Fröling T4 boiler circulation pump turns continuously on and off in standby mode.

Wir betreiben einen Fröling T4 Hackschnitzelkessel mit einer drehzahlgeregelten Wilo Stratos Umwälzpumpe (0–10 V Steuersignal). Der Kessel ist außerdem mit einem Rücklaufmischer ausgestattet. Wir verfügen über einen 3000-Liter-Pufferspeicher. Wir verwenden Hydrauliksystem 4 mit vier Pufferspeichertemperatursensoren.

Der Kessel läuft beim Heizen einwandfrei, aber im Leerlauf/bei ausgeschalteter Feuerung spielt die Pumpensteuerung verrückt. Kurz nach der Feuerung schaltet sich das Pumpensteuersignal kontinuierlich ein und aus, oft nur für etwa eine Sekunde. Die beigefügten Oszilloskop-Fotos zeigen das Pumpendrehzahlsteuersignal über einen längeren Zeitraum und ein vergrößertes Beispiel. Beachten Sie, dass die Pumpe ständig mit Strom versorgt wird, aber über das Drehzahlsteuersignal ein- und ausgeschaltet wird.

In Beispiel 2 decken die kleinen Abbildungen den Kesselbetrieb vom Heizbeginn über „Stöber aus“, „Wartezeit Abschaltung“ bis hin zum „Standby“ ab. Die kleinen Abbildungen 51 bis 54 decken das Pumpensteuersignal etwa eine Stunde nach der Feuerung ab. Man sieht, dass das Drehzahlregelsignal der Pumpe sehr schnell ein- und ausgeschaltet wird – viel schneller, als die Wilo Stratos nach Aktivierung des Drehzahlregelsignals starten soll. Nach einigen Stunden beruhigt sich das Regelsignal, die Kesseltemperatur sinkt auf die Puffertemperatur und darunter.

Diese Art der Pumpenregelung widerspricht den Spezifikationen von Wilo. Ich vermute, dass dieses ständige Ein- und Ausschalten die Pumpe ausfallen lässt. Wir haben innerhalb von sechs Betriebsjahren zwei Pumpen verloren, während andere Pumpen seit 45 Jahren problemlos laufen.

Seit dem Ausfall der zweiten Pumpe habe ich das Pumpenregelsignal vorübergehend angepasst. Auf den Fotos ist zu sehen, dass das Regelsignal immer bei 3 V oder darüber liegt. Das bedeutet, dass die Pumpe tatsächlich dauerhaft mit Mindestdrehzahl oder mit höherer Drehzahl läuft, wenn der Kessel eine höhere Drehzahl anfordert. Ohne diese Mindestdrehzahl von 3 V würde das Regelsignal noch stärker schwanken. Ziel war es, das ständige Ein- und Ausschalten der Pumpe zu vermeiden. Später habe ich diese Schaltung leicht modifiziert, sodass die Pumpe abschaltet, wenn das Kesselsteuersignal für etwa 20 Minuten oder länger ausfällt. Dies ist jedoch nur eine vorübergehende Lösung zum Schutz der Pumpe.

Soweit ich weiß, schaltet der Kessel die Pumpe im Standby-Modus ein, sobald die Kesseltemperatur die Pufferspeichertemperatur um etwa 2 Grad überschreitet. Das Problem scheint zu sein, dass überhaupt keine Hysterese vorhanden ist. Unser Techniker vor Ort weiß nicht, warum sich die Pumpendrehzahlregelung so verhält. Das Werk hat bisher keine Hilfe geleistet (ich habe vor zwei Jahren zum ersten Mal um Hilfe gebeten). Mir wurde gesagt, dass der Hydraulikmodus mit zwei Pufferspeichersensoren gut funktioniert. Ich verstehe nur nicht, warum sich der Standby-Modus vom Hydrauliksystem 4-Modus unterscheiden sollte. Soweit wir wissen, scheint es keinen Parameter zu geben, der das Verhalten beeinflusst, weder einen Hystereseparameter noch sonst etwas, das helfen könnte (weder im Standard-Benutzermodus noch im -x-Benutzermodus).

