OK. Got it.
Lets assume would take two caps (5.4V altogether), one balancing board.
Do I charge them altogether with 5.4V through this balancing board? or separately with 2.7V?
What charging current should I use?
How do I know when it is charged? Is there any fix time for charging or CC/CV board stops charging automatically?
Yes, those are AGD DA-M1 modules in balanced push/pull config, 8 pieces.
Your dining table indeed looks interesting, I like the coffee and muffins
Nice project!
You charge them in total, 5.4V max, preferably a little lower to extend lifespan.
Charge current depends on the balancing boards and the used ultracaps.
In most cases you can charge them with more amps than your DC-DC CC/CV converter can provide….
CC/CV converters stop charging when reached the preset voltage.
I use these:
Max 5A Adjustable CC CV Step Down Receiver Charge Module Digital Voltmeter Ammeter Display LED Driver for Arduino Non isolated in Max 5A Adjustable CC CV Step Down Receiver Charge Module Digital Voltmeter Ammeter Display LED Driver for Arduino Non-isolated van Geintegreerde Schakelingen op AliExpress.com | Alibaba Groep
These can charge with 5A max, but, setting voltage below 2V is not stable.
It would be nice to have a charger that changes from cc to cv when the first balancing circuit activates. This can be done via some optocouplers. One optocoupler connected in parallel with each of the balacing resistors of each cap.
I have done something similar with lipos that are known to get slowly damaged in the cv part of the charge. So even if I charge a little bit less of the max capacity I get a longer life for the cells.
I have done something similar with lipos that are known to get slowly damaged in the cv part of the charge. So even if I charge a little bit less of the max capacity I get a longer life for the cells.
Ultracaps can be charged with their continuous decharge amperage.
My Bussmann Cooper ones, 100F, can be charged continuously with 11 Amps,
and peak charged 61 Amps. No need to throttle here...
The CC/CV converter noted earlier does charge CC to CV.
Yes you are right but if you have an imbalance between the voltage on the caps the balancing circuit won’t manage to discharge 11 amps to keep steady the voltage of the cap and you will overcharge it.
With very good caps/cells this is very less to happen but count in that with aging the impedance increases unequal for each cap/cell and sooner or latter the condition I mention above could happen. Another thing to mention is that even the best cells/caps on the market don’t have perfect matching impedances.
So when used at high currents(charge or discharge) it’s always best to have a bms that manages the overcharge/overdischarge conditions.
Ian,
My ultracaps project you saw (powering ALL dac-voltage rails separately, and ALL other voltage rails for fifo II/Isolator/dualXO II/ etc. as well) is also based on COOPER BUSSMANN HV1635 but the 100F version. They are not the best ultracaps, the Maxwells are much better, so can you imagine how yours perform
Btw, please do NOT go the LPS-1 route, since they still use regulation which ruins the perfect ultracaps specs!!!!!
My wish for you to design
Design a circuitry, microprocessor driven, with the possibility to set a LOW threshold Voltage to switch between two banks of ultracapacitors. Switching devices based on VERY LOW RDS-ON MOSFETS, capable of switching several Amperes for fast charging.
It also would be nice if design also could handle voltages up to 60V for powering power amps, AND ESPECIALLY for DRIVER AMP circuitry!!!
This way we also could use this design to power the DRIVER AMP section in our power amps, which will make a HUGE difference in SQ!!!
And last, but definitely not least, this could be the holy grail to power output stages.....
About your concern overpowering the ultracaps while switching, when microprocessor controlled, you could build-in a very short delay time, which will be compensated by a good buffer elco at it's output, that is what I use to prevent hick-ups in power delivery while switching the banks.
I am VERY sure you would have a HUGE hit on the DIY audio market, and MANY people would be interested in buying such a design, including myself!!
Before I forget, since Class-D designs are "the new thing", it is very good possible that in the near future there will be new Class-D designs which CAN surpass Class-A Mosfet amplifiers SQ. ALL Class-D amplifiers do prosper from GOOD power source. So, can you imagine how a GOOD Class-D amp would sound if powered by Ultracaps?
About Mosfets for switching, ST for example has a 60V 120A 0.0024 RDS-ON mosfet which would be a very good choice!
http://www.st.com/content/ccc/resou...df/jcr:content/translations/en.CD00272622.pdf
Don't forget using them for positive AND ground circuitry!
Btw, Powerex Incorporated has switching mosfet modules with 177 micro ohm rds-on,,,,, so maybe there are other contenders on the market too….
Keep up the good work Ian
Best regards,
Alex
Hi Alex
For the low RDS-ON MOSFET, the biggest problem would be the isolation. It's almost impassible to get isolated from charger when switches to circuit because of the body diode.
In this case, traditional relay still seems a better option for switching ultracapacitor banks. But It would be challenge to find relays with really low audible noise.
Ian
Ian,
If using relays to connect the charger to the caps I think that a pre charge circuit will be needed. Counting in that the caps have very low impedances and high charging currents are needed, sooner or later the relay contacts will weld together. It may seem a joke the welding stuff but my daily job implies lithium batteries and inverters so I assure you I have seen enough high current contactors with the contacts welded because of the high inrush currents that the capacitors from the inverters present to the contactor when are completely discharged. I solved this problem adding a second contactor with a power resistor in series with the contacts. This network stays on for few seconds after which is shunted by the contactor without the resistor and after I disable the contactor with the resistor.
