probably saying the same thing differently. but

R2 is the manual temp control (w/o thermostat) pot which I changed for a 10-turn and relocated - the values I'm using will be measured
TILT
my present heat load is only 175W, far from the max of ~400W
ok, I have a dial counter on it, I'll estimate something (how slop enters the process, bleh)

R1 to me seems the factory 'trim' pot to get the 22°C spot on

I have no idea what the temp range is over that 500 ohm pot, but using this pot will enable use of the existing thermostat control circuity - no ?
-> I should determine experimentally the range available by adjusting that pot ?

Ahh, a heater. The circuit makes a bit more sense now. The way it is set up, the heater can be on or off for any given half cycle of the power line. I dare say that 120 Hz granularity is the limiting factor. A digital pot may let you set your temperature with Labview, but I doubt the repeatability or control will be better than the ordinary pot.

repeatability is not an issue, nor is the temp variation of the pot itself given how it would be used (in a PID controller using an 'external' RTD input)

my principal concern is the resolution/steps available, that is the control limitation

with a 10-turn pot my adjustments now are often no more than a feather's touch of the dial,
I want to get rid of this 'tweaking' stuff

We seem to be speak at cross purposes. Perhaps a concise, coherent statement of your goals? What ranges of temperatures? What resolution? How accurately the chamber must hold that temperature?

hmmm

I wish to define a set point, and the incubator to maintain that temp - also as the internal heat 'load' changes

the unit will now maintain the temp only if using the factory set point of 22°C

more better ?

addendum

range: 20 - 35°C
resolution: 0.01°C
acceptable variation: ±0.05°C (the machine will do better on the 22°C fixed set point)
heat load range: 0 - 400W

Using the switch to manual and the pot, the thermostat should still be maintaining the temperature, just at a different temperature than the 22° set point. If it isn't doing that, then either your controller board has a problem, or the schematic you have isn't correct.

How exactly does it behave when you manually control?

well, looks like I was blowin' smoke with that 400W heat load capability
went to 250W and the unit cannot hold the temp
probably come to equilibrium ~ 27 or 28°C (based on previous work)

coolant 38.59, die 78.61 with a small rad and Delta M fan (not at equilibrium)
in spec for this project

this is the DIY 'hangup', inability to accept high silicon temps (within the mfgr's thermal envelope)

ah, finally
long have I suspected that it should work 'in lieu of' - but am too ignorant to read the schematic (which I only got yesterday)

when I put SW1 to the manual position I can set the temp with the pot R2
but as the heat load changes so does the temp
obviously, to me, the thermostat function is not working

the unit has two 2-wire RTDs and I had to fuss with their connections to get the unit working

this is a small "Proportional Temperature Control" board, its connections are as follows:
S1, S2: the legs of a 1k RTD
G: to a triac, 25A
T1,L1: AC line - but between a cap and the triac
L2: AC power common

the second 1k RTD is connected to another board, the "Performance Monitor" board and is used for the lcd display

no scanner or I'd put it up

when I shut down I'll match the components to the schematic to check

How are T1, L1, cap, triac, and AC line connected? An ASCII diagram or a Paint gif would be cool.

You have a paper schematics? If so, knowing the component values would be handy in figuring out what's up (some of the values in the gif are borderline indecipherable).

which are unclear ?
R1: 500 ohms
R2: 5k pot
R3: 1.27k 1%
R4: 1.18k 1%
R5: 13.7k 1%
R6: 6.34k 1%
most of the others seem ok

re the triac, it has 3 connections
the control leg is "G"
L1 and T1 are connected to the 'input' side of the triac
between the line 120VAC and the L1 and T1 connection is a capacitor looking symbol (contacts ? -| |- ) labeled "I-CR"
output of triac goes to heater, heater indicator, etc.

looking at S1 and S1; connected to a 1k RTD
R19 1k 1% is across these terminals, effectively halving its resistance
S1 goes to pin 3 of U3, and this line is grounded through R1 plus R3
pin 3 is positive so the effect of changing the resistance to ground . . . .
- would seem to be to change the voltage applied to the RTD ??

I still think I should crank on it and see what happens

ideas ?

Voltage to power the RTD come from pin 4 'Vref' of U1, current goes to ground throught the RTD || R19, R1, R3; your basic voltage-divider. R1 is so that the factory set point of 22° can be trimmed to allow for component tolerances. You can crank on R1, it will raise/lower the set point, let me look up the tempco for RTDs and I'll tell you how much you can change the set point there.

I-CR, weird. Not sure what that is. I implies inductor, but what up with the cap symbol?

R9 is other one I can't read well, "2001 C" can't be right.