Google hat mir mitgeteilt, dass mindestens zwei weitere Fröling-Kunden (US-Markt) das gleiche Problem haben und beide den Hydraulikmodus 4 nutzen. Sie haben das Problem mit einem Pumpenzeitrelais gelöst, verwenden aber eine Kesselumwälzpumpe mit fester Drehzahl. In Kombination mit dem 3-Wege-Mischer sollte es eine bessere Lösung geben.

Ich frage mich, ob jemand aus dem Hauptmarktgebiet von Fröling dieses ähnliche Problem bei der Nutzung des Hydraulikmodus 4 beobachtet und eine Lösung dafür gefunden hat.

In English if you prefer:

We have a Fröling T4 wood chip boiler with a Wilo Stratos variable speed circulation pump (0-10 V control signal). The boiler is also equipped with a return feed mixer. We have a 3000 litre buffer tank. We use hydraulic system 4 with 4 buffer tank temperature sensors.

The boiler runs perfectly when heating but in idle/fire off state the pump control is crazy. Soon after fire off the pump control signal turns on and off continuously, often on for about once second only. The attached oscilloscope photos show the pump speed control signal over a longer period and a zoomed in example. Note that the pump has power all the time but is switched on and off via the speed control signal.

On example2 the small figures cover the boiler operation from start of heating to "stoker off", "shutdown wait" and finally "standby". Small figures 51 to 54 cover the pump control signal about an hour after fire off. One can see that pump speed control signal turns on and off very rapidly, much faster than the Wilo Stratos is supposed to start after speed control signal activation. After a few hours the control signal calms down, the boiler temperature drops to the buffer top temperature level and below.

This kind of pump control is against the specifications Wilo gives for the pump. I believe this continuous on and off control makes the pump fail. We have lost two pumps within 6 years of operation while we have other pumps that have been running continuously for 45 years without any problem.

Since we lost the second pump, I have modified the pump control signal temporarily. On the photos one can see that the control signal always stays at 3V or above. This means the pump actually runs continuously at minimum speed or at a higher speed if the boiler asks for a higher speed. If this 3V minimum speed was not added, the control signal would be fluctuating even more on and off. The intention was to avoid switching the pump continuously on and off. Later on I have modified this circuit a bit so that the pump turns off if the boiler control signal is off for some 20 minutes or more. Still just a temporary fix trying to protect the pump.

My understanding is that the boiler starts to turn the pump on in standby mode when the boiler temperature exceeds the buffer tank top temperature by some 2 degrees. The problem seems to be that there is no hysteresis at all. Our local dealer technician does not know why the pump speed control behaves like this. The factory has not provided any help at least so far (I have requested help 2 years ago for the first time). I've been told the two buffer tank sensor hydraulic mode works well. I just don't understand why the standby mode operation should be any different from that in hydraulic system 4 mode. There does not appear to be any parameter that affects the behaviour, no hysteresis parameter or anything else that might help as far as we understand (not in standard user mode or the -x user mode).

Google tells me at least a couple of other Fröling customers (US-market) suffer from the same issue, both using hydraulic mode 4. They have handled the issue with a pump timer relay but they use a fixed speed boiler circulation pump and together with the 3-way mixer a better solution should exist.

I wonder if anyone from the main Fröling market area has observed this similar issue when using hydraulic mode 4 and if anyone has found a solution to fix the issue?

Thank You for your reply.

The T4 seems very similar to other Fröling boilers as far as I've seen wiring diagrams. Return sensor is connected to the core module (wood chip module) as well as the boiler pump, buffer tank sensors are connected to the hydraulic module. Buffer tank temperature sensors are less than 10 metre long. I have measured the AC-voltage at the boiler control boards with a basic multi-meter and found it to be negligible. I cannot remember the exact figures, I have to measure this again with my better multi-meter and report the exact figures.

Also temperature readings at the H3200 controller screen appear very stable but the screen does not update measurements very fast. I would assume a properly built control software implementing decent time hysteresis too to reduce the control errors from input noise. Then again, an acceptable implementation may assume proper input signal quality. Better to double check this signal noise anyway.

