a side question - are you intending to use the peltiers to cool your photomultipliers - or is this a different project?
- Howard
Hi Howard,
Except for an alarm I made for my house, all my Propeller projects have been aimed at this one sprawling research project. In this case, the main driving force behind the Pelt is to cycle the biologicals. Right now the prototype has a big honking fridge that occasionally warms up the samples via a heating element. I'm attempting to shrink everything down to a table top unit that is cheaper and faster to make, and easy to modify on the fly. Cooling the photomultipliers (PMTs) helps a lot as it gets rid of some of the thermionic noise, so I'm interested in extending this Pelt control to them as well, when/if I get the Pelts to work. The PMTs are photon counting modules rated down to 32 F, and they have been reaching that temperature not so much by design as by accident, incidental to the cold "leakage" through the freezer window. But I definitely want to try to use the Peltiers to cool the photomultipliers once I get the Pelts under control. The past few days I've been trying to re-invent the wheel on water-cooled heat sinks for the hot side of the Pelts - looks like it might be best to just buy them off the shelf, though. Copper sheet is a lot harder to neatly solder together than I thought.
If you are using fairly low voltage(<25V), TI makes some AMAZING MOSFETs. These things have an Rds_on of about 1.3mOhms, and are logic level compatible. However, you really should use a MOSFET driver if you are operating above 1kHz or so. The only draw back, is that they are rated for only 25V.
ElectricAye said... I definitely want to try to use the Peltiers to cool the photomultipliers once I get the Pelts under control.·.... Copper sheet is a lot harder to neatly solder together than I thought.
I think you can also reduce the therm.noise electrically/electronically, but unfortunately as I might have mentioned, I don't remember how this was done. I do remember that it was just tweeking / adding resistors and caps in the right places.
RE the copper sheet. Copper's used for heat sinks for a good reason - same one that makes it·a PITA to solder. You have to·braze copper sheet. It's still soldering, in that the copper is not melted (i.e. welded), but you have to have a hot torch and the right kind of brazing rod. The best rods are about 3 feet long with the flux around it. For a torch, you need MAPP gas w/ oxygen, Acetylene w/oxy,·or Propane w/oxy - those little 'plumbers propane torches' or a big soldering iron won't do. But I guess I don't need to tell you that now, eh ?
(Some hardware stores have the little oxy/acet. torches for <100USD and you can get the brazing rods at a decent welding store where they refill the tanks.)
What on Earth kind of microbes are these little critters?
Philldapill said...
If you are using fairly low voltage(<25V), TI makes some AMAZING MOSFETs. These things have an Rds_on of about 1.3mOhms, and are logic level compatible. However, you really should use a MOSFET driver if you are operating above 1kHz or so. The only draw back, is that they are rated for only 25V.
thanks, Phil, I'm just now starting to learn that I have a lot of great options out there. This post has really paid off thanks to the general brilliance of this forum.
CounterRotatingProps said...
....
What on Earth kind of microbes are these little critters?
They're some kind of gooey slimy sometimes smelly bacteria that have been genetically modified to glow in the dark when undergoing a specific reaction. To me it's all freaky frankenstein stuff but I'm told it's kinda safe, kinda off-the-shelf application of glow-in-the-dark genes. Nevertheless, I periodically stand in front of the mirror in total darkness, check my eyeballs, open up and say AHHHH just to be sure I'm not picking up the plasmids they used. Insofar as I can tell, the brains behind this project do not glow in the dark either, but I'm not so sure they haven't already been inadvertently crossed with other organisms from somewhere weird - like Aldebaran maybe or somewhere east of M101.
▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
Watching the world pass me by, one photon at a time.
> Copper's used for heat sinks for a good reason - same one that makes it a PITA to solder.
First, a little definition:
Soldering and brazing both do not melt the base-material. If the base material is molten, it is called welding.
Soft-soldering works by definition below 450°C, brazing or silver-soldering above that temperature.
