So I was thinking of how I could build my own temperature controller, and it would appear that the temperature sensor port has 2 pins which simply
short closed when the hotplate is enabled. Thus if I turn the hotplate temp to maximum and turn control over to an external PID controller with a
thermocouple in the right spot, I should get pretty accurate long term heating.
The one part that I am struggling with is where to find the connector for the hotplate. In the ikamag manual it is listed as a 'DIN 12878' connector,
but google only refers to Ikamag hotplates with this term, and nothing else. Does anyone know the true name?Oxydro - 4-7-2010 at 16:20
Could you post a picture of the connector? I've never looked at an Ikamag hotplate. I've had good luck finding connectors by tearing apart consumer electronics 'till I find the right thing.aonomus - 4-7-2010 at 17:50
Strangely enough I think I have found the connector. I went to a local electronics shop with a sketch of the connector and found it. Its a standard
DIN connector, but I believe that Ikamag refers to it by a non-standard name.
Edit: wikipedia is failing today, but its a standard 5 pin DIN connector (ie: not a 180 degree connector), center pin appears to be unused, and the
'12V' present on the connector only appears to be 10.5V, which puts a kink in my plans to run the device completely stand-alone from the hotplate
power only.
[Edited on 5-7-2010 by aonomus]peach - 5-7-2010 at 16:20
I think you're probably right.
And I KNOW that IKA is run by a bunch of annoying turds.
Their gear is nicely made, but it costs a ridiculous amount compared to imports offering similar or better features.
Also, when something breaks, they won't discuss repairing it. They'll only sell you new boards and such, with a seriously fat price tag for a few
cheap components. One of the staff got quite angry with me when I suggested it wasn't fair they wouldn't tell me the value of a 0.001 cent SMD
resistor that'd blown (all zero's correct).
I would not put it past them at all to have purposefully renamed the socket so as to avoid third parties producing competing thermometers, as you're
now doing; not willing to pay out more $$$$ for their own. Or to make it hard for customers to know what will work with which plate and require them
to replace the plugs on other options. I'd also bet they've used DIN because making a proprietary socket would cost them money and make their devious
scheming obvious.
The 10.5V could be another part of their kinky nature.
To get round it, have a look online for PCB sized DC to DC converters. There are some that are fully encapsulated power supplies the size of a big
chip. Sold the input on, the output out, set the voltage you want and it'll regenerate it.
If you go to power.nationalsemi they have a supply designer that'll produce a schematic for a whole range of applications and converter formats.
I know a lot of people on here or reading probably have an IKA, or want one. I have a big box full of components and was hoping there'd be a converter
in there for you, but couldn't find the one I was looking for. If you need anything like a voltage regulator though, let me know and I'll post you one
for free as part of my "Say no to IKA" campaign.
IKA is one of those big brand names that has some questionable policies to me.
[Edited on 6-7-2010 by peach]aonomus - 5-7-2010 at 20:22
I suspect that a PID controller will want more current than can be supplied by that port, and since there is no specific current limit listed... tough
luck.
I did however turn up more information from the Heidolph hotplates that I use at work, the manual appears to reference the same DIN 12878 and actually
describes some of the pinout. Apparently of those 5 pins, 3 control the hotplate: 1 common, and 2 contacts. 1 contact opens/closes due to temperature,
and the other is a thermometer breakage sensor (ie: glass breakage opens the circuit and disables, this pin pair is the same as the shorting pin in
the blank plug on the back of the hotplate)
It would appear the specification is for 9-12V, and
I might just make a 120VAC powered version and isolate the device completely by using an old fashioned relay instead of an SSR, etc. Makes it easier
since I'm unsure as to which direction current is supposed to flow through, etc. Alternatively, a 12V wall-wart powered version is just as good (and
safe due to my wiring being low voltage only, the potential for a failure inducing 120VAC into the thermocouple is nil).
I'd probably also make a direct 120VAC version so I can plug a heating mantle into it and use a thermocouple to control temperature. I might make a
unit where you can either use an internal thermocouple or an external thermocouple.... it beats using a light dimmer to control power.
This would be really nice because you can maintain stirring while refluxing at a certain temperature, or, distill at a fairly precisely controlled
temperature, and make cleaner cuts by preventing overshooting. Total cost probably 1/4 if not less than what Ika or Heidolph charge for their
temperature controllers.
