Post by thread:[EE] DC drive to aquarium diaphragm pump ideas?

I have an application where I need to drive a diaphragm pump (like those used in aquariums) from low voltage DC (battery). Here in the US, these pumps are mechanically tuned to run directly off the 60 Hz line. A single coil on a core moves an armature that has a couple magnets on it. The armature moves back and forth at 60 Hz, vibrating the diaphragm. So, in this application, I need to run the pump on DC. Ideas I'm considering are: 1. Rewind the coil to have fewer turns and a center tap. Have a PIC drive transistors that alternately ground one end of the coil or the other with the center tap going to + supply. 2. Rewind the coil to have fewer turns without a center tap. Connect one end to + supply, the other to a single transistor to ground that is switched. 3. Electrically resonate the existing coil using a circuit similar to that on page 4 of http://www.edn.com/contents/images/21705di.pdf ("Sine-wave step-up converter uses Class E concept") where the inductor is the coil in the pump. Any ideas? THANKS! Harold -- FCC Rules Updated Daily at http://www.hallikainen.com

> .... > >> Any ideas? > > a small DC motor that you run at 60*60 RPM that moves a piston rod back > and > forth that physically toggles a switch between positive and negative > voltage. > > - Marcel Pretty elegant! Actually, if I had the DC motor, a crank on it could run the diaphragm directly. Turns out this pump is incredibly cheap (cheaper than a motor), so I'm trying to figure out a cleaver way to drive it.. Thanks for the ideas... Any more? THANKS! Harold -- FCC Rules Updated Daily at http://www.hallikainen.com

Any idea on the power level / onductance / resistance of the coil? RP On 30/09/05, Rob Robson <spam_OUTrobTakeThisOuTsilk.net> wrote: > I'd be inclined to hang the pump off a variac to see how little AC amplitude > you could get away with. Also, the pump's coil might not be too thrilled > with square waves, so you might have to do what the cheaper commercial > 12V-to-120VAC inverters do and simulate a sine wave with a two-step > staircase. Just some thoughts. > > RR > > > {Original Message removed}

I'm waiting to get a sample. I understand it takes about 6W at 120VAC 60Hz. Harold {Quote hidden}> Any idea on the power level / onductance / resistance of the coil? > RP > > On 30/09/05, Rob Robson <.....robKILLspam@spam@silk.net> wrote: >> I'd be inclined to hang the pump off a variac to see how little AC >> amplitude >> you could get away with. Also, the pump's coil might not be too >> thrilled >> with square waves, so you might have to do what the cheaper commercial >> 12V-to-120VAC inverters do and simulate a sine wave with a two-step >> staircase. Just some thoughts. >> >> RR >> >> >> {Original Message removed}

OK. I was looking at the ringing choke option & was trying to figure out what size capacitor may be required for 60Hz resonance. I think that it could be a reasonable option - and wouldn't require a coil rewind which is another plus. You may have to tune the capacitor value so that the resonant frequency matches the mechanical resonance. Other drive alternatives (using a rewound coil) could be a 1/2 or full bridge. If you can find a centre tapped wall-wart or similar transformer a simple push-pull may be an option too. RP On 30/09/05, Harold Hallikainen <haroldKILLspamhallikainen.com> wrote: {Quote hidden}> I'm waiting to get a sample. I understand it takes about 6W at 120VAC 60Hz. > > Harold > > > > Any idea on the power level / onductance / resistance of the coil? > > RP > > > > On 30/09/05, Rob Robson <.....robKILLspam.....silk.net> wrote: > >> I'd be inclined to hang the pump off a variac to see how little AC > >> amplitude > >> you could get away with. Also, the pump's coil might not be too > >> thrilled > >> with square waves, so you might have to do what the cheaper commercial > >> 12V-to-120VAC inverters do and simulate a sine wave with a two-step > >> staircase. Just some thoughts. > >> > >> RR > >> > >> > >> {Original Message removed}

On Thu, 2005-09-29 at 15:23 -0700, Harold Hallikainen wrote: > I have an application where I need to drive a diaphragm pump (like those > used in aquariums) from low voltage DC (battery). Here in the US, these > pumps are mechanically tuned to run directly off the 60 Hz line. A single > coil on a core moves an armature that has a couple magnets on it. The > armature moves back and forth at 60 Hz, vibrating the diaphragm. This may be too simple, but why not just purchase an inverter? It sounds like this device doesn't use that much power, and inverters of smaller wattages are very cheap these days. TTYL ----------------------------- Herbert's PIC Stuff: http://repatch.dyndns.org:8383/pic_stuff/

