Drayton Digistat 3+ Accuracy issue?

[stormyuk]

Hi,

I wonder if someone can give me some quick advice, I have just replaced old Drayton Digistat 1 which decided to give up the ghost with a new Digistat 3+ and have noticed something weird with the new thermostat.

It appears that when the temperature is rising with the demand on that the thermostat sometimes switches off the demand 1c - 2c before the programmed temperature is reached. Say the program is set for 16c then sometimes the demand will be switched off at 14.5c - 15.5c and say the demand was 21c the unit sometimes switches off at 19.5-20.5c.

The Worchester 27CDi has no controller of its own and is wired directly to the Digistat 3+ to handle the on/off of the central heating. The old Digistat seemed to be more accurate than the new 3+ which "supposedly" measures to 0.1 of a degree. The actual temperature displayed on the unit seems very accurate to other thermometers as I wanted to check the display was "actual" what the temperature was so that does not seem to be the issue.

Is this normal tolerance, a "known" issue with the software or something which is a fault?

There is a temperature offset function in the menu's but that is currently disabled.

Cheers,

Mike

 

[greco77]

Drayton have told me that this is due to the design. It's called Pulse Width Modulation and triggers the boiler in 15 minute cycles. At the beginning of each 15 minute cycle it determines how long to run the boiler/ heating based on the set point and current temperature.

What it does is learn over time how your home responds to running the heating i.e. determining the heating capacity of the system.

What I noticed with mine is that it initially only ran the heating for 2 minutes or so but if the room temperature was not responding adequately, it began to increase the run times up to around 5 minutes.

So it tries to find the most efficient run duration to achieve the desired temperature so that you don't get an overshoot and by also giving the radiators time to actually affect the ambient temperature. This is ingenious if you ask me, when it works of course. One thing that would have made a huge difference would have been the ability to switch between this and standard on off or perhaps have a configuration option where you could set a minimum run time provided the set point was not reached.

 

[greco77]

Oh. I realize that I may have misunderstood the question.

And if what you're actually asking is the temperature offset, then yes, you can take it up or down a couple of degrees. The instructions are online on the Drayton controls web site.

 

[stormyuk]

No, I dont think you have misunderstood.

Thanks for the response.

I have no offset set on the thermostat and I don't want one.

I was just trying to understand why sometimes even though my program is set to achieve 16c overnight, it will sometimes switch off the demand at 14-15c. I would have thought your explaination can't really account for a change in temperature of 2 degrees?

I have only had the unit installed a couple of weeks though so will see if it settles down.

 

[greco77]

The documentation I have seen does day that if there is a 2 degree difference between the set point and temperature, either way, that it will stay on or off permanently, depending on the case.

 

[Bratters]

Sorry to resurrect an old thread, but I'm having the same problem with a Drayton Digistat +2. (In fact, I reported it elsewhere on this forum years ago, but I've just started using the thermostat again).

I contacted Drayton at the time, and they did say the same thing about Pulse Width Modulation on a 15 min cycle. I think if the set point is say 16C, it stops at maybe 14.5/15.0 for a while, as long as the temperature is still rising, to make sure it doesn't overshoot.

The strange thing with mine is that it *does* overshoot, and often quite badly. I currently have it set to 19.0, the current temperature is showing 20.2, but it is still calling for heat for a few mins in each 15 min cycle. So the house is getting warmer and warmer, currently 1.2C over set point.

I wonder how the 'adaptive' thing learns. My thermostat is set to 19.0C 24h a day, but a separate timer cuts the heating out at night, so the temperature falls even though the thermostat is calling for heat all night. I wonder if that's confusing the adaptive part (i.e. at night it thinks it's heating the house but it isn't getting any warmer). I may try switching the timer to 24h operation, and using the Drayton to control the timing, and see if that helps. I'll feed back here for anyone else finding this thread in the future.

 

[D_Hailsham]

Found this in the FAQ section of Drayton's website:

Q.My Digistat +2/3 appears to switch off before my set point is reached and sometimes remains on after my set point is reached, is this normal?

A. Yes, this thermostat on a energy saving algorithm.


Unfortunately you do not seem to be able to modify the algorithm to meet the specific requirements of a house.