I will keep this thread going as I actually develop this device and make it more lab-friendly.watson.fawkes - 5-7-2010 at 21:43
I suspect that a PID controller will want more current than can be supplied by that port, and since there is no specific current limit listed... tough
luck.
I looked up the power specs for a Watlow 93; I have the document handy because I've got a couple of
them. It also seems to be model that a bunch of the Chinese controllers are using as clone source. Power consumption is stated at 5 W max. Supply
range is nominal 12-24V, but specified as 10-26 V. That's admittedly about 500 mA, which it may or may not put out, but it's easy enough to test with
an ammeter and a 20 Ω power resistor.peach - 6-7-2010 at 06:18
The mantle idea is excellent! I'd be willing to help out with that, sourcing cheap components and checking over schematics.
There are a ton of mantles around with zero temperature control built in, which is something of a joke considering where we are with solid state now.
I was going to suggest something like the LR8. I have some of those. They're linear regulators you can run direct from the rectified mains but the
output is a measly 10mA. ST Micro also do one called the VB408, which will do up to 60mA. But running directly from the unisolated mains is not a good
plan at all, as you undoubtedly know.
A board mount transformer would do fine, and you can get tiny board mounts really cheap that'll do for powering the controller.
Contrary to that, I wouldn't bother using any magnetics to control the mantle output. I'd just rectify it with an all in one chip, dump some cheap
electrolytic capacitance on it and use a 555 style PWM circuit to regulate the temperature. Or, if that's not safe enough, stick another transformer
on there to isolate it. I think at least some of them are single or double insulated, making it safer to use the direct mains on them. And others have
a conductive linear, which is no doubt earthed.
There is one factor that concerns me about external temperature controllers, and it's that a lot of people may simply set the controller to their
reflux / distillation point and hit go. Where upon the controller will throttle the heat source to it's maximum.
With constant PWM feedback from the thermometer, overshooting might be too much of an issue, as the power input will drop off as it approaches the
correct temperature. But I worry about that kind of rapid heating with larger flasks or things under vacuum. There's a greater chance of them popping
as the mantle initially slams it with it's maximum amount of heat.
Maybe some ramp up delays could be built into the controller for various flask sizes.
This is probably why the IKA unit uses fuzzy logic. But delays would likely work okay.aonomus - 6-7-2010 at 13:22
You raise a good point about going full-on with the mantle, although with a traditional hotplate as well it would go full on but have more heat loss
to the environment making it less risky. Perhaps turning the PID controller on once the hotplate is already at operating temperature? A small
internal-external switch?peach - 8-7-2010 at 08:00
That's what I was thinking as well.
People might like a switch and delays, so they can either do it themselves or just power it up and walk away. Working out the delays would mean asking
for opinions on warm up times for various flasks and possibly trying to burst a few with different times; starting off reaaaaally slow and working up
to full on, instant heating. Doing it under vacuum would also be an idea.
It wouldn't need expensive ground glass for those tests. The cheapo rubber bung glass would work. Using inferior glass for the tests would be a
benefit, as it'd give a safety margin for the expensive glass.
I don't think it'll be such an issue for the 50, 100ml scale of things. But for 500 - 2l it would. Above 2l, it certainly will. I've burst numerous
multi-liter borosilicate flasks simply trying to warm them up to 100C from room temperature.
The switch and delays will cost next to nothing to add. A switch? A dollar? The ramps, cents in components?
There could be a switch on the controller that simply says "50, 100, 2l, 10l" etc that swaps between the RC ramps. If I were getting really smart, I'd
couple that flask size switch to another that selects the rough heat capacity of the solution; e.g. DCM will heat up far more quickly than water for
the same energy input.
Keep me updated on this one, any ideas you have or progress you make. I'd be really interested in getting a cheaper option out there. The western
commercial options are far too much. Even producing it on a small scale, I'm sure we could easily beat the catalog prices.
I'm working on my own polar (har har har) alternative to this, a recirculating TEM chiller for condensers, flasks, traps and such. The current options
are thousands. I think I can get that down to hundreds, whilst maintaining a similar quality; I think I could actually better on a number of them.
