EMC Testing

I am developing a solar thermal (hot water) system with a control system using the PICAXE 20X2 chip. I’ve done plenty of user trials and now I’m gearing up to commercialise.
To get my CE marking I believe the system needs to pass some EMC tests – so that we know it won’t be upset by other systems in a house and it does not in turn produce undue interference, eg to radios etc.
A test lab recommends initial testing on Radiated Emissions, Common mode RF immunity and Fast Transient Immunity, as a quick “check on design strategy”. This will cost and I don’t want to have too many iterations before getting a compliant circuit.

I’ve not found any thread on this kind of issue.

So my questions:
Has anyone experience of getting a product using a PICAXE controller to market?
Are there tests you know that will identify problems in the areas above?
Are there any datasheets I can send to the lab?
Are there any known issues with PICAXE circuits which I can address first? In particular I am using hpwm which might produce a noisy signal when its amplified with a FET to power a 12V pump motor. I also use the DS18B20 temperature sensors.

Any help, leads, links, guidance, etc will be very welcome.

You're a brave man rolling out such a scheme in British weather! Additionally solar PV based water heating has just become cost effective as I'm sure you know. UK/EU needs may well be costly too.

No doubt you're up with EMC books,but have you contacted Rev. Ed themselves in Bath? Their "Hippy" may well be the man to assist.

Many years ago I was head of type approval for UK marine radio and radar systems and ran a series of labs which did, amongst other things, EMC testing (we were a Notified Body in EU-speak, so could CE mark equipment). EMC testing can be fairly expensive, but the extent to which it is required varies from product to product. Some items, like marine radio sets, could cost a few tens of thousands to test and approve, others, like ship radar systems could end up costing £100k or more to test.

The main issues with EMC are the housing and filtering on any cables going in or out of it. 99% of all the EMC problems we ever saw with equipment were fixed by improved screening and cable filtering, it was very rare for there to be something deep in a circuit inside the equipment to need any remedial work.

The first problem you face is finding out how much testing and what standards your particular device needs in order for you to CE mark it. My guess is that this device would be in the same category as domestic heating thermostats and programmers, so would probably need to be tested to the same protocol as them. You really need to get hold of a current copy of the EMC Directive and have a look, as it may be that the testing isn't that onerous or costly. As a guide, I can suggest that getting a test house to run a full spectrum radiated emissions test in an anechoic chamber is probably going to take about a day (including chamber set up time). Similarly, a full spectrum radiated susceptibility test in a screened room would probably take about the same time. It's years since I've been involved in this, but IIRC we used to charge around £2000 to £3000 per day for either the screened room or the anechoic chamber, test gear and a technician to run the tests.

You have quite a task on your hands because as with most regulations, a large amount of common sense has been lost along the way.
Having said that, with experience (and I'm not sure of yours) you get a 'feel' for what level of screening/suppression is required for each noise source and (as you correctly point out) also immunity.

Only the correct calibrated test equipment will give you the numbers required but there some basics you can do yourself before submitting for formal testing.

The AM radio trick is a very good indicator of emmisions. Simply place an AM radio that is not tuned to a station close to your cable runs and see if you can 'hear' the noise. You can get a level of calibration by doing a similar test near some known good equipment that is likely to work in a similar way to yours.
To test for immunity, try your mobile phone near the electronics to see if it resets. Also try switching large (ideally inductive) loads on and off that run from the same ring-main as your power supply.

Even when a design sticks to all the design rules, it can sometimes fail and extra precautions need to be implemented.
That is why good electronic/electrical design engineers earn good money. It's the same reason lawers do. They know things that other people don't that cannot easily be passed on with just a few words. (otherwise we'd all know everything).

ed1066 said:

I am developing a solar thermal (hot water) system with a control system using the PICAXE 20X2 chip. I’ve done plenty of user trials and now I’m gearing up to commercialise.
To get my CE marking I believe the system needs to pass some EMC tests – so that we know it won’t be upset by other systems in a house and it does not in turn produce undue interference, eg to radios etc.
A test lab recommends initial testing on Radiated Emissions, Common mode RF immunity and Fast Transient Immunity, as a quick “check on design strategy”. This will cost and I don’t want to have too many iterations before getting a compliant circuit.

I’ve not found any thread on this kind of issue.

So my questions:
Has anyone experience of getting a product using a PICAXE controller to market?
Are there tests you know that will identify problems in the areas above?
Are there any datasheets I can send to the lab?
Are there any known issues with PICAXE circuits which I can address first? In particular I am using hpwm which might produce a noisy signal when its amplified with a FET to power a 12V pump motor. I also use the DS18B20 temperature sensors.

Any help, leads, links, guidance, etc will be very welcome.

Click to expand...

CE marking is about more than just the EMC Directive. You will also need to be compliant with the Low Voltage Directive (mains safety if applicable) and also the Restriction of Hazardous Substances Directive, although you may be able to argue a Category 9 exemption (Monitoring and Control Equipment) but that exemption ends in 2014, at least for non-industrial equipment. There may be other relevant Directives, although those are the ones that come to immediately to mind.

