Duet 3.3V supply quality December 27, 2013 11:31AM |
Registered: 10 years ago Posts: 300 |
Re: Duet 3.3V supply quality December 27, 2013 11:46AM |
Registered: 13 years ago Posts: 1,611 |
Re: Duet 3.3V supply quality December 27, 2013 11:49AM |
Registered: 10 years ago Posts: 135 |
Re: Duet 3.3V supply quality December 27, 2013 12:06PM |
Registered: 10 years ago Posts: 300 |
Re: Duet 3.3V supply quality December 27, 2013 12:51PM |
Registered: 13 years ago Posts: 1,611 |
Re: Duet 3.3V supply quality December 27, 2013 01:14PM |
Registered: 10 years ago Posts: 103 |
Re: Duet 3.3V supply quality December 27, 2013 01:18PM |
Registered: 10 years ago Posts: 14,672 |
Re: Duet 3.3V supply quality December 27, 2013 01:34PM |
Registered: 10 years ago Posts: 135 |
Re: Duet 3.3V supply quality December 27, 2013 01:53PM |
Registered: 10 years ago Posts: 14,672 |
Quote
The output capacitor determines the output voltage ripple and is used to close the control loop. As outlined in the Control Loop section, the bandwidth has been optimized for an output capacitance of 20 μF. If a particular application requires an extremely low output voltage, the output capacitor can be increased. Any increase will tend to reduce the bandwidth and therefore compromise the transient response performance. In general the output capacitance should not exceed 1000 μF or be less than 10 μF, as this may cause a loop instability to occur.
Re: Duet 3.3V supply quality December 27, 2013 02:09PM |
Registered: 10 years ago Posts: 135 |
Re: Duet 3.3V supply quality December 27, 2013 03:16PM |
Registered: 13 years ago Posts: 1,611 |
Re: Duet 3.3V supply quality December 27, 2013 04:04PM |
Registered: 10 years ago Posts: 14,672 |
Quote
Rory166
I like Ian's post above which I only read after fitting the cap. Feed the 5v rail from a 5v output of the psu. Let the poor old 12v take a right bashing from the switching transients. For that matter the psu also produces up to 30A of 3.3V so why not use that ?
Re: Duet 3.3V supply quality December 27, 2013 05:03PM |
Registered: 10 years ago Posts: 300 |
Sure, but the regulator sees a pretty constant load and is "finessed" down to 3.3 by the LDO. I was shooting for around 1000uF with low ESR and 2200uF was the nearest thing I had to hand. I checked the 5V ripple with a scope and it's below 70mV - still, there's certainly more than one way to skin this cat.Quote
dc42
I've been thinking about this a little more:
1. Adding extra capacitance to the output of a switching regulator should be done with care. From the A4403 datasheet:
Quote
The output capacitor determines the output voltage ripple and is used to close the control loop. As outlined in the Control Loop section, the bandwidth has been optimized for an output capacitance of 20 μF. If a particular application requires an extremely low output voltage, the output capacitor can be increased. Any increase will tend to reduce the bandwidth and therefore compromise the transient response performance. In general the output capacitance should not exceed 1000 μF or be less than 10 μF, as this may cause a loop instability to occur.
Yes but of course it can't supply volts that go missing at the input. Both 100uF caps discharge in around 1 to 2mS then it's all over. I have yet to catch it in the act but I suspect the PSU goes offline for a a few cyces now and then. It seems to be a very slow switcher.Quote
dc42
2. The switching frequency of the A4403 is high, so it should be able to react to changes in load very quickly, even more so if capacitor C6 is added.
I get a feeling it's a fair bit more than that. I can see 500mV of ripple @ 1KHz on the 12V. The amount of current drawn is quite a challenge for the interconnects used. I resorted to using the bootlace ferrules as solder cups and pinched them square to fit the clamp terminals. I suspect this may be an area that others might have to check carefully to avoid volt-drops.Quote
dc42
3. There are some heavy loads on the 12V supply, especially the bed heater and the extruder heater (I think I metered one at 3 ohms, so 4A current draw). When these heaters switch on and off, there are likely to be some hefty transients on the 12V rail.
