1. Home
  2. Electronics Design
  3. How about it? Experiments of the third kind , take 999999.

How about it? Experiments of the third kind , take 999999.

How about it? Experiments of the third kind , take 999999. Update Hotplate design.

After entering the results of the heating experiment into the neural net (mine), it became clear to that neural net that things should be done differently than a simple resistor heating the air in a box as a means to keep component temperature stable.

So then, it was shown to me, that if we mount all the components that we want to be at a fixed temperature on a 'hotplate', in this case a small aluminum plate, and heat it with a nice TO220 type transistor, then if we keep 'hotplate' free from the box, we can make a nice even temperature with very little input power for the heater. So after some sawing and drilling I changed pictures of solution presented by the neural net into something real in the material world:

formatting link
The components will be glued on the alu plate, there will be no PCB, as leakage would ruin everything anyways with 120 MOhm resistors, would need guard bands, what not. from left to right tritium tube, one of the photo cells (those will be directly connected to the opamp next to it and in thermal contact with the tritium tube), the TLC274 quad opamp, a 10 MHz crystal, a PIC 18F14K22 microcomputer with multi channel ADC and PWM output, and 2 24LC256 EEPROMs for data storage (dark and light values, plus some more, no need to hold these at a fixed temp, maybe even keep them cooler, these chips in the pictures are smaller ones, still waiting for the big ones), so maybe keep them out of the box? [1]

Anyways, below the hotplate is the hotplate heater:

formatting link
it has 2 functions, heating the hotplate, and keeping it separated from the black plastic box. This shows the bottom of the box, the washer will keep it separated from the PCB it is mounted on.
formatting link
Also it shows some sketch of the circuit diagram mentioned neural net came up with, so to run the heater from the PIC PWM but with practically DC so we have no interference, the hotplate is floating here, but if collected to the drain of the IRZ43N MOSFET it would be at +4 V, and that would be stabilized and fine too, give better thermal contact. We will see. I will use a plastic bolt, as soon as I get near a hardware shop that has those in M3 size, bit difficult on Sunday. Plastic will give better thermal insulation.

So next is superglue, or / and maybe 2 components glue, we have that at hand, There are of course more components, but mainly some small capacitors and some very small SMD resistors that can easily fit between IC pins etc.. I will not make a guess how much less power we will need in this setup to keep say the tritium tube at 40C, but ... OK. You can do and place bets of course, if somebody takes them. I think this is a cool circuit.

[1] The big thermal leakage is now likely to come from the wires going in and out of the black box. we need: ground power +4V (heater) power +3V (chips) I2C data perhaps I2C clock perhaps RS232 logic level in (MAX232 is on main board RS232 logic level out Temperature warning LED out (can be done with I/O expander on main board via same I2C). So basically if I put the EEPROMs in the black box, I can save 1 copper wire. And of course what I forgot, but I do not know that yet :-)
Reply to
Jan Panteltje
Loading thread data ...

How about it? Experiments of the third kind , take 999999. Update Hotplate design.

After entering the results of the heating experiment into the neural net (mine), it became clear to that neural net that things should be done differently than a simple resistor heating the air in a box as a means to keep component temperature stable.

So then, it was shown to me, that if we mount all the components that we want to be at a fixed temperature on a 'hotplate', in this case a small aluminum plate, and heat it with a nice TO220 type transistor, then if we keep 'hotplate' free from the box, we can make a nice even temperature with very little input power for the heater. So after some sawing and drilling I changed pictures of solution presented by the neural net into something real in the material world:

formatting link
The components will be glued on the alu plate, there will be no PCB, as leakage would ruin everything anyways with 120 MOhm resistors, would need guard bands, what not. from left to right tritium tube, one of the photo cells (those will be directly connected to the opamp next to it and in thermal contact with the tritium tube), the TLC274 quad opamp, a 10 MHz crystal, a PIC 18F14K22 microcomputer with multi channel ADC and PWM output, and 2 24LC256 EEPROMs for data storage (dark and light values, plus some more, no need to hold these at a fixed temp, maybe even keep them cooler, these chips in the pictures are smaller ones, still waiting for the big ones), so maybe keep them out of the box? [1]