So far as EMC is concerned I think it unlikely that a PICAXE of itself will cause a problem but switching power regulators may well unless care has been taken with the design. At the company where I work, we send equipment to a UKAS Test House to show that the equipment meets the requirements of the EMC Directive. As you will have found, this is not cheap, costing £1K+ per day. You can also take the "Technical Construction File" route, which is to compile a document showing why the equipment meets the requirements of the Directive without having it tested. For emissions, BS EN 55022 will be the standard.

Solar panels are becoming notorious for interfering with amateur radio communication due to emissions from the switching regulators.

A very useful freebie magazine is the "EMC Journal" - this has articles about Standards and also about good design practice - you can download back issues. "EMC for Product Designers" by Tim Williams is a useful book and goes into the Standards and tests in some depth.

Richard

Jeremy Harris said:

Many years ago I was head of type approval for UK marine radio and radar systems and ran a series of labs which did, amongst other things, EMC testing (we were a Notified Body in EU-speak, so could CE mark equipment). EMC testing can be fairly expensive, but the extent to which it is required varies from product to product. Some items, like marine radio sets, could cost a few tens of thousands to test and approve, others, like ship radar systems could end up costing £100k or more to test.

Click to expand...

Interesting you mention that.

I have been asked by a few people if I would make and sell one of the Lost Model Locators, that I posted in the projects section. Although not very cheap to make (in the style of things from China) there is nothing similar on the market, and nothing that would be as easy for the user to modify as the PICAXE based product.

However its an RF transmitter, capable of 100mW and on any frequency from 240Mhz all the way to 960Mhz, so quite a range to test for spurious emissions. It would only ever be made in small quantity, so the cost of EMC testing is out of proportion to the sales. And the need to use lead free solder - yuk.

Shame really, but I guess we all sleep better knowing these regulations are protecting us from rogue devices ........

srnet said:

Shame really, but I guess we all sleep better knowing these regulations are protecting us from rogue devices ........

Click to expand...

Or possibly not..............

Back when I was involved in this (1997 to 1999) the EU test houses were increasingly finding that more and more manufacturers were getting their testing done in the Far East, as it was cheaper. We once had occasion to test a device (IRRC it was an EPIRB) by Trading Standards, as they suspected that, although it bore a CE mark and TA number, it might not be compliant with the regs. It wasn't. In fact I am reasonably sure it hadn't been anywhere near a test house, as it failed on so many counts that we had grave doubts as to whether it would actually ever have worked. A while later, one of the Japanese radio/radar manufacturers reps let slip that you could buy an EMC certificate from one or two Far Eastern registered Notified Bodies without the need for any testing. He believed that one or two of their competitors were doing just that, as it was cheaper than getting kit tested and going through the inevitable round of product improvements that would be needed (I don't think we ever saw a bit of kit pass the EMC testing at the first go).

ed1066 said:

Has anyone experience of getting a product using a PICAXE controller to market?
Are there tests you know that will identify problems in the areas above?

Click to expand...

Getting a product through EMC testing should be no more difficult with a PICAXE than with any other microcontroller and there is nothing specific about a PICAXE that would cause any undue problems.

Most radiated EMI issues would stem from signal switching and I would expect there to be minimal issues with digital signal lines from the PICAXE itself in the great majority of cases.

THANKS

Thanks folks. This is a great response with plenty for me to work on.

The lab will charge me £400 + VAT for the half day initial tests, so that sounds about right.

A couple of clarifications on solar as its my thing:

(only (a) is of relevance regarding PICAXE & EMC etc)

(a) This is solar hot water, not PV, so the scary inverters etc are not part of it. In fact its all 12V or 5V DC, supplied by a plug in power supply. (I’ve already found that a switched mode power supply interacts really badly with the brushless pump motor. It generates loads of RF noise. I’m expecting to need a linear one, which appears to resolve the issue completely.)

(b) The use of solar PV to heat water (post #2) gets me very angry. As electricity bill payers, we subsidise folks massively to put PV panels on their roof and we expect in return to have a reduced need for UK electricity generation. Instead they divert the electricity to their hot water tank, when gas heating would cause a fraction of the carbon footprint to do the same job. It’s the government’s mistake, they have perversely incentivised people to waste the precious PV generated electricity, if only they paid people for actual export to the grid rather than deemed export it would be fine. And to make it worse there are no incentive schemes for solar hot water – the proposed scheme has just been delayed yet again. (It does give me time to get my super low cost, robust and EMC safe system established tho’)

So, keep the info coming. Meanwhile I'll do the tests suggested and look up the books / links etc and report back ... Cheers, Ed

Can't help with EMS testing but recently I have been dismantling various bits of old electronics to salvage bits. One thing I have noticed is that every modern circuit seems to be awash with ferrite bead inductors in all the transmission lines. AFAIK these can only have the function to suppress EMC so I would advise researching this and probably adding some to your circuit