I have now checked temperature sensor readings with a 4 digit multi-meter. Voltage figures were in the 2.635 V range and the mV digit was absolutely stable. 0 V in AC-mode, even peak hold min/max zero. But this was a stationary situation, no pump switching or such. Someone at the Hearth.com forum reported his top buffer tank sensor showing a one degree drop when the pump starts, then one degree up when the pump stops. Obviously need to try to monitor sensors during these transient cases. Our sensors have a common ground at the boiler connectors, the pump or the mixer should not affect sensor readings but one never knows. Reading from the console though sensors remain stable at the 0.5 degree resolution but the slower update rate may hide some spikes.

I already tried to capture sensors with my memory oscilloscope but realized the cheap PC-scope has an 8 bit ADC and no option for analogue DC-offset to increase gain and the scope is not capable of reading sensors at sub-degree accuracy, more like a couple of degrees resolution which is a no go.

Somewhere at this forum I found discussion about access modes other than user and -7, I wonder if these give better access to hysteresis parameters and such.

Meanwhile I also got a response from the factory (via our dealer, the factory does not want to talk directly with a customer) where they suggested the boiler running quite hot and not able to transfer heat the buffer properly. This is nonsense, at the start of standby mode, the lower buffer temperature is less than 70 degrees while the boiler temperature is in the 81 to 82 degrees range. However the boiler does not open the mixer and run the pump properly but instead keeps the return temperature very close to the boiler temperature.

I should have mentioned that we have a check-valve after the pump in the return feed line. The check valve was included with the boiler installation parts and intended to allow a cascade setup. The check-valve prevents incorrect water flow when a second boiler is heating and the one in question is off. I'm afraid the check-valve prevents natural water circulation from temperature difference between the boiler and lower part of the buffer. The pump might not need to come on to cool the boiler if there was no check valve. The mixer however would need to be opened by the boiler. No big deal to run the pump too at the same time for a short while (but longer than a second or so as currently).

I'll do some more measurements and report later. Hoping also feedback from other hydraulic mode 4 users as well as any other ideas.

Thank You Holzheizer for all your ideas.

I believe my Fluke multi-meter is quite fast in peak measurement mode but probably difficult to use when trying to capture transients from switching pumps and mixers etc. specifically when the actual temperature also slowly changes.

Very good idea to use AC-coupled mode with the oscilloscope. I just hope my cheap Chinese USB-scope does have a true AC-coupled input option instead of implementing AC-coupled input via digital signal processing at the laptop.

I'm hesitating to touch sensor wiring, also the boiler gives a fault code and switches off if a sensor is disconnected. Adding a parallel resistor to either of the temperature sensors would be possible on-the-fly and without disconnecting existing wiring. In this case I guess reducing the boiler temperature reading by a couple of degrees with a parallel resistor.

Very nice that you have looked at different parameters being available with different access modes. A pity if there does not appear to be any adjustable hysteresis for pump control.

I agree earth loops and inductive or capacitive coupling may be an issue even if this feels unlikely when our boiler house does not really have other devices except those for boiler operation and all wiring there is coming directly from the boiler circuit boards. Should have a common ground but one never knows.

I'll think about all the proposals but will probably first try the scope with AC-coupling to capture any temperature sensor transients during the period where the pump control signal misbehaves.

Our local dealer has also suggested capturing a log with their add-on logging device. This should be informative as it would give exact readings of all measurements exactly as the boiler controller sees them. I just don't know if log analysis needs to be done at the factory or if the local dealer can decode the data.

My comments below inserted to your text:

Quote from Holzheizer12345

I don't even think that the problem is caused by disturbances, but the method to narrow down the source of the problems should be "from outside to the inside". So the regulator has it's input signals (temperature of boiler and upper tank) and this will generate some sort of output signal for the pump control signal. So my suggestion would be to first check the input signals for the regulation loop before looking deeper into the controller settings.

NF: My noise measurements on the temperature sensors are ongoing. My scope does support a true AC-mode and I can read transients at some 10 mV level. First observation is that there are some very rare spikes of 20 mV but those do not seem to appear at the time the pump control signal oscillation occurs. The duration of these spices is extremely short, less than 100 microseconds. Unless bad design, I would assume the boiler having a low pass filter that attenuates these spikes completely away (if those even are real, I would not be surprised if the scope picks up some internal noise).