Anyways, below the hotplate is the hotplate heater:

formatting link
it has 2 functions, heating the hotplate, and keeping it separated from the black plastic box. This shows the bottom of the box, the washer will keep it separated from the PCB it is mounted on.
formatting link
Also it shows some sketch of the circuit diagram mentioned neural net came up with, so to run the heater from the PIC PWM but with practically DC so we have no interference, the hotplate is floating here, but if collected to the drain of the IRZ43N MOSFET it would be at +4 V, and that would be stabilized and fine too, give better thermal contact. We will see. I will use a plastic bolt, as soon as I get near a hardware shop that has those in M3 size, bit difficult on Sunday. Plastic will give better thermal insulation.

So next is superglue, or / and maybe 2 components glue, we have that at hand, There are of course more components, but mainly some small capacitors and some very small SMD resistors that can easily fit between IC pins etc.. I will not make a guess how much less power we will need in this setup to keep say the tritium tube at 40C, but ... OK. You can do and place bets of course, if somebody takes them. I think this is a cool circuit.

[1] The big thermal leakage is now likely to come from the wires going in and out of the black box. we need: ground power +4V (heater) power +3V (chips) I2C data perhaps I2C clock perhaps RS232 logic level in (MAX232 is on main board RS232 logic level out Temperature warning LED out (can be done with I/O expander on main board via same I2C). So basically if I put the EEPROMs in the black box, I can save 1 copper wire. And of course what I forgot, but I do not know that yet :-)
Reply to
Jan Panteltje

Double posting retard!

Reply to
WoolyBully

How about it? Experiments of the third kind , take 999999. Update Hotplate design, stabilizing the heater MOSFET

formatting link
shows the quick sketch diagram of the MOSFET driver. The idea is that the current through the MOSFET is exactly proportional to the input voltage. As we all know, MOSFETs like to do their thing too, and in such a configuration like to sing (oscillate). And so also this circuit.

Playing with it and the oscilloscope a bit did lead to this very quiet circuit, that within a fraction of a millivolt has the MOSFET drain current follow the input. There were 2 issues here, 1) MOSFET likes to oscillate, and 2) opamp too. So sort of separating both from each other worked.

+4V R1 | 470k + |------ ---===--------------|\ 100k 1k |--- | | | - | >-------===--------===----| |
Reply to
Jan Panteltje

On a sunny day (Sun, 01 Apr 2012 08:04:38 -0700) it happened WoolyBully wrote in :

The only retard is you and some here who endlessy meouw about nothing going on in your head. Lets bring some sanity into usenet, go away!

Reply to
Jan Panteltje

Try a cap from op-amp output to -input, 100pF is enough. Add a series 10k resistor between this node (-in with cap) and the shunt resistor, so the cap has something to work into. Get rid of the 100k series resistor and 1uF filter cap.

As shown, it'll something between oscillate and motorboat, depending on the position of the planets. The op-amp is an integrator with 90 degree phase shift, and the RC following it does the same (for a more limited range of frequencies). 180 degree phase shift at most any frequency means you're guaranteed to oscillate somewhere.

You might consider making both series input resistors (the 470k control voltage filter and the shunt voltage feedback) the same value (100k?) so the input bias (if any) generates a matching offset on both inputs. As a CMOS amp, offset won't be the greatest (a few mV, plus input offset bias), but if this is within a temperature control loop, offset and 1/f noise won't be a problem.

Tim

-- Deep Friar: a very philosophical monk. Website:

formatting link

formatting link

Reply to
Tim Williams

(mine),

aluminum

the opamp next to

output,

black plastic box.

PCB it is mounted on.

formatting link

Early on, my thought about the heater was multiple resistors, 10, or

  1. What ever it took for uniform temperature. You could capture resistors between the aluminum plate and the plastic housing with or without heatsink compound. That was early on, maybe heating the aluminum plate in the center is good enough. Cool experiment, thanks for keeping us posted. Mikek
Reply to
amdx

On a sunny day (Sun, 1 Apr 2012 10:48:46 -0500) it happened "Tim Williams" wrote in :

I have tried several configurations, of course the first one that came to mind is 'integrator' but I had problems with huge oscillation at some input voltages. After many test I settled for this, as I cannot get it unstable in any way. I reduced the source resistor to 1.1 Ohm (2 x 2.2 parallel) to get more of the generated heat into the hotplate.