Hi Ed,

You haven't said what quantities are involved, nor what the PICaxe is actually doing (e.g. driving PWM?). If the quanties are sufficiently large to justify RFI testing, perhaps you should be considering developing the program code for a raw PIC and not PICaxe (since any changes later would very probably require re-testing/certification). As far as CE marking is concerned, I believe (but am not an expert) that there are exclusions for small quantities and "prototypes", etc., and was told that CE marking itself is basically a "Self-Certification" issue. However, you might need to "defend" yourself if accused of non-compliance, e.g. of the (RFI) regulations, etc..

matherp said:

I have noticed is that every modern circuit seems to be awash with ferrite bead inductors in all the transmission lines. AFAIK these can only have the function to suppress EMC so I would advise researching this and probably adding some to your circuit

Click to expand...

A few months ago there was a thread on RFI issues where I believe the conclusion was that you should NOT just sprinkle ferrite beads about, they MUST be used/designed to address specific issues (i.e. specific RFI frequencies). However, IMHO a "safe" solution (where practical) is to mount the electronics inside a closed metal box (a Faraday screen) with all connections to the "outside world" via "feedthrough capacitors".

Cheers, Alan.

In addition to what everyone else has mentioned, another sure way to reduce emissions is to reduce the clock frequency. If your application does not require near-instantaneous response -and I believe that in a solar water heater a delay of a few seconds could be of no consequence-, you could go realy down in frequency......just an idea, which may or may not be applicable.

Even if you need to run full blast, you can slow down the rising/falling edges of any pulse train that runs to the outside world. Of course, at the receiving end, a schmitt trigger buffer would be required.

Another idea is to employ communications via differential signals over twisted pairs. Again, I ignore whether it applies to your circuit.

And last but not least...decouple, decouple, decouple the supply. All ICs should have its own decoupling cap.


Anyways, as Judas Priest, that quintessential British heavy metal band would say: rehearsals are free, studio time is expensive..... do as much as possible and identify potential pitfalls before you go to the certifying lab.

If you want my gut feeling, based on nothing more than a guess as to what your circuit is doing, and with the assumption that it will be inside a reasonably well screened case, with filtering on any leads going in and out, then I don't think you're very likely to have any significant EMC issues.

The Picaxe itself presents an extremely low risk of creating any spurious emissions, or of being overly susceptible to emissions from other equipment. It'll be things like power supplies (switched mode supplies are a bit notorious), switching devices and sensors that might cause one or two minor problems. The latter can be quite challenging, as the susceptibility testing can induce some fairly high voltages into unscreened wires running at a fairly high impedance, for example.

My solar hw system uses a 12v pump which is in turn powered by a small pv panel. No light=no heat=no pumping. I use a lead-acid battery supply so that the Picaxe controller stays up and can pump out the residual heat at dusk.

So, in principle there is no need for 240v at all and so no switching noise, no cabling, no voltage risk. The only thing with any power in it and thus the potential to cause much interference is the pump.

Ed: This may be tending a tad off topic but my extensive solar experiences (largely here in New Zealand) may be worth reflection.

Your "very angry" comment is perhaps justified given FiT (Feed in Tariff) sweeteners diverse governments have offered (often regretably & briefly!). Even without these incentives (NZ has none!),collapsing PV prices mean installation of PV based water heating can now be cheaper than classic piped thermal SWH (solar water heating). Furthermore wired SWH is free of wintery issues, can be easily adjusted/maintained/enhanced, & also performs admirably in cold conditions.

The fundamental issue with domestic solar thermal water heating is that it's likely to produce too much hot water in high summer-just when you don't need it. What CAN you do with excess hot water,especially if you're away!?. Surplus PV derived electrical energy is much more versatile, and at today's bargain PV prices "wired water heating" looks the way ahead. I'm sure you're up with Fanney's et.al mid-late 1990s US Govt. study on this => http://fire.nist.gov/bfrlpubs/build02/PDF/b02012.pdf

------------------------------------------------------------------------------------------------
Background- over the last 40 years here in latitude 41 Sth sunny coastal NZ (~2200 bright sun hours p.a, largely frost free - snow is virtually unknown) I've lived with DIY & two commercial SWH systems on family homes. All have given major problems within 5-10 years (broken glass, connection leaks,dislodged sensors, corrosion, blown tempering valves, noisy pumps, solar warrenties not honoured due to firms going belly up, mounting concerns with convectional flow tanks &-sigh- visual eyesores) & were eventually removed. Although my thermal involvement was extremely educational,all up none of these SWH systems gave a cost effective RoI (return on investment) either, especially when $$$ plumbers call outs were factored in.

Even modern evacuated tubes have failed to sway me as a result of such experiences! We retain a dead simple simple backyard solar system however for outdoor showering after summer swims etc - this gets used ~6 months of the year, & is valued for it's "no dirty feet indoors" savings!

