P4 settings for chiller.
- Thread starter Ming_
- Start date
- Tagged users None
I just ran a test with my settings and the outlet turned on at ~79.0F then back off once it dropped below ~79.0F. I say approximate, give or take 0.1F. It was definitely off well before reaching nominal of 77.18F.
T = 77.18 + (0.83)(0.396) + 1.44
I´ll think i have figured it out now. If you use cooling differences - you adjust up the nominal value with the set value. If your nominal value is 25.4 - the on nominal value for cooling will be 25.9 degre and the cooling on/off will take place with the actual hysteresis around that figure.
Sincerely Lasse
Sincerely Lasse
Last edited:
Keeping the settings the same, a Cooling Difference of 0.00F resulted in the following:
On: 77.7F
Off: 77.3F - 77.4F (the number jumped from 77.5F to 77.3F when the outlet went off)
On: 77.7F
Off: 77.3F - 77.4F (the number jumped from 77.5F to 77.3F when the outlet went off)
- Joined
- Jul 16, 2009
- Messages
- 5,071
- Reaction score
- 8,108
I can't confirm how the P4 works - as I don't think it operates as described.
But in theory a PID has no "set point" but a set UPPER and LOWER control point. The range in between is the hysteresis band.
So you want 100 and you set +/- 2 then :
For HEAT the control turns ON at 99 and OFF at 101.
For COOL the control turns on at 101 and OFF at 99.
I think you can see the problem here - the system will want to bounce between HEAT and COOL.
In an ideal world we would heat to 100 and heat would start at 99 or below. We would cool to 100 and COOLING would start at 101. This is two degrees of hysteresis. with no overlap. We would also build logic to say that in case of overshoot (our PID is not good) that NO HEAT or COOL cycle can happen with in X amount of time of the opposite. This prevents slight overshoot and subsequent "hunting" if the set point by osicallations between heating and cooling.
I (honestly) gave up on trying to figure out how the P4 works. The settings and instructions are convoluted and very poorly worded. The nominal value appears to be the middle of the band (100 in our example) and the cooling delay maybe the the hunting prevention timer.
But in theory a PID has no "set point" but a set UPPER and LOWER control point. The range in between is the hysteresis band.
So you want 100 and you set +/- 2 then :
For HEAT the control turns ON at 99 and OFF at 101.
For COOL the control turns on at 101 and OFF at 99.
I think you can see the problem here - the system will want to bounce between HEAT and COOL.
In an ideal world we would heat to 100 and heat would start at 99 or below. We would cool to 100 and COOLING would start at 101. This is two degrees of hysteresis. with no overlap. We would also build logic to say that in case of overshoot (our PID is not good) that NO HEAT or COOL cycle can happen with in X amount of time of the opposite. This prevents slight overshoot and subsequent "hunting" if the set point by osicallations between heating and cooling.
I (honestly) gave up on trying to figure out how the P4 works. The settings and instructions are convoluted and very poorly worded. The nominal value appears to be the middle of the band (100 in our example) and the cooling delay maybe the the hunting prevention timer.
What we talking about here is not a PID programming - PID stands for proportional–integral–derivative controller and its a system that is rather good described here With loops - this system make the real value to get closer and closer to the nominal value and you can with help of the three variables P, I and D adjust for both slow processes (heating water to a certain value) or very fast processes like adjust the input of heated water from a heat exchanger to a tap water system. The system is a little bit self learning
GHL two position controller is a standard on/off system there the hysteresis tell you how much deviation that is allowed. The hysteresis is divided in an upper and lower deviation depending on if the you want a rising or a decreasing effect. Hence adjusting upwards - upper endpoint is nominal value +1/2 hysteresis - adjusting downwards - lower endpoint is nominal value - 1/2 hysteresis. For P4 is valid for all sensors but the temp sensor is a little bit more complicated because the use of different type of heaters that was common when the first profilux system was constructed plus that both heating and cooling managed by the same probe was very common during time with MH lighting. IMO - GHL regulation system more or less follow European standards according to on/off regulation
An on/off system will always be like the tooth of a saw - just around the wanted value and is IMO not so good to use in slow processes or where there is a large distance between the point of action and the measurement probe. There is one situation there you IMO need to use a on - off system and that is with compressor-driven coolers. The compressor does not work well if it is to many stop and starts per time unit
However GHL also offer other regulating systems that some times are very practical to use. Pulse fixed/Pause fixed; Variable pulse/fixed pause and Fixed pulse/variable pause. For cooling - there is also a 1-10V controller (for fans) but can also be used for regulating valves.
In my aquarium I use Variable pulse for my heater and 1-10 V regulation for my cooling fan.
Pulse - max 4 min - variable depending to the distance to nominal value Pause always 30 sec if not real value = nominal value. This happens below 26 C Above 26 C - tha fan is 100% at 26.5
Resulting graph. The blue bars are my heater working - 24 Hours
12 Hours
Sincerely Lasse
- Joined
- Jul 16, 2009
- Messages
- 5,071
- Reaction score
- 8,108
HI - Lasse!
