LM5145: constant current loop stability problems

user5069983

Part Number: LM5145
Other Parts Discussed in Thread: LM5117, LM5141

Hello,

we have designed a CV/CC charger for lead acid batteries (Vout=41,4 V (3x12V block), Ioutmax=6A). We used the LM5145 standard circuit for CV charging and added a current limiting circuit for constant current charging.

If we are charging big lead acid batteries with low internal resistance (<5mOhm) we have problems with loop stability in constant current charging mode. The output current is oscillating.

A simplified schematic of our circuit is attached.

We don't know how to stabilize the circuit. Can you give us advice here?

over 3 years ago

0 Timothy Hegarty over 3 years ago

TI__Guru 73752 points

Hello,

Please send a completed quickstart calculator so we can review compensation for constant voltage (CV) operation first. Don't forget to derate the ceramic output capacitance for voltage. You can use PSPICE to model the CC loop and improve the response by reducing the pole frequencies (RC time constants) in that path.

Regards,

Tim

0 user5069983 over 3 years ago in reply to Timothy Hegarty

Prodigy 60 points

Here is the quickstart calculator file.

3531.LM(2)5145, LM5146 contorller design tool - revA1.xlsm

We already considered a derating of the output ceramic capacitors of -60%.

The CC loop is already very slow. If it gets too slow, the overcurrent Limit of the controller will be triggered at start up and hot plugging of the batteries.

0 user5069983 over 3 years ago in reply to Timothy Hegarty

Prodigy 60 points

There is a 100µF electrolytic capacitor at the output of the battery charger. This cap is placed relativly far from the regulator output (there is an ideal diode circuit between the regulator output an this cap). This 100µF cap was not considered in the quickcalculator. Is it necessary to consider this cap. If yes, how to calculate the equivalent capacity and ESR of the ceramic an electroyltic caps in parallel.

0 Denislav Petkov over 3 years ago in reply to user5069983

TI__Mastermind 19891 points

Hello,

The team in the US is on public holiday. Please expect another reply tomorrow.

Regards,
Denislav

0 Timothy Hegarty over 3 years ago in reply to Denislav Petkov

TI__Guru 73752 points

Yes, even though the cap is downstream, it still needs to be considered in the loop. The best way to consider the effective capacitance and effective ESR with multiple caps in parallel is with simulation.

Regards,

Tim

0 user5069983 over 3 years ago in reply to Timothy Hegarty

Prodigy 60 points

Thanks for your help and quick answer. As we are loading a battery block with very low impedance, whitch is something like a big cap , this will also have an impact ot the loop stability and has to be considered. Is this statement right?

It is difficult for us to model the battery blocks, also because there are different batterys in size we want to charge.

Can you support us with functioning simulations for loop stability (simulation of open loop bode plot)?

0 Timothy Hegarty over 3 years ago in reply to user5069983

TI__Guru 73752 points

Correct, you need a small-signal model for the battery. Try setting something up in PSPICE for TI. I haven't got anything readily available unfortunately. However, if you have SIMPLIS, I may be able to get something running.

Regards,

Tim

0 user5069983 over 3 years ago in reply to Timothy Hegarty

Prodigy 60 points

It would be great if you can support us with a running SIMPLIS simulation.

0 Timothy Hegarty over 3 years ago in reply to user5069983

TI__Guru 73752 points

No problem, please give us a few days to work on a model.

0 user5069983 over 3 years ago in reply to Timothy Hegarty

Prodigy 60 points

Thanks for the outstanding support. In the meantime i will learn more about modeling batteries.

0 Timothy Hegarty over 3 years ago in reply to user5069983

TI__Guru 73752 points

Ok, thanks. One important point to note is that the circuit is currently setup to limit current (i.e. brickwall current limit), effectively by reducing Vout at a certain max current threshold. This is know as CV-CC operation.

I think you need CC-CV for battery charging, which gives constant current (CC) up to a certain voltage (CV) clamp setpoint. This is an important distinction - let me know which one is required.

Regards,

Tim

0 user5069983 over 3 years ago in reply to Timothy Hegarty

Prodigy 60 points

Ok, I think you are right. What we really need is CC-CV regulation. We didn't know how the CC circuit looks like with LM5145 and how to calculate it, so we took the CV circuit and added CC regulation. Is there an applicaton note or a quick calculator for CC regulation.

