Paging Dr.Acula, Paging Dr. Acual! advice needed on the UC3906 please.

B

BrendanP

Senior Member
Hi Dr. I'm using the UC3906 to manage the gell battery Im using in the alarm system I mentioned a few weeks back. Im running few tests before I have the final ( I hope) proto pcb made.

I ran the unit until the 12V gell cell was just about dead flat. I then plugged in the external power supply but the battery charging system based on the UC3906 didn't kick in.

I ripped off the design from the kit based on the UC3906 sold by Altronics. I think it was based on a sil chip article. I remember you said you had used the part in your projects.

Is there any reason the IC wont charge a dead flat battery? Ive looked through the TI data sheet and cant see any indication that it wont charge a dead flat battery.

Once I realised the circuit wouldnt charge the gel cell I connected it up to my bench top power supply and the gell cell took a charge no problem.

Whats your diagnosis Dr.?
 
M

moxhamj

New Member
This is like being back at the hospital:
Paging Dr Jordan Frame, please go to orthopaedics.
Paging Professor Billy Rubin, please return to haematology.

Occasionally when Switchboard were feeling in the mood we would put out these pages. Trouble is, only the doctors understood the joke.

Yes, it won't charge a dead battery. It is buried away in the specification sheet - actually it does charge a flat battery but only at the trickle rate until the battery reaches about 10V. This protects the charger against trying to charge a shorted battery at high current. If the battery is fully flat it would be a very sick battery and the lead plates may well be shorting internally and even if this is only partial then the self discharge rate may be higher than the trickle rate of the UC3906.

Bottom line is that lead acid cells hate being discharged below about 11V and you will only get away with doing this a few times before the battery will die. A low battery cutout circuit can be designed to disconnect the battery as a disconnected battery is better than a dead one.
 
B

BrendanP

Senior Member
I was reading the data sheet again over lunch at work today and saw something about it trickle charging until it got to a certain V level. So now I understand what that means thanks.

Ive got a potential divider /adc set up on the pcb it so its easy to monitor V level. It sends the users a sms when it gets belore a certain level, Ill add a message to the eeprom/lcd to advise users that it will take a long time to charge if full discharged and advise them to avoid doing so.

Any suggestions on a low batter cut out circuit? I'll do asearch but I thought you might be able to recommended one. Ill have a look for IC that does it.

Do you think I should put a some decoupling caps on the +Vin pin? Does the part have any 'smarts' in it or is it just analogue?

If you ever come over to Vic. to visit we should meet. My girlfriend is a speech pathologist at the Monash Medical Center in Clayton. I met Stan?Manuka when he visisted Australia last. Its great to meet another enthusiast.

Most peoples eyes just glaze over when you start talking electronics.
 
M

moxhamj

New Member
Sure, be good to catch up. I go to Melb every few years.

Re the circuit - I bought the kit from Altronics - it worked out cheaper than buying the chip and making a board etc. So not sure about decoupling but can't hurt.

Low cutout - I'm sure there are kits for that as well - people add them to the second battery on their 4WDs for the fridge etc. You don't want a circuit that wastes power when it is on ( a relay) nor one that drops too many volts (transistor) nor one that wastes power when it is low (a relay) though a picaxe running at low power would not be too much of a worry as it would be less than self leakage current. If I were designing one I'd use a latching relay and a picaxe, and run the picaxe off a low quiescent 5V reg. I'm sure there is a mosfet solution but have had a beer and brain not functioning clearly...
 
I

inglewoodpete

Senior Member
Silicon Chip had a battery isolator project 2 or 3 years ago. It used a high side N-channel MOSFET. The circuit had a clever inverter to produce the 15 or 17v wrt ground needed to turn the MOSFET fully on.
 
You must log in or register to reply here.
Top