:19 -0700, Jim Thompson ><To-Email-Use-The-Envelope-Icon@On-My-Web-Site.com> wrote: > >>On Tue, 9 Apr 2013 21:59:16 -0700 (PDT), Bill Sloman >><bill.sloman@ieee.org> wrote: >> >>>On 10 Apr, 13:08, John Larkin >>><jjlar...@highNOTlandTHIStechnologyPART.com> wrote: >>>> On Tue, 09 Apr 2013 21:43:44 -0500, Tim Wescott =
<t...@seemywebsite.please> wrote:
>>[snip]. >>>> >>>> >When the motor starts to draw power the motor rail drops: when it =
drops
>>>> >down to match the power supply voltage things are switched on again=
and
>>>> >it draws power from the supply. >>>> >>>> That sounds a little tricky. You could just hang an active zener =
across the
>>>> power supply output to clamp at, say, 25 volts. That might keep the =
ps from
>>>> crowbarring, too. >>>> >>>> I'm sure Jim would be delighted to design an accurate, 10 amp active=
zener
>>>> circuit for you. >> >>More demented snarkiness. >> >>But rather trivial to design, even with off-the-shelf discrete's... >>think a TL431 plus a power PNP. > >Show us.
Crikey Larkin, even i can see how to do that. Just how dumb are you? ?-)
Reply by josephkk April 14, 20132013-04-14
On Tue, 09 Apr 2013 17:37:22 -0500, Tim Wescott <tim@seemywebsite.com>
wrote:

>So, I'm working on a motor drive board for a customer. Due to a desire=20 >for performance improvements within the existing space, we're going from=
=20
>the traditionally-accepted brute-force power supply to a switcher. > >I have this nice board and software set that's carefully designed to=20 >treat the power supply with kid gloves -- and it works very well with =
the=20
>power supply that we used for prototyping. > >But for space reasons they went out and found what appears to be the=20 >smallest possible power supply with the apparent specifications needed=20 >(24V and 10A): > >http://www.powerconversion.com/assets/lps200-m_ds_1226129797.pdf > >It's not that this power supply sucks, per se. It's just that things=20 >that work just fine with the original supply (A TDK/Lambda SWS300-24)=20 >make the Emerson supply either exceedingly unhappy or downright dead. =20 >Worse, I'm not sure what's taking out the power supplies. > >So -- > >Does anyone have any mileage with these Emerson supplies? After I told=20 >the customer "why yes, that should do" I began to notice the name=20 >"Emerson" on this group, and the view of their stuff seems to be pretty=20 >dim. I'd like to know if you feel these are OK, or comprehensively bad,=
=20
>or what. > >Also, I've carefully designed things so that any time the motor is=20 >generating the power goes into a resistor and not into raising the motor=
=20
>supply rail (and thus backfeeding the supply). But I did notice that=20 >doing so kills one of these Emerson supplies (the TDK supply mostly =
takes=20
>it in stride, or if you're really egregious about things it turns off=20 >until you cycle power). > >Do you know of anything else that would kill one of these supplies? =20 >While I've eliminated the backfeeding problem, the board still has a=20 >current draw that can vary from a few tens of milliamps to ten amps over=
=20
>the space of less than a second -- if it would make the supply happier I=
=20
>can change this so that whenever the motor is turning the board is =
always=20
>drawing some minimum power. Efficiency will suffer -- but we're not=20 >after efficiency, we're after performance in a small space. > >Also, if anyone knows of a supply that's the same size or not much=20 >bigger, but comes from a manufacturer that's known to make bulletproof=20 >stuff, I'd appreciate it.
