My third SmartGauge...

[nicknorman]

6 hours ago, jonesthenuke said:

I wonder if what you say is really representative of how a SmartGauge works. These units do not measure anything but the battery voltage. They they cannot infer SoC from just the voltage at a point in time as you would also need to measure the battery temperature. The algorithm must therefore calculate SoC from the changes in voltage over time. Consequently,a fixed error will not matter, its much the same as operating the battery at a different temperature.

 

 

 

No you do not need to know the battery temperature to measure the SoC using voltage. A battery’s voltage for a given SoC with zero load barely varies with temperature.

 

A fixed error does matter because the SG both measures changes in voltage to calculate the SoC during discharge, and also measures SoC directly from voltage when there is no load, using this in part to tweak its coefficients for the particular battery model.

[Slim]

Rather than start another Smartgauge topic I have a question that I have about my smartgauge. It is regarding which battery type to set it for. When I first fitted the gauge I had Multicell AGMs and Gibbo told me to consider them FLA and set the battery choice to 1 (if I remember correctly) Whatever I set it on the gauge seemed to work O/K. Sometime down the line I replaced the Multicells with Lucas AGMs and left the settings the same. Again the gauge seemed to work O/K. A couple of months ago I replaced the Lucas batteries with Leoch AGMs

and since then the Smartgauge is reluctant to read less than 100% SoC. Example overnight 50 ah out of a 520 ah bank (according to the Mastervo;t MICC) Smartgauge still shows 100% SoC. Another night with 67 ah out Smartgauge showed 98% SoC. Obviously the SG is reading incorrectly. Since fitting the Leoch batteries the at rest voltage seems slightly higher than it was previously. I'm not on the boat so can't give more precise information. I contacted the UK agents for Leoch asking about 'at rest' voltage and they sent me the attached spec sheet. I can't see anything specifically about  an 'at rest voltage. Has anyone got any idea what battery type I should set the SG for?   

 

LAGM-130.pdf

 

Many thanks

 

Frank

 

Edited July 4, 2018 by Slim

[WotEver]

6 hours ago, Slim said:

Rather than start another Smartgauge topic I have a question that I have about my smartgauge. It is regarding which battery type to set it for.

I’d try Type 3 and see how you get on. Don’t forget to do a factory reset too. 

[Dr Bob]

9 hours ago, dmr said:

You are missing a crucial point here though.

Yes the smartgage is most likely using rate of change of voltage as well as voltage to deduce battery state of charge, I also suspect its curve fitting (oops doing regression analysis) to the the exponential voltage changes in response to current changes, But, the reason it struggles with charging is not because Gibbo did not have enough charging data, its because charging voltage is controlled by the alternator regular Not by any battery characteristics so there is simply no real battery related information available, the instruction book actually says this. So its just a best guess from basic generic information..."at 14.4 volts a 400Ah bank will take 3 hours to charge" sort of thing.

 

..............Dave

I agree with the point, but still would think Gibbondidnt measure enough charging types...because there will be too many. I have five methods on our boat. There will be battery info available as an alternator shoves in the power at 14.4v as the voltage will have a rate of change and the SG will learn (as Nick says.......I knew all that but left it out for simplicity). The SG should be able to work out the full charge more intelligently than just time. The problem will be with my battery charger, my AtoB, my solar charger where one minute it is 14.4V and rising slowly then 13.4V and steady. The voltage will go up and down depending on current draw and that will then mean the SG can work out the unit is in Float. Does the SG then report 100%? My BMV went into float far too soon till I changed the parameters and I always knew I had more charging to do. I can envisage how an SG can respond to all sorts of changes in V, dv/dt etc but not the variation in float voltage especially if two devices are providing that input, i.e. My AtoB and two different solar controllers ( that makes 3 devices). I hope the SG doesn't go to 100% when the charger goes to float otherwise it is at the mercy of the charger and no wonder people have had problems under charging. 

 

Eta....I used the term multi linear regression earlier (MLR). It could of course be using Partial Least Squares (PLS) as the method. Both similar modeling routes.

Edited July 5, 2018 by Dr Bob

[rusty69]

22 minutes ago, Dr Bob said:

I agree with the point, but still would think Gibbondidnt measure enough charging types

Was he related to tree monkey? 

[dmr]

46 minutes ago, Dr Bob said:

I agree with the point, but still would think Gibbondidnt measure enough charging types...because there will be too many. I have five methods on our boat. There will be battery info available as an alternator shoves in the power at 14.4v as the voltage will have a rate of change and the SG will learn (as Nick says.......I knew all that but left it out for simplicity). The SG should be able to work out the full charge more intelligently than just time. The problem will be with my battery charger, my AtoB, my solar charger where one minute it is 14.4V and rising slowly then 13.4V and steady. The voltage will go up and down depending on current draw and that will then mean the SG can work out the unit is in Float. Does the SG then report 100%? My BMV went into float far too soon till I changed the parameters and I always knew I had more charging to do. I can envisage how an SG can respond to all sorts of changes in V, dv/dt etc but not the variation in float voltage especially if two devices are providing that input, i.e. My AtoB and two different solar controllers ( that makes 3 devices). I hope the SG doesn't go to 100% when the charger goes to float otherwise it is at the mercy of the charger and no wonder people have had problems under charging. 

