Fan & Servo Questions for New Build with RD3


 
D

Dominick A

TVWBB Fan
I got my RD3 all put together last night and have been combing the forum for tips on these items and could use some additional guidance, please.

With the RD3 on a table and a TC connected as the pit probe:

1) My fan is running non stop. I have all of the fan output values listed at 0%. The probe is reading room temp, the set point is 20 degrees below that. When the auto fan mode is disabled, I can control the servo. If I leave it open, then enable auto fan mode, it closes fully, so that seems to be working. Initially, the graph shows that the fan is at 100% and then after a few minutes it shows fan at 0% but it is clearly still running. What am I missing here with the fan?

2)Servo is fully closed when it shows 20% or less and is not quite fully open at 100%. I have been tinkering with the outside damper plate and the servo pulse duration settings to try to get it right. I'm still working through it, but is this pretty standard to work out the kinks and get it right or did I mess something up from the beginning in putting the servo and the damper plate together?

3)Ultimately, I want to get the servo and fan functioning where the fan only kicks on after the servo has been fully opened and the set temp is not yet achieved. I followed this for the firmware I am running and no longer see the "On at Max only" option that I had before.

"To install the snapshot firmware, just go to System -> Backup / Flash Firmware and put the snapshot url in the "Image URL (.gz)" box. This will also update your HeaterMeter's AVR Firmware to the snapshot as well.
http://heatermeter.com/devel/snapshots/bcm2708/openwrt-rpi.gz"

Thanks!
 
You have a couple issues here, let me address them one at a time....

First, RD3 calibration.

In your HM Config for Servo Pulse Duration set numbers 700 - 2100 as ballpark starting numbers. Do not select Inverted or Open/Closed only.

Now make your HM go to 100% output

When servo is at 100% position mate the damper disc on top of the RD3 body so the openings are aligned on top of each other, or as close to that as possible.

Unplug the CAT5 cable from the RD3 and make sure the damper disc can still rotate at least a tiny bit in the open direction. If not then remove the damper disc and rotate it one tooth in the open direction and re-seat it.

Now make your HM go to 0%

Your RD3 should move to somewhere close to closed. Adjust the lower SPD number (in small amounts) to adjust the closed position until it is completely closed.

Unplug the CAT5 cable and make sure you can rotate the damper disc at least a small amount further in the closed direction (you do not want your servo to be pegged at the end of its range in either the 0% or 100% position)

Now make your HM go to 100% again, adjust the upper SPD number until the valve is perfectly open.

Now enter a high setpoint in the HM, it should go to 100% and RD3 should fully open. Toggle Lid Mode and the RD3 should completely close. Bounce it back and forth a couple times to make sure the motion range is right, also make sure the servo is not pegged in either direction.

Tweak upper or lower SPD numbers if needed, remove and rotate the damper disk a tooth or two in the appropriate direction if the servo is reaching the end of its range in either 0% or 100% position...
 
Last edited:
On the fan operation mode, that is something that is actually evolving at the moment. What you are asking for is not quite available right now but it should be coming soon I think...

Right now with the snapshot release you are running you have the Fan "On Above" option, which you can set a value there and the blower will start up (at MIN value) when the HM output reaches that %, and the fan will ramp up from there to the MAX value when the HM output is 100%. So in this scenario the servo will not fully open before the fan starts running (again, there has been some discussion recently and I think Bryan decided we will move in that direction soon)

Previously we had Fan On at "Max Only", which would have the fan off while the HM output was from 0-99%, and the fan would come on at MAX value when the HM output is 100%. You can still do this by selecting "On Above" 99%. This setup, however, is problematic in my experience because the fan comes on at MAX speed and the air flow in the pit changes very abruptly. Typically this results in a rapid dip in pit temp followed by a large amount of overshoot as the HM stokes the pit to compensate for the rapid temperature drop. With the current firmware I would suggest you select "On Above" something like 50% and then set the MAX fan speed to something reasonable for your pit.
 
Finally, on your fan running all the time, that sounds like a hardware malfunction to me.