I have seen problems before when a load has a serious content at the switcher frequency, or one of them is a near harmonic of the other. Both tend to misbehave. The available fixes are a different switcher, a different motor controller frequency, or a heavy LC trap. YMMV ?-)
Reply by Jamie April 10, 20132013-04-10
mroberds@att.net wrote:

> Tim Wescott <tim@seemywebsite.com> wrote: > >>Due to a desire for performance improvements within the existing >>space, we're going from the traditionally-accepted brute-force power >>supply to a switcher. > > > A few years back, I was doing software for a device that had some small > (50 W or so) motors with integrated controllers; one end of the motor > had a shaft and the other end had connections for 24 V DC and a serial > port. These motors would try to regenerate into the power supply when > slowing. > > Version 0 of the assembly, for bench testing only, had a 120 VAC to > 24 VDC switching power supply, which I think was a Triad AWSP series, > a foot or two of wire, and the motors. The motors "liked" this power > supply; I could command them from 0 to full speed to 0 again as fast as > I liked without any trouble. > > Version 1 of the assembly had a 120 VAC to 30-60 VDC adjustable > rackmount supply by Sorensen, a long wire, a DC-DC converter brick to > make 24 V DC, a foot or two of wire, and the motors. I don't remember > the brand of the brick; it was about 0.5" x 2" x 3", and rated about > 150 W or so. (Because of where the long wire had to live, there were > safety concerns with bringing line voltage AC directly to the device.) > At first, the motors were just connected directly to the output of the > brick, but that didn't work very well; starting the motors worked, but > slowing or stopping them would cause the DC-DC converter brick to shut > down. Cycling the input power to the DC-DC brick would bring it back. > > The short-term fix, so I could keep working on the software, was to put > a few 1000 uF or so capacitors on the 24 V DC rail. I know for sure > there was one directly on the output pins of the brick and I think > there was one at each motor as well. These seemed to soak up enough of > the regeneration energy that the DC-DC brick *usually* didn't shut down. > I don't know what the final production fix was. > > >>http://www.powerconversion.com/assets/lps200-m_ds_1226129797.pdf >>[...] >>Does anyone have any mileage with these Emerson supplies? > > > I have no experience with them. > > >>While I've eliminated the backfeeding problem, the board still has a >>current draw that can vary from a few tens of milliamps to ten amps >>over the space of less than a second -- if it would make the supply >>happier I can change this so that whenever the motor is turning the >>board is always drawing some minimum power. > > > The Emerson data sheet says "minimum load 0 A" but data sheets have been > known to lie before. What happens if you just slap a plain old power > resistor across the output of the supply, such that it always draws 5 > or 10 W or so at 24 V DC? This is more to find out if the supply needs > a minimum load than a production fix. On the other hand, maybe you can > provide the minimum load with a lamp and sell it as a high-visibility > power indicator. :) > > The data sheet also says it has a 12 V DC, 0.5 A fan power output. > Maybe loading that output helps the regulation or behavior? > > This also appears to be a "medical" power supply. I *think* this mostly > has to do with isolation voltages (higher), leakage currents (lower), > and more numbers and squiggles on the rating plate. These probably > don't have a direct effect on your problem, but if there is a non- > medical version of the same supply, it might be worth a try. > > Matt Roberds >
2 things, regenerative technology and a DBr circuit. Of course if you don't have a full bridge and regen circuit I guess you would have to rely on a DBr circuit to switch in a burden R on the DC buss when it over does it. A basic active clamp with R in series with it works well. Jamie
Reply by John Larkin April 10, 20132013-04-10
On Wed, 10 Apr 2013 11:44:00 -0500, Tim Wescott <tim@seemywebsite.com>
wrote:

>On Tue, 09 Apr 2013 20:08:37 -0700, John Larkin wrote: > >> On Tue, 09 Apr 2013 21:43:44 -0500, Tim Wescott >> <tim@seemywebsite.please> wrote: >> >>>On Tue, 09 Apr 2013 16:51:46 -0700, John Larkin wrote: >>> >>>> On Tue, 09 Apr 2013 18:17:07 -0500, Tim Wescott <tim@seemywebsite.com> >>>> wrote: >>>> >>>>>On Tue, 09 Apr 2013 15:52:40 -0700, John Larkin wrote: >>>>> >>>>>> On Tue, 09 Apr 2013 17:37:22 -0500, Tim Wescott >>>>>> <tim@seemywebsite.com> wrote: >>>>>> >>>>>>>So, I'm working on a motor drive board for a customer. Due to a >>>>>>>desire for performance improvements within the existing space, we're >>>>>>>going from the traditionally-accepted brute-force power supply to a >>>>>>>switcher. >>>>>>> >>>>>>>I have this nice board and software set that's carefully designed to >>>>>>>treat the power supply with kid gloves -- and it works very well >>>>>>>with the power supply that we used for prototyping. >>>>>>> >>>>>>>But for space reasons they went out and found what appears to be the >>>>>>>smallest possible power supply with the apparent specifications >>>>>>>needed (24V and 10A): >>>>>>> >>>>>>>http://www.powerconversion.com/assets/lps200-m_ds_1226129797.pdf >>>>>>> >>>>>>>It's not that this power supply sucks, per se. It's just that >>>>>>>things that work just fine with the original supply (A TDK/Lambda >>>>>>>SWS300-24) make the Emerson supply either exceedingly unhappy or >>>>>>>downright dead. Worse, I'm not sure what's taking out the power >>>>>>>supplies. >>>>>>> >>>>>>>So -- >>>>>>> >>>>>>>Does anyone have any mileage with these Emerson supplies? After I >>>>>>>told the customer "why yes, that should do" I began to notice the >>>>>>>name "Emerson" on this group, and the view of their stuff seems to >>>>>>>be pretty dim. I'd like to know if you feel these are OK, or >>>>>>>comprehensively bad, >>>>>>>or what. >>>>>>> >>>>>>>Also, I've carefully designed things so that any time the motor is >>>>>>>generating the power goes into a resistor and not into raising the >>>>>>>motor supply rail (and thus backfeeding the supply). But I did >>>>>>>notice that doing so kills one of these Emerson supplies (the TDK >>>>>>>supply mostly takes it in stride, or if you're really egregious >>>>>>>about things it turns off until you cycle power). >>>>>>> >>>>>>>Do you know of anything else that would kill one of these supplies? >>>>>>>While I've eliminated the backfeeding problem, the board still has a >>>>>>>current draw that can vary from a few tens of milliamps to ten amps >>>>>>>over the space of less than a second -- if it would make the supply >>>>>>>happier I can change this so that whenever the motor is turning the >>>>>>>board is always drawing some minimum power. Efficiency will suffer >>>>>>>-- but we're not after efficiency, we're after performance in a >>>>>>>small space. >>>>>>> >>>>>>>Also, if anyone knows of a supply that's the same size or not much >>>>>>>bigger, but comes from a manufacturer that's known to make >>>>>>>bulletproof stuff, I'd appreciate it. >>>>>> >>>>>> The Emerson has OVP, namely a crowbar. Maybe you're glitching that, >>>>>> or it trips from a big downward load step. Maybe it's a suicide load >>>>>> protector. >>>>> >>>>>I did have a bug in the software that was allowing the motor to >>>>>back-feed into the power supply rail without disconnecting as it >>>>>should. >>>>> So it could have been that. >>>>> >>>>>I've got more supplies on the way, I may do some testing with >>>>>resistors and a switch to see if it can stand simply switching a load >>>>>on and off. One would _hope_ that would be the case. >>>> >>>> Don't count on it! >>>> >>>> If you are decelerating a motor, where does the energy go? Mechanical >>>> systems can store a bunch of joules. >>> >>>If the system is working right, it all goes straight into a resistor. >>>As soon as the motor goes into regeneration the motor rail gets >>>disconnected from the power supply (via a FET) and is allowed to float >>>up. Then as soon as it gets above a set voltage the resistor gets >>>switched in. There's a healthy amount of capacitance at the motor rail >>>and the whole shebang gets sampled at the motor PWM rate, so it makes a >>>self- oscillating voltage regulator that holds the motor rail close >>>enough to the set point for things to work smoothly. >>> >>>When the motor starts to draw power the motor rail drops: when it drops >>>down to match the power supply voltage things are switched on again and >>>it draws power from the supply. >> >> >> That sounds a little tricky. You could just hang an active zener across >> the power supply output to clamp at, say, 25 volts. That might keep the >> ps from crowbarring, too. >> >> I'm sure Jim would be delighted to design an accurate, 10 amp active >> zener circuit for you. > >I'm sure I could design an accurate, 10 amp active voltage clamp circuit, >too -- it just wouldn't all be on one piece of silicon. > >And with such a part -- whether it came from me or from Jim -- I'd still >have the issue that if it didn't work then the power supply wouldn't be >protected. It is, to quote, "tricky".
Well, design a circuit that always works! So I may as well fix what I have,
>and make sure that I'm using a power supply that doesn't have any >unexpected gotchas (which is what this thread is trying to find out about >-- whether my power supply is going to cut me off at the knees).