> On Thu, 2005-09-29 at 15:23 -0700, Harold Hallikainen wrote: >> I have an application where I need to drive a diaphragm pump (like those >> used in aquariums) from low voltage DC (battery). Here in the US, these >> pumps are mechanically tuned to run directly off the 60 Hz line. A >> single >> coil on a core moves an armature that has a couple magnets on it. The >> armature moves back and forth at 60 Hz, vibrating the diaphragm. > > This may be too simple, but why not just purchase an inverter? It sounds > like this device doesn't use that much power, and inverters of smaller > wattages are very cheap these days. > We're going to be shipping several thousand units a month. The pump is about $5.00 I'm trying to keep costs down. Thanks for the idea, though! Keep 'em coming! Harold -- FCC Rules Updated Daily at http://www.hallikainen.com

How about using a Wien bridge or something to generate the wave? Alternatively, what about a PWM of some sort? Would just strobing a pulse through the coil at regular intervals be enough to cause a vibration? In that case you could just use a simple RC circuit or something. - Marcel "Harold Hallikainen" <EraseMEharoldspam_OUTTakeThisOuThallikainen.com> wrote: {Quote hidden}> > > On Thu, 2005-09-29 at 15:23 -0700, Harold Hallikainen wrote: > >> I have an application where I need to drive a diaphragm pump (like those > >> used in aquariums) from low voltage DC (battery). Here in the US, these > >> pumps are mechanically tuned to run directly off the 60 Hz line. A > >> single > >> coil on a core moves an armature that has a couple magnets on it. The > >> armature moves back and forth at 60 Hz, vibrating the diaphragm. > > > > This may be too simple, but why not just purchase an inverter? It sounds > > like this device doesn't use that much power, and inverters of smaller > > wattages are very cheap these days. > > > > > We're going to be shipping several thousand units a month. The pump is > about $5.00 I'm trying to keep costs down. > > Thanks for the idea, though! Keep 'em coming! > > Harold > >

> How about using a Wien bridge or something to generate the wave? > Alternatively, what about a PWM of some sort? Would just strobing a pulse > through the coil at regular intervals be enough to cause a vibration? In > that > case you could just use a simple RC circuit or something. > - Marcel > Thanks for the comments! Since there's going to be a PIC in the product to handle other stuff, I'm planning on having it drive a transistor or two to drive the pump. The trick is, should I: 1. Rewind the pump using a center tapped coil and pulse each end of the coil to ground, or 2. Rewind the pump with a single ended coil and pulse just one end to ground (mechanical resonance possibly would take it back the other way). 3. Don't rewind the pump, but make the coil resonant so I can get enough current with low voltage drive (as mentioned in the article I posted previously in this thread). Again, THANKS for the comments! Harold -- FCC Rules Updated Daily at http://www.hallikainen.com

On Sep 30, 2005, at 1:11 PM, Harold Hallikainen wrote: > > We're going to be shipping several thousand units a month. The pump is > about $5.00 I'm trying to keep costs down. > If you have a cheap source that will do things like rewind the pump magnet (!), perhaps you should just investigate what it would take to implement a DC-based pump (either along the same lines as a DC buzzer, or using a motor to drive a cylinder or diaphragm. I have such a motor driven pump surplus from Electronics Goldmine - it just uses a rotating cam on the motor to drive a the diaphragm.) BillW

Harold, I'm curious why you don't just try your idea. It looks reasonably sound and a quick test with a 555 (or suitable substitute) where it's easy to alter the frequency and duty cycle would get you a good answer in under an hour. It seems that if you can use the existing pumps "as is", that's the way to go. Go with plan 3 if there's any reasonable way to make it work. Dave {Quote hidden}>Thanks for the comments! Since there's going to be a PIC in the product to >handle other stuff, I'm planning on having it drive a transistor or two to >drive the pump. The trick is, should I: > >1. Rewind the pump using a center tapped coil and pulse each end of the >coil to ground, or > >2. Rewind the pump with a single ended coil and pulse just one end to >ground (mechanical resonance possibly would take it back the other way). > >3. Don't rewind the pump, but make the coil resonant so I can get enough >current with low voltage drive (as mentioned in the article I posted >previously in this thread). > >Again, THANKS for the comments! > >Harold > > > >

