chipamp lifepo4 powered monoblocks

[nige2000]

ive been thinking about doin this for quite a while i think there is might be a bit of sq potential, jk often says/rants:) about the influence of improper gnding this might resolve some of those issues along with the added advantage of shortening the signal length
this project should be reasonably low rent and fast to put together a working prototype (the way i like it :)

the theory is to have a main linear bi-polar ps say circa 13v +/-
feeding into the two monoblock units each unit containing 8 cells (yikes more expensive than i thought)

any ideas what chip to use was kinda looking at lm1875?
http://www.elisanet.fi/ch/gainx.html

[jkeny]

ive been thinking about doin this for quite a while i think there is might be a bit of sq potential, jk often says/rants:) about the influence of improper gnding this might resolve some of those issues along with the added advantage of shortening the signal length
this project should be reasonably low rent and fast to put together a working prototype (the way i like it :)

the theory is to have a main linear bi-polar ps say circa 13v +/-
feeding into the two monoblock units each unit containing 8 cells (yikes more expensive than i thought)

any ideas what chip to use was kinda looking at lm1875?

Remember I brought a battery powered chipamp to your place a while ago - everybody liked the sound of it & it fulfilled my rant quota :)

However, one problem emerged (& this is why I'm wary of batteries in series) - two of the series of 5 cells (~17V) I had powering the amp died after a while.
It wasn't because these were abused - I had a switch that change the config of the cells from 5 in series to 5 in parallel & started charging them @ 3.3v so each cell got the correct voltage. The problem seemed to be that more power was drawn from the pos (or neg, I can't remember now) end of the battery pack & the cells at this end tended to become drained but the amp still played at <17V so the current draw on these cells continued & killed them over time.

Trickle charging @ 17V would not work in this situation.

[nige2000]

What chip was in that chip amp ?
I think having some cells in parallel and some in series would exacerbate the voltage balancing issue

Long term we need to find a way to charge these cells
I've a couple of ideas I need to test out before investing in a heap of new battery cells I'm going to destroy ;)

[jkeny]

What chip was in that chip amp ?

Can't remember now - it was car audio chipamp - TDAsomething - sounded great on battery but soundstage collapsed when run on linear

I think having some cells in parallel and some in series would exacerbate the voltage balancing issue

No, that wasn't the config - when powering chipamp, 5 cells were in series for ~17V - when charging was needed, I had an arrangement that disconnected the cells & connected them in parallel for charging at 3.4V - this is the ideal arrangement, I think, as all cells are balanced & each gets 3.4V charge

Long term we need to find a way to charge these cells
I've a couple of ideas I need to test out before investing in a heap of new battery cells I'm going to destroy ;)

The above is the best way to charge them - disconnect & charge in parallel but that's not the problem - the uneven current draw on the cells in series is the issue - the cells at the positive end get drained quicker than negative end (or visa versa). As I said this leads to cell failure after a while - I had two dead cells so I didn't repeat the experience.

It seemed the only solution would be battery management circuitry on every battery but it would need to be one that passes a highish current. One of these on each battery to stop it overcharging & stop it undervolting would work. That way we could constantly charge @17 v or whatever voltage needed.

I didn't have any high current battery management circuits available at the time to try - I only had low current versions & these badly affected the sound

[nige2000]

this is ps scheme i cant see why it wouldnt work
every cell should be maintained at 3.3v
4 windings four bridge rectifiers and 4 regs

it looks more complicated than it really is

this is the start of the pre testing we have 0-3.3,6.6,9.9 and 13.2 v dc absolutely no problem
ill gather up 4 cells and put the test into practice with some car bulbs tomorrow

[nige2000]

ok test is on and all looks good
this bulb is hard on power 4.5 amp @ 13.2v
should be a good test


i think the cells on either end seem to be doing most of the heavy lifting
this is our issue with the in series system and the charge and discharge system, but in this system this is controlled :)
so this is success at least so far

now this gave me an other idea :)

[jkeny]

Yea, this looks good so far - is that the schematic for the linear regulators you're using?
Each leg is isolated & floating - there is no connection to mains ground
I wonder if this was powering your chipamp & you connected a grounded device to it, would it cause a problem?

[nige2000]

Yea, this looks good so far - is that the schematic for the linear regulators you're using?

no its only a similar one for explanation purposes, i just google images for it actually think it happened to be from twisted pear

Each leg is isolated & floating - there is no connection to mains ground

yes this is why it works, and the best thing is you can connect gnd to any point to get the negative voltages

I wonder if this was powering your chipamp & you connected a grounded device to it, would it cause a problem?

im not sure its certainly going to be interesting as the transformers are isolators (from mains) it should really act the same as a battery pack, however its normal for dc gnd to be connected to mains earth and our friend gnd loop can appear with the many links in the audio chain gnding to earth.
one gnd earthing or no gnds earthing might sound best but that's probably not safe or compliant ?

[Mike19]

This looks very interesting, wish I new more about batteries. Wonder if you could add paralled LiFePOs to each end of an SLA - SLA would give voltage lift and LiFePOs current. LiFePOs would provide 90% of their current at their working voltage of 3.3 - 3.2V and then be recharged off-line. Maybe no need for battery management between recharging?

[jkeny]

Yea, this looks good so far - is that the schematic for the linear regulators you're using?

no its only a similar one for explanation purposes, i just google images for it actually think it happened to be from twisted pear

Right - the reason I asked was because that schematic shows protection diodes around the voltage regulators which probably helps the whole arrangement work. I'm not sure if it would work without or at least it would be much more precarious without them? Using floating voltages can be very tricky & great care needs to be taken to ensure no voltage feedback into points where there shouldn't be any - that's why the diodes help

Are there diodes on the actual boards?

Each leg is isolated & floating - there is no connection to mains ground

yes this is why it works, and the best thing is you can connect gnd to any point to get the negative voltages

I wonder if this was powering your chipamp & you connected a grounded device to it, would it cause a problem?

im not sure its certainly going to be interesting as the transformers are isolators (from mains) it should really act the same as a battery pack, however its normal for dc gnd to be connected to mains earth and our friend gnd loop can appear with the many links in the audio chain gnding to earth.
one gnd earthing or no gnds earthing might sound best but that's probably not safe or compliant ?

Yep, ground is the biggest forgotten area in audio & keeping it as noise free as possible is crucial where sensitive, low voltage circuitry is concerned. Unfortunately ground is often used as a dumping area for noise in the belief that it is a infinite sink-hole.