Kamstrup Multical 403 Erratic flow readings

i m geting a 403 installed in a couple of weeks and by the looks of it, reading through this thread, i am perfectly setup for a world of pain … space is tight, bends are plenty, i have a plate exchanger and i m not running all that high pressure in the circuit, 1.5bar at most.

does anyone know what s the minimum distance to the recirculation pump i should leave? would a vertical installation be better? :frowning:

I’ve attached a document “The Challenge of correct flow sensor installation with respect to flow disturbances” I believe I got this from Glynn which gives remomended min distance for a variety of locations. I’ve given you a snapshot of that table as well as the whole document.

The challenge of correct flow sensor installation with respect to flow disturbances.pdf (517.7 KB)

The biggest problem most of us have faced is purging the air from the system as this leads to erroneous flow readings. Trapped air and location are the two hurdles to overcome.


I would recommend fitting a Sontex if you’re using a plate heat exchanger, as @Zarch found out getting all the air out o a plate is a challenge.

Vertical with fluid flowing upwards is probably your best option.

you guys are amazing, thank you. unfortunately choices are scarce in my country, i was barely able to find either a 403 or a qualcosonic e3 that would also have the m-bus interface so i’m pretty much stuck with one of those two.
will try to mount it vertical, on the return, with the fluid flowing towards a 90 degrees bend , then circulation pump, then plate exchanger… this will be “fun” :expressionless:
thanks again @JohnT and @glyn.hudson

1 Like

I’ve just ordered a Multical 403 from Stockshed so here’s hoping we can get monitoring without too many issues.

We can at least compare notes.

Good luck @alext !

1 Like

That’s a Kamstrup document @JohnT - not all meters are as nice as those with respect separateion distances from distrbances and tolerance of swirling flows.

I recommend mounting the flow sensing elements on the FLOW side not the RETURN side. The system pressure will be higher there. You’re less likely to have cavitation issues when the working fluid gets depressurised as it is accelerated through the samller diameter part of the heat meter where the flow is sensed.

The outlet of the monobloc (all of which have plates in - the existence of a plate doesn’t mean that you can’t bleed a system) is pretty much the perfect place. Mark at JMR actually fitted this one perhaps without realising quite why I was insistent it went on the flow, vertically, upwards, pretty much as soon as the primaries came up out the floor:

It isn’t entirely crazy to be putting these OUTSIDE either. Most are IP65 as standard so can tolerate being outdoors. Use the 1" temperature probe ball valves as your isolation for the monobloc. You can seal the insulation to the flow sensor body etc where the wires come through using self amalgamating tape.

Where are you based @alext? It looks like you’re having run running refriegerant through the plate there. I’m back in Lithuania later this month and can stick a meter in the mail fairly easily if it’s Europe you’re trying to pick one up in.

I’ve got a new but slightly damaged (calculator has a clip broken; but otherwise ok) DN25 Axioma E4 here. Yours for £125/€125 if you like?

Don’t put it on the suction side of the pump. That’s a really bad place with the lowest possible static pressure / most likely place for air to be drawn out of the water.


Do heat meters need reconfiguring if they are on the flow rather than the return?

ie, I see this on descriptions of various heat meters:

  • Return / Outlet Pipe Install (Flow install available on request).

As the calcs are just flow rate, flow temp, return temp (and calculated DT) to get heat output, you can get this from wherever you stick the meter?

What changes in the meter for “Flow install available on request”? :man_shrugging:

1 Like

Strictly speaking yes…

Ultrasonic measures sp* eed of fluid

  • Speed plus area gives volumetric flowrate
  • Volumeric flowrate plus density (known by fluid type and the temperature measured at that point) gives mass flowrate
  • Mass flowrate plus heat capacity gives energy

For heat pumps though…

…where the dT is small between supply and return, the density error is small and it won’t make a major difference if you density correct based on the return temperature even through the flowrate sensor is mounted in th supply.

Diehl meters cannot be reconfigured onnce set.