I am very familiar with PID - and am aware that the P4 does not implement PID control loop logic - but alas that is my point, the instruction manual and wording however are not that of a two position controller, but that of a PID controller. Operationally, even though it is NOT PID, the hysteresis is setup to mimic PID (+/- hysteresis around a point, not hysteresis TO a point, and that is NOT common. See below if I am being confusing.
And - this as far as I can tell is NOT implemented in a standard way and that is what is causing great confusion to most people. It is implemented to mimic PID between TWO limits and not from a control point TO a set point.
In a normal two position temperature control there are TWO CONTROL POINTS, the hysteresis is between those two points. What is confusing here is that GHL has opted to complicate the entire setup by using a +/- hysteresis with a centered "nominal" temperature. This (is at best confusing) but would not be so bad if there was only a single mode (HEAT or COOL).
So an IDEAL and EASY setup would be (just easy numbers, not real for anything)
For HEAT:
SET POINT = 100
HYSTERESIS = 10
That makes the call for heat CONTROL POINT = 90
The controller turns ON at 90 or LESS and turns OFF at 100 or more.
For COOL:
SET POINT = 100
HYSTERESIS = 10
That makes the call for cool CONTROL POINT = 110
The controller turns ON at 110 or more and turns OFF at 100 or less.
This would allow a VERY SIMPLE setup with 20 total degrees of hysteresis.
Having individual hysteresis settings for HEAT and COOL allows for customizations. Say the heater stops heating instantly, but the chiller overshoots due to the large cold coil, etc.
It is over complicated for absolutely no reason and honestly, I have struggled to get it to behave as described in the manual.
We don't disagree at all with this point and (see above) exactly why HEAT and COOL should have their OWN hysteresis settings TO the set point, not a nominal value (set point) surrounded by the hysteresis points. It is confusing for all involved.
I have absolutely no delusion that GHL will ever change or improve this part of the interface, there are so many other things that need to be improved (and likely never will be) that I am not going to hold my breath for something like this... and assume Matthias would disagree that there is a problem to begin with.
When I was setting up my P4 I had a thread on the operation I observed. From my testing in that thread and this thread, I do not think it is possible to have heating and cooling on at the same time (when using the same sensor). The Programming Guide states as much, but also seems to contradict itself.
5.2 Nominal Value Section
5.2 Nominal Value Section
The Programming Guide also does not do a good job at describing the difference between the Heater (Main/Tubular) and Substrate Heater Functions. It should, since they use the hysteresis differently.
I like what it does a lot, as I have my primary heater (125W) set to "Substrate Heater" and my backup heater (50W) set to "Heater". If my primary heater cannot keep up, then my backup heater will also come on.
Here's the operation I observed with my settings:
Nominal = 77.2F
Hysterysis = 0.4F
Decreasing Temperature
77.1F - Substrate Heater On (my primary heater)
77.0F - Heater On (my backup heater)
Increasing Temperature
77.3F - Heater Off (my backup heater)
77.4F - Substrate Heater Off (my primary heater)
So as you can see, the on/off points of each heater function is not +/- 0.2F as most would think for a 0.4F hysteresis.
I haven't tried it with anything other than a 0.4F hysteresis, so I'm not sure where those on/off points would be for each function with a different setting. The Programming Guide doesn't really say.
Here are the excerpts I could find:
5.2 Nominal Value Section
5.4 Hysterysis Section
There is more in the Summer Switching section, but I have that function off.
I like what it does a lot, as I have my primary heater (125W) set to "Substrate Heater" and my backup heater (50W) set to "Heater". If my primary heater cannot keep up, then my backup heater will also come on.
Here's the operation I observed with my settings:
Nominal = 77.2F
Hysterysis = 0.4F
Decreasing Temperature
77.1F - Substrate Heater On (my primary heater)
77.0F - Heater On (my backup heater)
Increasing Temperature
77.3F - Heater Off (my backup heater)
77.4F - Substrate Heater Off (my primary heater)
So as you can see, the on/off points of each heater function is not +/- 0.2F as most would think for a 0.4F hysteresis.
I haven't tried it with anything other than a 0.4F hysteresis, so I'm not sure where those on/off points would be for each function with a different setting. The Programming Guide doesn't really say.
Here are the excerpts I could find:
5.2 Nominal Value Section
5.4 Hysterysis Section
There is more in the Summer Switching section, but I have that function off.
FYI nominal value and set point is the same. In the German manuals use the word Sollwert - that is set point in English - why it has been nominal value when translated - I have no clue
(from here)
And here
And by the way - have been working with PID systems (both European and American) for many years - never have been using a hysteresis in these system - I have use constants for P, I and D - but never a hysteresis. With this - I do not mean that there can not be systems that use both PID and hysteresis but the normal explained differences is these below - from here
I will say that GHL´s definition of a hysteresis band is correct compared with both my old knowledge and my new.
