PD or Iota for my upgrade?
I'm thinking about a converter and battery upgrade. The batteries will be wet cell lead acid 6 volt Possible T-105s or the 230AH, EGC2 from Sam's club. I'll probably go for four. (Does nayone have...
Forum Discussion
I spent some more time studying the PD circuit. It doesn't show the internals of the UC3846, so I studied it, too. If I'd done this more thoroughly, I'm pretty sure I would have saved some time, as I now think I know what's going on. I made certain assumptions about how the UC3846 was implemented that were not warranted. I'm not sure why it was done the way it was, but ....
I currently think this is correct.
I currently think this is wrong and BFL is correct.
This is all wrong, too. (I think.)
Also wrong. :(
Over the weekend, I tweaked the output voltage to 14.8 by adding a 10 turn trimpot and a safety resistor in parallel with the boost resistor. It performed somewhat better, but the current still drooped.
Studying the circuits (I posted them here in an earlier post and they are in the patents) I now think I know why. (Tech warning here - read at your own risk :))
The UC3846 is voltage controlled current source. It has a voltage input (op amp with inverting and non inverting inputs) that produces an error signal - error voltage. It also has a current sense input that tells it the current flowing from the output. The error signal from the voltage error amp is compared to a voltage signal produced by the current sense. The difference controls the duty cycle. The duty cycle controls the output current and that controls the output voltage.
The output of the voltage error amp is connected to an adjustable voltage limiter. It can't exceed that limit regardless of how much the error signal is. That limits the output current to a selected maximum.
The PD design has a voltage feedback circuit that produces its own error signal and feeds that to the non-inverting input of the error voltage op amp. The error voltage op amp is connected as a voltage follower of unity gain. (inverting input connected to output.)
I think that unity gain is the problem. Or, to put it another way, the lack of gain in the PD's error voltage circuit is the problem. When the output voltage has risen to near the set voltage (14.4 for boost, measured by me to be 14.5 stock, and modified by me to be 14.8 for my batteries) the error signal is reduced. The closer the output voltage gets to the output voltage setting, the lower the error signal. The lower the error signal, the lower the duty cycle and the lower the output current.
When the battery voltage is low, the error signal between the measured voltage and the voltage setting the PD is trying to reach will be high enough to equal or exceed the current limit voltage and you'll get full output. However, if the battery voltage is higher, the difference between the two will be less than the current limit setting and we'll see less current. That's why it tapers the current - the error voltage is too low as the output voltage rises close to the target voltage. The error voltage needs to be at or above the voltage set by the current limit circuit.
There's a simple fix. :) Change the gain of the error voltage amp. Cutting one lead and adding two resistors will change it from unity gain to any desired gain. One resistor Rf from output to inverting input and one Rg from the inverting input to ground produces a gain of 1+Rf/Rg.
Increasing the gain of the error signal will produce a longer period of current limited action followed by tapering, but the tapering will occur at a point much closer to the natural point where the battery is limiting the current.
Time to open the PD case again I suppose. I've already added an output voltage adjust resistor, I might as well make the gain adjustable. The DW is not going to smile if it goes up in smoke.
More later. This is all just speculation at this point and I'm open to any suggestions of other circuit design changes (I might modify the PD error circuit gain instead of the gain in the control chip, or perhaps adjust the current sense, which is an adjustment that I think PD uses.)
If the PD were working "properly" it would stay at the same amps for the bulk stage until the battery rising voltage makes the amps start to taper.
I currently think this is correct.
DryCamper11 wrote:
I think we're seeing it in voltage limited conditions, not current limited due to the resistance of my cable, the battery resistance and the low limited Boost voltage,.
I currently think this is wrong and BFL is correct.
I think the PD is doing exactly what other converters would do with my cables/batteries if they had the same boost voltage and the same 80 amp current limit. If they had a lower current limit - say 60 amps - then I suspect they would perform worse and would be 60A limited for the first 3 hours instead of starting at 80 and dropping to 60 over that period. If they had the same 80A current limit as the PD9280 plus had a higher Boost voltage than the PD, then I'd expect better performance - current limited at 80A for even longer.
This is all wrong, too. (I think.)
I really don't think the decaying current has anything to do with the PD design beyond its boost voltage limit.
Also wrong. :(
Over the weekend, I tweaked the output voltage to 14.8 by adding a 10 turn trimpot and a safety resistor in parallel with the boost resistor. It performed somewhat better, but the current still drooped.
Studying the circuits (I posted them here in an earlier post and they are in the patents) I now think I know why. (Tech warning here - read at your own risk :))
The UC3846 is voltage controlled current source. It has a voltage input (op amp with inverting and non inverting inputs) that produces an error signal - error voltage. It also has a current sense input that tells it the current flowing from the output. The error signal from the voltage error amp is compared to a voltage signal produced by the current sense. The difference controls the duty cycle. The duty cycle controls the output current and that controls the output voltage.
The output of the voltage error amp is connected to an adjustable voltage limiter. It can't exceed that limit regardless of how much the error signal is. That limits the output current to a selected maximum.
The PD design has a voltage feedback circuit that produces its own error signal and feeds that to the non-inverting input of the error voltage op amp. The error voltage op amp is connected as a voltage follower of unity gain. (inverting input connected to output.)
I think that unity gain is the problem. Or, to put it another way, the lack of gain in the PD's error voltage circuit is the problem. When the output voltage has risen to near the set voltage (14.4 for boost, measured by me to be 14.5 stock, and modified by me to be 14.8 for my batteries) the error signal is reduced. The closer the output voltage gets to the output voltage setting, the lower the error signal. The lower the error signal, the lower the duty cycle and the lower the output current.
When the battery voltage is low, the error signal between the measured voltage and the voltage setting the PD is trying to reach will be high enough to equal or exceed the current limit voltage and you'll get full output. However, if the battery voltage is higher, the difference between the two will be less than the current limit setting and we'll see less current. That's why it tapers the current - the error voltage is too low as the output voltage rises close to the target voltage. The error voltage needs to be at or above the voltage set by the current limit circuit.
There's a simple fix. :) Change the gain of the error voltage amp. Cutting one lead and adding two resistors will change it from unity gain to any desired gain. One resistor Rf from output to inverting input and one Rg from the inverting input to ground produces a gain of 1+Rf/Rg.
Increasing the gain of the error signal will produce a longer period of current limited action followed by tapering, but the tapering will occur at a point much closer to the natural point where the battery is limiting the current.
Time to open the PD case again I suppose. I've already added an output voltage adjust resistor, I might as well make the gain adjustable. The DW is not going to smile if it goes up in smoke.
More later. This is all just speculation at this point and I'm open to any suggestions of other circuit design changes (I might modify the PD error circuit gain instead of the gain in the control chip, or perhaps adjust the current sense, which is an adjustment that I think PD uses.)
