It didn't explode....... YET (aka niges amp)

[nige2000]

..............

THE BIG QUESTION remains: do I simply wire up two assemblies - a 26650 with a 3.3 volts regulator/charger along with a supercaps at approximately approx. 1.9 volts regulator/charger and then put them in series? This is where I get concerned.

I do not see a DC-DC converter at less than 3.3 volts - so i guess one uses this on the 26650 and MRAVICA for the remainder, it can be set to 1.9 volts?

It is really that simple? Will the regulators allow this series connection? Please give me a few details.

Yes, supercap + battery in series - supercap at the gnd end of the series with a 1.9V reg charging it & 3.3V ISOLATED DC-DC converter on the battery. The isolation is the critical characteristic - it blocks the positive voltage on the supercap from connecting to the ground of the charging PS. Because the setup is being powered from one PS, without this isolation you would have a short from pos of supercap to gnd of supercap & fireworks!!

Nige, achieves the same isolation by using a transformer (on each bank of 6 batteries?) with (6?) independent secondaries i.e 6 floating outputs which each charging a battery through Vregs.

Actually 20 secondaries rectified powering 20 regs with 20 diodes on the reg + powering 20 cells in series

So do you enjoy the sound from the supercaps as much as the 26650?

It's a different sound - more bite to it - jury is still out, time will tell.
I'm thinking it's time for a tour of these different PS devices - I have 2 PS products now battery PS & supercap PS - hybrid PS to follow - anyone interested in comparing the sound of these PSes? I should start another thread, not to hijack this one any further.

I would suggest you try 2 supercaps in series as a first test config (it's probably simpler?) & see what you think

Yea I think id compare the supercaps to the batteries first, its simpler to implement
youll have to current limit the regs id guess
Although im fairly sure if u balance charge the lifepo4 at 2.5v it wont die.
Any in series charged 5v between two cells one cell always eventually dies. i believe its to do with different impedance in the two cells one cell will take all the charge and at 2.5v min vol there in a compromised state
although the rep in the wina batt factory says if you impedance match the cells you can charge em in series

[rickmcinnis]

Thanks to you both, John and Nigel.

I should have mentioned that I knew that separate secondaries were required.

So I guess, as usual on Earth, compromise is in order no matter which path you take.

So, if one went with the hybrid solution WITH two secondaries could one get by without the DC-DC converter? Old biases die hard.

I am a relative newcomer to LTSpice and thought I drew this up correctly - separate power supplies, etc., but it would not double the voltages. I have to admit I was unsure about my battery specification. It is a great tool but it takes lots of looking around to get all that you need.

I have played with the TINA one, also, but it seems to not know what to do with a choke, so I quit using it. It is easier to get your drawing mounted but LTSpice is easier to work for the graphs. Both of them are great toys - and I mean that they are toys. They are fun and if you play around with enough values you get a better idea of what happens but it is well known one should not get too sure of what it tells you.

The MRAVICA is heat sinked well and he and TI say it can deliver 1 amp with good heat sinks. Nonetheless, I figure one should build a 1085 to charge it back to 3.3 or maybe a resistor on the MRAVICA to the battery could make it work? Am I thinking correctly?

I am sorry that I veered off subject but since this article in many ways is a reference I thought it should be filled out a little more completely. Plus I could not think of a good title!

Back to the batteries at 2.5 volts. One figures they would lose much of their low ESR and then being in series makes this even worse. I was going to use four to counteract this but the more i consider it the hybrid might be just fine for these purely digital tasks - other than running those motors in the disk drive, of course.

Take care,

[jkeny]

.....Actually 20 secondaries rectified powering 20 regs with 20 diodes on the reg + powering 20 cells in series

I was close though - just missed out on some small details, haha :)

Yea I think id compare the supercaps to the batteries first, its simpler to implement
youll have to current limit the regs id guess
Although im fairly sure if u balance charge the lifepo4 at 2.5v it wont die.
Any in series charged 5v between two cells one cell always eventually dies. i believe its to do with different impedance in the two cells one cell will take all the charge and at 2.5v min vol there in a compromised state
although the rep in the wina batt factory says if you impedance match the cells you can charge em in series

Yes impedance difference is the main issue with balancing batteries & supercapacitors

Impedance matching the batteries is a good start but not very practical - how do you do this? That's probably OK when you have a large pool of batteries to choose from. Do they drift apart over time & usage, though? 2.5V is just on the edge of falling off the cliff to depletion - they drop voltage very quickly from 2.7V downwards - not like when they are fully charged where they hold their voltage at 3.2V for a long time when not being float charged.

I reckon It's best to have a charging setup which deals with any impedance mismatch & any mismatch drift over time.

[jkeny]

@Rick
Yes, no need for DC-DC ISOLATED converter when using transformer with isolated secondaries but you probably need a diode on the neg (-) leg of each charger line to the (-) pole of each battery - this prevents reverse current from getting back into whatever regulator you are using - the DC-DC isolated converters automatically stop this back current.