Our current family home is strongly passive solar (enhanced for low angle winter sun), & is all mains electric (no gas),with a wood fire & heat pump for winter space heating. Annual energy bills are ~US$1500, with about 20% of that being a "dollar a day" utility company supply charge. Although some energy companies offer sweeteners,even without FiT incentives I'm increasingly conscious that a few kW of grid tied PVs (costing all up now US$5000-$10,000) could offer an appealing return.

manuka said:

Even without these incentives (NZ has none!),collapsing PV prices mean installation of PV based water heating can now be cheaper than classic piped thermal SWH (solar water heating). Furthermore wired SWH is free of wintery issues, can be easily adjusted/maintained/enhanced, & also performs admirably in cold conditions.

Click to expand...

Very true, in fact I've been modelling just such a system for the new self-build passive house we're just starting to build. PV is now so cheap (even allowing for the new EU tariff on Chinese imports) that it doesn't make sense (if you have the roof area) to fit wet solar thermal systems. Not only is PV used solely for electric water heating cheaper, but it also gives far better performance if you need to raise the hot water temperature by more than about 20 degrees above ambient. What's more, if you use a thermal store, PV allows you to extract useful solar energy even when the thermal store temperature is high. I also tend to believe that a PV system like this will be more reliable than a wet system. All it needs is a control system that can match panel output under varying conditions to an immersion heater in the tank, not hard to do with a multi-element heater.

ed1066 said:

The use of solar PV to heat water (post #2) gets me very angry. As electricity bill payers, we subsidise folks massively to put PV panels on their roof and we expect in return to have a reduced need for UK electricity generation. Instead they divert the electricity to their hot water tank, when gas heating would cause a fraction of the carbon footprint to do the same job. It’s the government’s mistake, they have perversely incentivised people to waste the precious PV generated electricity, if only they paid people for actual export to the grid rather than deemed export it would be fine. And to make it worse there are no incentive schemes for solar hot water – the proposed scheme has just been delayed yet again. (It does give me time to get my super low cost, robust and EMC safe system established tho&#8217

Click to expand...

I'm one of those you're angry with, because it makes me money, it's that simple. If you had investsed ~£10k on PV would you not want to get the maximum return possible? (and it also makes a great PICAXE project)

Just for fun (because it really doesn't bother me) we could turn that arguement around and say that people who fit SWH are selfish and NOT being green because they have wasted their roof space by not fitting PV which can not only produce hot water but also reduce their electrical consumption from the grid.

My grid tied PV generates ~2.5kW and my house idles at an average of ~500W.
On a 'good' day, that gives me ~2kW 'spare' for MAKING HOT WATER. (for about 5 hours)
From cold (~25C), I can use about 7kWhrs to get the tank to 60C
From warm (and it often is) I can only put about 3kWhrs into the hot water tank.
That leaves on average about 5kWhrs that goes back onto the grid to help out those who didn't make a similar investment
If I had only SWH, then what would I do with the surplus?

Getting back on track with EMC.
As several others have mentioned, I doubt you will have any problem with the PICAXE side of things as long as you take a little care. The only issue is likely to be with your PWM driven pump. If it does prove to be a problem (the radio test will be a good indicator) you could consider utilising a buck converter style output from the PWM drive (kept in a metal box) which would present a near DC voltage down the wires to the motor.
I had the same concerns with my immersion heater power control. Rather than use phase angle fire I opted to use distributed pulse fire which switches the immersion on/off for complete mains cycles. Much more complex to do but for me it's a hobby and a challenge rather anything to do with 'going green' or even saving money.

Solar Diversion (None of this is EMC)

Thanks BeanieBots, you are an ideal case to explain my position. On a good day you generate 12.5kWh and have 10kWh of spare electricity. On average you can use 5kW to heat the tank.

Under the current rules you are deemed to export only a little and are paid 3p per kWh on top of the FiT – 38p. I assume you have a gas boiler that’s reasonably efficient, it would have cost you about 6p per kWh to use gas, so you’re saving another 6*5 = 30p. 68p in all plus FiT. The UK gets a carbon saving of 5kWh. If you exported all the spare you’d be 30p worse off.

Under the sensible rules there should be no deeming. You are paid for actual export, effectively your electricity meter is running in reverse. Retail electricity costs 13p per kWh or so. So you export all your spare electricity and gain 130p – double the present. The UK gets a carbon saving of 10kWh, less a tiny amount for the extra gas which has a carbon footprint of a third, ie about 8.3 kWh electricity equivalent.

I’ve even written to the minister at DECC on this. As I said in the earlier post – its not you, it’s the civil servants I’m angry with.

The comments on solar thermal are helpful and put into focus what I’m up against. I believe still that with a level playing field solar thermal has its place. Clearly PV is more important overall. The system to be tested is low pressure, low temperature (up to about 80 deg C), low voltage, low maintenance, low cost and should be reliable due to the lower stresses on the system. I hope to halve costs, with the proposed government incentive you’d get a system for a little over £1000. The efficiency of solar thermal is about 80% in good conditions and so little roof space is needed, freeing up the rest of the roof for PV. Overheating is avoided in the new system by drainback – the panel just gets hot.

