I guess this would be primarily addressed to Gus or support. I would like a definitive realistic current handling capacity for the analog output. The datasheet for the chip says 100ma per pin. The only other reference I saw was that the AO3 is weaker than the digital pins.
The second piece of test equipment I am building with the Panda requires a slow analog ramp to drive DC motors from 0 to 2.5 volts. I know this is easily accomplished with pwm and an h-bridge, but that will not work in this app.
I was using the AO from my DAQ’s to drive an emitter follower power amp but when I tried that with the Panda II it worked for several hours but the chip was running warm and about 2.5 hours in the AO burned out. Fortunately the rest of the AI lines still work. Unfortunately I did not have a meter on it to see what it was drawing.
I am going to set it up tomorrow with an opto coupled darlington which will work for this app even though they are very slow. I would like a realistic number for the output if I have an app that needs faster response than a darlington. Or would I be better off to add a dedicated DAC an run it from the digital lines?
aout is not for driving motors, PWM is. It is current is not 100ma, much less…in micro amps. See datasheet for exact numbers but overall, this is a “signal” so you should never rely in any significant current
I am not using the aout to drive the motors. I am using the signal to drive an emitter follower preamp to dive a power amp to drive the motors. Just like I do with the ttl aout from my DAQ’s.
You cannot drive a motor from the pwm lines directly either. I cannot use pwm in this amp because it is too noisy and does not represent the power flow from the circuit it is supposed to emulate.
The data sheet for the LPC2388 clearly states 100ma per pin with no differentiation for the aout. not microamps.
I would just like a clear indication of what it’s real limits are.
Just a suggestion - How about a FET linearly biased preamp?
That way you’ll have near to none current, and if you bias it well you’ll get a linear response representative of the input voltage… And with a couple more components you could get feedback onto that and get it perfect, eliminating any need for drawing high currents from the panda.
Thanks for the input. I expect that is where I am going to end up. The optocoupled darlingtons draw a minimum current but without knowing what the limitations of the device are it is very difficult to design around them. I really don’t want to sacrifice any more Panda’s to figure it what the limit is. I would have thought that number would be available.
I have 4 different models of DAQ’s here and they had no trouble driving the existing circuit from the aout but the Panda died trying.
It is essential that the motors be driven from as clean a dc voltage as I can feed them. It is also essential that the power share a common ground with the Panda so I can use it to make the necessary measurements of the life test. If I use a pwm driven h-bridge or mosfet curcuit and filter it I lose the common ground.
I may also use a 16 bit dedicated DAC as I would like better than the Panda aout.
Certainly not 100mA. Page 43 of the datasheet (Figures 10, 11) show the output voltage vs. output current for port pins. Above 10-12 mA you’ll be outside the ranges for reliable CMOS usage.
Thank you. I missed that graph although we must have a different version of the datasheet as the page numbers are off.
I was going by the limiting values section of the document.
It looks like anything beyond 6ma is going to pull down the supply V.
It is much easier to design something without dealing with a ‘black box’ in the equation!
Found another reference specific to the onboard DAC. Minimum load resistance 1K. So they are wanting less than 3.3 ma flow from the circuit.
Sorry - I should have posted the version I’m using: Rev. 4 — 10 February 2011
I’ll check if there’s something newer.
<edit: seems to be the newest http://www.nxp.com/documents/data_sheet/LPC2387.pdf >
Thanks again. That is newer than the one I had originally and is where I found the DAC 1k load minimum numbers.