Extreme ammeter geekery.

[MoominPapa]

Please can someone check my calculations and reasoning?

 

I have a 0-100A moving-coil ammeter measuring the output from my alternator via a shunt. The shunt was supplied as matching the meter.

 

I want to use this meter to measure the output from my newly-installed solar panel when the engine isn't running. Clearly, 100A FSD is a bit OTT for a panel, so the plan is to use a suitable shunt to scale it to 0-10A.

 

The meter dial has "75mV" written on it by hand. I'm assuming that means that full scale deflection occurs when there is 75mV across the meter coil, ie the product of the current sensitivity and the coil resistance is 75mV. Sound reasonable?

 

So, for my solar shunt, I need something which drops 75 mV when 10A is flowing, ignoring the parallel resistance of the meter coil which will be much larger. R=V/I, so 0.0075 ohms. I happen to have bag of 0.1ohm 3W resistors, and I recon that 13 of those in parallel will be .0076 ohms which is a pretty good approximation. Total dissipation in all 13 will be less that 1W, so no problems there.

 

So that's a plan. New shunt in solar panel line and a DPDT relay controlled from the engine ignition which switches the meter across the alternator shunt when the engine is on and the solar shunt when it's off. But! The current through the meter coil and therefore the relay contacts is going to be small - tens or hundreds of microamps? I seem to remember that there's something called "contact wetting" and that switch and relay contacts need a certain minimum current to connect reliably. Said minimum is more than a hundred microamps.

 

My solution to this problem is to move one if the 0.1 ohm resistors from the solar shunt and connect it directly across the meter coil. Said resistor will pass 0.75A at full-scale, plenty to wet the relay contacts. Since the resistance of the shunt in the alternator circuit is a couple of orders magnitude lower than 0.1 ohms, the loss of accuracy from this resistor when in alternator mode is less than one percent, and I'm not worried about that.

 

Gibbo, Keeping Up, Nibble, anyone else, what have I missed?

 

MP.

Edited August 23, 2011 by MoominPapa

[Keeping Up]

At first thought (as usual, posting before the arrival of any second thoughts) you could significantly affect the accuracy in normal usage because the extra current taken by your new setup would cause a significant voltage drop in the relatively thin wires from the shunt to the meter. If they have a resistance of 0.05 ohm each for example, then you would drop half of your applied 75mV in the wiring and your meter will be 50% out.

 

I think a better solution would be to use a relay that is designed for very low currents, such as one with gold-plated contacts.

[MoominPapa]

At first thought (as usual, posting before the arrival of any second thoughts) you could significantly affect the accuracy in normal usage because the extra current taken by your new setup would cause a significant voltage drop in the relatively thin wires from the shunt to the meter. If they have a resistance of 0.05 ohm each for example, then you would drop half of your applied 75mV in the wiring and your meter will be 50% out.

Well, that's a good first thought. Those wires are reasonably chunky, but they're three or four metres long. Damn.

 

I think a better solution would be to use a relay that is designed for very low currents, such as one with gold-plated contacts.

More Googling found something that suggested that to punch-through the oxide layer what matters is not the current when closed, but the voltage across the contacts before they close; 10V being mentioned. Clearly there's no way to achieve that here.

 

The relay from the bits-box slated for this role came from Maplins, I think it's this one: Maplin. The specs there don't give contact material but it looks like gold in the picture. (I don't have it with me to actually look).

 

If the contacts are gold, is that sufficient, or are low-current relays special in other ways?

 

 

Thanks for your help.

 

MP.

Edited August 23, 2011 by MoominPapa

[Keeping Up]

I remember from my old old GPO days, they called it "DC wetting"; supplying a little bit of DC down the cables so that the joints (which were merely twisted together) remained conductive. But that was driven by 50v dc from the exchange.

 

I do not think that the relay you linked to would have gold contacts. I believe that simply using gold contacts would be enough (gold does not oxidise) but I am prepared to be corrected in this. I can't see one in the Maplins catalogue though

[MoominPapa]

I do not think that the relay you linked to would have gold contacts. I believe that simply using gold contacts would be enough (gold does not oxidise) but I am prepared to be corrected in this. I can't see one in the Maplins catalogue though

OK, I'll source a relay with gold-plated contacts. Better to get this right first time!