How are you disconnecting the PS from the h-bridge rail? That needs a bidirectional switch, not just one mosfet, I think. -- John Larkin Highland Technology, Inc jlarkin at highlandtechnology dot com http://www.highlandtechnology.com Precision electronic instrumentation Picosecond-resolution Digital Delay and Pulse generators Custom laser drivers and controllers Photonics and fiberoptic TTL data links VME thermocouple, LVDT, synchro acquisition and simulation
Reply by John Larkin April 10, 20132013-04-10
On Wed, 10 Apr 2013 08:37:19 -0700, Jim Thompson
<To-Email-Use-The-Envelope-Icon@On-My-Web-Site.com> wrote:

>On Tue, 9 Apr 2013 21:59:16 -0700 (PDT), Bill Sloman ><bill.sloman@ieee.org> wrote: > >>On 10 Apr, 13:08, John Larkin >><jjlar...@highNOTlandTHIStechnologyPART.com> wrote: >>> On Tue, 09 Apr 2013 21:43:44 -0500, Tim Wescott <t...@seemywebsite.please> wrote: >[snip]. >>> >>> >When the motor starts to draw power the motor rail drops: when it drops >>> >down to match the power supply voltage things are switched on again and >>> >it draws power from the supply. >>> >>> That sounds a little tricky. You could just hang an active zener across the >>> power supply output to clamp at, say, 25 volts. That might keep the ps from >>> crowbarring, too. >>> >>> I'm sure Jim would be delighted to design an accurate, 10 amp active zener >>> circuit for you. > >More demented snarkiness. > >But rather trivial to design, even with off-the-shelf discrete's... >think a TL431 plus a power PNP.
Show us. -- John Larkin Highland Technology, Inc jlarkin at highlandtechnology dot com http://www.highlandtechnology.com Precision electronic instrumentation Picosecond-resolution Digital Delay and Pulse generators Custom laser drivers and controllers Photonics and fiberoptic TTL data links VME thermocouple, LVDT, synchro acquisition and simulation
Reply by Tim Wescott April 10, 20132013-04-10
On Wed, 10 Apr 2013 09:17:35 -0700, Mark wrote:


>> >> That sounds a little tricky. You could just hang an active zener >> >> across the power supply output to clamp at, say, 25 volts. That >> >> might keep the ps from crowbarring, too. >> >> > agreed, if a small SW error is fatal to the hardware, you are asking > for trouble..
Well, it was only small in physical size, spanning perhaps four or five words of flash. Once I noticed it, it was about as large and as glaring as two diodes in parallel, where the obvious circuit intent was two diodes in anti-parallel.
> you said efficiency is not an issue, why not a big old diode, Shottky > if you prefer, in series with the PS output, that is fail safe. and I > guess you need the zener also
I may do that. If the power supply is well behaved I don't see the need for the clamp, but I'm going to at least see what the series diode does to me. The power consumption isn't an issue, true -- but if I introduce a 10W hot-spot into the design it'll stick out of the overall thermal management landscape like a sore thumb. -- My liberal friends think I'm a conservative kook. My conservative friends think I'm a liberal kook. Why am I not happy that they have found common ground? Tim Wescott, Communications, Control, Circuits & Software http://www.wescottdesign.com
Reply by Tim Wescott April 10, 20132013-04-10
On Tue, 09 Apr 2013 20:08:37 -0700, John Larkin wrote:

> On Tue, 09 Apr 2013 21:43:44 -0500, Tim Wescott > <tim@seemywebsite.please> wrote: > >>On Tue, 09 Apr 2013 16:51:46 -0700, John Larkin wrote: >> >>> On Tue, 09 Apr 2013 18:17:07 -0500, Tim Wescott <tim@seemywebsite.com> >>> wrote: >>> >>>>On Tue, 09 Apr 2013 15:52:40 -0700, John Larkin wrote: >>>> >>>>> On Tue, 09 Apr 2013 17:37:22 -0500, Tim Wescott >>>>> <tim@seemywebsite.com> wrote: >>>>> >>>>>>So, I'm working on a motor drive board for a customer. Due to a >>>>>>desire for performance improvements within the existing space, we're >>>>>>going from the traditionally-accepted brute-force power supply to a >>>>>>switcher. >>>>>> >>>>>>I have this nice board and software set that's carefully designed to >>>>>>treat the power supply with kid gloves -- and it works very well >>>>>>with the power supply that we used for prototyping. >>>>>> >>>>>>But for space reasons they went out and found what appears to be the >>>>>>smallest possible power supply with the apparent specifications >>>>>>needed (24V and 