Current system is able to achieve 2% of full scale value instead of set value. For eg., in 360V, 100A, for 100A (Full scale) 2% is 2A regulation could be achieved. But if I set 10A (set value) to flow in same 100A charger, current regulation should be 0.01 (1% of set value) to be achieved. How to do this?
Presently, system is based on SCR topology.
If any supporting docs available, please attach.
1 % of set value is impossible for a random set value: imagine set value = 0 !
Define the lowest set value for which the condition shall apply, convert it into % FSR and check, whether this specification is reasonable. If so, you already have the numbers (% FSR) that have to be established.
Hope this clarifies the issue
With SCR technology you would need to change ranges when moving between high and low. You would have to increase filter inductor size and change the sensing shunt. An alternate would be to use IGBTs and PWM as a supplementary low current low power charger/discharger.
Agree with James: with SCR you can only talk about mean current ( for precisione,ripple is a big issue), so you have to increase output filter if you want to keep you power structure) or go to igbt/mosfets with pwm controller.
Anyway, If you have a good sensor, and you are used to deal with analog signals, it is not a problem to regulate less than half an amp with a 100Amps FS, but it would be a problem to regulate 10mA: but t this point I hardly ununderstand why...!
Thanks people. @Dreher Companies like Bitrode, Jinfan etc., gives 0.1%, 0.5% of FS value respectively in datasheet but in practical, their chargers gives even
@james and Guido, do you have literature where inductive filter calculation is mentioned for this application? FYI, capacitive filter is avoided here considering battery as a capacitor. Presently, shunt is having accuracy of 0.5%, what do you suggest?
Higher precision? Via either multiplexing some components or using higher precision components (eg. ADCs with >= 16 Bits resolution, resistors with .1 % tolerance etc.)
Alternatively it is possible to 'calibrate' such a unit - meaning you measure the tolerance(s) and take them into account in the control loop.
Regards
Shunt used has 0.5% tolerance. Is it possible to still calibrate (Adding all tolerances from shunt to micro controller) to reach less than 0.1% tolerance?
The accuracy of the shunt is not much of a problem it can be calibrated in software provided you have instruments with traceable accuracy better than 0.1% The issue is dynamic range and zero offset, and temperature coefficient. The shunt will heat up at high currents, at 100A, a 100mV shunt will dissipate 10W. 0.1% at 10A would be 100mV /10,000 = 10uV.. This requires an amplifier with a stable zero offset less than this. A 16 bit adc would also be required to achieve this resolution.
Is this charger discharger required to operate accurately over this full range when working on a single batch of batteries, or is it used at high currents on large capacity batteries and low currents on smaller batteries.
Tolerance values are production tolerances. Thus - if you have an instrument capable to measure the individual resistance (not worse than 0.01 % precision) - you can adjust your software to use the exact value.
But this is only part of the story: as James Brown already outlined, temperature coefficient (not only of the shunt), dynamic range, zero offset - as well as aging effects etc. matter. Regarding temperature effects, it might be necessary to measure the circuit temperature to achieve the precision targeted.
Overall: such precision can be achieved. But this is quite a challenge.
Regards
Thanks James. This would be used for QA testing wherein high and low capacity batteries would be tested for wide ranges of currents.
Adding to the question, I'm in need of design calculation for the same system.
There is a choke/Inductor after scr bridge before connecting to battery. Can you tell how to calculate the value of it? If you have any article, pls share.
Regards,
jaikumar
- Badges
- Science topic
- Similar topics
- Engineering
- Electrical Engineering