Kamstrup can be reconfigured in the field with an optical eye and MeterTool (only available to distributors/they get rather pissy if shared further so best ask Kamstrup directly)

Sontex can be reconfigured in the the field with an optical eye and Prog531 etc (available to distributors, can be reshared if promise not to share with all and sundry)

Axioma can be reconfigured in the the field with an optical eye and E3 configurator (available to distributors, can be reshared if promise not to share with all and sundry)


oh boy, i so so wish i would of found this topic and posted sooner!!
thank you very much for the offer @Marko_Cosic , that is one extremly sweet deal, about half price for an E4 here, but I ordered the kamstrup 403 already and it should be here tomorrow(gonna take a while to install tho).
Unfortunately not knowing all you guys wrote here I also asked for it to be configured for return :frowning: I actually had more space at the top, on the flow, could of just went up after the plate exchanger, damn it.
I’m in Romania, and that is indeed a refrigerant + water plate exchanger, got a zubadan split system without the pre built internal unit, just the parts added to various walls (not sure what the name for this installation method is)
once i see how big it is, i l ltry to make a drawing on intended installation :stuck_out_tongue: , but right now the only option seems this section of pipe with red… would it be too far from the exchanger? ~1.5m of pipe+pump+y filter

@alext what flow rate is your surplus DN25 E4 designed for and what outputs does it have?

I’ve got a second system which I was intending to monitor but can’t warrant £500 for a Kamstrup when I’m planning to decomission it next year.

Thanks a lot for sharing your insight, installing on the Flow does make a lot of sense. Is the manufacturer’s general guidance to install on the Return is more applicable to systems with a high flow temperature? Since installing on the return will shield them from the highest temperatures, obviously this is not really applicable for an ASHP.

We’re familiar with Prog531 to reprogram Sontex heat meters, but we’ve not explored doing this with Kamstrup.

Nice, did you specify MBUS module? This is not the standard option

That’s a fair price for one that a has not been broken by a clumsy idiot. :wink:

They are 260 mm long with G 1 1/4" threads. You also need to add some space for the ball valve with the temperature probe pocket; or a fitting with the M10 x 1.0 mm thread in it for the temperature probe.

Romania was a close guess but no prixe for me. People in Poland also seem to like to buy cheap minisplit air conditioners and replace the indoor unit with a heat exchanger to their underfloor heating.

Look at the front of the meter and see if it shows arrows ↔ anywhere on the display. If they are showing then the meter is still in a configurable state and with a lot of swearing at the insturctions you can still change it by long and short pressing of the button.

It may already be locked. They will automatically lock after water has been through them.

1 Like

(loggin in as the original user - sorry about two usernames)

The manufacturers usually defer to the distributors for advice. They’re given a margin so that they can provide advice. And also so that it isn’t the manufacturers’s fault. :smiley:

For domestic heat pumps Iwe would advise putting the heat meter on the supply side and at the bottom of the house as far away from where all the air can gather as reasonably practicable. Vertical or horizontal with the flow sensors/mirrors at the 3 and 9 o’clock positions to avoid static air/dirt on them. The example in the earlier post, where the heat meter flow sensor is the first thing on the primaries as they rise out of the floor void and up into the cylinder cupboard, is blob on. Up in the roof space on the return is the worst possible place to put the flow sensor; both from air tending to want to be there and from a static pressure perspective. Inbetween is in-between.

As far as the manufacturer is concerned the datasheets tell you:

  • What the min/max temperatures are for each element (the calculator will have different ratings to the flow sensor)
  • What the min/max pressures are

Practice varies between countries and meter types / application types.

Clockwork (mechanical) meters like clean water. They often need to be oriented a certain way around too so that friction doesn’t cause issues. That often dictates where they are mounted.

On high temperature (superheated high pressure water over 100C) district heating systems the flow could be too hot for a flow sensor to live a happy life so these were mounted on the return by neccessity.

On medium or low temperature systems (below 80C) you can mount the flow sensor on the flow. Pressures are higher here so you avoid needing such high fill pressures. That can matter in higher rise buildings. (top floor plus 2 bar plus system differential pressure ends up quite a lot by the time you’re at the bottom)

If the flow sensor is on the supply, and the heating is isolated, but you still see flow, then the district heating company knows that you have a leak. Sometimes there are flow sensors on BOTH the supply and the return for high value assets where you want to double check there is no leak / double check that the flow is being measured correctly.