 

 

Thanks Allan.

 

MP.

[FadeToScarlet]

Why not put in a second ammeter? Do you just like try challenge of making it work?

[rjasmith]

OK, I'll source a relay with gold-plated contacts. Better to get this right first time!

 

 

Thanks Allan.

 

MP.

 

Even better might be a mercury wetted reed relay (I've got quite a few lying round here from past working days and could let you have a couple - they probably aren't changeover reeds so you'd need two).

 

Will just have to find a couple with 12v coils.

 

CEGB/Nat Grid used to use vast nos of these to switch small analogue voltage signals representing flows on the Grid into the ADCs of the telemetering gear at substations.

 

Instead of needing dc voltage "wetting", the reeds have a tiny amount of mercury inside that "wicks" up the reeds so that when the reed blades close the mercury gives a relatively large conducting area instead of a single point. Also the atmosphere inside the glass reed tube is inert to avoid oxide formation.

 

Let me know if you're interested (No fee - just returning a favour in respect of past help with libraries!!)

 

Richard

[MoominPapa]

Even better might be a mercury wetted reed relay (I've got quite a few lying round here from past working days and could let you have a couple - they probably aren't changeover reeds so you'd need two).

 

Many thanks for the offer. I can see a couple of potential problems with these. The first is that I need to switch both ends of the meter coil - because of the split charge arrangements there's no sensible way to arrange a common point at one end of the shunts. Using two relays risks one side making before the other breaks, and instant magic-smoke escape from the meter coil. The second may not be a problem, but I've never seen a reed relay with normally closed contacts. I really don't want to have to pay the power consumption cost of keeping a relay energized all the time the engine _isn't_ running. With a conventional DPDT relay I can use the NC contacts for the solar side and just energise the relay when the engine is on. Is there an equivalent option with reeds?

 

Cheers,

 

MP.

[smileypete]

This'll do.

 

cheers,

Pete.

[rjasmith]

Many thanks for the offer. I can see a couple of potential problems with these. The first is that I need to switch both ends of the meter coil - because of the split charge arrangements there's no sensible way to arrange a common point at one end of the shunts. Using two relays risks one side making before the other breaks, and instant magic-smoke escape from the meter coil. The second may not be a problem, but I've never seen a reed relay with normally closed contacts. I really don't want to have to pay the power consumption cost of keeping a relay energized all the time the engine _isn't_ running. With a conventional DPDT relay I can use the NC contacts for the solar side and just energise the relay when the engine is on. Is there an equivalent option with reeds?

 

Cheers,

 

MP.

 

Yes but I think most of my relays are twin reed ie double pole switching and some have c/o reeds - shall I go and have a look in the shed?

 

Richard

[Gibbo]

The idea is sound and will work as is without moving one of the resistors into the meter.

 

Don't worry about the minimum current through the switch. It's theoretical bollocks that has no bearing on reality. Think about it...

 

If a switch needed a certain current through it in order to work correctly, they would have a very short shelf life. Just sitting on the shelf at the stockholders, passing no current, would wear them out. We all know this doesn't happen. In over 30 years in this job I have never, not once, come across a switch that failed due to too small a current passing through it.

Edited August 23, 2011 by Gibbo

[MoominPapa]

shall I go and have a look in the shed?

You need an excuse to visit the shed?

 

Seriously, yes please, that sounds like a very Rolls-Royce solution.

 

The idea is sound and will work as is without moving one of the resistors into the meter.

 

Don't worry about the minimum current through the switch. It's theoretical bollocks that has no bearing on reality. Think about it...

 

If a switch needed a certain current through it in order to work correctly, they would have a very short shelf life. Just sitting on the shelf at the stockholders, passing no current, would wear them out. We all know this doesn't happen. In over 30 years in this job I have never, not once, come across a switch that failed due to too small a current passing through it.

 

I think the mechanism is that without enough load, the oxide film can persist between the contacts and the relay is high-resistance. A bit or arcing is enough to remove the oxide and it all works fine.