The planets seemed favorable today, you will have to show the math why these should influence - and in what way- the circuit. I have read about dark energy, but am not that much of a believer, especially not if it interacts with the electrons in this circuit. Once someone told be he had a system for the stock market, and in that system you should buy if this planet went up and the other down, so I asked him: Why not this planet down and the other up? That shorted his neural net enough to drop the idea...

The 1k is just a gate resistor to stop the MOSFET from interfering with local FM and short wave, the 100k isolates the rather low output impedance from whatever the MOSFET wants to put out on its gate, it is basically a 2 way 'T' filter.

The circuit is now like this: PIC runs on 3V supply, and the PWM switches between 0 and 3V. With an 1.1 Ohm resistor for 330 mA max (for now, based on previous plastic box burning test) no more than 600 mV should be at the + input of the TCL274, making 330 mV about mid range (it can become -5 °C). This requires a voltage divider of ratio of .6 to (3 - .6) makes .6 to 2.4, and to not load that opamp I have now 470 k to the PIC PWM output and 120 k to ground, normal E series, high values preferred so I can use my 1 uF caps.

+4 V R1 | 470k + |------ ---===---------------|\ 100k 1k |--- | | | | - | >-------===--------===----| |
Reply to
Jan Panteltje

No, like this.

formatting link
It is easy to show the speed, stability, accuracy and phase margin are greatly improved. If you cannot show why this is better than the circuit you've drawn, you shouldn't be working with op-amps at all, do it in software.

Tim

-- Deep Friar: a very philosophical monk. Website:

formatting link

Reply to
Tim Williams

formatting link

Assuming he is serious about this experiment then the least bad way to do it would be with the core electronics inside a hefty block of hermetically sealed aluminium heated by Raychem heating tape on the outside and wrapped in a very thick layer of polystyrene. The aim would be to keep it at a very steady temperature - although it might be worth making that well below room temperature. There are published patents on TEC designs for isotope ratio mass specs that exploit the unusually low thermal noise characteristics of certain opamps operated at appropriate sub-zero temperatures with 100G feedback resistors.

The electronic gain wants to be as low as possible and the optics need considerable work to maximise signal to noise. Around 10x optical gain is possible by non-focussing flux concentrators. It is routine in all difficult photon detection scenarios for high energy physics where neutrinos and other elusive particles are being sought.

You can even buy the requisite optics of the shelf from Edmunds.

I would also monitor humidity, air pressure and external temperature as I am sure that whatever "effects" have been "measured" so far correlate far more closely with ambient temperature than with anything else.

--
Regards,
Martin Brown
Reply to
Martin Brown

On a sunny day (Sun, 1 Apr 2012 15:12:05 -0500) it happened "Tim Williams" wrote in :

Yes Tim, but that is basically an integrator, as I mentioned.

Well, a bit bold statement, you obviously did not read about capacitive loading of opamp outputs, to return the political statement ;-)

But anyways, I will try your negative feedback for high frequencies, alias 'integrator', for a totally different reason: The PWM filter, the RC combination that smoothes the PWM to DC, is not perfect. This causes a slight RF ripple of the drive voltage on the opamp + input, resulting in a lot of output swing in my circuit.

It is less with higher PWM frequency, but for this PIC microcomputer a higher PWM frequency means 6 bits resolution, a medium PWM frequency has 8 bits resolution, and the lowest PWM frequency has 11 bits resolution, all set by the internal PIC hardware. So doing the integrator thing would also filter the PWM better, and quiet things down some more. So I will leave in the T filter, as it is proven to work, but do a test with your circuit added, to see how it reacts to PWM ripple. If you like I can engrave your name on the alu bottom of the box if it works[1]. I bought a nice engraving tool for about 5 $ on ebay, and have played with it. Would that heal any wounds I may have made by being not totally cooperative in immediately agreeing and sending flowers?

[1] Will report back.
Reply to
Jan Panteltje

I wrote: [1] Will report back.

+4V | R1 ---- | 470k + R2 | | PWM ---===----------------|\ 100k |--- | 0-3V | | - | >---------===----| |
Reply to
Jan Panteltje

[snip]

George posted > I wrote:

100k.

