Nut upsuser - Jan 2007 - APC Smartups protocol - Hidden Programming Mode

Hej!

Hope nut-upsuser is an appropriate place to post additional info about
the smartups protocol, please tell me if not.

I got an APC Smartups 2200 (5U Rack Mount, European/230V version, 48V
batteries) a couple of days ago, was very happy to find the protocol
description at http://www.networkupstools.org/protocols/apcsmart.html,
and started to play. At the point when everything was working fine, I
pressed a couple of random keys with my minicom open and managed to
completely fubar the UPS...

In the end, I discovered a couple of new commands and got the UPS
working again. I'd like to share my findings, so here they are.


Commands (char \n meaning \n typical result \n other info)
=========================================================
1  
Enter programming mode
PROG
Send twice with 3s delay
** warning ** Don't try this at home - your might damage your UPS when
changing some parameters **
The output of the ^Z command does not change. You can leave this mode
with the turn dumb ("R") command.


^Y
Clear Everything
Every possible alarm
** Don't do it. ** (Programming mode only)
Resets everything: Voltmeter calibration, model information,
region code, manufacturing date, serial no... Chances are that your UPS
will become unusable.
True stroy: After typing ^Y, my UPS was unhappy with the line frequency
(50Hz instead of 60Hz), the voltage of both line (0V instead of 110V)
and battery. Consequence: it switched itself off - and refused to be
switched
on again. No way to talk over the serial line, no way to fix (I didn't
know how to do so anyway) the config. The solution was to hold the
On-Button pressed - I jammed a wooden stick between the button and a
battery after recharging the batteries with a power supply.


???
0000, 3C00, 3C67, ...
???


Variables writeable in programming mode
======================================
0,3,4,5: Didn't notice any effect

B: Calibrate the battery voltmeter. 0 is maximum gain, correct setting
for my box is FA.

L: Calibrate the line voltmeter. If set to 0, the ups won't detect any
input power. Correct setting for me is EA.

P: Calibrate the power load measurement.

b: Change model and country/region code, e.g.:
Value  Revision  Model string (^A)
00     40.9.D    SMART-UPS 450
06     50.9.A    SMART-UPS 700
13     80.9.I    SMART-UPS 2200
B7     91.9.I    SMART-UPS 3000 XL
D1     82.9.M    SMART-UPS 2200 RM
Note: The voltmeter in the ups depend on the correct setting of the
region code. After the "Big Reset" (^Y), my European 230V SmartUPS
2200
thought it were an US Smartups 450; suddenly the battery voltage read
27V instead of 54V, the expected input voltage became 110V etc... I
suspect that an overloaded converter might go undetected if the model
version is incorrectly set (haven't verified that theory, for obvious
reasons).

m: Manufacturing date, actually arbitrary string (8 char)

n: Serial number, actually arbitrary string (12 char)

o: on battery output voltage. Possible values: 220, 225, 230, 240V on my
model - actually works.



Using car batteries
==================
Btw, I'm using car batteries in my UPS. Since I couldn't find any
experience report about that in google, here's mine:

The SMART-UPS 2200 (European version, 110V models may be different)
operates at 48V. So I connected four el cheapo batteries (36Ah) in
series. They served almost five hours with 17% load on my SmartUPS 2200.
That sounds good, but I have no idea of the lifetime yet.

I'm a bit concerned about the different capacity of the batteries: Since
the batteries are connected in series, the weakest one will be
discharged the most, which in turn hurts its capacity. I'm thinking
about connecting a resistor (e.g. 220 Ohm, 1W) in parallel to each
battery, so that they at least get the same tension when being charged.

One thing to keep in mind is that car batteries produce emit hydrogen
gas,  which may explode violently. That's dangerous enough by itself,
but when a couple of batteries full with sulfuric acid explode together,
that's real fun. Car batteries need fresh (circulating) air - and keep
fire and sparks away.

I agree with those who argue that car batteries are built for delivery
of high currents during a short period of time, while they are not very
good at delivering moderate currents for hours. But they are four or
more times cheaper than gel batteries, so I don't care if their lifetime
is a bit shorter.



Greetings from Vienna,

  Alexander

> Hope nut-upsuser is an appropriate place to post additional info about
> the smartups protocol, please tell me if not.
We would prefer nut-devel I think, since that's more about the development
of drivers and such.

[...]
> Using car batteries
> ==================
[...]
> I'm a bit concerned about the different capacity of the batteries:
Since
> the batteries are connected in series, the weakest one will be
> discharged the most, which in turn hurts its capacity. I'm thinking
> about connecting a resistor (e.g. 220 Ohm, 1W) in parallel to each
> battery, so that they at least get the same tension when being charged.
What do you think a car battery actually is? Six individual 2V cells
connected in series. What you're essentially doing with connecting four
car batteries in series, is just adding more cells in the battery string.
As long as you use the same make/type/age of batteries, there is no need
for equalizing with a resistor, the capacity will be practically the same.

Besides wasting energy, you might prevent the charging circuit from
detecting full battery state (since the charge current will never go to
virtually zero). Only in *very* large installations it is worthwhile (and
cost effective) to charge and monitor each cell separately to make the
most out of the lifetime of them. But you'd be looking at the scale of
batteries used in submarines, rather than at a home/small office UPS.

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