 

Eta....I used the term multi linear regression earlier (MLR). It could of course be using Partial Least Squares (PLS) as the method. Both similar modeling routes.

The number and type of charging device is pretty much irrelevant, the only variable is charging Voltage and I believe that smartgage takes at least some account of this.

Smartgage is a discharge capacity meter based on voltage, it is quick and easy to install because it has no shunt. Its not too good during charge because no battery voltage is available, only charge source voltage.  Now maybe Gibbo didn't want to include a charge measurement at all but thought the gauge would be criticised if he didn't, so he included an approximate one and gets criticised, who knows????. He could have done a bit better but it would never have been more than an approximation. My own view is that allowing the guage to reach 100% too soon has turned out to be quite a problem and he should have made it get slower and slower after about 90% so that it almost never got beyond 99%, but then this would have caused trouble too; a not so good approximation is possibly better than a better but still inaccurate approximation which people would take as gospel.  Actually the smartgage is pretty good at indicating a very good "daily charge" but not good at indicating the "weekly real 100% charge". Maybe this was the plan ? 

 

...............Dave

 

[nicknorman]

3 minutes ago, dmr said:

The number and type of charging device is pretty much irrelevant, the only variable is charging Voltage and I believe that smartgage takes at least some account of this.

Smartgage is a discharge capacity meter based on voltage, it is quick and easy to install because it has no shunt. Its not too good during charge because no battery voltage is available, only charge source voltage.  Now maybe Gibbo didn't want to include a charge measurement at all but thought the gauge would be criticised if he didn't, so he included an approximate one and gets criticised, who knows????. He could have done a bit better but it would never have been more than an approximation. My own view is that allowing the guage to reach 100% too soon has turned out to be quite a problem and he should have made it get slower and slower after about 90% so that it almost never got beyond 99%, but then this would have caused trouble too; a not so good approximation is possibly better than a better but still inaccurate approximation which people would take as gospel.  Actually the smartgage is pretty good at indicating a very good "daily charge" but not good at indicating the "weekly real 100% charge". Maybe this was the plan ? 

 

...............Dave

 

Yes the SoC ramps up linearly or nearly so, which could never be right.

[dmr]

9 minutes ago, nicknorman said:

Yes the SoC ramps up linearly or nearly so, which could never be right.

Its just ahead of its time and "Lithium ready".

I think it ramps up slower if the charge voltage is lower????.

As I said, a simple approximation that is an obvious approximation is probably better than one which looks accurate but is still based on a guess.

....but this is not correct because lots of quite clever have believed this very crude approximation!!!

I'm sure Gibbo had the knowledge and data to do a much better approximation to the charge curve but he chose not to.

If the Smartgage was wrong or no good then I would support the criticism, but it works really well and is an almost essential instrument on a liveaboard boat, so I get a bit frustrated by the people here who keep on criticising it for not doing something it is not intended to do.

So can we have a new forum rule, you can only criticise the Smartgage if you own one ?

 

..............Dave

[MtB]

43 minutes ago, dmr said:

If the Smartgage was wrong or no good then I would support the criticism, but it works really well and is an almost essential instrument on a liveaboard boat, so I get a bit frustrated by the people here who keep on criticising it for not doing something it is not intended to do.

So can we have a new forum rule, you can only criticise the Smartgage if you own one ?

 

 

The criticism is of the marketing not matching the reality. Smartgauge is "A fuel gauge for batteries" apparently. This is what fooled me initially. 

 

Moving on, the documentation does not alert one to the fact that this is not true until you get to Page 28 of the manual, and I fell asleep on Page 27 so densely and boringly is it written. 

 

So if the nature of the instrument had been accurately and honestly promoted, and the BIG DRAWBACK mentioned on page 1 of the manual, along with some advice to check calibration before relying on the marketing hype, I'd have had no criticism to offer. SO I'm not criticisng the instrument but the misleading marketing and documentation surrounding it. 

 

Edited July 5, 2018 by Mike the Boilerman

[dmr]

NEVER let the bloke who designed anything also write the user manual. Also never let the marketing department write the manual.

 

Don't think the manual should mention the calibration issue, because the calibration issue really should never have happened.

 

It would have been much better to have designed the instrument not to need any calibration adjustments but I suspect Gibbo was struggling to keep the cost down (and folk do criticise the smartgage cost) and getting very accurate electronic components can ramp the price up quite a bit.

 

...............Dave

[dmr]

55 minutes ago, Mike the Boilerman said:

 

The criticism is of the marketing not matching the reality. Smartgauge is "A fuel gauge for batteries" apparently. This is what fooled me initially. 

 

 

 

<smartarse comment warning>

 

In the car I use my fuel gauge only to monitor fuel during discharge and to decide when to recharge, I almost never use it during charging ?