Whatever way you do it, the isolation is what is important in charging each battery/supercap in series. In other words you need to isolate charge each battery/supercap in the series - if you have 20 in series, you need 19 of them isolated (you don't need to isolate the one at the ground end).

BTW, you should be fine running the motor off the setup - it really depends on how you factor in the continuous & max current handling needed for this - the transformers, voltage regulators, rectifier, diodes & charging PS need to be up to the current delivery

[rickmcinnis]

I hope the impedance mismatch is not too big an issue with a simple two element supply. Certainly with more elements this becomes far more ungainly. Then again, it might not make much difference at all. If Nigel's amps are still running properly with their original batteries that would seem to prove that there is reliability. I hope I have not cursed two hundred batteries into oblivion.

Just for more clarification - Nigel, are you using the diode across the regulator or across the batteries?

If I decide to us the MRAVICAs would it be a good idea to use the diode the same you have with LT1085 - however it is you used it. I assume across but please let me know.

AND assuming the TP7S4700 (?) is tolerant of a diode there ...

Thanks,

[rickmcinnis]

Saw your post before making this post. THANKS for answering the diode question!

I monitored the voltage and then the current from the switcher to the motherboard and it is not moving around at all.

Using this for REDBOOK CD seems to not require much power at all. I guess when you engage the video stuff things change!

Since I will only have 1.9 volts (the MRAVICA makes this very simple) across those SUPERCAPS one can get extravagantly large value ones. I think one was 100F. Should one go as high as they want or are there reasons for keeping these to a certain capacity?

Took another look and see one can get 3000F! Is too much just enough?

THANKS again,

[jkeny]

I hope the impedance mismatch is not too big an issue with a simple two element supply. Certainly with more elements this becomes far more ungainly. Then again, it might not make much difference at all. If Nigel's amps are still running properly with their original batteries that would seem to prove that there is reliability. I hope I have not cursed two hundred batteries into oblivion.

Ah, sorry we confused you - the impedance matching is only needed if you are putting single voltage across two (or more) batteries/supercaps in series, not if you are charging each batt/cap individually, as you intend. The impedance match means that the batteries will share a single voltage across them equally.

Just for more clarification - Nigel, are you using the diode across the regulator or across the batteries?

If I decide to us the MRAVICAs would it be a good idea to use the diode the same you have with LT1085 - however it is you used it. I assume across but please let me know.

AND assuming the TP7S4700 (?) is tolerant of a diode there ...

Thanks,

Nige can answer this but the diode is to stop back current reaching the regulator - he might have a diagram to show this? Yes, you can use a diode with any V reg.

In answer to your next post, be careful with supercaps - unlike batteries, they usually start at 0V & so present a short to any regulator - not a good idea & will probably burn out a regulator. One way around this is to initially charge the cap up slowly using a controllable PS & then never take it off charge - it will deplete fairly quickly (48 hours or so) when off charge (even if all connections to it are disconnected.

[nige2000]

Just to be clear the diode is in series of the positive out of the reg
Voltage setting must allow for the voltage drop on the diode
Lower voltage drop diodes are preferable

Also this voltage drop acts as as a current limiter of sorts but as caps start at 0v it will draw max current until the voltage rises

I'd keep an eye on the esr of the supercaps as well as the capacity
The larger regs 1085 etc would be better suited to charge the supercaps

I'm getting some supercaps I'll let ye know how I get on

[jkeny]

As you go up in Farads value the supercapacitor esr usually gets lower (in the same brand/model)
The esr of the 350F supercaps Nige is getting are 3.2mOhm - two in series gives 175F & 6.4mOhm - very competitive, if not better than 26650 LiFePO4 batteries - it will be good to hear what you think of them

The ideal charging scheme for both supercaps & batteries is CC/CV charger - constant current/ constant voltage with smart charging pre-conditioning built-in - in other words CC/CV chargers that are made for Lithium batteries use a small current to start charging when it senses the connected battery/supercap is below ~2V - this prevents the charger seeing a virtual short & trying to output very high current.

[rickmcinnis]

Thanks so much to the both of you.

My plan was to use this supply in the old hallowed TIR NA tradition of keeping it on all of the time.

I figure for initial start up i would need to use a 10 Ohms resistor in series with the positive lead of either device and afterwards hope the power stays on which is not something I have to worry about here. After charging the resistor would be removed (might even keep it there). I am not using this regulator for any reason other than it is here and it works. All of its vaunted qualities of quietness will not be of use in this application.

Is the quality of the diode important? Any need for an SIC diode? One part of me thinks I should try without the diode.

Going to order that 3000F cap - mainly because I like the idea of having screw posts and it has them.

Hope to have it working over the weekend.

Now if anyone could help me figure out that mystery of how to get the thing to actually play a CD with an outboard 15 volts supply ...

Take care,