Sometimes they want to put the calculator and the flow sensor together when they are mounted. Batteries powering calculators are rated for N years at 30 or 35C ambient. If you’re in Germany and you’re forced to bin or recalibrate the meter every 5 years you don’t care. If you’re in the UK and you’re expecting 15 years on a batteyr and 30 years on the meter then it matters. So you have to mount the flow sensor on the return if the battety is to stand a chance of lasting the course. Better is to remote mount the flow sensor BUT better doesn’t always win over easier.

Air is less of a concern on big systems designed well. (disappears off elsewhere to be vented or they use vacuum degassers to literlally suck it out) Dirt is more of a concern. If you want to avoid the meter filling with muck then you might put it on the return so that there’s nothing downstream of the filter that could cause muck to be trapped within the meter. Example on this from HMG study:

And links:

Most small meters in the UK are for district heating schemes and they are return mounted so that the calculators can be clipped directly onto the pipework.

Hop over to the Netherlands and the small meters tend to be supply mounted with remotely mounted calculators.

If you’re buying generic stock in the UK it’s probably going ot be configured for return mounting because it’ll be from a distributor that mostly serves district heating. You can ask for either. You can change it for many meter types. Sometimes the meters ship in an “unlocked” state too where it can be flipped from one to the other using cryptic butotn presses on site. We tend to avoid this ourselves. Better to ship the plumber something as idiot proof as possible!


I would look at this from a pressure drop. I have a suspicion that you’ll want numbers:

Ultrasonic Metering.xlsx (715.4 KB)

The slightly dog eared unit is a DN25 E4 with 260 mm x G 1 1/4" body; so has a Kv value aaround 12.

Pressure drop in service will depend on flowrate; flowrate will depend on power output / sytem dT. If youre system is running with a dT of 5 degC or above this is fine for up to 12 kW before the pressure drop becomes noticeable. 16 kW at dT of 7 degC also ok. 16 kW at 5 degC will work but is burning more pupm head than is ideal (spending £s on pumping electricity) so I’d step up a size if you’re on a really big unit.

M-Bus and wireless M-Bus on this unit. Battery power. Limit on battery for these meters is 80 hours per month. You’re fine to read every minute at 9600 baud. You can’t read every 10 seconds. I can chuck in a mains supply for +£25 if you like. Then no limit.

Kamstrup units are nicer in this regard (every 10 sec on battery is fine), but you expect this at £500 vs £175 new (or £200 with mains). The Kamstrup unit also have a lower pressure drop but you can mitigate that by stepping up a size in the Axioma units if you’re fussed, as low flowrate measurement isn’t a think on heat pumps perticualrly.

Sontex excel when you’re full of glycol rather than water, and on the smaller meters because they draw power from the M-Bus so can be read as often as you like without killing the battery or needing mains supply.

Diehl are like the Axioma. Every 5 minutes for them on battery. The reason the Kamstrup is so good is that the previous generation was bloody awful (couldn’t read it every 5 minutes, which is all folks using them for non R&D applications care for) so they went way over the top on the next generation!

Landis are every minute for the basic battery powered meters.

It all gets a bit academic.

For the curious:

On battery most meters either use 32 second measurement intervals, and even longer filtering windows, or they use “adaptive” measurement intervals, where they measure say the flow every 2 seconds but only bother measuring the temperature every 60 seconds unless there has been a sudden change in flowrate that the meter has been configured to interpret as meaning that the temperature might be expected to change quickly. So when you read them every 10 seconds thats lovely but the data you’re getting is going to be measured less often than this and then “filtered” (usually moving average) before it is spat out on the M-Bus.

(and they can usually be configured in any one of these modes in software)

If you take a look at slide 32 / 33 here you can see that the meter shows a delayed rise and a delayed fall in the flowrate vs what we knew the flowrate and valve position to be from independent measurements:

You’ll see the same thing for most heat meters. Probably the flowrate change lags the pump/compressor electrical power change due to moving-average filters in the OEM data? Internal meter sampling intervals / filtering behaviour will be why. All comes out in the wash in the final energy figures but don’t assume that the ramp rates you see are really ramps. They’re probably steps. :slight_smile:

1 Like

@glyn.hudson yes I went for MBus and optical outputs. I was planning to order a QALCOSONIC E3 but Adrian @ Stockshed told me he’s no longer stocking these due to supply issues.