If your HM display Output bar (or % on the HM LCD) is showing 0% output then your fan should not be blowing, if it is then you have a hardware problem.

The RDTC board has direct traces from the CAT5 jack to the blower/servo header, there are no electronic components in the circuit there. You should not have a problem on the RDTC board with the blower/servo as long as you have attached the wires to the right holes on the header, unless there is a big blob of solder on the board causing a short or something.

So that means your problem is likely on the HM board in the blower driver circuit (unfortunately)

You can use your MultiMeter in DC Voltage mode, ground the black lead on the power jack ground and touch the red lead to the BLOW pin on the HM CAT5 jack while the HM output is at 0%. If you see voltage on the BLOW pin when the HM output is 0%, "Houston, we've got a problem"... Unplug the CAT5 cable while testing just to be sure nothing outside the HM could be influencing your measurements.
 
RalphTrimble said:
Finally, on your fan running all the time, that sounds like a hardware malfunction to me.

If your HM display Output bar (or % on the HM LCD) is showing 0% output then your fan should not be blowing, if it is then you have a hardware problem.

The RDTC board has direct traces from the CAT5 jack to the blower/servo header, there are no electronic components in the circuit there. You should not have a problem on the RDTC board with the blower/servo as long as you have attached the wires to the right holes on the header, unless there is a big blob of solder on the board causing a short or something.

So that means your problem is likely on the HM board in the blower driver circuit (unfortunately)

You can use your MultiMeter in DC Voltage mode, ground the black lead on the power jack ground and touch the red lead to the BLOW pin on the HM CAT5 jack while the HM output is at 0%. If you see voltage on the BLOW pin when the HM output is 0%, "Houston, we've got a problem"... Unplug the CAT5 cable while testing just to be sure nothing outside the HM could be influencing your measurements.
Click to expand...

I adjusted the on above to 50% and with the output reading 0%, the BLOW pin was showing 12V. I then unplugged power and checked for continuity to the surrounding pins. I don't know if there is such thing as a partial short, but on my multimeter, I have a 4 digit readout. If there is no continuity, it just shows a 1 as the left-most digit. If there is continuity, it generally shows 002 or 001. In this case, I am getting 170 from BLOW to GND and 170 from +5V to GND. Does this indicate anything? Visually, it looks fine. What now?
 
Ok, there is a problem with the blower circuit.
The blower drive is fairly simple, just a couple components, the feedback circuit however is a bit larger... So first thing to do is in your HM config change the Fan Output mode from Voltage to PULSE. Pulse mode does not need the feedback circuit. If the fan is still blowing after that change then we need to look at the big MOSFET that drives the blower, and the smaller MOSFET that is between the ATMega and the large MOSFET. I think closeup pics of that end of the board around the large MOSFET would be in order at that point. (both sides of the board)
 
If changing to voltage mode doesnt make the blower go off then you can take a look at what is going on with the bit MOSFET Transistor that drives the blower
With the HM sitting as you would normally use it (Display facing you), locate the 3 legs from the large MOSFET just under the Boxy Robot...
WIth your meter in DC Voltage mode ground your black lead to the power jack GND, then on the MOSFET the pin on the right should be 12VDC supply voltage. The pin on the Left is the control voltage, when it reads the same as the pin on the right (~12VDC) the Output Pin in the Middle should have no voltage on it (when HM is at 0%). As the voltage on the left pin drops the output voltage on the center pin should rise upward....
So there is some testing that can be done, but my guess is the pictures may tell the story, you probably have a solder bridge under the larg MOSFET legs if I had to guess...
 
Thanks Ralph, Pulse didn't stop it, I will test and report back tomorrow with pictures if need be.
 