It comes a bit complicated if you want something reliable in time.
If using relays to connect the charger to the caps I think that a pre charge circuit will be needed. Counting in that the caps have very low impedances and high charging currents are needed, sooner or later the relay contacts will weld together. It may seem a joke the welding stuff but my daily job implies lithium batteries and inverters so I assure you I have seen enough high current contactors with the contacts welded because of the high inrush currents that the capacitors from the inverters present to the contactor when are completely discharged. I solved this problem adding a second contactor with a power resistor in series with the contacts. This network stays on for few seconds after which is shunted by the contactor without the resistor and after I disable the contactor with the resistor.
It comes a bit complicated if you want something reliable in time.
Hi schultzsch，
Thank you for sharing your experience. To use relays for ultracapacitor power supply, the current has to be reduced below rated level of relay at the moment of switching.
You are right, additional relay might be needed for pre-charge or pre-discharge. And also we don't really need that high charging current though it is capable for. There will be a lot of considerations to take.
Regards,
Ian
@iancanada
Ian,
How about the following scenario:
What is your typical listening session time? Hours?
What is the acceptable time for your DAC to power up? Minutes?
What if you have Super Capacitor PSU for your DAC that would charge quickly (minutes) but only from time to time and then discharge for hours …
Would two BIG 3000F Maxwell cells meet those requirements? Those two could be charged with dozen of Amperes in minutes and discharge with hundreds of miliAmpers in hours with assumption to allow some small voltage drop ~0.5V. In such case discharge to charge ratio is around ~1:60 (one minute charging to one hour of discharging in this case). I assume voltage drop here from 5.4V to 4.9V, but could be lower down to acceptable local LDOs input levels.
No switching here. The use case here is like with Power Bank, but with quick automatic charging only from time to time. Most of the time (98% ?) the DAC would be powered with SuperCaps. Short charging periods could be indicated with some LED, so you would know when to plug the ears
For me something like this seems to be far easier to build, if one accepts those short charging periods.
What do you think guys?
Ian,
How about the following scenario:
What is your typical listening session time? Hours?
What is the acceptable time for your DAC to power up? Minutes?
What if you have Super Capacitor PSU for your DAC that would charge quickly (minutes) but only from time to time and then discharge for hours …
Would two BIG 3000F Maxwell cells meet those requirements? Those two could be charged with dozen of Amperes in minutes and discharge with hundreds of miliAmpers in hours with assumption to allow some small voltage drop ~0.5V. In such case discharge to charge ratio is around ~1:60 (one minute charging to one hour of discharging in this case). I assume voltage drop here from 5.4V to 4.9V, but could be lower down to acceptable local LDOs input levels.
No switching here. The use case here is like with Power Bank, but with quick automatic charging only from time to time. Most of the time (98% ?) the DAC would be powered with SuperCaps. Short charging periods could be indicated with some LED, so you would know when to plug the ears
For me something like this seems to be far easier to build, if one accepts those short charging periods.
What do you think guys?
@iancanada
Ian,
How about the following scenario:
What is your typical listening session time? Hours?
What is the acceptable time for your DAC to power up? Minutes?
What if you have Super Capacitor PSU for your DAC that would charge quickly (minutes) but only from time to time and then discharge for hours …
Would two BIG 3000F Maxwell cells meet those requirements? Those two could be charged with dozen of Amperes in minutes and discharge with hundreds of miliAmpers in hours with assumption to allow some small voltage drop ~0.5V. In such case discharge to charge ratio is around ~1:60 (one minute charging to one hour of discharging in this case). I assume voltage drop here from 5.4V to 4.9V, but could be lower down to acceptable local LDOs input levels.
No switching here. The use case here is like with Power Bank, but with quick automatic charging only from time to time. Most of the time (98% ?) the DAC would be powered with SuperCaps. Short charging periods could be indicated with some LED, so you would know when to plug the ears
For me something like this seems to be far easier to build, if one accepts those short charging periods.
What do you think guys?
Thanks TioFrancotirador, it's a very good point.
Still need relays to turn on/off the charger. The key difference would be switching every 10 minutes or turning on/off every one hour if you don't want to stop music.
But this arises me another idea: Design the charger with smart, to charge the ultracapacitor only during silence time between music, and keep ultracapacitor disconnected from charger while music is playing.
Please let me know what do you think.
Regards,
Ian
Well I thought in this case switching could be done not at capacitors level where relay could be welded because of high amperage, but at 230V, where amperage is lower. Kind of plug and unplug mechanism or maybe at DC/DC converter level at least. Unless not amperage, but wattage matters here so it does not matter
Your idea about DAC informing charger to work on: mute, no PCM nor DSD signal or simply no music is also interesting … however this narrows down usage of such PSU to only unit capable of sending such information … needs more thought on this one …
Well I thought in this case switching could be done not at capacitors level where relay could be welded because of high amperage, but at 230V, where amperage is lower. Kind of plug and unplug mechanism or maybe at DC/DC converter level at least. Unless not amperage, but wattage matters here so it does not matter
Your idea about DAC informing charger to work on: mute, no PCM nor DSD signal or simply no music is also interesting … however this narrows down usage of such PSU to only unit capable of sending such information … needs more thought on this one …
We don't need to worry about relay contacts at all, I have tons of way to protect them from welded.