I hope the collapsing PV prices are technology related and not a surplus of supply. I’ve heard in some circles that it’s the latter.

So, while PV has the potential eventually to save the world, …. Just in the meantime let’s use the precious electricity to do real valuable work and leave water heating to the heating technologies

The circuit

Thanks again for all the comments.

Here is my circuit, somewhat simplified (eg no download cable and only one of multiple sensors shown).



Some notes:

The control circuit PCB is currently in a plastic box, but a metal one won’t be much more expensive.

The pump has a brushless motor it does not seem to produce back emf.. It is surrounded by thermal insulation. The insulation material consists in multiple layers of foil and a polymer wool, making a faraday cage. The lead from the control box to the cage is 20cm, but can be straight through if required. It is simple two core speaker cable.

About 4m of the same speaker cable powers the system from a switched mode power supply or a linear power supply if that proves necessary.

The three DS18B20 temperature sensors are on 4m – 6m or so leads, simple 4 core ‘alarm cable’ with one core not connected to anything (3 core not available). One sensor shares a lead with the OLED for some of the way. One capacitor covers all the sensors (the OLED has its own)

The 27 ohm resistor covers power to all the temperature sensors & OLED and is intended to protect the circuit if a short occurs in a lead or connection.

The OLED is the AXE133Y with an additional output pin on its chip used to power a piezo. The piezo only sounds when there is an error and so the pump has stopped. The OLED lead is 3 – 5m long and its in a plastic box, not easy to make it metal.

The two resistance measurements are to sense whether there is water between 2 copper objects. It is powered by an output pin rather than the 5V because the voltage can be raised just for a measurement and so produce almost nil electrolytic corrosion. The measurement lead is into the foil insulation faraday cage, just like the pump’s lead.

The frequency is reduced to 4 MHz which is that used by the ReadTemp command. As I need ReadTemp I can’t see how a lower frequency will help.

If you see any EMC weak spots or suggestions, please post

Cheers

Ed

No catch diode on your pump.
Although you say it's BLDC I find it hard to beleive that there is no back EMF!
Is it really suitable for PWM control?

It would be a good idea to put some uF on the power out line just after the 27R (which is a very good idea BTW).
Having had noise issues with a few recent PICAXE projects myself recently, I've found the culprit to be the OLED.
This came as quite a surprise but replacing with LCD cured the problem so it's definitely the OLED causing it.
If you want to stick with OLED (I did), then put a similar resistor (27R) at the OLED end of the lead and fit 100uF + 100nF each side of the resistor.

A little more about PV.
My electricity is more like 18p/kWh
I get 45p/kWhr for generation. (FiT)
I get a further 3p/kWhr for what goes back to the grid (based on 50% of generation assumption)
Effectively 46.5p/kWhr irresepective of what I do with it.
Gas costs are about 6p/kWhr equivalent.
So, if I was paid for what I put back, I'd be out of pocket by 3p/kWhr for the ~5kWhrs that I can store as hot water. (about 15p/day)
What I'd really like to do is find a way to store more. Batteries are not cost effective to buy but I get mine from old server room UPS swap outs.
I only store 100Whrs or so each day which is used for LED lighting at night.
I dream one day of converting all the surplus into hydrogen and using that to feed a fuel cell but the technology is far too expensive right now.

There's never any back EMF on BLDC. Absolutely no need for a catch diode there, as they are inside the BLDC controller, across each of the three motor phase wires (technically they aren't separate diodes, they are usually the body diodes inside the commutation FETs).

Hi Ed,

You don't appear to have a decoupling capacitor on the input to the 78L05. Also, you may need to review the power dissipation within a 78L05, since you presumably have a long target lifetime for the system (power = heat = higher failure rate). Also, does the OLED emit 24/7? If so, you might need to review the expected lifetime (manufacturer's specification).

However, my main concern is with the "PWM" drive to the motor. Assuming the motor is even responsive to voltage/speed control, I would have expected a choke in series with the FET (and probably a catching diode to the supply rail), but the inductance will depend very much on the motor characteristics (i.e. just for RFI reduction, or maybe full load current filtering). However, as mentioned above, with a high (thermal) inertia system such as this, might it be better to run the motor (at full speed) in bursts of some seconds or minutes?

Cheers, Alan.

AllyCat said:

You don't appear to have a decoupling capacitor on the input to the 78L05. Also, you may need to review the power dissipation within a 78L05, since you presumably have a long target lifetime for the system (power = heat = higher failure rate). Also, does the OLED emit 24/7? If so, you might need to review the expected lifetime (manufacturer's specification).

Click to expand...

Thanks, so its decoupling capacitors everywhere. Yes, I'll upgrade the regulator anyway. The main 5V power drain is the OLED. Yes, the OLED is on 24/7, the datasheet on TechSupplies has no info on its life - where might I find that info?