Last edited:
RalphTrimble said:
If changing to voltage mode doesnt make the blower go off then you can take a look at what is going on with the bit MOSFET Transistor that drives the blower
With the HM sitting as you would normally use it (Display facing you), locate the 3 legs from the large MOSFET just under the Boxy Robot...
WIth your meter in DC Voltage mode ground your black lead to the power jack GND, then on the MOSFET the pin on the right should be 12VDC supply voltage. The pin on the Left is the control voltage, when it reads the same as the pin on the right (~12VDC) the Output Pin in the Middle should have no voltage on it (when HM is at 0%). As the voltage on the left pin drops the output voltage on the center pin should rise upward....
So there is some testing that can be done, but my guess is the pictures may tell the story, you probably have a solder bridge under the larg MOSFET legs if I had to guess...
Click to expand...

The right 2 pins on the MOSFET have been reading the same 12V when the output should be 0%.

I thought I saw a possible solder bridge between those 2 legs, so I used a fluxed dipped solder wick to clean out between those 2 pins. I wound up fully resoldering the right pin. The fan still runs when it should be off. The pic here shows the resolder, and I think just a reflection from my flash. However, on the closeup, there looks like there might be something between the left and middle pin. I guess a clean resolder on that left pin wouldn't hurt.

https://drive.google.com/file/d/0BzSBmZGY0VqYd0JxaUpaQmRyVzA/view?usp=sharing

In the assembly instructions on the github site, it says "Soldering the tab to the PCB is not required for the standard blower" Would that hold true for this blower? I did not solder it. http://www.digikey.com/product-search/en?keywords=603-1370-ND

Here is the other end of the board. I think the smaller MOSFET there looks ok.
https://drive.google.com/file/d/0BzSBmZGY0VqYY1hGbXNaQ2dtbzg/view?usp=sharing

Here's a look at the component side of my board:
https://drive.google.com/file/d/0BzSBmZGY0VqYUngzbW9XWjhVOE0/view?usp=sharing
 
I never solder the large blower driver MOSFET to the board, it is not required. Typically mounting a MOSFET is done to dissipate heat, but I have run a 25CFM blower without it and haven't blown the MOSFET.... so don't bother with it is my advice...
Looking at your pics now....
One comment though... you reported that the 12V supply voltage is there on the right leg, and 12V is also on the output leg (Center leg). You neglected to state what voltage is on the control (left) leg, which is really the important part.....
 
The second pic is useless, the smaller mosfet we are talking about is right above the larger mosfet, not on the other end of the board.
To be clear, PIN 5 on the ATMega is where the control signal for the blower comes out, the trace from that pin runs across the board right under the tactile switch, and then to the BS170 MOSFET that is right above the larger MOSFET. The smaller MOSFET connect to the larger MOSFET's left pin to control it. IF the voltage on that pin is 12V like the right pin the center pin should have no voltage on it. As the voltage on the left pin goes downward from 12V the voltage on the center pin should go up.

Looking at your pics I cant see anything that looks like a solder bridge, but you can't always see everything in a 2 dimensional pic.

Take a voltage measurement on the left pin of the large MOSFET and let me know what it reads....

The one thing I took away from your pics is the large MOSFET looks like it has been very hot... More than half the surface looks kinda melted like it's been over heated either in soldering or perhaps by being shorted out and burned up.... So you may have a bad MOSFET there, IDK WHY, but I'm just sayin... Looks like its been really hot...
 
Last edited:
Just fyi, testing the voltage on the left leg of the large MOSFET with the blower Min/Max set to 0/100 in PULSE mode the value goes from the supply voltage (~12v) down to no volts, the lower the voltage on this control leg the higher the voltage on the center (output) leg. As one goes down the other goes up if things are working properly.

In VOLTAGE mode, however, I only see the control leg dipping down into the 9V range, though I am testing without a blower connected so perhaps that throws off the feedback control circuit. At any rate, checking in PULSE mode seems to be an easier way to verify what is going on.

The control output for the blower comes from pin 5 of the ATMega and connects to the top pin of the small transistor right above the big blower MOSFET. The voltage on that pin should range from the VCC of the ATMega (3.3v) when the blower is at 100% down to no voltage when the blower is at 0%.