1 Like

so much information. can t say thank you enough, will check for the arrows before mounting.

what i have installed is a proper Zubadan PUHZ-SHW112YHA + FCT6 but the installers either on purpose or simply because of lack of interest/knowledge have been less than ideal. No power meter, no heat meter, no flow meter, 80l buffer tank i seem to not need, no secondary temp sensors on the buffer tank. Only found out about this reading and learning myself, trying and doing changes, monitoring as much as i could (thanks to my trusty and very very very old emonTx v3!!).

my next steps on this voyage are installing the heat meter, adding a bypass for the buffer tank, removing the UHF manifold mixing pipe/thermostat, maybe pump as well, simplifying zones if not removing completely, lots to do and try.

1 Like

[puts UK distributor hat on]

There are no supply issues. (covid has passed)

Axioma manufacture ultrasonic water meters by the squillions.

They also manufacture heat meters…if you make it worth their while to setup the line for a run of these. (it’s not just a case of making them - each and every one of these needs to be calibrated on a giant rig that does dozens at a time)

You can buy them. You pay when you place the order. The order gets manufactured later. During covid this could be 6-9 months later due to component lead times; or indeed “no deliveries int eh next 12 months; please check back later” etc. Now it’s usually 6-8 weeks for composite (E4, moulded in house on-demand on the production line) or 12-16 weeks (E3, brass, ordered in from Italy). You order by the pallet. (256 if small meters) You sell them some time later.

Can confirm it hurts almost as much to look at piles of cash sat on the shelf as it does to look at piles of cash sat as a pending order. The paying out to being paid cycle is 6-18 months. :face_with_peeking_eye:

I should imagine that he doesn’t want to shell out upfront for what can be quite large minimum orders that take long enough to sell that customers might start to ask why they’re receiving 2021 dated meters (batteries) in 2023. That and it beign quite liekly that Diehl/Kamstrup offer you better rates if you list the meters (for eyeballs) but then flip the sales over to Diehl/Kamstrup when folks enquire. I don’t think the latter is the case with Adrian - I think it’s just not worth him stocking slow moving items now that the volumes of Diehl in particular have gone up.

But yes you can absolutely buy these.

The E4s make more sense for domestic heat pumps. Rated for up to 90C and 16 bar is fine. Don’t go green if they get wet etc.

We used the on the low temperature (<90C) / low pressure district schemes too. Only where you’ve got yikes temperatures/pressures OR they’re being cut into pipework by 800 lb gorillas that will cross-thread anything at the slightest provocation do we go brass for +15% cost uplift.

Sounds like quite the journey! Simpler is usually better. :slight_smile:

Installers possibly didn’t know what to do. Are (pretty idiot proof) air:air units not more common over in Romania?

I would leave the buffer tank if it is a 2-pipe type (adding volume to the system) and you have open zones. It doens’t hurt you other than space.

I still need to get my own units up and running before the snow comes. (no other heating) Nothing like the weather gods to impose a project deadline. :smiley:

Very simple 6 kW fixed speed unit with a 200 litre 2-pipe buffer (for minimum runtimes) and radiators (not enough water in those compared with underfloor heating in a slab to satisfy minimum runtimes)

Sorry - the arrows are for the Axioma units.

On Kamstrup you need to get into the setup loop:

MULTICAL® 603 - Installation and User Guide - English.pdf (3.2 MB)

Item 7 (3-007) is whether the flow sensing element is mounted in the supply or the return.

403 is probably the same. Try unclipping the front of the meter (to power it down for 1 minute) then clipping it back on and getting into this setup loop before the meter locks again.

IF the meter is new and has been shipped in an unlocked state you SHOULD be able to change this.

IF the meter has has water through it then you’re screwed without the special Kamstrup software.

Don’t fit it until you’re ready.

@Dan_Nichols you might also be able to make use of this to flip your 403. If they work the same way as the 603s.

1 Like