AllyCat said:

However, my main concern is with the "PWM" drive to the motor. Assuming the motor is even responsive to voltage/speed control, I would have expected a choke in series with the FET (and probably a catching diode to the supply rail), but the inductance will depend very much on the motor characteristics (i.e. just for RFI reduction, or maybe full load current filtering). However, as mentioned above, with a high (thermal) inertia system such as this, might it be better to run the motor (at full speed) in bursts of some seconds or minutes?

Click to expand...

The motor works well with pwm. I have applications that have been running for a year now with no problems. I assume the choke is for reducing radio frequency emissions, rather than extending life, is that correct? I can't let the motor stop as the system would drain back and would need to re-fill on start up. The power take also drops massively when pwm is implemented and so I only run the motor at full power to prime the system and when the sun is full on the panels.

Cheers, Ed

BeanieBots said:

It would be a good idea to put some uF on the power out line just after the 27R .....................
If you want to stick with OLED (I did), then put a similar resistor (27R) at the OLED end of the lead and fit 100uF + 100nF each side of the resistor.

Click to expand...

Thanks - very helpful

BeanieBots said:

A little more about PV.
My electricity is more like 18p/kWh
I get 45p/kWhr for generation. (FiT)
I get a further 3p/kWhr for what goes back to the grid (based on 50% of generation assumption)
Effectively 46.5p/kWhr irresepective of what I do with it.
Gas costs are about 6p/kWhr equivalent.
So, if I was paid for what I put back, I'd be out of pocket by 3p/kWhr for the ~5kWhrs that I can store as hot water. (about 15p/day)
What I'd really like to do is find a way to store more. Batteries are not cost effective to buy but I get mine from old server room UPS swap outs.
I only store 100Whrs or so each day which is used for LED lighting at night.
I dream one day of converting all the surplus into hydrogen and using that to feed a fuel cell but the technology is far too expensive right now.

Click to expand...

The government is crazy! Why do they pay you 45p to generate it and at the same time give you no incentive to export. You need to be paid heaps more to export - the 18p/kWh in fact. You can get an old night storage heater and waste your kWhs that way too. That's the perverse incentive they've set up for you.

I agree with the dream for storage, but with the wind turbines etc the UK needs more storage on the grid, which solar will complement. And large scale storage is cost effective. You should be given the incentives to use it.

I dream one day of converting all the surplus into hydrogen and using that to feed a fuel cell but the technology is far too expensive right now.

Click to expand...

Unless you have NASA level budgets & support,static fuel cells can be costly & cantankerous,& pretty inefficient overall,especially if using locally produced hydrogen.Significant electrical energy is wasted just heating the water as it's electrolysed of course- fine minds have long pondered improvements !

I've been involved with datalogging a hybrid scheme (wind/solar/gene/fuel cell) on an offshore NZ island (Somes/Matiu-Wellington harbour),which uses surplus PV/wind generated capacity to electrolyse water & store the hydrogen in underground pipes.

Best I don't get too emotive here about the fuel cell outcome,but will relate that the hydrogen has merely been used to run a gas BBQ ! For the money involved a room full of storage batteries (long life LiFeYPO4 perhaps) could have been used instead. Some modest battery storage has now been provided in fact.

Ed: Best of British with your thermal SWH venture- perhaps look for most sales on the UK South Coast!

ed1066 said:

the OLED is on 24/7, the datasheet on TechSupplies has no info on its life - where might I find that info?

Click to expand...

Hi Ed,

Possibly in a data sheet from the manufacturer of the original display panel, but not necessarily even there. It's a long time since I had any "inside information" but one reason that OLEDs took (and are still taking?) a long time to become a widely-used technology was because of premature "wear out". There are of course other display techologies which (also) potentially exhibit limited lifetimes such as Vacuum Fluorescent and Plasma, etc.

ed1066 said:

I assume the choke is for reducing radio frequency emissions, rather than extending life, is that correct?

Click to expand...

Primarily yes. Each time the FET switches on, there may be a pulse of current (= RFI) as it charges the capacitor across the motor. An inductor potentially reduces the current spike and also the peak power dissipated in the FET. You haven't said at what PWM frequency the system operates, but I suggest that you should be evaluating both power dissipation (losses) and RFI at various PWM frequencies, before submitting to any formal testing.

Predicting long-term system reliability might be an important issue for this project. I guess the "general environment" may involve quite high temperatures and the lack of ventilation if enclosing components within sealed boxes (and insulation) will cause them to heat further. Bear in mind that the normal method for "accelerated reliability testing" of electronic components is to heat them up!

Cheers, Alan.

Hi again,

I thought that a search of the forum might identify the original manufactuer of Techsupplies' OLED panel, but only found a few other members expressing concern about the "longevity" of OLEDs. However a web search soon produced a data sheet for a "branded" panel which is moderately encouraging. Page 51 quotes 100,000 hours (about 10 years continuous operation to save calculating) to 50% brightness at 25 degrees C (I believe a rule of thumb is that lifetime halves for each 10 degrees rise in temperature).