 

[Bratters]

Thought I'd just follow up, in case this helps anyone else in future.

Basically, I have had to give up on my Drayton Digistat +2, as the accuracy of the temperature control was awful. It's a shame, as in my old house I had a Drayton Digistat 2 (not the +2), and it was superb, holding the temperature to within about 0.2C normally.

I gave the +2 a month to 'bed in' (it claims to need up to a fortnight to learn the thermal characteristics of the room).

After a month, it was still over-shooting the temperature by up to 1.5C. Even when the temperature was say 1.2C over the set point *and climbing*, sometimes the thermostat would suddenly decide to call for heat. I don't see anything "energy saving" about that - in fact, the old bi-metalic thing I took off the wall in the first place was probably better.

I'm not a total stranger to control system theory. I understand that if the temperature is over the set point but falling, a call for heat might make sense. But when it is significantly over and rising, a call for heat doesn't seem like a good idea, and doesn't tie in with Drayton's claim that with each 15 minute cycle, the temperature should get closer to the set point.

It's a shame there don't seem to be thermostats based on a PID controller. No PID controller would behave in the way the Drayton does (or if so, it's a very badly tuned PID controller, and has failed to adjust itself even after a month of operation).

Anyway, I'm now on a Seitron, which has adjustable hysteresis - set to 0. It works perfectly, switching on when the temperature goes below the set point (measureable to 0.1C), and switching off when it rises back to the setpoint - it usually overshoots by about 0.2C due to residual heat in the radiators, which is fine.

I don't know if anyone is having more luck with Drayton, but I'll steer clear in future.

 

[PaulTheOnlyOne]

I think the algorithm is very strange. Must admit I am using the thermostat in an unusual way - to control an industrial hot air heater - but the previous model worked fine. This one is very odd. Comes on at 6:30 am and rockets up to 19.5 deg in a couple of hours but then slows right down. For the rest of the first half of the day the room is too cold but the temperature gradually builds up to 1 deg over the set point (22.5) which it gets to by 3 pm. It then maintains this temperature until the shut off time at 6 pm.

I have uploaded the graph into an album. The relevant graph is 1C

Untitled

• PaulTheOnlyOne
• 26 Nov 2013
• 1

 

[PaulTheOnlyOne]

I have worked out for myself now what must be happening. When the first on time of the day is hit the unit goes in to bang-bang mode for a maximum of 2 hours. Then it goes into PID mode. This is gives the quickest warm up to the set point and for a normal house will result in no oddities due to building up an integral term. The problem for me is the air space is large and the building very poorly insulated. The heating system needs more like 3 - 4 hours to get to the set point. Because the unit is well below the set point when switched into PID mode (after 2 hours) a very large integral term then builds up as it heads for the set point. This results in a substantial overshoot which lasts for hours - indeed until the end of day at 6 pm. I don't think it can ever self correct because of the 2 hour limitation.

The conclusion is that this controller is unsuitable for large poorly insulated buildings. I have corresponded with Drayton and they have not confirmed the above but they say that the algorithm can be disabled in Option 5 of the Installer menu in later models. For earlier models like mine this option only disables the optimum start feature.

 

[Steve Rawlings]

Some great analysis recorded here by PaulTheOnlyOne.

This is an old thread, but it's worth confirming that the latest model of the Drayton Digistat +3RF still has the same problem (Software V4.3).

The predictive algorithm in the Drayton Digistat +3RF is riddled with bugs that causes the boiler to be 'off' when it should be heating; and 'on' when the room is above the set temperature. This causes mayhem at the boiler, with very frequent 'heating on'/'heating off' commands that prevent proper cycling (modulation) of modern boilers.

Bizarrely, these spurious on/off signals to the boiler also happen in 'manual' mode. There is no option for switching off the adaptive features, or for emulating a classic 'bimetallic strip' thermostat - which is all a modern boiler requires from a standalone controller.

Setting Installer Options 12 and 13 to '1' may provide some improvement, but the best solution is to ditch the Drayton entirely and buy a proper thermostat. You'll save money in the long run.