The only thing need to be focused is how to get the best sound quality form the performance of the ultracapacitor power supply.
I like the idea for using big ultracapacitor (>3000F) to eliminate the bank switching, but how to charge it has to be very smart.
Regards,
Ian
Last edited:
So maybe hybrid solution, switching between BIG Caps (6000F) 98% of time and some linear or switching 5V regulator for the rest 2% of time with some information to the user whether is on battery or on regulator.
Ultracapacitor is very sensitive to voltage changes, you will still need LDOs to achieve a stable voltage.
How about going for a simple solution? Switch between regular psu and “silent running” for quality listening?
I use my Pi for more than just serious listening and for watching tv shows or movies, it’s not overly critical to have super clean power.
This way you could also toggle the power while you go and make coffee, using the down time for charging.
P.S. In my case it’ll probably be more like 1.5A continuous draw since my RPi has a hdd hooked up.
I use my Pi for more than just serious listening and for watching tv shows or movies, it’s not overly critical to have super clean power.
This way you could also toggle the power while you go and make coffee, using the down time for charging.
P.S. In my case it’ll probably be more like 1.5A continuous draw since my RPi has a hdd hooked up.
Last edited:
Guys,
I have used my principle in my dac/streamer 3/4 year and was very pleased with it.
Switching between the banks is not a problem at all, and charging is VERY simple when using DC-DC converter which converts 19V laptop adapter to the needed hosts voltage. Since it does CC first, then CV, and it stops charging when ready, even with 100F you can listen about half an hour, so switching once half an hour.
Don't make too big a deal of it, it's just that I had to preset runtime with a timer for switching the banks because of no voltage sensing circuit, that is the thing I would want to add at my system.
About best sound quality:
Choose relays with the lowest contact resistance, I chose relays with bifurcated contacts, with two sets of contacts which I use in parallel, so half the resistance of the initial (50mOhm) contact resistance.
I would still prefer mosfets, not only because of mechanical silence, but for low RDS-ON ( resistance). and I can't imagine that Collector Emitter can't galvanic isolated from each other, not sure how to switch GND with mosfet though….
I have 50 relays switching every 30 minutes, it is not a huge clicking sound, but my listening distance is 7 meters. I can imagine if listening distance is 3 meters switching noise can be a pita.....
Just my 5 cents....
Alex
I have used my principle in my dac/streamer 3/4 year and was very pleased with it.
Switching between the banks is not a problem at all, and charging is VERY simple when using DC-DC converter which converts 19V laptop adapter to the needed hosts voltage. Since it does CC first, then CV, and it stops charging when ready, even with 100F you can listen about half an hour, so switching once half an hour.
Don't make too big a deal of it, it's just that I had to preset runtime with a timer for switching the banks because of no voltage sensing circuit, that is the thing I would want to add at my system.
About best sound quality:
Choose relays with the lowest contact resistance, I chose relays with bifurcated contacts, with two sets of contacts which I use in parallel, so half the resistance of the initial (50mOhm) contact resistance.
I would still prefer mosfets, not only because of mechanical silence, but for low RDS-ON ( resistance). and I can't imagine that Collector Emitter can't galvanic isolated from each other, not sure how to switch GND with mosfet though….
I have 50 relays switching every 30 minutes, it is not a huge clicking sound, but my listening distance is 7 meters. I can imagine if listening distance is 3 meters switching noise can be a pita.....
Just my 5 cents....
Alex
All you need is a DC-DC CC/CV charging circuit, with finetuning V and A option.
I showed in previous posts what kind of converter I use. Simple $4 converters.
One needed for every voltage needed.
Of course, CC/CV charger has LDO on board.....
Just charge the ultracap with preset voltage of CC/CV converter.
@Ian
Not sure if you want to incorporate a CC/CV charger on board, would be nice, but not neccessary though.
You could design a pcb with microcontroller controlled voltage sensing circuit with input terminals leading to CC/CV charger, and terminals for connecting the ultracaps, and of course terminals for output voltage.
Another thing not to forget:
The PCB needs a ON/OFF switch option which disconnects ALL ultracaps from output. In my case this was not possible, thus, each time when starting up dac i had down-time because ultracaps were drained, or at least one bank, and possibly the other for lets say 95%, depending on the time when switched off.
The problem was, when switching once in 30 minutes, I had to wait at least 30 minutes before switching to charged bank.....
That is something to keep in mind when designing!
Cheers guys!
Alex
Last edited:
I think what you need is a micro processor based ultracapacitor PSU controller with isolated voltage sensing and isolated ON/OFF control signal.
Have to have a display to show the status, also some consideration of protecting relays from damaging by the inrush current...
Regards,
Ian