I heartily agree! The blocking diode D1 will prevent the additional capacitance seeing the 12V loads and the higher working voltage will keep the 5V in regulation for longer for a given capacitance, although a 25V cap may be advisable in this position. Unfortunately, when I looked at doing this originally I couldn't see a neat way of adding a big leaded part to beef-up C3.Quote
dc42
So maybe it is the 12V rail that most needs more capacitance, along with the addition of C6 in the regulator circuit as I mentioned in my previous post?
Re: Duet 3.3V supply quality December 27, 2013 05:37PM |
Registered: 10 years ago Posts: 14,672 |
Re: Duet 3.3V supply quality December 28, 2013 06:18AM |
Registered: 10 years ago Posts: 300 |
Re: Duet 3.3V supply quality December 28, 2013 07:57AM |
Registered: 10 years ago Posts: 135 |
Re: Duet 3.3V supply quality December 28, 2013 09:08AM |
Registered: 10 years ago Posts: 314 |
Re: Duet 3.3V supply quality December 28, 2013 09:29AM |
Registered: 10 years ago Posts: 14,672 |
Re: Duet 3.3V supply quality December 28, 2013 03:54PM |
Registered: 10 years ago Posts: 300 |
Re: Duet 3.3V supply quality December 28, 2013 04:30PM |
Registered: 10 years ago Posts: 256 |
Re: Duet 3.3V supply quality December 28, 2013 04:39PM |
Registered: 10 years ago Posts: 300 |
Not Dynamo, but Alternator I think you'll findQuote
kwikius
Dunno if its related and apologise if this is just noise in this thread or simply explainable, but heres a "fun" thing. ( NB there is no power applied to the board anywhere during this experiment) If, with all wiring and everything else (except the atx psu) connected, I wind the z-axis rod fast by hand either up or down, the extruder fan rotates at a good speed. (Hey I'm generating 12volts somehow!) However if I disconnect the z-motor connector at the board and do the same thing I get no rotation whatsoever from the extruder fan.
Bear in mind I havent verified my connections yet. I'll check everything but it seems that The Ormerod is quite a good dynamo?
regards
Andy
Re: Duet 3.3V supply quality December 28, 2013 04:41PM |
Registered: 13 years ago Posts: 1,611 |
Re: Duet 3.3V supply quality December 28, 2013 04:43PM |
Registered: 10 years ago Posts: 135 |
Re: Duet 3.3V supply quality December 28, 2013 04:52PM |
Registered: 10 years ago Posts: 300 |
Quote
droftarts
@Radian: thanks for your continued detective work. It would be good to get someone else to verify this, just in case it's a fault with your board.
Re: Duet 3.3V supply quality December 28, 2013 05:07PM |
Registered: 10 years ago Posts: 14,672 |
Quote
Rory166
I would like to examine the circuit diagram of the duet but am unable to access through github, possibly due to my use of a Mac. Are you able to direct me to a copy? I need this to use my scope onth eboard to confirm your results are universal.
Re: Duet 3.3V supply quality December 28, 2013 05:17PM |
Registered: 12 years ago Posts: 1,236 |
Re: Duet 3.3V supply quality December 28, 2013 05:27PM |
Registered: 10 years ago Posts: 300 |
Quote
bobc
With those sort of transients, I am surprised it works as long as it does... that said I better have a look at my own design!
Re: Duet 3.3V supply quality December 28, 2013 05:40PM |
Registered: 10 years ago Posts: 14,672 |
Re: Duet 3.3V supply quality December 28, 2013 05:57PM |
Registered: 12 years ago Posts: 1,236 |
Quote
dc42
Radian, I attached a scope to my Duet board to see if I can replicate your observations. This is what I found on the mosfet drain:
Re: Duet 3.3V supply quality December 28, 2013 06:04PM |
Registered: 10 years ago Posts: 14,672 |