24v->5v PICAXE power conditioner - thoughts?

[Eclectica]

Hi,

I have drawn out a circuit diagram to take a standard isolated 24Vdc control panel power supply input and enable it to be used in a DIN rail box mounted PICAXE circuit to control temperature of a heated bed for horticultural seed germination (and probably other things later on).

The main idea is to avoid the annoying resets typically encountered where the power supply (and/or) inputs are noisy with spikes. We have an extensive electrical system to which are connect many 3-phase water pumps, vent motors, sodium lamps, and solenoids. Where possible I have added suppression at the source of inteference with: X2 caps, MOVs, and snubbers, but it is not perfect. Where possible too, the obvious step of keeping clear on mains voltage cable runs is taken.

The thing is, there is not a great deal of space in my 3wide din rail mounting box and so if anything deemed unnecessary could be left out then it would help putting it all together.

I was wondering what anyone with a good deal of experience with PICAXE in electrically noisy environments could cast an eye over the diagram and offer any thoughts/advice - polite of course.

Thanks in advance.

M.

 

[MartinM57]

Depending on the spec of the 24v transorb it might be to low for a nominal 24v input. Got a proposed part number? Maybe 28v or so would be better?

I'd seriously consider putting the inductor before the regulator - anything nasty it filters will have already gone through the regulator

EDIT: Worth considering a reverse biased diode across the converter input and output (pointy end to input) to lose any high voltage spikes induced in the downstream circuitry back into the power supply.

What DC-DC converter is it? Part no?

 

[Eclectica]

Hi Martin

Thanks for the reply, and suggestions.

Transorbs are 1.5KE24A 24V (Vbr 22.8-25.2 @ It)... (could start having an effect even with ~0.6 volt drop across 1N4004?)

...and SA5.OA 5V (Vbr 6.4-7.0 @ It).

DC-DC convertor is a Recom R-785.0-0.5, (Vin 6.5-34V)

Regards
M.

 

[MartinM57]

Well I would't choose a 24v transorb IMHO - you want to absorb occasional transients some way above 24v, not have a transorb that starts conducting at 22.8 volts when the standing supply voltage it sees is 23.4 volts.

 

[Dippy]

I agree.
I suggest some deep and meaningful reading of 'transorb' or any transient suppressor studying it's response .

I would put the choke on the input side.
I can't suggest a vlaue without knowing the characteristics of your noise problem.

You haven't specified capacitor type for the 100nF caps.
Why the cap in series with 100R. I know the principle but why have you put it there?

Input Cap for reg.
You have a 1uF tant.
Is this a guess?
I see the recom data sheet suggests 10uF (type unknown).

Is your design just a belt & braces idea , or actually based on a known 'good' circuit.

 

[BeanieBots]

Transorbs are generally used to catch massive surges that might destroy devices rather than to filter "noise". Without knowing more specific details about your issues, it's hard to advise.
36v transorbs are typical for protecting a 24v line.
The 5v transorb won't do much for you unless the switcher fails and gives a high voltage output. That will only happen if you've got its input wrong.

The datasheet for the switcher should give suggestions on how to eliminate switcher noise. That inductor looks too big IMHO for such a job. DON'T try to filter out input noise by putting components on the OUTPUT.

Does your application need to have it's 0v closely referenced to supply 0v?
If not, consider the use of a common mode choke on the input.
The 0.1uF+100R (snubber?) won't do much for you here.
Consider a few ohms of resistance (and/or inductance) between the first input cap and switcher input cap. Talking of which, the input cap needs to be much bigger. At least several 100uF and low ESR. This is true simply because it's a switcher irrespective of any noise issues.

If you can supply more details (numbers) about the nature of the "noise", we can probably be more specific.

 

[Eclectica]

Thanks for all the informed replies!

Regarding the transorbs, I usually just use the 5v ones as a matter of course in virtually all circumstances. Here it looks like I extrapolated the logic of '5v transorb for 5v supply' to 24v without reading the specific manufacturer's spec - will change this.

Nature of interference - I probably should take my scope up to the glasshouse and see the extent of the spikes, but I don't fancy taking my aged steel framed Telequipment DM63 up to such a moist environment with all the drips from the glasshouse roof etc!

Taking a guess at the inteference, it would be short (<20ns?) spikes/oscillating bursts originating from collapsing fields from solenoids, motor coils, and iron core ballast sodium lamp ignitions. There are also about 20 computers and other IT equipment on the system near the source dist board which I imagine tend to transmit lots of spikes and harmonics through the neutral all over the rest of the installation.

[The typical induced AC (with attendent random spikes) picked up by me in the office (horrible mobile phone picture of my scope screen attached!) causes me to be concerned.]

Belt and braces approach (loosely based on several ideas) - yes! I hate having to go back and add on MOVs, caps, and snubbers all over the place to make a circuit work as reliably in the field as it did on my bench.

[As a typical example, we have a 1 year old portable fan assisted gas heater with no suppression across the fan, and it sends back sufficient inteference down the mains and 240v thermostat switch lines to reset a PICAXE if you run a 5V switchmode powered unit on the same mains circuit. (Adding a 100R+0.1 X2 cap across the heater mains input, and clipping a ferrite on the thermostat cable, and adding a 0.22uF X2 cap across the thermostat switch terminals stops this.]