Can you support us with both circuits? I don't know if we really want so start from scratch with a new CC-CV cirucit.

0 Timothy Hegarty over 3 years ago in reply to user5069983

TI__Guru 73752 points

I can work on getting something to you for CC-CV operation as that seems to be the main requirement.

Regards,

Tim

0 user5069983 over 3 years ago in reply to Timothy Hegarty

Prodigy 60 points

OK thanks. That would be great.

0 Timothy Hegarty over 3 years ago in reply to user5069983

TI__Guru 73752 points

Give me until Friday to work on this - thanks for your patience.

0 Timothy Hegarty over 3 years ago in reply to Timothy Hegarty

TI__Guru 73752 points

Are you still interested in the CC-CV operation? I can get you a schematic / SIMPLIS model by tomorrow.

0 user5069983 over 3 years ago in reply to Timothy Hegarty

Prodigy 60 points

Yes, we are still interested.

0 user5069983 over 2 years ago in reply to Timothy Hegarty

Prodigy 60 points

Do you have already prepared the SIMPLIS model for us?

0 Timothy Hegarty over 2 years ago in reply to user5069983

TI__Guru 73752 points

Ok, let me get this over to you ASAP.

0 Timothy Hegarty over 2 years ago in reply to Timothy Hegarty

TI__Guru 73752 points

Here is the voltage-mode loop. Take a look at the component values. I went with 10uH inductor for optimal ripple current given the 41.4V, 6A, 232kHz spec. The file plots the bode plot and load transient response.

I'll follow up shortly with the external current loop included.

Regards,

Tim

https://e2e.ti.com/cfs-file/__key/communityserver-discussions-components-files/196/LM5145-48Vin-41.4Vout-6A-232kHz-regulator.sxsch

0 Timothy Hegarty over 2 years ago in reply to Timothy Hegarty

TI__Guru 73752 points

Here is the file for CV-CC, similar to your original circuit, but it has 100k in series with FB and 100k/100pF around the amplifier (this improves stability).

https://e2e.ti.com/cfs-file/__key/communityserver-discussions-components-files/196/LM5145-48Vin-41.4Vout-6A-232kHz-regulator-_2D00_-with-external-CC-loop-_2D00_-E2E2.sxsch

0 Timothy Hegarty over 2 years ago in reply to Timothy Hegarty

TI__Guru 73752 points

Note that, in general, a current-mode controller is much easier to compensate than a voltage-mode design in terms of CC operation.

I recommend using the LM5117 as it has an integrated current monitor (CM) pin that is ideal for these type of circuits. It's a 65V controller, so it should meet you Vin requirement. See also app note SNVA829 for more detail.

Regards,

Tim

0 user5069983 over 2 years ago in reply to Timothy Hegarty

Prodigy 60 points

We have considered using the LM5117 but TI told us to use the LM5145 because we are at the duty cyle limit if we want to adjust the output voltage up to 44V with a minimum input voltag of 46V.

0 user5069983 over 2 years ago in reply to Timothy Hegarty

Prodigy 60 points

Thanks for your support, we will evaluate if the circuit would meet our considerations.

0 user5069983 over 2 years ago in reply to Timothy Hegarty

Prodigy 60 points

As I see, you prepared a working CV-CC simulation for us, which is very helpful becaus this is our current circuit topology.

Would it also be possible to get a CC-CV circuit with the LM5145?

0 Timothy Hegarty over 2 years ago in reply to user5069983

TI__Guru 73752 points

Yes, the LM5117 has a higher toff-min as it uses emulated current-mode control where the current is sampled-and-held at the end of the switching period. Thus, the max duty cycle capability is not as high. If the max Vin is < 65V, you can also consider the LM5141 65V peak current-mode controller, which is easier to implement from a CC-CV standpoint.

Regards,

Tim

0 user5069983 over 2 years ago in reply to Timothy Hegarty

Prodigy 60 points

Is there a possibility to get output voltages higher than 15V with the LM5141 with external components? The internal current sense amplifierer has a max. input voltage of only 15V.

0 Timothy Hegarty over 2 years ago in reply to user5069983

TI__Guru 73752 points

Correct, the common-mode range of the CS amp only allows a Vout-max of 15V for the LM5141. We have a new controller close to release that may be suitable here - please contact your local TI FAE for more detail on that if needed as I can't share the details here.

Regards,

Tim

Power management

Power management forum

LM5145: constant current loop stability problems