Lester ferro-resonant charger?

[MikeMl]

(also posted on AAC)

My neighbor's golf cart is showing signs of chronic undercharging of its flooded lead-acid deep-cycle batteries (uneven voltages between groups of cells, low capacity). It has 18 cells; nominally 36V. Under charge using the **broken link removed**, the batteries barely reach 38.0V (2.12V/cell), which is not enough to top them up. The end-of-charge voltage should be at least 45V (2.5V/cell). The charging current tapers to zero prematurely (too soon)

The Lester charger is dirt simple. Here is the schematic. Big transformer, 115Vac primary, center-tapped secondary, two rectifiers with common cathodes which make the +; - taken from the c.t., i.e. full-wave rectifier. The tranny has an extra winding which is paralleled with an 6uF 660Vac AC capacitor, with no other connections. AFIK, it is some sort of ferro-resonant regulator.

I have checked continuity, and everything appears good. Diodes good, fuse good, transformer windings all show continuity, no shorts between windings. The only thing that looks suspicious is that the resonating capacitor measures 2.94uF instead of the marked 6uF.

Is this the reason why it is not charging the batteries? Not resonating properly? How does this regulate?

 

[jpanhalt]

Mike, Your diagnosis certainly makes sense. If the capacitor is too small, the resonate frequency is higher and clipping should occur at a lower voltage. Wikipedia does mention (voltage regulators) that such designs are sensitive to capacitance. These references describe the operation n more detail. The first link is essentially the simplified design you have. The second describes the operation.
https://www.ustpower.com/comparing-...-guide-constant-voltage-transformer-operation
**broken link removed**


John

 

[dr pepper]

That makes sense too, also your measurement is probably at 9v or less on a multimeter, when the cap is operating in circuit passing large currents the capacitance is probably different still, maybe even zero.
A motor start capacitor would probably make a good replacement, obviously same capacitance.

 

[MikeMl]

After reading the posted references, here is what I gleaned:

In this application, the ferro-resonant transformer is not used because it acts as a voltage regulator; rather it is used to regulate charging current, making the charger act more like a constant-current source, rather than a constant-voltage source. If you use a well-regulated voltage source to charge a lead-acid battery bank, the problem is that the initial charging current is huge (and it blows up the power supply). In the Lester charger, they are relying on transformer core saturation to keep that initial charging current down.

The core saturation also has another benefit; it squares the waveform that comes out of the full-wave rectifier such that the current flow is more like filtered DC. There is no filter capacitor in these chargers, but unlike a regular transformer-rectifier combo that would deliver a series of short high amplitude current pulses to the battery, the ferro-resonant charger delivers wider, flatter pulses, closer to pure DC.

Now here is something I haven't yet figured out. Talking to an old ham buddy about it, he suggests that the resonant circuit is tuned to the third harmonic of the power line frequency (180Hz). The third harmonic added to the fundamental makes a waveform which looks more like a square wave (by filling in the corners, knocking down the peak in the middle). This action occurs after the charging current has dropped no longer causing the core to saturate. The Q of the resonant circuit increases as the core becomes more linear (non-saturated), boosting the output voltage to finalize the charging cycle. Is this true?

In the charger I'm trying to fix, the capacitor has changed value significantly, preventing that freewheeling voltage boost at the end. I'm ordering a replacement, and will report back after I put it in...

 

[NorthGuy]

The first stage is constant current (probably somewhere around 20-30A). Once a certain voltage (43-44V) is reached it should start limiting the voltage and to do that, it needs to decrease current. It then outputs constant voltage until current goes below some pre-set value. At this point it may consider the battery charged. Some chargers (e.g. most fork lift chargers) have an extra constant-current stage at the end. They keep relatively low current for pre-set amount of time or until voltage rises above certain level (around 48V), whichever comes first. There are lots of variations to the process.

 

[nsaspook]

After reading the posted references, here is what I gleaned:

In this application, the ferro-resonant transformer is not used because it acts as a voltage regulator; rather it is used to regulate charging current, making the charger act more like a constant-current source, rather than a constant-voltage source. If you use a well-regulated voltage source to charge a lead-acid battery bank, the problem is that the initial charging current is huge (and it blows up the power supply). In the Lester charger, they are relying on transformer core saturation to keep that initial charging current down.

