I am constructing a Ni-Cd battery charger using an LM2576-adj regulator and LMC660 quad Op-Amp based on the circuitry described in Texas Instruments application report SNOA327B http://www.ti.com/general/docs/litabsmultiplefilelist.tsp?literatureNumber=snoa327b This is their circuit diagram:
My requirements were for a charging current of 2.16amps on a 14.4volt 1800mAh battery which would need 18 or 19 volts to achieve this so I ran the regulator from a 24volt dc supply.
I started by building the power supply section first and in place of the battery I used a piece of spiral electric fire element that measured 8.6 Ohms. At that stage I did not use the Op-Amp to provide feedback but used a voltage divider from top to bottom rails consisting of 10k resistor, a 10k trimmer pot and a 1k2 resistor at the bottom. I could easily control the current to the required 2.16 amps and that gave a voltage drop of 124mV across the 0.05 Ohm resistor R12.
Things were going fine and I had a sizeable 3mm Al sheet heatsink about 40mm square on the regulator but I was not happy with the temperature at which it was running so I removed it with the intention of making a larger heat sink and in the process I dropped the regulator on the floor. When I reassembled everything the charging current was very high and I could no longer control it. I suspected E.S.D. and took stringent precautions to prevent this happening again to a replacement LM2576.
The new regulator ran very cool at 2.16 amps and was switched to ON when pin 5 was grounded but would not shut OFF completely when pin 5 was taken high, leaving a residual current of around 200mA if I remember correctly. This was problem No.1. Why could I not shut down the regulator completely?
Hoping to find an answer to the shutdown problem later on, I carried on using a second board for both the overcharge protection/current regulating circuitry. This is the first time I've ever used an Op-Amp so I read up on them first and came to the conclusion that U5B was configured to be non inverting with a gain of 1+ (75/10) = 8.5. To maintain the required 1.23V feedback I needed my 124mV to be amplified by a factor of 9.92 so I chose R9 to be 10K which meant R7 would have to be 89.2K. That made it rather handy as I could use an 82K resistor for R7 with the possibility of trimming it later if required.
Having got that stage of the circuit completed I found that I was not getting any control of the charging current. Indeed, the Op-Amp output was only 2.6mV and the potential between the non-inverting and inverting pins was 0.8mV.
I had been using channel A on the Op-amp and not channel B as in the Texas circuit so I changed to channel B and found no difference. I swapped the LMC 660 for another spare that I had bought and still no difference. Then I swapped the LMC660 for a TLV274 (which is configured the same) but still no difference. I'm sure that I'm doing something wrong but check and check as I may I cannot find what. Does anyone have any ideas?
To put you as completely in the picture as I can I have to say that I did not provide 5V from a 78L05 but instead I used a an LM317T and adjusted it for a 6V output. Also when swapping the LMC660 for a TLV274 I went back to the breadboard stage just in case I had made any wiring errors on the little punched matrix board on which was I had built this feedback controller.
My requirements were for a charging current of 2.16amps on a 14.4volt 1800mAh battery which would need 18 or 19 volts to achieve this so I ran the regulator from a 24volt dc supply.
I started by building the power supply section first and in place of the battery I used a piece of spiral electric fire element that measured 8.6 Ohms. At that stage I did not use the Op-Amp to provide feedback but used a voltage divider from top to bottom rails consisting of 10k resistor, a 10k trimmer pot and a 1k2 resistor at the bottom. I could easily control the current to the required 2.16 amps and that gave a voltage drop of 124mV across the 0.05 Ohm resistor R12.
Things were going fine and I had a sizeable 3mm Al sheet heatsink about 40mm square on the regulator but I was not happy with the temperature at which it was running so I removed it with the intention of making a larger heat sink and in the process I dropped the regulator on the floor. When I reassembled everything the charging current was very high and I could no longer control it. I suspected E.S.D. and took stringent precautions to prevent this happening again to a replacement LM2576.
The new regulator ran very cool at 2.16 amps and was switched to ON when pin 5 was grounded but would not shut OFF completely when pin 5 was taken high, leaving a residual current of around 200mA if I remember correctly. This was problem No.1. Why could I not shut down the regulator completely?
Hoping to find an answer to the shutdown problem later on, I carried on using a second board for both the overcharge protection/current regulating circuitry. This is the first time I've ever used an Op-Amp so I read up on them first and came to the conclusion that U5B was configured to be non inverting with a gain of 1+ (75/10) = 8.5. To maintain the required 1.23V feedback I needed my 124mV to be amplified by a factor of 9.92 so I chose R9 to be 10K which meant R7 would have to be 89.2K. That made it rather handy as I could use an 82K resistor for R7 with the possibility of trimming it later if required.
Having got that stage of the circuit completed I found that I was not getting any control of the charging current. Indeed, the Op-Amp output was only 2.6mV and the potential between the non-inverting and inverting pins was 0.8mV.
I had been using channel A on the Op-amp and not channel B as in the Texas circuit so I changed to channel B and found no difference. I swapped the LMC 660 for another spare that I had bought and still no difference. Then I swapped the LMC660 for a TLV274 (which is configured the same) but still no difference. I'm sure that I'm doing something wrong but check and check as I may I cannot find what. Does anyone have any ideas?
To put you as completely in the picture as I can I have to say that I did not provide 5V from a 78L05 but instead I used a an LM317T and adjusted it for a 6V output. Also when swapping the LMC660 for a TLV274 I went back to the breadboard stage just in case I had made any wiring errors on the little punched matrix board on which was I had built this feedback controller.