But even when there would be a parameter in development level, which could change the 2K Hysteresis, why should it be changed at your controller since you never accessed the development level?

NF: The boiler has behaved like this since new. I assume the 2 degree hysteresis has never existed. Perhaps a bug with the hydraulic system 4 factory defaults or even code. Only owners who use hydraulic system 4 seem to face this issue.

When you observe this oscillating pump control signal, how are the temperature values at this time?

NF: In standby state the boiler temperature stays at about 82 degrees, the buffer tank top temperature at about 1.5 degrees less as read from the boiler controller screen. Looks like the difference hits 2 degrees only very shortly, difficult to catch from the screen. The pump then oscillates at the 1.5 degrees even if the buffer bottom temperature would easily allow cooling the boiler down to less than the buffer tank top. Here I would have expected the pump running until the boiler temperature drops at least by 1 degree. I guess it drops by 2 degrees with hydraulic system 1 configuration (no personal experience). If the temperature sensor signal had some noise of the order of 2 degrees, the pump output could oscillate even with a hysteresis of 2 degrees. This is why I intend do continue monitoring the sensors, specifically trying to catch the period of pump control signal oscillation (even if so far I have not seen signal noise that could cause the issue). I should also test those capacitors for the temperature sensors.

Maybe try to disconnect the pump control signal cable, which carries the 0-10 V signal, and then observe the signal if it is oscillating further (only with scope as load on the analog output). Just to check if the output is working properly and the load (cable and pump control input impedance) is not causing any problems with the output. I mean the output is just working fine in "normal heating mode", but I think that control parameters for the output signal could be different in heating and standby mode (in standby mode, output signal generation is maybe more "agressive" and less smooth than in heating mode - but I don't know it exactly).

NF: I have not monitored the pump control signal when disconnected from the pump but I have figured out the control signal having a pretty low drive impedance and the pump input having a very high input impedance. I'm concluding this based on my current RC-filter, how the boiler loads the RC-filter capacitor and a 1 kohm resistor drives easily the pump via the parallel path without any signal drop at the resistor.

The pump control indeed is more aggressive in standby mode. When heating, the pump mostly runs at less than 40%, in standby and already at waiting for shutdown the pump runs at max speed even if minimum speed would be sufficient every now and then shortly to pump residual heat from the boiler.

Logging is another good idea, decoding the data shouldn't be a problem for you and/or your local dealer. As far as I know it is some sort of CSV data which can easily be read with Excel or something simllar.

Best regards

Edit: Last solution would be to buy a PWM controlled pump next time but before this, you should measure in PWM mode if that does the job properly

Edit2: Another idea would be to buy a cheap PWM to 0-10 V adapter module and use the PWM output of the controller with the actual pump (implied that PWM doesn't show this faulty behavior)

NF: changing the pump to a PWM type would be challenging. Trying a PWM to linear voltage drive more feasible. So far I have assumed hydraulic mode being the issue, I should ask our dealer if those who are fine with hydraulic mode 1 are using linear 0-10 V drive or PWM. Those US Fröling owners who had similar problems with hydraulic mode 4, were actually using a fixed speed pump. Makes me lean towards hydraulic mode 4 but need to keep an open mind here too.

Display More

Agreed, any noise in the sensor signals defeat the hysteresis if the controller does not have any time hysteresis (not even for a couple of seconds). Difficult to be sure if the hysteresis is missing or if some noise of the order equal to about one degree on both sensors kill the hysteresis.

I should be able to test with some self introduced hysteresis without detaching sensor wires. A parallel resistor on the boiler temperature sensor connectors would make the controller see "low enough" boiler temperature relative to the buffer tank top temperature. I've just been a bit busy lately and need to find the time to test this shortly after boiler shutdown, at standby state.

Our boiler SW is from November 2020, actually thought it was updated later. This is when the older SW stopped supporting Fröling Connect application.

Now I've been handed contact info for a factory expert too. I will soon explain them the issue hoping they could provide a simple solution. Will report of course if it is something that could benefit others.

A progress update here: we are going to have the logging device installed to the boiler. It should show if sensor readings are noisy or misbehaving and a lot of other useful info. I'll give an update when the log is available and has been analysed.