Yes, the ferro-resonant transformer acts as a current limiter that usually matched closely to the Ah charging load in this application. We used them in some old computer power systems I installed on ships before electronic voltage regulation was common.

Design for voltage regulation: http://www.elect.mrt.ac.lk/CVT_ICIIS06.pdf

For a battery charger (no need for an AC output filter) the Is (short circuit current) is selected for the Charge rate (C5 to C20 for large traction batteries) of total Ah capacity and Vo is selected for the battery absorption voltage. We usually could tell if one was defective by seeing how noisy and hot it was. If it was cool and not sounding like a mower it was defective (usually a bad cap circuit and/or bad batteries).
**broken link removed**

 

[dr pepper]

At royal ordnance we had some very old lansing mobile plant, I dont know if they were ferro resonant, but you could tell when one of the cells was down by the sound of the charger as above, if it sounded like the treble was turned right down there was a dead cell in the truck, the chargers were so reliable I dont remember any of them going down.

 

[MikeMl]

...
Design for voltage regulation: http://www.elect.mrt.ac.lk/CVT_ICIIS06.pdf ...

I especially liked the Conclusions Section of this paper

 

[cowboybob]

Mike, were you able to check the specific gravity of the various cell's electrolyte (and, of course their levels)?

At my last job we regularly used "Club Car" buggies for transportation around a 78 acre facility and eventually ended up having to replace the sulphuric acid in the batteries as they exhibited identical charging (and voltage level) problems as you posted (see below):

My neighbor's golf cart is showing signs of chronic undercharging of its flooded lead-acid deep-cycle batteries (uneven voltages between groups of cells, low capacity).

To be sure, the lead in the batteries was also severly reduced but the acid replacement gave us nominal charge/voltage levels and about an additional 6-8 months use before replacment batteries had to be installed (at no small expense).

Just a thought and not meant to impugn anyone's expertise.

 

[MikeMl]

I'm no stranger to maintaining flooded and sealed lead-acid batteries. These all show the symptoms of chronic undercharge, and the V/I output of the charger falls short.

First thing I did was measure the resting voltage (after 24hours of isolation): batteries discharged. Second, I measured S.G. which shows some low cells. I then equalized the batteries by passing a 2A constant-current through them for 36hours, during which time the cells bubbled slowly. Lots of scum floated to the top. After another 24hour rest, they show a resting voltage that is consistent with being fully charged, and the S.G. is higher and more uniform, cell-to-cell.

I expect that the owner will get a few more months out of the batteries. I'm still waiting for the parts to repair the charger...

 

[cowboybob]

Very thorough, Mike. I expected as much.

Just had to ask, is all...

 

[MikeMl]

... I'm still waiting for the parts to repair the charger...

Part came and here is what I learned:

With no cap installed, the battery terminal voltage reached 40.3V (batteries were previously fully-charged, after 36 hours of equalization).
With the old 2.9uF cap installed, the battery terminal voltage reached 42.9V .
With the new 6.0uF cap installed, the battery terminal voltage reached 46.9V .
With both caps in parallel, the battery terminal voltage reached 49.3V .

The correct end-0f-charge voltage for these flooded-cell batteries at room temperature is 2.45V/cell, so 18cells would be 44.1V.

With partially charged batteries, the initial charging current is >20A, while with the bad capacitor is was only ~14A. If the charger runs an hour past when ammeter tapers to near zero, with the new capacitor, the ammeter reading is still >1A, while with the bad capacitor, the ammeter would taper much faster, and go completely to zero...

 

[johansen]

I have a 25 amp Clark 36 volt charger and it sucks donkey balls.

It outputs far too high a voltage (which it must do because the transformer is not nearly large enough to actually deliver 1 kilowatt without overheating) and it is very loud.

But anyhow the resonant capacitor isn't really critical, while it does play a role in setting the output voltage, the core is designed to leak, not saturate for constant current purposes at the beginning of the charge. a larger capacitor will raise the output voltage because it is in competition with the transformer's shunts which reduce the voltage, in effect shorting out the core. too large however and it will reduce the voltage and the magnetic shunts will be in complete saturation.

