[time-nuts] Re: HP Z3801A project update

[ed breya]

The battery control system uses relays to ultimately control the 
connect/disconnect operations. The front panel battery control switch 
selects the external, off, or internal battery mode, mutually exclusive 
(although they don't necessarily have to be).

The internal battery operation will be fully automatic. When AC power is 
on, the relay connects the internal battery and provides the charging. 
If the AC drops out, the relay is held on by the battery power, and the 
blocking diode provides Vs from the battery. During discharge, the 
charge control circuit monitors the battery voltage and the balance 
between the two SLAs. If the total voltage drops to the minimum level 
(about 22 V), or the balance goes out of range (about 0.6 V difference), 
the tripper shuts it all down, totally disconnecting the battery. The 
battery won't reconnect until AC power is restored, starting the process 
again.

Normally, line dropouts will be short in duration, so the battery just 
keeps everything going until normal operation is resumed. With full 
charge, it should back up for several hours. When AC is restored, a 
front panel indicator lights to show that line has been lost at some 
point. If power is off beyond the backup capacity, and the battery trips 
out, everything is off until power is restored. Then, another warning 
light comes on, indicating that the battery has disconnected, and 
everything had to restart from scratch - the Z3801A and the clock, 
mostly. The warning lights stay on until manually reset, to show what 
happened.

In this cold start condition, the Z3801A needs extra power for a while, 
to get the ovens warmed up. At the same time, the battery would be fully 
depleted, needing maximum recharging current available. The circuit 
automatically reduces the current limit available for charging, and 
gradually allows more as the Z3801A gets through its warm up process, 
and demand decreases.

The charger will initially be set up for temperature compensated float 
mode. Unfortunately, it could take up to a week to recharge a fully 
spent battery. It may be possible to add a "fast" mode with higher 
voltage and tempco, but there's only so much current available, so it 
may not be worth it. After I get everything fully working, I'll be able 
to assess this option.

I built the external battery/source control system to go with the 
original DC-DC converter scenario, but it will work essentially the same 
with the new boost converter. The new converter offers some possible 
other options, to be considered later.

It is manually operated, with no provision for charging. It's also 
independent of the AC power. It's main purpose is control and 
protection, while taking as little current as possible, until activated. 
The "hookup" indicator on the back panel, and one on the front, light 
red if the external source polarity is reversed, and no operation is 
allowed. If the polarity is right, the back indicator goes green, then 
shuts off after about 20 sec. Another LED on the front lights to show 
the external source is available. It's a high efficiency type, showing 
nice and bright at less than 1 mA, which is most of the current draw 
until activation. The idea is to keep the quiescent current very low, to 
not run down an external battery - it's typically much less than the 
self discharge rate, so relatively insignificant.

If the source voltage is above 15 V, no operation is allowed. If less, 
then it's OK to go. When the battery control switch enables the external 
mode, the "start" push button can force the relay on, connecting the 
source to the load - the boost converter, in this case. If the source 
voltage holds up or recovers to over 11 V, the circuit will latch on, 
and continue operation, and an indicator shows that it is running. If 
the source drops below 11 V, from discharge or insufficient current, 
then it won't latch, or it will trip out if it was running. It can only 
be activated by the button. If the source voltage rises above 15 V, it 
will trip out, and allow no operation. The whole thing can be turned 
off, of course, by the battery control switch.

Once the button is pushed, and latch-on is a success, the external 
supply is committed, and will be loaded to somewhere around 50-60 mA, 
(mostly for the relay coil) plus the load current. If AC is on, the 
boost converter will be idling, around maybe 10 mA, and ready for 
action. If AC drops out, the Vs line falls until it reaches the setpoint 
of the converter. For now, I have it set for a little over 24 V. The 
feedback for the converter is actually Vs, DC-wise, so when Vs is lifted 
above the setpoint, the converter goes to zero duty cycle, essentially 
off except for idling current. As Vs falls, the converter smoothly takes 
over and holds it.

If the AC is off when the button is pushed, this is the worst case 
loading, and possibly a cold start. It takes a stout source to get 
things moving. The LT1070 activates at around 2.5 V, and tries like hell 
to charge everything up and make 24 V from it. The first thing needed 
was under-voltage lockout, and then soft start. The UVLO is set for 
about 10 V, after which the soft start is allowed up, maybe about 50 
mSec duration.

The moment the relay connects, the source has to charge about 20 mF 
total Vs bus capacitance (not to worry, there's plenty of ESR and wiring 
R to soften it up) to the UVLO level, where the soft start is enabled, 
letting the converter run it up the rest of the way.

It gets more complicated when running the actual loads. There are three 
(four if you count the boost converter itself) DC-DC converters in the 
system - two in the Z3801A (the main load), and the one I added for the 
12 V supply (fairly small load). Each has its own UVLO point and 
negative resistance characteristics that will come into play. I found 
the data sheets for them, and will have some more to say about this 
after some study.

That's all for now. More later.

Ed