I shut down the thermal stuff for the moment, am building a sim of your controller card.

"am building a sim of your controller card"
oh lord

R9: 200k 1%
0.00375 ohm/ohm/°C, 1000 ohms @ 0°C

when I pull it apart I will locate "I-CR", it connects to a unique wire

A change of only 50 ohms on R1 would cover your entire 15° C range. It'd be very touchy unless you added another smaller, multi-turn pot in series -- a coarse and a fine adjustment.

Circuit doesn't seem to turn the triac/heater on and off as quickly as I had thought. I don't quite understand how it manages as fine of control as it apparently does. Time for bed anyway, this gives me something to sleep on.

Groth
sorry, I wasted some of your time
my description of the 'problem' was wrong, thermostat function works fine (I had the sensing RTD confused with the indicating RTD, AND therefore in the wrong location - duh)
- this being the case, no point in messing with R1

the R2 value at 25°C with a nominal 70W load is:
W1-W2: 3.49k, W1-W3: 1.46k, and 4.95k across W2-W3
the dial counter reads ~700 (0-999), increase to decrease temp;
the max range might be ±150, 550 - 850

the max heater on period is ~1 sec., with a max ~40% duty cycle (under max heating off period a bit longer)

think my task is to setup a 5k digital pot with 1024 steps and see if the resolution is sufficient

Bill,
Are you sure about those measurements? Could there be some voltage in the circuit when measured? It would seem that, since they are in series, w1-W2 should be greater than W1-W3. Maybe I'm confused on how the resisters are acting in parallel, let me think....On a more constructive note, since it looks like it's all around 5k ohms, perhaps you should replace your 5 k pot with a 4.9k 1% resistor and a 100 ohm, 8 bit digital pot? This will improve precision, but decrease range, although if Groth has it right, it won't matter.

Sorry about not getting back sooner. Brewing beer today :) and catching up with kids.

no, I too think the reference is wrong, the pot was un-pluged when measured

by the color coding in my diagram:
W1-W2: 4.95k
W1-W3: 3.49k
W2-W3: 1.46k

better ?

no, think Groth was referring to R1, not R2
8 bits yields only 256 steps, or ~0.4% resolution

the AD5231 is 10k and 1024 steps (run 2 in parallel ?)
do you know of a 5k with 1024 steps (or better ?)

Ok, I played with Excel to get the resistances right and I think it's about 0.488k per degree of change if your readings at 25 C are right. This sounds high given Groths comments(?). But if correct, then to get 0.01C resolution, you will need 5 ohm precision on the digital pot with this configuration, or at least two of the 10k, 1024 states pots in parallel to get the presicision. I'd change R5 to a larger value, but I'm worried that the current flow on the wiper won't be enough to turn on the optocoupler circuit. So 2 of the AD devices seems like a winner. Unless murray13 has a better plan.....

Not wasted time, I rather like their CA3059 IC. I had wondered about sensor location, and if it had been anyone but you Bill, I'd have asked. Oh well.

If you still want digital control, I still advise doing it at the opto. A single digital signal will be easier than dealing with a SPI or I2C bus, surely one of your existing boards has a spare general purpose pin that can supply 5mA to the opto. Quite simply, import your chamber temp into Labview, feed it into a virtual PID, PID says heater on/off, signal sent to opto, repeat. You'll be able to set up your PID however you want--minimum on and off times for the heater, hysteresis, etc.

But if you want the digi-pot, here's how. At your set point, the output of W1 into the comparator section wants to be ~2.102 V. Given the output range of buffer U3 over your temp range, and the 13.7K resistor from W2 to ground, your full range of interest will be about 120 ohms (a quarter turn of your existing pot, sound about right?). To cover that 120 ohm range and allow for a bit of slop, parallel your digital pot with a 150 ohm resistor. To properly position your window, rewire the existing 5K pot.
http://pages.sbcglobal.net/water.gro...s/env-cham.gif

Bill,

I'm not big on the electrical side, so I haven't bothered looking at your schematic, but I've got a hunch there's no integrator on that thing. As such, you'll have some steady-state error and it'll get bigger with added load. I suspect you know the gist of PID control, but if not I'll say that without an "I" you'll have steady-state error. Can anyone tell if this circuit board has an integrator? I also suspect you can forget about "D" as things aren't "moving" that quickly.

They have a part time integrator, a capacitor fed by a voltage divider fed by an amp. Cap charges at a rate (mostly) proportional to the temperature error, discharges proportional to the heater duty cycle plus some leakage.

But yeah, it has some load dependence to both steady state temperature (poor integrator) and temperature variance (coarse control of heater).