10A): >>>>>> >>>>>>http://www.powerconversion.com/assets/lps200-m_ds_1226129797.pdf >>>>>> >>>>>>It's not that this power supply sucks, per se. It's just that >>>>>>things that work just fine with the original supply (A TDK/Lambda >>>>>>SWS300-24) make the Emerson supply either exceedingly unhappy or >>>>>>downright dead. Worse, I'm not sure what's taking out the power >>>>>>supplies. >>>>>> >>>>>>So -- >>>>>> >>>>>>Does anyone have any mileage with these Emerson supplies? After I >>>>>>told the customer "why yes, that should do" I began to notice the >>>>>>name "Emerson" on this group, and the view of their stuff seems to >>>>>>be pretty dim. I'd like to know if you feel these are OK, or >>>>>>comprehensively bad, >>>>>>or what. >>>>>> >>>>>>Also, I've carefully designed things so that any time the motor is >>>>>>generating the power goes into a resistor and not into raising the >>>>>>motor supply rail (and thus backfeeding the supply). But I did >>>>>>notice that doing so kills one of these Emerson supplies (the TDK >>>>>>supply mostly takes it in stride, or if you're really egregious >>>>>>about things it turns off until you cycle power). >>>>>> >>>>>>Do you know of anything else that would kill one of these supplies? >>>>>>While I've eliminated the backfeeding problem, the board still has a >>>>>>current draw that can vary from a few tens of milliamps to ten amps >>>>>>over the space of less than a second -- if it would make the supply >>>>>>happier I can change this so that whenever the motor is turning the >>>>>>board is always drawing some minimum power. Efficiency will suffer >>>>>>-- but we're not after efficiency, we're after performance in a >>>>>>small space. >>>>>> >>>>>>Also, if anyone knows of a supply that's the same size or not much >>>>>>bigger, but comes from a manufacturer that's known to make >>>>>>bulletproof stuff, I'd appreciate it. >>>>> >>>>> The Emerson has OVP, namely a crowbar. Maybe you're glitching that, >>>>> or it trips from a big downward load step. Maybe it's a suicide load >>>>> protector. >>>> >>>>I did have a bug in the software that was allowing the motor to >>>>back-feed into the power supply rail without disconnecting as it >>>>should. >>>> So it could have been that. >>>> >>>>I've got more supplies on the way, I may do some testing with >>>>resistors and a switch to see if it can stand simply switching a load >>>>on and off. One would _hope_ that would be the case. >>> >>> Don't count on it! >>> >>> If you are decelerating a motor, where does the energy go? Mechanical >>> systems can store a bunch of joules. >> >>If the system is working right, it all goes straight into a resistor. >>As soon as the motor goes into regeneration the motor rail gets >>disconnected from the power supply (via a FET) and is allowed to float >>up. Then as soon as it gets above a set voltage the resistor gets >>switched in. There's a healthy amount of capacitance at the motor rail >>and the whole shebang gets sampled at the motor PWM rate, so it makes a >>self- oscillating voltage regulator that holds the motor rail close >>enough to the set point for things to work smoothly. >> >>When the motor starts to draw power the motor rail drops: when it drops >>down to match the power supply voltage things are switched on again and >>it draws power from the supply. > > > That sounds a little tricky. You could just hang an active zener across > the power supply output to clamp at, say, 25 volts. That might keep the > ps from crowbarring, too. > > I'm sure Jim would be delighted to design an accurate, 10 amp active > zener circuit for you.
I'm sure I could design an accurate, 10 amp active voltage clamp circuit, too -- it just wouldn't all be on one piece of silicon. And with such a part -- whether it came from me or from Jim -- I'd still have the issue that if it didn't work then the power supply wouldn't be protected. It is, to quote, "tricky". So I may as well fix what I have, and make sure that I'm using a power supply that doesn't have any unexpected gotchas (which is what this thread is trying to find out about -- whether my power supply is going to cut me off at the knees). -- My liberal friends think I'm a conservative kook. My conservative friends think I'm a liberal kook. Why am I not happy that they have found common ground? Tim Wescott, Communications, Control, Circuits & Software http://www.wescottdesign.com