Copper is a good heat conductor, so it requires a good heat source. If you want to braze copper, the best rods are *without* flux. This only works with copper, but here like a charm. As long as your copper is clean. Depending on size and thickness of the sheet, a propane torch can be enough.
Even soft-soldering with a soldering iron works perfect. As long as the sheet is thin enough and/or the iron is strong enough. Or soldering a PCB would not work.
Nick
▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
Never use force, just go for a bigger hammer!
The DIY Digital-Readout for mills, lathes etc.: YADRO
because my Peltier is only about 40mm square and I want the heatsink contact surface to be no bigger than the Peltier, most of my difficulties with copper construction stem from the way the solder flows and leaves larger-than-desired fillets. Solder itself is not so heat conductive (compared to copper), so you really don't want the solder blobs to be part of the heat conductive path. So I have found this process to be hard to control. I'm great with soldering pipes for the kitchen sink, but this was a whole new ball game. I suppose that is why brazing might be best, since I think the brazing rod material has more copper, but like Howard pointed out, that takes some special equipment and probably even greater skill. Also, even the propane torch I used tends to heat the entire part either not enough or suddenly almost all at once (thanks to the incredible thermal conductivity of copper), so things that have already been soldered often come loose and wiggle around once there is enough heat. If you clamp the parts in place, the clamps act like heat sinks and the part will not heat, etc.
Long story short, the only material that really needs to be copper is the contact surface between the Pelt and water, so I'm looking at ways of doing that.
I like making things myself for a number of reasons: it forces me to understand the problem down to fundamental details and, once I learn a way to do it, it gives me the ability to modify things on the fly when plan B, or plan C, or.... plan ZZa and plan ZZb are still not working at 3 AM.
> so things that have already been soldered often come loose and wiggle around once there is enough heat.
Screw them together, rivet them together (you can drill a blind hole, make a dent onto the rivet and drive it in). Then solder it. You could also use some stainless steel wire to keep the parts together. You can even use solder mask (some special paste available in jeweler's tool stores) if you have difficulties controlling it.
If that thing looks like a pan or box, put the bottom part onto some ceramic tile, place the wall onto the bottom and heat from the center. Add solder from the inside onto the joint. -> no solder outside. The wall should be one single part. You can hammer that around some hardwood-block.
HTH,
Nick
▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
Never use force, just go for a bigger hammer!
The DIY Digital-Readout for mills, lathes etc.: YADRO
First of all, let me say that I have NOT been ignoring all the great advice I have been given so far on interfacing Mosfets, etc; it's just that I have a lag time between getting new parts so I've been putting together the circuits I had thought would work with parts on hand so I could further understand why my first designs were wrong. Having said that....
Once upon a time, I thought I knew enough about FETs to make them work like simple switches. But I confess that my latest experiments have left me feeling lost at sea. Exactly as Ariba predicted, the circuit (see attachment) yielded an output of only 1 volt.
Ariba said...
I fear your circuit will not work. The BS170 acts as source follower and will bring a very low voltage on the IGBT gate (3.3V - BS170-GateThreshold Voltage =~ 1V). I think you expected to switch the 15V to the gate, but this needs another circuit....
My problem is that I don't understand why this is the case. Once upon a time I thought that to use a FET like a switch, all you had to do was to get the gate voltage up above the gate threshold voltage (so long as you were cool with the current and voltage ratings passing from drain to source.) Since the Propeller outputs 3.3 volts and I didn't need much current to flow through the BS170, I figured the Propeller was good enough to work on the gate of the BS170 (data sheet attached), which is an N-channel FET, enhancement mode gizmo. So I don't see how Ariba figured out that my pathetic circuit was not going to output roughly 5 volts. What am I missing here?
Not certain, but it can be tricky trying to use an N channel device as a high side driver.· If you flip that upside down so that the BS170 is switching the ground, it should work but the logic will be inverted.
edit - Note that while fet's are voltage operated devices, gate capacitance can cause current to flow through the connection to the gate input.· It is fairly common to use a low value resistor such as 100 ohm in series with the io pin to guard against damage to the io pin.