However, the inclusion of two footnotes is interesting:

2. You can use the display off mode to make long life.
3.The average operating lifetime at room temperature is estimated by the accelerated operation at high temperature conditions.

Also, that does seem quite an "expensive" display, you might not get the same lifetime from a "noname" product from China via Ebay.

Cheers, Alan.

I don't think it's any secret that the OLEDs are from Winstar, who are, I believe, one of the biggest manufacturers of OLED displays. In terms of life, I have one sitting next to me that has been on continuously since around January last year and it still seems to be exactly the same as it was when new.

Interesting steer on OLED's. The principal feedback I got from my pioneer solar installations was that the LCD's were a pain. The displays were in airing cupboards by the hot tank, usually not that well lit, and most people needed a torch to read them. Hence the change to OLED's. I need a life of more than 10 years, so I guess I need a push button to switch them on to read the output. Airing cupboards are typically 25 to 30 degrees C.

Cheers, Ed

On the subject of OLED's in darkened [rarely visited by humans] places, I happen to have the same issue on my current project. The solution I'm adopting is a cheap and simple PIR on a spare (digital) input pin. The PIR is on an interrupt and simple 'enables' the code displaying the OLED data when it's active. After a 10 second 'lack of PIR input' the display switches off.
I'm an automation nut - but in this case it saves the user finding the push button in the dark as well!

A few suggestions:

Rather than alarm cable, you could use twin-screened (two cores in a single screen). I use that for remote temp sensing though I use serial from a remote 08M with the sensor rather than direct connection. It would increase noise immunity, should that be a problem.

Birds seem to like picking at anything exposed on the roof. I have had to cover pipe insulation with EDPM membrane, and I run the panel sensor cable inside 10mm copper from the hole in the roof tile to the panel.

My algorithm starts pumping when the panel is 10C hotter than the tank bottom, and stops when the arriving water is less than 1C higher than the departing water. There is a startup delay to allow the first hot water to arrive. So I have four sensors - panel, tank bottom, arriving temp, departing temp. There is no risk of boiling where I live, freezing is however an issue so the thing shuts down when any temperature is below 2C. The system itself is safe when frozen so just uses plain water in a closed loop with a top-up tank.

I haven't found any reason to use PWM on the pump, I just use on/off. As the lag in the system is rather long and water takes several minutes to arrive (via 8mm silicone rubber pipe), subtle control seemed a bit redundant, with nothing obvious to optimise.

The one real weakness that I haven't yet solved is how to measure the actual heat flow, to get some idea of actual kWhs transported. This would need flow rate, specific heat and many temp samples, and calibration which I haven't rried to tackle. All I can say is that the cold is preheated thus giving the gas less to do, and the pre-heating in my tank can be by 30-40C in the summer.

I also have grid-linked PV, generating up to about 3.5kW at peak, and have just passed 4MWh in total. The thing about using PV for water heating is that we don't yet have smart meters in the UK so export can't be measured. Hence the 'deemed' figure. Once we do I would expect the tariff structure to change to reflect real-time demand and that will likely affect behaviour. Interestingly, in the UK we currently have around 1.5GWhp of sub-50kWhp PV installed, or around the capacity of several small or one mega power station. All built and running within a couple of years. And paid for by voters out of their own pockets.

AllyCat said:

However, the inclusion of two footnotes is interesting:

2. You can use the display off mode to make long life.
3.The average operating lifetime at room temperature is estimated by the accelerated operation at high temperature conditions.

Click to expand...

You could also dim the display to increase the life.

The Winstar OLED used in PICAXE Computer Fan Controller with speed and temperature measurement for over a year, probably on for about 5-8 hours a day, has absolutely no visible screen burn when all the pixels are lit using the 'Test OLED' facility in the menu. In that time the display has not been dimmed and some of it has been displaying the exact same thing all that time.

Winstar OLEDs certainly don't seem to be plagued by the same problems that AMOLED displays supposedly are but then again in the latter case it's in a phone that a manufacturer wants to sell the customer a new one every year (on a 12 month contract) so they can manufacture AMOLED displays down to a low price and increase profits.

AllyCat said:

Also, that does seem quite an "expensive" display, you might not get the same lifetime from a "noname" product from China via Ebay.

Click to expand...

Rev-Ed use Winstar OLEDs and even say so on the AXE133 product page. Rapid overcharge for their sales of Winstar OLEDs - Technical has even said so.

geoff07 said:

A few suggestions:

Rather than alarm cable, you could use twin-screened (two cores in a single screen). I use that for remote temp sensing though I use serial from a remote 08M with the sensor rather than direct connection. It would increase noise immunity, should that be a problem.

Birds seem to like picking at anything exposed on the roof. I have had to cover pipe insulation with EDPM membrane, and I run the panel sensor cable inside 10mm copper from the hole in the roof tile to the panel.

Click to expand...

