Nut upsdev - May 2007 - Re: Changes to megatec.c

Hi, Charles,
On Wed, 2 May 2007 07:58:08 -0400, you wrote:
> > Will that be enough to make it useable as a contribution? I believe
> > the more sensible values for 36V UPS'es will be useful to other
models
> > than the one I have, too.
> 
> Sounds good, but I don't think I saw it hit the list yet...
I now have the following in megatec.c:

static BatteryVolts batteries[] = {
	{ 12,  9.0, 16.0,  9.7, 13.7 },   	/* Mustek PowerMust
600VA Plus */
	{ 12, 18.0, 30.0, 18.8, 26.8 },   /* PowerWalker
Line-Interactive VI 1000 (LB at 22.3V) */
	{ 24, 18.0, 30.0, 19.4, 27.4 },   /* Mustek PowerMust 1000VA
Plus (LB at 22.2V) */
	{ 36,  1.5,  3.0, 1.64, 2.31 },   /* Mustek PowerMust 1000VA
On-Line (LB at 1.88V) */
	{ 36, 32.0, 42.0, 33.0, 40.8 },   /* Mecer / Mustek 2kVA
ME-2000 - Added FvW 1 May 2007 */
	{ 96,  1.5,  3.0, 1.63, 2.29 },   /* Ablerex MS3000RT (LB at
1.8V, 25% charge) */
	{  0,  0.0,  0.0,  0.0,  0.0 }};  /* END OF DATA */

You will notice that I made the following guesses:

- A 36V UPS should have a battery clamping voltage of between 32 and
42V. Assuming the possibility of a 'low' reading being reported by the
UPS, it should at least be 32V or more. The float charge voltage of
13.62-13.64V as recommended by sealed lead acid battery manufacturers
should keep it under 42V. Overcharged or worn batteries might have a
highere clamping voltage, but more than 42V will move it into the
regions of 48V power supplies.

- Battery manufacturers recommend a minimum of 10.8V for a 12V battery
as the cutoff discharge voltage (further discharge will cause battery
damage) but the UPS is likely to shut off before that, so I
guesstimated 11V for a 12V cell to be a reasonable "empty battery"
voltage. 

- The float charge (clamping) voltage for a fully charged 36V battery
if 40.9V but I've subtracted 100mV to allow for tolerances



Kind regards,
    Frank van Wensveen

---
 * Origin: www.vanwensveen.nl (80.253.112.29)

> You will notice that I made the following guesses:
[...]
> - Battery manufacturers recommend a minimum of 10.8V for a 12V battery
> as the cutoff discharge voltage (further discharge will cause battery
> damage) but the UPS is likely to shut off before that, so I
> guesstimated 11V for a 12V cell to be a reasonable "empty
battery"
> voltage.
The cutoff charge depends heavily on the application of a VRLA battery. If
you're using a 1/20 C (or 20 hour) discharge rate, 10.8V may be a
reasonable cutoff voltage. However, in most UPS systems, discharge
currents will be well in excess of 2 C. It is not unreasonable to use a
lower cutoff voltage in that case, to account for internal losses in the
battery and/or wiring losses. Common cutoff voltages for a 12V battery in
UPS systems will be as low as 9.7 or 9.8V. You'll get practically half the
specified runtime with a cutoff voltage as high as 11V.

Best regards, Arjen
-- 
Eindhoven - The Netherlands
Key fingerprint - 66 4E 03 2C 9D B5 CB 9B  7A FE 7E C1 EE 88 BC 57

Hi, Carlos,
On Fri, 4 May 2007 09:11:21 +0100, you wrote:
> >  { 36,  1.5,  3.0, 1.64, 2.31 },
> >
> > So that would mean that a 36V UPS will report a current voltage of
> > between 1.5 and 3.0V?
> 
> That's right.
To be honest, I've never seen a 36V UPS do that. All UPSes I've ever
used report their actual battery voltage, not the voltage per cell. As
a detection method the latter wouldn't make any sense, either, because
you cannot use the per-cell voltage to determine how many cells there
are in a battery...

[Using config file for battery voltages]
> That's exactly what you can do with the "battvolts" option
supported
> by the latest version of the driver.
But that's a percentage, right? It still leaves you with the problem
of NUT not being able to detect a 36V UPS. Or have I completely
misunderstood you?


Kind regards,
    Frank van Wensveen

---
 * Origin: www.vanwensveen.nl (80.253.112.29)

On 5/4/07, Arjen de Korte <nut+devel@de-korte.org>
wrote:
> I have to disagree here. You won't be the only one with a PowerWalker
600,
> so it's just a matter of time that we can expect the reports about
> erroneous battery charge levels flooding the support mailing list. And we
> mechanically repeat the same solution over and over again, that people
> need to change some settings in 'ups.conf' to correct the values.
The same can be said for not having default values at all. People
would either accept the fact that they won't be getting any indication
of how charged is their UPS (for most people, the battery voltage that
the driver always reports is meaningless), or they would keep on
asking how they can obtain these values that the driver is asking
about.

In either case, the solution is to point them to the "megatec(8)"
manpage, which (for the trunk version) reads:

"The battery charge (%) value is only a best-effort calculation and may be
incorrect, or non-existant, for some UPS models. In any case, you can always
resort to finding the voltages at which the battery becomes fully discharged
and fully charged, and specify them manually using the "battvolts"
option.
If you do this, please notify the author of this driver."

-- 
Carlos Rodrigues

On 5/4/07, Frank van Wensveen <frankvw@myconnection.co.za>
wrote:
> That's what I understood them to be. Therefore the 36V line in the 2.0
> stable distribution doesn't make sense:
>
>  { 36,  1.5,  3.0, 1.64, 2.31 },
>
> So that would mean that a 36V UPS will report a current voltage of
> between 1.5 and 3.0V?
That's right.
> That's why I would consider it a very good idea to have these values
> in a config file rather than hardcoded in the driver's source code.
> That way you could override the hardcoded defaults without breaking
> your legacy support for other users, _and_ users could change voltage
> thresholds to suit your particular needs without having to edit C
> source code and do a recompile (things that most users won't be too
> comfortable with).
That's exactly what you can do with the "battvolts" option
supported
by the latest version of the driver.

-- 
Carlos Rodrigues