What counts is the Voltage between Gate and Source of a FET. When you apply 3.3V to the Gate of the BS170 in your circuit, the FET tries to switch on. This increases the voltage on the 1k resistor and the voltage difference between gate and source decreases. With 1 V at the 1k resistor, you have 2.3V between gate and source left (3.3V - 1V) and this is the needed Gate threshold voltage. That's why the FET regulates itself to this point.
To switch 15V to the resistor the FET would need about 17.3V at the gate.
This citcuit with the Load at the source is known as source follower, and is only practicable to amplify the current but not the voltage.
If you connect the Load at the Drain of the FET, like in my proposed circuits, the source stays at 0V and the gate at 3.3V -> enough voltage to switch full on.
What counts is the Voltage between Gate and Source of a FET. When you apply 3.3V to the Gate of the BS170 in your circuit, the FET tries to switch on. This increases the voltage on the 1k resistor and the voltage difference between gate and source decreases. With 1 V at the 1k resistor, you have 2.3V between gate and source left (3.3V - 1V) and this is the needed Gate threshold voltage. That's why the FET regulates itself to this point.
To switch 15V to the resistor the FET would need about 17.3V at the gate.
This citcuit with the Load at the source is known as source follower, and is only practicable to amplify the current but not the voltage.
If you connect the Load at the Drain of the FET, like in my proposed circuits, the source stays at 0V and the gate at 3.3V -> enough voltage to switch full on.
Andy
Andy,
okay, now I get it! Your explanation makes it all so clear now. I've spent all day trying to figure out what I was doing wrong. Thank you for saving my sanity on this!
Thanks, too, to Agent for advice on protecting the pin.
There are simple ways to make a cooler for your Peltiers. In the attached .jpg method A was to solder the tubing to a copper plate. I used that for cooling a CPU chip and it worked very well. The copper plate was 16 gauge or 1/16" (don't recall which), and the tubing was 1/16 id. I soldered it using a propane torch and plumbing flux (acid) and solder.
Method B and C I have seen in equipment I service, but have not made them myself. The copper block is thicker and has holes drilled through it with 180 degree tubes or milled slots between the drilled holes. They also seem to work very well.
Ariba said...
What counts is the Voltage between Gate and Source of a FET....
As a longtime software guy who's new to hardware, I found this the most valuable post of my week! Eureka moment -- that's what was wrong with my circuit! Thanks, Ariba (and everyone else on these forums) for taking the time to share your expertise...
There are simple ways to make a cooler for your Peltiers. In the attached .jpg method A was to solder the tubing to a copper plate. I used that for cooling a CPU chip and it worked very well. The copper plate was 16 gauge or 1/16" (don't recall which), and the tubing was 1/16 id. I soldered it using a propane torch and plumbing flux (acid) and solder.
Method B and C I have seen in equipment I service, but have not made them myself. The copper block is thicker and has holes drilled through it with 180 degree tubes or milled slots between the drilled holes. They also seem to work very well.
kwinn,
I've been meaning to get back to you on your very nice reply. It's much appreciated. My whole approach to this problem got "burned" however by other considerations, most notably the cost of cooling the water that cools the Peltier. When you're pumping 150 watts or so into the cooler's cooling system, even the cooling water heats up after a while, then you have to cool that, and that requires radiators and fans and such. And the cost and complexity got to be too much for my comrades, so, sniffle sniffle, I'm now using my Peltier PWM PID to drive some resistance heaters deep inside our big honking freezers. Such is life.