The core is designed to saturate for the purpose of limiting the no load voltage, and the larger the capacitor, the more current will be flowing in it.. the copper losses on the resonant coil is not negligible.. mine is wound over the top of the 12 awg (iirc) secondary coils.

 

[MikeMl]

...
The core is designed to saturate for the purpose of limiting the no load voltage, and the larger the capacitor, the more current will be flowing in it.. the copper losses on the resonant coil is not negligible.. mine is wound over the top of the 12 awg (iirc) secondary coils.

Hi Johansen,

No, you have it backwards. The core saturation limits the charging current during the first part of the charging cycle when the current into the battery is high. Because core saturation lowers the Q and/or the coupling to the resonant winding, it is effectively out of the circuit at this time. This is causes the AC rms voltage across the capacitor to be less than 100V during the initial charging cycle.

As the battery voltage comes up, the current into the battery naturally decreases, and the transformer core comes out of saturation, and the resonant winding takes over. The voltage across the resonating capacitor comes up to 660Vac rms, and that causes the output voltage from the full-wave rectifier to boost the output voltage across the output winding so that the battery can reach the fully-charged potential of 2.6V per cell. The failure of the capacitor prevented this last part of the cycle from happening.

The Lester I was working on has a timer clock. You set the duration of the charge cycle to about double the time you expect that it will take the charging current to taper to less than a couple of amps. The resonant circuit keeps the voltage across the 18cell battery at about 46V while the current has tapered to a 1A. This drives the battery to the region where it is gassing lightly, which guarantees that it is fully charged. It would not be good to keep the charger on more than a couple of hours after the current tapers (i.e., use it as a "float" charger.

The timer is simple, and it prevents the battery from being "overcharged" too long. It would be quite easy to build an automatic voltage detector that would switch off the Lester after the battery voltage first reaches a preset voltage, say 47V (2.61V/cell).

 

[johansen]

there are two ways to build these ferroresonant transformers.
the way cheap battery chargers do it, and the right way.

without the magnetic shunts and the resonant capacitor disconnected the core is not supposed to be saturated.
there will be minimal current limiting. shorting out the secondary should fry the diodes/fuses

when you put the shunts in, you'll notice that they total 1/3rd of the core area. its' that way for a reason. perhaps they shunt more, up to about half the core area is sent through the shunts.

the capacitor which is connected across the resonant winding is directly over the top of the nominal 36 volt battery winding.. thus, capacitor volts = battery volts+ r copper loss voltage drop - Rcopper loss voltage drop in the resonant winding(not insignificant either).
my charger pulls about 5.5 amps from the outlet, under no load with magnetic shunts and the capacitor.. without the capacitor or the magnetic shunts inserted it is 1.2 amps iirc.

the capacitor is of course selected to saturate the magnetic shunts and its half of the core.. excition provided by the primary coil.
open circuit voltage is limited to 1.7T through the secondary coil, which when connected to a fully charged battery should push a safe equalization level charge, into a 16 volts*3. yes, its that high..
the primary coil maintains its 1.3T in its section of core.
the magnetic shunts take the difference. third harmonic is present because the iron is saturated.. that's just how it works, its not designed to maximize the third harmonic..

when the output is shorted, the current will be limited according to the flux density curves of the shunts and their airgap, not the core. suppose they saturate at 1.7T, which takes 30-50% of the flux away from the primary. so you'll have about half the nominal volts being sent through the secondary coil, which will be consumed in its own resistance.. limiting the current to perhaps 50% higher than its nominally rated "25 amp" battery charger.. which should take about 30 seconds to 2 minutes to blow the fuse.

if i'm not mistaken, the shunts and their air gap is designed to run at the knee of their saturation curve.
open circuit voltage regulation is provided by the knee in the core under the secondary winding.

under output shorted conditions, the highest flux density is in the magnetic shunts, it will be reduced everywhere else.

 

[MikeMl]

No external shunts on the Lester. Just three windings on an E| type laminated core.