I/O and 1wire over TCP/IP

[Michael McSharry]

Uncheck the "Web Login Enabled" checkbox on the WebControl Network Configuration for at least testing. I forgot all about this when I wrote the manual.

[Pete]

That was it. (web login enabled)

[Michael McSharry]

Do you want a password/login to be added to xapmcsWebControl? I also noticed that you have 3 decimals on the humidity reading. I belive this should be only an integer value so I need to look into that.

[Pete]

I don't think its an issue to shut off the login so its really not needed. The humidity reading should be fine as an integer. I was just testing the three decimal points.

Thanks Michael for your help with this.

[Pete]

One thing that I have noticed though refreshing the web control status page is that the percentage of the humidity varies each time I update it from 37% to 49% which is different from the readings that I see with the Temp0X or 9097 1-wire modules. The average and most likely accurate percentage is about 42%.
In the documentation see:

The filter for the continuous data is implemented as a moving average where the weight of the average is a number between 0.0 and 1.0. A value of 0 provides no filtering. A value of 1.0 provides infinite filtering. A value of 0.25 will produce a value where 25% is determined by the latest raw reading and the remaining 75% is determined by the prior filtered value.

So will set up mine to .25.

Currently NTP is:

WebControl™ automatically obtains its time over Internet.
Normally you only need to set its timezone to reflect where you are.
However, if it has no Internet connection, you may set its clock here.

Ideally though it would be beneficial to specify where it gets it time from as I get mine internally from an internal NTP server and how often it checks and syncs its time. Wondering if the device is just reading NTP broadcast messages or pro actively going out looking at anything specifically. This makes me think about how easy is it to update the firmware on this device.


I think I will need to play some more with the filtering. Here are two reads seconds apart.

[hult]

micro is under a heat sink so not observable without disassembly. Marc has had contact with CAI so llikely he will know.

It's a 10.5 MIPS PIC18F66J65 with 96KB Flash program memory, 3.8KB RAM, 11 channels 10-bit analog in, 10base-T with a -40 to 85C and 2 to 3.6 vdc operating range. The PCB also has a SST25VF040B-50 4Mbit serial flash. I/O is buffered and level-converted through two LVX4245 dual-voltage octal transceivers. The crystal is a 25.000H8B ( dunno temperature range, but typically -20 to 70c).

HTH ... Marc

[Pete]

Marc,
Curious regarding the fluctuations on the humdity sensor with every screen refresh of the Web Gui?

-pete

[hult]

Marc,
Curious regarding the fluctuations on the humdity sensor with every screen refresh of the Web Gui?

The humidity readings on my units toggle between two values ( eg 42 % and 43 %) but are not erratic.

The advantage of an voltage-output analog sensor ( not digital, not capacitive) becomes apparent here.

1) Retighten the connections to the sensor.
2) Find your volt meter (if not successful, buy/get one ;-)
3) Measure the nominal 5vdc supply and output voltages at the sensor leads and the screw-terminal block. You should read about 4.6 vdc between pins 1 and 3 and about one-half that (1.8 to 3.0vdc ) between pins 1 and 2 if RH is in the range of 40-60% .
4) If the nominal 5vdc changes more than about 0.05 vdc as you watch it, try a different power supply, The one that you are using may be noisy.

( The HHIH-xxxx humidity sensor is powered by the 5vdc from the LM2940 regulator + 200uf electroytic + small ceramic capacitor. There is an input diode for reverse-polarity protection but no inductor on the input to reject noise from switching power supplies. Normally I would not expect this to be a problem -- but you have a problem ;-)

HTH ... Marc

[Pete]

Hello Marc,

Using a 6VDC 200Mw transformer. Will look for a 9VDC higher amperage transformer. Meanwhile will check voltages with my Voltmeter.

Between pins 1 and 3 on the humidity sensor getting 4.3 VDC and between 1 and 2 getting 1.8 VDC.

Will look for a 9VDC transformer and post voltages shortly. Found numerous 5, 6 and 12VDC transformers....

Found a multivolt switching regulated 1000Mw transformer (3-12VDC) and set
to
7.5 VDC:
1-3 and 1-2 readings are now: 4.4VDC and 1.9VDC
9 VDC:
1-3 and 1-2 readings are: 4.4 and 2.0VDC - will leave it there for now.

Humidity is bouncing between 34% to 36% - better than before.

After a few minutes it seems to have stabilized a bit going between 33-34%. Maybe it was me breathing over the board??

I have a spare Temp08 and was going to solder a humidity sensor to the board to do side by side humidity value comparisions.

I've configured a CAI Humidity variable in MCS Temperature to get some graphs. (the device is sitting in the basement).

Thanks Marc.

-Pete

[Pete]

Attached are a few hours readings and settings via Michael's XaP application.

[Pete]

Attached are a few hours readings and Michael's XaP plugin settings.

[Michael McSharry]

It does look as if it has settled down. The sensor was probably just excited to be taken out of its cage and able to breath some good air.