I am looking to build a panacea 5V PICAXE psu conditioner which can be run from a normal isolated 24vDC panel supply without having any kind of issues - sorry perfectioninst side of me coming through here I think.

As for the filter components on the 5v side - there might be a possibility of stray coming back from the 5v powered NTC cable runs (this is another issue), and I don't want induced spikes upsetting 5v rail and subsequently the ADC10 readings. I'll likely use T-network capacitors with ferrite beads on the signal input and then 10k (or more) to the PICAXE pin).

Line of reasoning with the large inductor is to filter down to the lowest feasible frequencies possible - this is probably an oversimplification. Switcher runs at 380kHz, so could get away with a small inductor on output. Common mode on input sounds a good thought if I can squeeze one on - looks like Murata's PLA10 have a small footprint. Any idea what would be a good choice of value for common mode?

Switcher datasheet suggests input capacitor not needed, but felt something should be there! Also max output is into 220uF, so I'll change this back (original circuit design had 100uF input and output caps!).

0.1uF,100V are WIMA polyester MKS02, so I thought they might be a good choice for filtering. Also, if the 100R + 0.1uF snubber is not going to do much then it should go.

Thanks again. If there is anything else worth bearing in mind then I would love to hear it - like everyone here, I hate it when a circuit fails due to lack of thought and planning for part of it.

Regards
M.

 

[Dippy]

Ideally, you should be investigating suppressing the noise at source.
Paying attention to good wiring methods and looking to see if you have a dodgy/knackered device switching something.

Then you should look at suppressing the mains input to your power supply.
Something along these lines. (See diagram)
Choose the component values and types with care!!#
A fuse and mains transient suppressor would also be good.
If in ANY doubt then just ask.
Maybe a filtered plug or distribution plug board may help.

Try and filter as MUCH as possible BEFORE your regulator.
Fast spikes will shoot straight through many regulators and sometimes also cause instability.
Nip it in the bud.

By all means add ceramic decouple/bypass caps.
A bit of LC filtering can help , but inductors have varying characterisitcs and varying results.

Don't just stuff 220uF caps in hope
You'd do far better with a 100uF Low ESR cap there rather than any old fat boy.
X7R or better ceramic in //l.

Be careful where you put capacitors and what size and what type.
Flinging them at a circuit and crossing your fingers is often a waste of time and money.
Size isn't always important

.... and then of course there is your skill at PCB design.


And ALWAYS , ALWAYS, ALWAYS decouple your PICAXe with a 22nF to 100nF ceramic cap as close to the PICAXE +v / 0V power pins as possible. Maybe add a 10uF tantalum cap in //l.
If you think noise is getting to an I/O pin or the reset pin then filter it.

 

[BeanieBots]

The sodium lamp ballasts will be a significant issue.
I've had trouble with these myself and had to resort to using enormous snubbers. Values in the order of a few ohms + a few microfarads were required to keep them quiet.
As Dippy states, "Choose the component values and types with care!!#"

 

[eclectic]

Snipped

Nature of interference - I probably should take my scope up to the glasshouse and see the extent of the spikes, but I don't fancy taking my aged steel framed Telequipment DM63 up to such a moist environment with all the drips from the glasshouse roof etc!

Taking a guess at the inteference,

A guess at the interference!

Just an opinion from the Junior Common room.

In that case, shouldn't you really
be thinking about

A. A small portable 'scope, or
B. A Picoscope type setup

or
C. anything that can use batteries and
can fit inside a big polythene bag?

As far as I can see, in your job, it's essential.
e

Addit.

I'm sure that there is a Fluke 'scope
that can do the job,
but the Boss may have to sell the greenhouse
to buy one.

 

[Dippy]

Is this for something commercial/important?

That circuit I posted was from a Data Sheet of a commercial suppression box.
It cost £12+VAT and we've been discussing for 2 days.

Unless you can measure it then we're guestimating.
If you aren't happy with being able to choose the right components then just go and buy a commercial unit. Surely your happiness is worth £20 and then you can get on with the job ... and not worrying whether your home-made electricals will be going up in smoke.

 

[Eclectica]

Thanks for all the suggestions - I'll attempt to bear it all in mind when I build these units.

Interestingly I have been building this kind of thing for many years using only MOVs and snubbers and the occassional mains filter and then transorb/caps and choke on the low voltage side. It is only recently that the effects of transients in this particular environment have caused a problem when using PICAXE devices as control elements.

Previous designs have been based around temperature control modules and triac switching modules etc. but these prove to lack funtionality achievable using the most excellent PICAXE system!! The typical electronic horticultural temperature control modules are relatively expensive (£100) and fail within 2 years.

My thought is to build bulletproof modular units with replaceable plug in/changeable PICAXE-08M processor board with rock steady operation even when inherent vent motor suppression fails.
(We do already try to suppress at source - I think I said that - sorry).

Again thanks for all the suggestions. I'll post a project file when it is all done and tested.

Sharing good 'stuff' is what it is all about, and the PICAXE community is so great at this.

Thanks again.

M.

 

[fernando_g]

Agree with Dippy, a commercial suppressor box is best.

Both Schurter and Schaffner are esperts in this field, and offer a WIDE array of ready made devices. Their offerings should be available from Farnell (I have not checked though), but I've seen them both from DigiKey and Mouser.