Comments
Hi Howard,
Except for an alarm I made for my house, all my Propeller projects have been aimed at this one sprawling research project. In this case, the main driving force behind the Pelt is to cycle the biologicals. Right now the prototype has a big honking fridge that occasionally warms up the samples via a heating element. I'm attempting to shrink everything down to a table top unit that is cheaper and faster to make, and easy to modify on the fly. Cooling the photomultipliers (PMTs) helps a lot as it gets rid of some of the thermionic noise, so I'm interested in extending this Pelt control to them as well, when/if I get the Pelts to work. The PMTs are photon counting modules rated down to 32 F, and they have been reaching that temperature not so much by design as by accident, incidental to the cold "leakage" through the freezer window. But I definitely want to try to use the Peltiers to cool the photomultipliers once I get the Pelts under control. The past few days I've been trying to re-invent the wheel on water-cooled heat sinks for the hot side of the Pelts - looks like it might be best to just buy them off the shelf, though. Copper sheet is a lot harder to neatly solder together than I thought.
thanks, evan, I've made a note of this.
RE the copper sheet. Copper's used for heat sinks for a good reason - same one that makes it·a PITA to solder. You have to·braze copper sheet. It's still soldering, in that the copper is not melted (i.e. welded), but you have to have a hot torch and the right kind of brazing rod. The best rods are about 3 feet long with the flux around it. For a torch, you need MAPP gas w/ oxygen, Acetylene w/oxy,·or Propane w/oxy - those little 'plumbers propane torches' or a big soldering iron won't do. But I guess I don't need to tell you that now, eh ?
(Some hardware stores have the little oxy/acet. torches for <100USD and you can get the brazing rods at a decent welding store where they refill the tanks.)
What on Earth kind of microbes are these little critters?
- H
▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
thanks, Phil, I'm just now starting to learn that I have a lot of great options out there. This post has really paid off thanks to the general brilliance of this forum.
They're some kind of gooey slimy sometimes smelly bacteria that have been genetically modified to glow in the dark when undergoing a specific reaction. To me it's all freaky frankenstein stuff but I'm told it's kinda safe, kinda off-the-shelf application of glow-in-the-dark genes. Nevertheless, I periodically stand in front of the mirror in total darkness, check my eyeballs, open up and say AHHHH just to be sure I'm not picking up the plasmids they used. Insofar as I can tell, the brains behind this project do not glow in the dark either, but I'm not so sure they haven't already been inadvertently crossed with other organisms from somewhere weird - like Aldebaran maybe or somewhere east of M101.
▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
Watching the world pass me by, one photon at a time.
First, a little definition:
Soldering and brazing both do not melt the base-material. If the base material is molten, it is called welding.
Soft-soldering works by definition below 450°C, brazing or silver-soldering above that temperature.
Copper is a good heat conductor, so it requires a good heat source. If you want to braze copper, the best rods are *without* flux. This only works with copper, but here like a charm. As long as your copper is clean. Depending on size and thickness of the sheet, a propane torch can be enough.
Even soft-soldering with a soldering iron works perfect. As long as the sheet is thin enough and/or the iron is strong enough. Or soldering a PCB would not work.
Nick
▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
Never use force, just go for a bigger hammer!
The DIY Digital-Readout for mills, lathes etc.:
YADRO
because my Peltier is only about 40mm square and I want the heatsink contact surface to be no bigger than the Peltier, most of my difficulties with copper construction stem from the way the solder flows and leaves larger-than-desired fillets. Solder itself is not so heat conductive (compared to copper), so you really don't want the solder blobs to be part of the heat conductive path. So I have found this process to be hard to control. I'm great with soldering pipes for the kitchen sink, but this was a whole new ball game. I suppose that is why brazing might be best, since I think the brazing rod material has more copper, but like Howard pointed out, that takes some special equipment and probably even greater skill. Also, even the propane torch I used tends to heat the entire part either not enough or suddenly almost all at once (thanks to the incredible thermal conductivity of copper), so things that have already been soldered often come loose and wiggle around once there is enough heat. If you clamp the parts in place, the clamps act like heat sinks and the part will not heat, etc.
Long story short, the only material that really needs to be copper is the contact surface between the Pelt and water, so I'm looking at ways of doing that.