Must this be low voltage powered ? If you have mains available you can pulse the coil with mains. Possibly with half waves to avoid kickback. An opto and a mains rated transistor or thyristor switch will likely be cheaper than rewinding the coil. Since you are into light controls I am going to guess that this is for some sort of light control (mechanical mirror drive etc). Then you could sync your effect to the mains and use the solenoid as is. In my limited experience the cheap aquarium solenoids have only 2-3 mm stroke and they achieve that only at resonance. But the resonance point is wide since they have a lot of damping (cheap thick rubber cap used as membrane). Driving the device with unipolar pulses should not be a problem because the moving armature is a permanent magnet (if it would not be then the armature would have to vibrate at 2x mains frequency and that is hard to do mechanically, plus the force would be lower). Just pick the current direction that attracts the armature when the coil is energised (has more force). You might be able to achieve resonance using single-stroke drive if you use a mains voltage doubler rectifier to provide the supply and a ~30% duty cycle for the switch. But the kickback will be horrible. It could be reduced using a RC damper of high value (say 3uF and 50 ohms if the coil resistance is not enough). 'Some work' will likely be required. Peter

Thanks for the comments! I'll probably do a little experimentation this week. On using some sort of inverter to generate line voltage 60 Hz, it doesn't seem cost effective. We already have a coil (in the pump), so why add two more (a transformer). Just my opinion... Again, THANKS for the comments. I'll experiment with one of the existing pumps this week. Harold {Quote hidden}> Harold, > > I'm curious why you don't just try your idea. It looks reasonably sound > and a quick test with a 555 (or suitable substitute) where it's easy to > alter the frequency and duty cycle would get you a good answer in under > an hour. It seems that if you can use the existing pumps "as is", > that's the way to go. Go with plan 3 if there's any reasonable way to > make it work. > > Dave > >>Thanks for the comments! Since there's going to be a PIC in the product >> to >>handle other stuff, I'm planning on having it drive a transistor or two >> to >>drive the pump. The trick is, should I: >> >>1. Rewind the pump using a center tapped coil and pulse each end of the >>coil to ground, or >> >>2. Rewind the pump with a single ended coil and pulse just one end to >>ground (mechanical resonance possibly would take it back the other way). >> >>3. Don't rewind the pump, but make the coil resonant so I can get enough >>current with low voltage drive (as mentioned in the article I posted >>previously in this thread). >> >>Again, THANKS for the comments! >> >>Harold >> >> >> >> > > > -

> On Sep 30, 2005, at 1:11 PM, Harold Hallikainen wrote: >> >> We're going to be shipping several thousand units a month. The pump is >> about $5.00 I'm trying to keep costs down. >> > If you have a cheap source that will do things like rewind the pump > magnet (!), perhaps you should just investigate what it would take > to implement a DC-based pump (either along the same lines as a DC > buzzer, or using a motor to drive a cylinder or diaphragm. I have > such a motor driven pump surplus from Electronics Goldmine - it just > uses a rotating cam on the motor to drive a the diaphragm.) > I'm currently thinking mechanical design is the lowest cost. Since I'll have a PIC in there anywhay to run other stuff, adding a transistor or two to drive the coil seems simple enough. Just trying to get any other ideas (and I'm getting them!). Again, thanks for all the comments! Keep 'em coming! Harold -- FCC Rules Updated Daily at http://www.hallikainen.com

Thanks for the great comments! We do contract design work for several companies, so this is not for the lighting industry. It is highly desirable to have the thing operable on low voltage DC, both a wall wart and secondary batteries. Driving the coil in only one direction and relying on mechanical resonance for the travel in the other direction is something I'm considering. I'll do some experimentation this week. THANKS! Harold {Quote hidden}> > Must this be low voltage powered ? If you have mains available you can > pulse the coil with mains. Possibly with half waves to avoid kickback. > An opto and a mains rated transistor or thyristor switch will likely be > cheaper than rewinding the coil. > > Since you are into light controls I am going to guess that this is for > some sort of light control (mechanical mirror drive etc). Then you could > sync your effect to the mains and use the solenoid as is. > > In my limited experience the cheap aquarium solenoids have only 2-3 mm > stroke and they achieve that only at resonance. But the resonance point > is wide since they have a lot of damping (cheap thick rubber cap used as > membrane). Driving the device with unipolar pulses should not be a > problem because the moving armature is a permanent magnet (if it would > not be then the armature would have to vibrate at 2x mains frequency and > that is hard to do mechanically, plus the force would be lower). Just > pick the current direction that attracts the armature when the coil is > energised (has more force). > > You might be able to achieve resonance using single-stroke drive if you > use a mains voltage doubler rectifier to provide the supply and a ~30% > duty cycle for the switch. But the kickback will be horrible. It could > be reduced using a RC damper of high value (say 3uF and 50 ohms if the > coil resistance is not enough). > > 'Some work' will likely be required. > > Peter > > -