So you can check what is going on in those two places, and also make sure you dont have something shorting to the BLOW pin on the CAT5 jack there, it looks like you've got quite a mess of bare exposed wires there... Perhaps the 5V jumper wire you added for the broken trace is shoring on the BLOW pin?
 
Last edited:
The burnt look on the large MOSFET was just oil from my fingerprint, it wiped right off.

The voltages I gave were with PULSE mode enabled.

Here's a better pic of the smaller MOSFET and my brutal CAT5 jack wiring:
https://drive.google.com/file/d/0BzSBmZGY0VqYY1ZDQlhreU82X0E/view?usp=sharing

All 3 pins of the smaller MOSFET showed 0V when the fan output should be 0%, same thing at pin 5 of the ATMega.
When the fan SHOULD be at 100%, the top pin was only showing .84V, same thing at pin 5 of the ATMega.

I removed the ATMega and reseated it, no change.

As for the BLOW pin on the CAT5 jack, I checked for continuity to the surrounding pins. I don't know if there is such thing as a partial short, but on my multimeter, I have a 4 digit readout. If there is no continuity, it just shows a 1 as the left-most digit. If there is continuity, it generally shows 002 or 001. In this case, I am getting 170 from BLOW to GND and 170 from +5V to GND. Between BLOW and +5V, it is showing roughly 1800, but I think when I checked this the other day, it was flashing between 170 and 1800. Does this indicate anything?

Something tells me I have 2 issues here. Likely a short on the BLOW pin, and something wrong with the ATMega since it should be outputting 3.3V on pin when the output should be 100%, right?
 
Last edited:
The output on pin5 of the ATMega should vary from 0 to 3.3v, yes.... But if something on the board was shorting that circuit to ground it wouldn't be able achieve that. You could pull the ATMega from the socket and check for a short to ground on pin 5 of the socket to see if that is the case.
Verify that the small transistor says BS170 on it.... Could be the wrong part there, or could be that BS170 is blown. The three connections on that BS170 are from the ATMega pin 5, to the blower MOSFET and GROUND.
 
In your most recent pic it kinda looks like the legs on the right and center of the large MOSFET have a solder bridge? I can't really see clear enough to say, but look at that yourself and make sure that is not the case.
 
Check the blue wire where it crosses the green wire and make sure it didn't melt the insulation on the green wire and short when you soldered it.
 
Yes, that should go without saying, check all those wires to be sure the insulation isn't melted off and the wire shorting on something else. It kinda looks like the shorter jumper you added on the OKI has a solder bridge on the trace below, but that would short 12v to 5V and I think you have both 12V and 5V present on the board so I guess that is not shorted, but check that out anyway, cause you've really got something odd going on here.

Since you shorted things out and made your ATMega flake out once upon a time, requiring a reflashing to get it working again, perhaps you do have something odd going on with the ATMega. I wouldn't normally suggest this, but one thing you can do is remove the ATMega from the socket and very carefully bend pin 5 outward, then reinstall the ATMega into the socket so that leg does not go into its socket. You dont need to seat the chip all the way down, just get it in far enough for all the legs to make connection (except pin 5). Now Pin 5 is isolated from everything and you can measure the voltage on that pin right as it comes out of the ATMega, completely unaffected by the rest of the circuits on the board. Now measure the voltage on that leg, it should go from 0v to 3.3v when the blower speed is 0% to 100% respectively (with the blower set to PULSE mode). This is where your control signal originates, so you will know if it is bad from the source or if you have a problem elsewhere on the board.
 
In order of suggested issues.......

-No short to ground on pin 5 of ATMega socket
-The small transistor DOES say BS170 on it
-The blue and green wires were NOT in contact
-ATMega pin 5 bent outward measured 0V at 0%, 1.67V at 50%, and 3.3V at 100%.

I also cleaned up the cat5 jack connections to eliminate my doubts there.

Given my earlier measurements on the BS170, would this indicate a bad BS170, or should I still be looking for a short?

Thanks again for all of the suggestions!
 
You must log in or register to reply here.

 

Back
Top