 

..............Dave

[cereal tiller]

2 hours ago, rusty69 said:

Was he related to tree monkey? 

Was he Funky?

[alan_fincher]

1 hour ago, Mike the Boilerman said:

 

The criticism is of the marketing not matching the reality. Smartgauge is "A fuel gauge for batteries" apparently. This is what fooled me initially.

 

So how accurate do you reckon the fuel gauges are in cars / vans you own then?

If it is anything like any of our current cars, a SmartGauge doesn't need to be particularly accurate to match it!

[Dr Bob]

12 minutes ago, dmr said:

<smartarse comment warning>

 

In the car I use my fuel gauge only to monitor fuel during discharge and to decide when to recharge, I almost never use it during charging ?

 

..............Dave

The car fuel tank analogy is not particularly good. For a car fuel tank you don't need to have it full. For a battery you do.

[WotEver]

4 minutes ago, Dr Bob said:

The car fuel tank analogy is not particularly good. For a car fuel tank you don't need to have it full. For a battery you do.

I always fill my tank when refuelling. Just like Dave though, I never use my fuel gauge to determine its fullness. I use another method. My fuel gauge isn’t very accurate when the tank approaches full. 

[Dr Bob]

2 hours ago, dmr said:

The number and type of charging device is pretty much irrelevant, the only variable is charging Voltage and I believe that smartgage takes at least some account of this. 

 

Yes, voltage measurement is key. The SG will monitor volts and time and return a SoC. It doesn't matter if it's charging or not. The question though is does the SG give up trying to predict from the algorithms if the system goes into float or does it still try to determine it. If the latter, then you must have sufficient data in the training set for good prediction. Different charge devices will have different characteristics and hence need to be tested. That is what these models are all about.

It is not good to try and extrapolate with this type of model. Interpolation can be ok but not exterpolation.

However if the SG just sees the change to a float voltage and assumes it's full, then that will mean under charging for many. 

[Richard10002]

3 minutes ago, WotEver said:

I always fill my tank when refuelling. Just like Dave though, I never use my fuel gauge to determine its fullness. I use another method. My fuel gauge isn’t very accurate when the tank approaches full. 

But your car fuel gauge doesn't show a percentage, so doesn't suggest a level of accuracy to a resolution of 1%. Generally, you get 3 labels, E 1/2 F. You might also get a bit of red and green, and a few extra lines.

 

Smartgauge comes with a display that, on the face of it, tells you that your battery bank is at a particular percentage SoC. To suggest that this isn't the case, only after 28 pages of small print, seems to be disingenuous.

 

You could argue that people shouldn't be so naive as to think it is accurate at all times to within a percentage point or two... but why shouldn't they?

[WotEver]

3 minutes ago, Richard10002 said:

But your car fuel gauge doesn't show a percentage, so doesn't suggest a level of accuracy to a resolution of 1%.

No, it’s much better than that. It shows me a maximum range of over 400 Miles when full, giving a resolution of <0.25%. 

[Dr Bob]

15 minutes ago, WotEver said:

I always fill my tank when refuelling. Just like Dave though, I never use my fuel gauge to determine its fullness. I use another method. My fuel gauge isn’t very accurate when the tank approaches full. 

Ok, you're right of course and the analogy is good. I was wrong...BUT the difference is that you need your battery full and that is not spelt out in the literature. Anyway, I am not knocking theSG here but providing info on how the modelling works. 

[WotEver]

Once I’ve been driving for 10 or 20 miles my fuel gauge more accurately represents the contents of the tank. 

[alan_fincher]

10 minutes ago, Dr Bob said:

Ok, you're right of course and the analogy is good. I was wrong...BUT the difference is that you need your battery full and that is not spelt out in the literature. Anyway, I am not knocking theSG here but providing info on how the modelling works. 

 

But also actually need to deplete the bank to what SG believes as being less that 75% SOC, before it will start to sensibly recalibrate itself.

If you don't take on board all the limitations, and plan around them,  it is hardly a "fit and forget" device, IMO.

[Jen-in-Wellies]

2 hours ago, dmr said:

 

So can we have a new forum rule, you can only criticise the Smartgage if you own one ?

 

..............Dave

Same goes for the Ec*fan! ?

[alan_fincher]

3 hours ago, dmr said:

So can we have a new forum rule, you can only criticise the Smartgage if you own one ?

 

18 minutes ago, Jen-in-Wellies said:

Same goes for the Ec*fan! ?

 

Please don't introduce this rule for pump-out toilets, (or pram hoods!).

[Iain_S]

3 hours ago, nicknorman said:

Yes the SoC ramps up linearly or nearly so, which could never be right.

If you mean linearly with time, then I have to say neither of mine do. As expected, displayed  S.O.C. rises fairly quickly at the beginning, and then slows right down after 90% or so.

[Jen-in-Wellies]

14 minutes ago, alan_fincher said:

 

 

Please don't introduce this rule for pump-out toilets, (or pram hoods!).

Agreed. You don't need to have owned either to realise what a bad idea they are. ?