 

I found an interesting anecdote online. Apparently it's common to make universal relays which both work at very low current and support a reasonable load. This is done by making the contacts of bulk silver/cadmium (for load) and plating them with gold (for low current). If the relay is used with a load the very thin gold is blasted away and the silver/cadmium comes into play. Of course this means that if a relay is re-purposed from switching a load to dry switching, the gold is gone and it may fail. A dual pole relay with a symetrical base was once plugged in backwards on an America submarine, connecting what had been loaded contacts to the low current logic and vice-versa. This resulted in a very embarrassing failure to submerge when required!

 

MP.

Edited August 23, 2011 by MoominPapa

[Gibbo]

I think the mechanism is that without enough load, the oxide film can persist between the contacts and the relay is high-resistance. A bit or arcing is enough to remove the oxide and it all works fine.

 

Indeed. The mechanism is well documented. The problem is, it simply doesn't happen in the real world. Almost every audio switch in the world is operated miles below its theoretical minimum current, yet almost every one of them is still working fine 50 years later.

[Robin2]

Please can someone check my calculations and reasoning?

 

I have a 0-100A moving-coil ammeter measuring the output from my alternator via a shunt. The shunt was supplied as matching the meter.

 

I want to use this meter to measure the output from my newly-installed solar panel when the engine isn't running. Clearly, 100A FSD is a bit OTT for a panel, so the plan is to use a suitable shunt to scale it to 0-10A.

 

What about buying a cheap digital multimeter that can measure 10amps - £6 ?

 

ETA (because I have been thinking of this subject for myself) what you really need with solar panels is an amp-hr counter because the amps can fluctuate so much that instantaneous readings on an ammeter are no more than interesting reassurance.

Edited August 23, 2011 by Robin2

[MoominPapa]

What about buying a cheap digital multimeter that can measure 10amps - £6 ?

 

ETA (because I have been thinking of this subject for myself) what you really need with solar panels is an amp-hr counter because the amps can fluctuate so much that instantaneous readings on an ammeter are no more than interesting reassurance.

 

I'd add another meter but there's nowhere to put it without is looking like a late addition:

 

 

Agreed about the counter, but I have on on the batteries, which is the most important place. "interesting reassurance" is mainly what I want.

 

Cheers,

 

MP.

[rjasmith]

Oh dear!

 

After all my splendid promises and much searching of the various sheds plus the loft I've now given up looking for the box of B & R Hg wetted relays that must still be lurking around here somewhere (but not, sadly, where I've been looking!). I've found a box of dry reed relays that are DP but not DT but I realise you don't want to do it that way!

 

Although I can't send you a freebie (unless the box I was after turns up), maybe this one on eBay here for a few quid would do for you?

 

The coil voltage is only 3.3v so you'd need a dropper resistor. Otherwise, perhaps an ordinary gold plated DPDT relay will work well enough as Gibbo suggests.

 

I have to say though that on my little PO 3000 based strowger house PAX I built just for fun out of old ex CEGB bits I did forget a bit of DC wetting in one part of it and had trouble with noisy contacts until I realised what I'd done. After a mod to include the wetting voltage in the transmission bridge - no more noisy contacts!

 

(But then these were rather ancient relays and uniselectors so a bit of contact cleaning might have improved it as well!)

 

Richard

[MoominPapa]

Oh dear!

 

After all my splendid promises and much searching of the various sheds plus the loft I've now given up looking for the box of B & R Hg wetted relays that must still be lurking around here somewhere (but not, sadly, where I've been looking!). I've found a box of dry reed relays that are DP but not DT but I realise you don't want to do it that way!

Not to worry Richard, and thanks for looking. As a teenager I could keep track of my extensive inventory of salvaged and reclaimed electronic components in my head, sadly these days I tend to forget I have stuff, buy another and then find the original or be convinced I have something and then fail to find it. You're not the only one.

 

Otherwise, perhaps an ordinary gold plated DPDT relay will work well enough as Gibbo suggests.

The original Maplin candidate is in my hand now, and the contacts sure look gold-plated to me, so I'll try that one as a first go, I think.