Hi Jan, I like the V-I converter. It gives a nice linear relation between voltage in and power out. Do you stick the 1.1 ohm current sesnse resistor on the hot plate too? Why not use a BJT for the pass element?

George H.

Reply to
JW

On a sunny day (Mon, 2 Apr 2012 06:40:32 -0700 (PDT)) it happened George

?????????????????

Never mind, need to add auto base64 codec to my newsreader one of those days.. panteltje10: ~ # base64 -d article.txt

The 1.1 (actually 2 x 2.2 Ohm in parallel) is soldered directly to the source of the MOSFET and in that way helps heat the MOSFET, and the MOSFET will heat the hotplate. I could use a BJT, but that needs driving power. I could get a darlington I suppose, but I have a bunch of these nice logic level IRLZ43N (on at about 2 V Vgs), and it is easy to drive from this CMOS opamp. I really want to keep the currents in the 4 TCL274 as low as possible, to avoid any cross-heating in the chip with the light and dark preamps. So now the output current is zero, this also improves output swing for this opamp. I am now working on the power supply with battery backup, so far the most difficult thing!

Reply to
Jan Panteltje

On a sunny day (Tue, 03 Apr 2012 06:33:58 -0400) it happened JW wrote in :

OK, got it, and decoded it with the 'base64' program in Linux. There is a content header that maybe I should one day have my newsreader parse...

Thanks!

Reply to
Jan Panteltje

On a sunny day (Tue, 03 Apr 2012 10:46:27 GMT) it happened Jan Panteltje wrote in :

oops, IRLZ34N

Reply to
Jan Panteltje

by=AD5j

oC=ADCg

Cj=AD4g

oC=ADCg

oC=ADAt

IK=ADAg

IK=ADAg

PT=AD09

IK=ADB8

IK=ADAg

oC=ADB8

fC=ADBU

oC=ADCg

oC=ADCg

oC=ADCg

oC=ADCg

IK=ADAg

IK=ADAg

ZC=ADBz

dm=ADVy

c3=ADRh

ID=ADEw

ZW=ADN0

ZS=ADBp

aW=AD5n

dm=ADlu

Z2=ADh0

ZX=ADIg

aW=ADNl

IH=ADlv

IH=ADBs

cm=ADdl

WTF is that? Did I do that... sorry.

George H.

Reply to
George Herold

Reply to
George Herold

looks like GG decided to convert half the spaces to non-breaking spaces and then base-64 encode the result. base64 wins over Quoted-Printabe once more than a quarter of the octets are high, and it wouldn't have been more legible as quoted-printable anyway.

Gotta wonder aboutr the space mangling thoough, mixing normal and non-breaking spaces is an html trick - there's no need to apply it to usenet.

--
?? 100% natural

--- Posted via news://freenews.netfront.net/ - Complaints to news@netfront.net
Reply to
Jasen Betts

On a sunny day (5 Apr 2012 11:11:41 GMT) it happened Jasen Betts wrote in :

Yes I noticed the messed up spaces, he probably had it in some webbrowser, maybe via google groups. I did notice the base 64 header, one of those days I fix my newsreader so it calls base64. Have not done any changes to it in many years, so that would take maybe a day to get back into the old code. Usenet should be just ASCII text, and maybe some binaries attached, but most ISPs already block binary groups out of fear of copy right issues.

These days I browse with 'pictures off' in seamonkey, only when I want to see something do I enable these. much faster :-) What worries me is all the lookups to 1.1.1.2 1.1.1.3 etc.. google this google that, facebook, as if they own your computer. For the adds I just do something like this in /etc/hosts/

10.0.0.190 ad.yieldmanager.com 10.0.0.190 adserver.adtech.de 10.0.0.190 doubleclick.net 10.0.0.190 ajax.googleapis.com 10.0.0.190 connect.facebook.net 10.0.0.190 ...... And 10.0.0.190 is NOT on the LAN, so that comes back really fast with connection refused. And I have 4444 lines of reject (some whole domains) in iptables too. Freedom!
Reply to
Jan Panteltje

ElectronDepot website is not affiliated with any of the manufacturers or service providers discussed here. All logos and trade names are the property of their respective owners.