There are a huge number of variables to work on first, like coolant selection, compatibility, pumping it, heatsinks, turbulence, expansion, toxicity,
flammability, complexity, cooling the TEMs themselves and so on. It all has to work indefinitely and it has to be cheap, down to every seal, every
component, bit of electronics, insulation and enclosure. I wouldn't feel right selling something I knew was going to be a pain, not perform well or
break, particularly to scientists. A cheap recirculating cooler that can manage sub zero would be great for vacuum distillation and might open up some
new doors for amateur and genuine lab chemistry / physics. At the moment, I'm looking at temperatures between -40 and -80C. Which is a handy range,
because that's around the cross over point between solvent vapors and elemental halide gases. Below that, it's economically worthless to cool with
heat pumps, liquid nitrogen is far cheaper and cooler (but impractical for many slightly above those temperatures, special glass, gloves, boil off
accounts...).
Lab suppliers abuse the fact they're selling to people funded by the government, I think. They know they'll essentially have a blank check from a lot
of them. I don't think they purposefully overcharge, but I don't think they spend much time optimizing either. There's nothing that exciting inside an
IKA plate. But the price tag is mind blowing.
I've leaned, the very expensive way, that slamming my plate to 10 to get to reflux is not a good idea; in terms of the glass and in terms of it's
contents. The most I'll do is turn it to 10 for about a minute, until it's about hot coffee temperature, then go back to 1 and work up. It's also
terrible lab practice. You're not going to get a clean product doing that kind of thing; may as well bin the vigreux and pour some tar into your
substrate first, cheaper and quicker.
[Edited on 8-7-2010 by peach]aonomus - 8-7-2010 at 13:53
My hotplate goes to 11...
All jokes aside, I was looking at the Omega site last night, and even though their products are pricier, I think it would be worth it in the long run.
They have $70 PID controllers with fuzzy logic which gets around this problem of heat ramping too quickly.densest - 8-7-2010 at 18:22
Some PID controllers have a maximum output setting for exactly the reasons mentioned.
TEMs are very useful but not a good idea for bulk cooling. They are about 1/4 to 1/10 as efficient as a conventional refrigeration unit: per unit of
heat energy moved, the TEM wastes a lot more. Up to 50W they are plausible (12 cal/sec, equivalent to condensing about 0.02g of water per second) the
simplicity of a TEM system gives them an advantage if the temperature differential is not high. They conduct heat themselves.
For a given application, consider how much heat is being applied with a hot plate, etc. or from the environment if the target vessel is below ambient
temperature. The cooler has to remove that much over the given temperature differential. Most macro scale lab setups will need to remove 100W or more.
There are a number of regimes to consider.
1) Cooled item much above ambient (over 100C) - glycol up to 150C or so (automobile antifreeze 50%) or automobile brake fluid (various types, some
glycols I -think-, some silicones) up to 250C or higher (400C???). Much above that molten metals take over. Dump heat to ambient air with a radiator
& fan. To remove large amounts of heat, spray water onto the radiator. 540 cal/g is a lot of energy.
2) Cooled item above ambient by 10C but below 100C - water to air like above - automobile antifreeze system works well. The antifreeze has
anticorrosion additives in it to protect pumps and metals in the cooling system. Active refrigeration is an option near ambient if large amounts of
heat must be removed and temperature controlled precisely.
3) Cooled item above ambient by less than 10C to below ambient but above -20C - automobile antifreeze to heat exchanger in water, water/ice,
water/ice/salts. Limited by availability of ice in bulk. Active refrigeration is an option if ice supply is limited or unattended operation is
required.
4) Cooled item below -20C - special attention must be paid to insulation. Heat leaks can overwhelm even large refrigeration units. Consider dry ice
(frozen CO2) systems. Special circulating fluids necessary below -30C.
My thoughts on peltiers are that they aren't for bulk heat rejection from a (relatively) closed system, but ok for maintenance.
Thats why you see them in DC powered coolers with heatsinks on both sides and fans to exchange air and move heat out of the cooler for those hot
summer days.
Sure you might be able to cool something down far enough with a big enough stack of peltiers to run reactions at -78degC, but they won't be able to
keep up with something as simple as dropwise addition of LDA during a reaction. At that point it truely is cubic money in the form of a cryocooler.
I'm sure one can be built using conventional refrigeration parts and the appropriate refrigerants (exotic ones, I recall a recent thread about liquid
N2 production using a multistage refrigeration system and exotic refrigerants, but I'm too lazy to UTFSE).
To be honest, it would almost be easier to use a PID controller with LN as a heatsink and an appropriate heat transfer fluid/gas to facilitate cooling
as absurd as it sounds.
. I've had good luck finding connectors by tearing apart consumer electronics 'till I find the right thing.