This unit should make a nice IP-based thermostat where the UI is provided on the PC and the inner loop temperature control is done by the logic in the PIC. The output logic based upon the temperature thresholds should allow this to be achieved. This should be quite simple for single stage control. Basic two stage control should also be easy for simplistic second stage logic. I will need to ponder a little more to determine if a rate-of-change control is feasible with only the WebControl. It should not be too much of an issue where the PC is able to augment the second stage control with setpoint updates based upon rate of temperature change and the fail-over being some static set point should the PC happen be be offline.

A System design issue will be where to get the power. When controlling a normal furnace or A/C the source power will be 24VAC and it would make sense to grab this power and step it down to around 7VDC. This would eliminate the need for walwart and always have power as long as the unit under control also has power. A simple diode-protected relay should be all that is needed on the output.

I will likely take the module with me when I leave tomorow and play with it in the evening for the PC-side UI to this. As is usually the case the software to interface to the hardware is trivial. The softare to interface to the human is what takes the time.

I would like to hear of of other ideas on effective use of the WebControl where advantage can be taken of the PIC logic and augmented by PC logic.

[Pete]

Curious relating to the message window; the XaP interface shows an initially onset of communications but nothing else. Is this intentional or is it supposed to show the TCPIP connection data reception / requests?

There is some documentation out there relating to the upper limits of both utilizing DC and AC power with the board. Think it mentioned an upper limit of 26VDC or AC and using heat sinks with a 12VDC power supply say for an automotive use.

Relating to thermostat control I you could also connect it to a number of dampers associated with strategically placed temperature sensors and have the unit control both the main furnace and dampers along with or in addition to humidity control.

My basement is unfinished today and in the summer (when the AC is on) I try to keep the humidity levels below 50% in the summer with a dehumidifier running all summer - its kind of a manual effort today but it would be nice to automate this a bit.

Unrelated and a bit of a tangent but I have a parrot (she's been with me now for 30 years) and I would like to figure out a way that she get along with my wife as they have never seen eye to eye and literally have shouting matches.....with the parrot usually getting the upper hand....

[mfisher]

It's a 10.5 MIPS PIC18F66J65 with 96KB Flash program memory, 3.8KB RAM, 11 channels 10-bit analog in, 10base-T with a -40 to 85C and 2 to 3.6 vdc operating range. The PCB also has a SST25VF040B-50 4Mbit serial flash. I/O is buffered and level-converted through two LVX4245 dual-voltage octal transceivers. The crystal is a 25.000H8B ( dunno temperature range, but typically -20 to 70c).

HTH ... Marc

Thanks for the info!

[hult]

It does look as if it has settled down. The sensor was probably just excited to be taken out of its cage and able to breath some good air.

This unit should make a nice IP-based thermostat where the UI is provided on the PC and the inner loop temperature control is done by the logic in the PIC. The output logic based upon the temperature thresholds should allow this to be achieved. This should be quite simple for single stage control. Basic two stage control should also be easy for simplistic second stage logic. I will need to ponder a little more to determine if a rate-of-change control is feasible with only the WebControl. It should not be too much of an issue where the PC is able to augment the second stage control with setpoint updates based upon rate of temperature change and the fail-over being some static set point should the PC happen be be offline.

A System design issue will be where to get the power. When controlling a normal furnace or A/C the source power will be 24VAC and it would make sense to grab this power and step it down to around 7VDC. This would eliminate the need for walwart and always have power as long as the unit under control also has power. A simple diode-protected relay should be all that is needed on the output.

I will likely take the module with me when I leave tomorow and play with it in the evening for the PC-side UI to this. As is usually the case the software to interface to the hardware is trivial. The softare to interface to the human is what takes the time.

I would like to hear of of other ideas on effective use of the WebControl where advantage can be taken of the PIC logic and augmented by PC logic.

Not coincidentally, I plan to incorporate WebControls into our multi-furnace, multi-zone, multi-loop home solar+geothermal HVAC system.

What would be extremely helpful would be if theWebControl logic provided < and > and a better way to create a deadband or hysteresis than timers.

One solution is to add external hardware to pre-process the data, but this can't be done with the DS18B20 temperature sensors.

I plan to design a stackable PCB board which could serve one or more WebControls pcbs:

The footprint of the board is large enough to provide Power Over Ethernet (POE) power supplies and at least some other functions. Candidates include:

24VAC or DC --> 5.6vdc DC-DC preregulator for WebControl
24VAC or DC --> 5VDC DC-DC power supply for auxilary equipment (WIFI, TCP/IP-->RS-xxx, hub)
24VAC or DC --> 13.8 vdc power supply ( eg motion detector) and charger for battery backup
I/O multiplexors ( 8- or 16-channel, analog or digital, input or output)
Differential amplifiers for signal conditioning ( transducers, signal differencing)
Relays for external control and isolation

Other suggestions welcomed.

I will post a picture of the WebControl in the factory case in a stack with similarly-sized pcbs for 5-port hub, WiFi, TCP/IP-> RS-xxx conversion ( from www.comtrol.com).

HTH ... Marc