I really appreciate your post and its great to share experiences .......

Yes, I use special tape from SolarShopEurope for bird protection. My applications have the sensor lead inside the pipe lagging, beside the silicone hose. Max temp is about 75 deg C and is rare - its controlled for that. Any hotter and the system drains. Very close to the panel the sensor lead changes to silicone insulated, as stagnation temps can be 170 degr C. Would any of this concern you?

geoff07 said:

My algorithm starts pumping when the panel is 10C hotter than the tank bottom, and stops when the arriving water is less than 1C higher than the departing water. There is a startup delay to allow the first hot water to arrive. So I have four sensors - panel, tank bottom, arriving temp, departing temp. There is no risk of boiling where I live, freezing is however an issue so the thing shuts down when any temperature is below 2C. The system itself is safe when frozen so just uses plain water in a closed loop with a top-up tank.

Click to expand...

It sounds like you have a SolarTwin type panel. Mine is not freeze-safe and I use drainback. My control is very similar, but the panel temperature and departing temperature double up. The sensor is held closely against the outlet pipe with thick adhesive heat shrink with a tiny dab of heat paste. This is exactly at the location where the silicone hose connects and I use the outside of the hose to help protect and separate the sensor connections. Its all covered with more soft silicone and tough heat shrink and then the lagging over all that. I found that the panel temperature should be rising as well to ensure a valid switch-on event. ... so easy to include when you have PICAXE.

geoff07 said:

I haven't found any reason to use PWM on the pump, I just use on/off. As the lag in the system is rather long and water takes several minutes to arrive (via 8mm silicone rubber pipe), subtle control seemed a bit redundant, with nothing obvious to optimise.

Click to expand...

Yes, sensible. I'm trying to minimise the drainback events as well as minimise power. Our big difference is the drainback - I need a decent pump to get the system primed, but then I want the ability to run it gently, especially in moderate sun. I find that the hose can have bubble build up which can stop it all functioning, so it goes full-whack for a few secs every 10mins. If pushed I can leave off pwm, with regret.

geoff07 said:

The one real weakness that I haven't yet solved is how to measure the actual heat flow, to get some idea of actual kWhs transported. This would need flow rate, specific heat and many temp samples, and calibration which I haven't rried to tackle. All I can say is that the cold is preheated thus giving the gas less to do, and the pre-heating in my tank can be by 30-40C in the summer.

Click to expand...

I use a simple flow meter bought from eBay which great with the COUNT command and runs from the 5V. Just add the resistor as they say and you're away.

http://www.ebay.co.uk/itm/Flowmeter-Switch-Pulsed-Water-Beer-Turbine-Flow-Meter-Sensor-1-/160749825954?refid=store

I only calibrate and then disconnect as its not designed for high temperatures and is something else to go wrong anyway. I derive the relationship between pwm output and actual flow rate and then use it in the equation

Watts = 4.18 * flow * Trise , where flow is g/s & Trise deg C

The display can then give Watts at all times and kWhr since commissioning. Fun to program.

geoff07 said:

I also have grid-linked PV, generating up to about 3.5kW at peak, and have just passed 4MWh in total. The thing about using PV for water heating is that we don't yet have smart meters in the UK so export can't be measured. Hence the 'deemed' figure. Once we do I would expect the tariff structure to change to reflect real-time demand and that will likely affect behaviour. Interestingly, in the UK we currently have around 1.5GWhp of sub-50kWhp PV installed, or around the capacity of several small or one mega power station. All built and running within a couple of years. And paid for by voters out of their own pockets.

Click to expand...

Yes, I'm with you on this. There was a debate in DECC as to the kind of meter that was acceptable for assessing export payments. They set the bar so high that they went for deeming, which is the least accurate possible and the source of my frustration.

geoff07 said:

Birds seem to like picking at anything exposed on the roof. I have had to cover pipe insulation with EDPM membrane, and I run the panel sensor cable inside 10mm copper from the hole in the roof tile to the panel.

Click to expand...

Anything installed outside brings with it the issue of lightning strike and protection which may be something else you need to add to your list of concerns.

Add corrosion too! Back in the late 1980s a "dirty connection" on a rugged roof top frost sensor lead to my commercial system failing to detect frost conditions. A rare (for NZ) overnight freeze caused pipes in several panels to rupture. The system had run "well" (although not cost effectively on RoI) for ~3 years up to that point, but the repair/replacement of the panel(s) pretty much chopped into any energy savings we'd made...

nick12ab said:

Rev-Ed use Winstar OLEDs and even say so on the AXE133 product page. Rapid overcharge for their sales of Winstar OLEDs - Technical has even said so.

Click to expand...

No we didn't. We said the quoted price at that time seemed excessive, which is very different to your overcharging interpretation. Rapid may be buying small quantities at a much higher input cost. A higher sale price does not automatically mean they are overcharging.

Again, thanks for everyone's help here. I have now been through the whole EMC Testing experience and have just now reported back using a new thread in this Forum.