 

I have to say though that on my little PO 3000 based strowger house PAX I built just for fun out of old ex CEGB bits I did forget a bit of DC wetting in one part of it and had trouble with noisy contacts until I realised what I'd done. After a mod to include the wetting voltage in the transmission bridge - no more noisy contacts!

 

(But then these were rather ancient relays and uniselectors so a bit of contact cleaning might have improved it as well!)

Ooh, that sounds like fun. Do you still have it? I love the Strowger exchange in the Kensington Science Museum.

 

 

MP.

 

Edited to add: I just Googled the manufacturers part number on the relay I have and the contact material is given as "AgNI gold flashed" so that looks hopeful.

Edited August 23, 2011 by MoominPapa

[rjasmith]

Ooh, that sounds like fun. Do you still have it? I love the Strowger exchange in the Kensington Science Museum.

snipped.

 

Certainly do! I can summon cups of tea from any corner of the house, garage, various sheds, even from inside the boat (while it's here outside in the garden) with this.

It's a 9 line PAX (well there's a "B" in there as well - guess what dialling 0 does - very norty!).

 

The ringing and tone generator at the top is one of two machines I have from the old East Grinstead Grid Control Centre which looked after the SE corner of the UK Grid.

 

The relay sets (I've taken their lids off for the photo) are from old electromechanical SCADA systems that used to control/gather data from substations during the 1950s through to the 1970s when we got our first computer things.

 

Sadly no two-motion selectors though so it's nothing like as exciting as a full size exchange or even the apparatus rooms of the Grid Control Centres.

 

I completely rebuilt the relay sets using a pair of uniselectors and relays from other sets to make up an exchange to my own design. Designing it relieved the boredom on train journeys to/from work for a while!

 

Richard

[Keeping Up]

Thanks for posting that Richard, I love it. Takes me right back to my apprenticeship days of adjusting relays in the exchanges, with a pair of long-nosed pliers and a spring balance (I won't mention my later job as an "attachment approvals officer")

[Gibbo]

A lost art and skill if ever I saw one. Its a delight to see stuff like that!

[WotEver]

Sadly no two-motion selectors though...

Oddly, I've never seen those available. You'd think they'd turn up at boot sales, or Army Surplus or something. Not that I'd know what to do with one if I had one, but I'm just surprised I've never seen one for sale.

 

Lovely job you've done.

 

Tony

[rjasmith]

Oddly, I've never seen those available. You'd think they'd turn up at boot sales, or Army Surplus or something. Not that I'd know what to do with one if I had one, but I'm just surprised I've never seen one for sale.

 

Lovely job you've done.

 

Tony

 

I'm blushing now - thank you chaps! Actually if a real GPO engineer (eg Keeping Up) saw it he'd have a good laugh at my rather average wiring tidiness, especially on the rebuilt relay groups. I was never properly trained in the art of PO 3000 style construction but tried to learn a few snippets at the feet of proper telecomms engineers now no longer with us.

 

It is rather fun though to pick up one of the house phones (all 746's except for the pride and joy in the sitting room which is a black bakelite 332 with proper plaited textile cords) and hear the linefinder uniselector hunt round to grab the line, the tone generator whirr into life and return "proper" 48Hz dial tone into your ear!

 

I think I have seen the odd 2 motion selector on eBay but when the strowger gear was decommissioned by BT (and on a smaller scale by Nat Grid, Railways etc) a lot of specialist scrap firms made a lot of money out of reclaiming the exotic metals from relay springsets. Our PABX7 at the office I'm happy to say was donated to the Watercress Line and should still be alive and well operating their little railway telephone network.

 

For a bit more nostalgia try this for a few long lost telephone noises!

 

Richard

[Gibbo]

...laugh at my rather average wiring tidiness

 

You're having a giggle right?

 

This proper wiring tidiness...

 

[WotEver]

This proper wiring tidiness...

But does it work?

 

Tony

 

For a bit more nostalgia try this for a few long lost telephone noises!

ooh, takes me right back to pressing Button B

[Gibbo]

But does it work?

 

Surprisingly yes. See the PCB at bottom right? That's a new toy that is, due out soon. Dogs bollocks