I like making things myself for a number of reasons: it forces me to understand the problem down to fundamental details and, once I learn a way to do it, it gives me the ability to modify things on the fly when plan B, or plan C, or.... plan ZZa and plan ZZb are still not working at 3 AM.
Screw them together, rivet them together (you can drill a blind hole, make a dent onto the rivet and drive it in). Then solder it. You could also use some stainless steel wire to keep the parts together. You can even use solder mask (some special paste available in jeweler's tool stores) if you have difficulties controlling it.
If that thing looks like a pan or box, put the bottom part onto some ceramic tile, place the wall onto the bottom and heat from the center. Add solder from the inside onto the joint. -> no solder outside. The wall should be one single part. You can hammer that around some hardwood-block.
HTH,
Nick
▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
Never use force, just go for a bigger hammer!
The DIY Digital-Readout for mills, lathes etc.:
YADRO
Once upon a time, I thought I knew enough about FETs to make them work like simple switches. But I confess that my latest experiments have left me feeling lost at sea. Exactly as Ariba predicted, the circuit (see attachment) yielded an output of only 1 volt.
My problem is that I don't understand why this is the case. Once upon a time I thought that to use a FET like a switch, all you had to do was to get the gate voltage up above the gate threshold voltage (so long as you were cool with the current and voltage ratings passing from drain to source.) Since the Propeller outputs 3.3 volts and I didn't need much current to flow through the BS170, I figured the Propeller was good enough to work on the gate of the BS170 (data sheet attached), which is an N-channel FET, enhancement mode gizmo. So I don't see how Ariba figured out that my pathetic circuit was not going to output roughly 5 volts. What am I missing here?
edit - Note that while fet's are voltage operated devices, gate capacitance can cause current to flow through the connection to the gate input.· It is fairly common to use a low value resistor such as 100 ohm in series with the io pin to guard against damage to the io pin.
▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
What counts is the Voltage between Gate and Source of a FET. When you apply 3.3V to the Gate of the BS170 in your circuit, the FET tries to switch on. This increases the voltage on the 1k resistor and the voltage difference between gate and source decreases. With 1 V at the 1k resistor, you have 2.3V between gate and source left (3.3V - 1V) and this is the needed Gate threshold voltage. That's why the FET regulates itself to this point.
To switch 15V to the resistor the FET would need about 17.3V at the gate.
This citcuit with the Load at the source is known as source follower, and is only practicable to amplify the current but not the voltage.
If you connect the Load at the Drain of the FET, like in my proposed circuits, the source stays at 0V and the gate at 3.3V -> enough voltage to switch full on.
Andy
Andy,
okay, now I get it! Your explanation makes it all so clear now. I've spent all day trying to figure out what I was doing wrong. Thank you for saving my sanity on this!
Thanks, too, to Agent for advice on protecting the pin.
There are simple ways to make a cooler for your Peltiers. In the attached .jpg method A was to solder the tubing to a copper plate. I used that for cooling a CPU chip and it worked very well. The copper plate was 16 gauge or 1/16" (don't recall which), and the tubing was 1/16 id. I soldered it using a propane torch and plumbing flux (acid) and solder.
Method B and C I have seen in equipment I service, but have not made them myself. The copper block is thicker and has holes drilled through it with 180 degree tubes or milled slots between the drilled holes. They also seem to work very well.
kwinn,
I've been meaning to get back to you on your very nice reply. It's much appreciated. My whole approach to this problem got "burned" however by other considerations, most notably the cost of cooling the water that cools the Peltier. When you're pumping 150 watts or so into the cooler's cooling system, even the cooling water heats up after a while, then you have to cool that, and that requires radiators and fans and such. And the cost and complexity got to be too much for my comrades, so, sniffle sniffle, I'm now using my Peltier PWM PID to drive some resistance heaters deep inside our big honking freezers. Such is life.
thanks tho,
Mark