@ingchr1 I think that the differences in switch on and off you see in your example is caused by the fact that GHL seems to use a higher resolution than our temperature scale allows. They control from a mV signal (analog probe) or frequency on/off (digital probe) I have seen this when I use pulse variable as control function
Below is a graph of my regulation of my conductivity probe. Because I measure KH each 4 hours and do not reuse the waste (around 0.6 L/day) from the sampling - my salinity will be lowered by time. I use my conductivity probe in order to add new salt when needed for keep up my salinity. Do not bother about seeing
around 32 psu in salinity - my probe is not exact in value but in variation. I have around 34.6 PSU in salinity with other methods.
You can see here that the controlling system use different amounts of salt adding in spite of the salinity probe show a steady 32.1 value
My settings for pulse variable function is as below, sample frequency = 5 min and my pump gives 20,5 ml/minute
10 minutes is the max pulse - as you can see in this example the highest pulse is 146 seconds (around 2.5 minutes) The min correction is in this example 26 sec. 8 minutes is set because it takes around 8 minutes for the added saltwater to mix and reach the sensor from the inlet point.
Sincerely Lasse
For me this is the same as nominal value (set point) 100 F and hysteresis ± 20 F with differences if heating or cooling is used. Because I always check if my knowledge is old or still valid - I did a search on the hysteresis concept in temperature controlling and most I have seen have this definition of the hysteresis in on/off switching
(from here)
And here
And by the way - have been working with PID systems (both European and American) for many years - never have been using a hysteresis in these system - I have use constants for P, I and D - but never a hysteresis. With this - I do not mean that there can not be systems that use both PID and hysteresis but the normal explained differences is these below - from here
I will say that GHL´s definition of a hysteresis band is correct compared with both my old knowledge and my new.
@ingchr1 I think that the differences in switch on and off you see in your example is caused by the fact that GHL seems to use a higher resolution than our temperature scale allows. They control from a mV signal (analog probe) or frequency on/off (digital probe) I have seen this when I use pulse variable as control function
Below is a graph of my regulation of my conductivity probe. Because I measure KH each 4 hours and do not reuse the waste (around 0.6 L/day) from the sampling - my salinity will be lowered by time. I use my conductivity probe in order to add new salt when needed for keep up my salinity. Do not bother about seeing
around 32 psu in salinity - my probe is not exact in value but in variation. I have around 34.6 PSU in salinity with other methods.
You can see here that the controlling system use different amounts of salt adding in spite of the salinity probe show a steady 32.1 value
My settings for pulse variable function is as below, sample frequency = 5 min and my pump gives 20,5 ml/minute
10 minutes is the max pulse - as you can see in this example the highest pulse is 146 seconds (around 2.5 minutes) The min correction is in this example 26 sec. 8 minutes is set because it takes around 8 minutes for the added saltwater to mix and reach the sensor from the inlet point.
Sincerely Lasse
No in post 33
Sincerely Lasse
Sincerely Lasse
I thought I had a pretty good idea on how the Heater and Substrate Heater Functions operated. So, I decided to perform another test to confirm. For this test I used a water bath and doubled my hysteresis from 0.4F to 0.8F. I performed the test three times, and the results were exactly the same each time.
Settings:
Nominal Value = 77.2F
Hysterysis = 0.8F
Decreasing Temperature
77.0F - Substrate Heater On (my primary heater)
76.8F - Heater On (my backup heater)
Increasing Temperature
77.3F - Heater Off (my backup heater)
77.7F - Substrate Heater Off (my primary heater)
Based on the data I was unsure of the Substrate Heater On point, so I performed additional tests on that function using 1.6F and 3.2F Hysterysis. The On points were 76.9F and 76.6F respectively.
From these results and the ones with my normal 0.4F hysteresis, I think it can be concluded that the heating functions of the Profilux operate as follows:
Heater Function ON = Nominal Value - (0.5)(Hysterysis)
Heater Function OFF = Upon exceeding the Nominal Value
Substrate Heater ON = Nominal Value - (~0.2)(Hysterysis)
Substrate Heater OFF = Nominal Value + (0.5)(Hysterysis)
@Lasse @vinny @Gaël #ghlusers
Settings:
Nominal Value = 77.2F
Hysterysis = 0.8F
Decreasing Temperature
77.0F - Substrate Heater On (my primary heater)
76.8F - Heater On (my backup heater)
Increasing Temperature
77.3F - Heater Off (my backup heater)
77.7F - Substrate Heater Off (my primary heater)
Based on the data I was unsure of the Substrate Heater On point, so I performed additional tests on that function using 1.6F and 3.2F Hysterysis. The On points were 76.9F and 76.6F respectively.
From these results and the ones with my normal 0.4F hysteresis, I think it can be concluded that the heating functions of the Profilux operate as follows:
Heater Function ON = Nominal Value - (0.5)(Hysterysis)
Heater Function OFF = Upon exceeding the Nominal Value
Substrate Heater ON = Nominal Value - (~0.2)(Hysterysis)
Substrate Heater OFF = Nominal Value + (0.5)(Hysterysis)
@Lasse @vinny @Gaël #ghlusers
