### Author Topic: buck converter power limit  (Read 857 times)

0 Members and 1 Guest are viewing this topic.

#### pitagoras

• Regular Contributor
• Posts: 115
• Country:
##### buck converter power limit
« on: March 13, 2018, 01:44:32 pm »
Hello,
In the endless design of my lab PSU I now plan to add a buck pre-regulator.
The architecture is: mains to 24V@3A transformer, rectifier+caps, followed by the buck converter, followed by the linear regulator.
My current plan is to use a TL494 based buck converter, controlled by the psu controller (e.g. if user sets 20V, it will ask say 24V from the buck converter).
If the efficiency of the converter results in ~70%, I can get 50W at all voltages. For example  5V@10A or 10V@5A. [This is the main objective, and secondarily reducing output stage heat].
Now, to protect the main transformer I should limit the power provided. Surely I can limit the CC control depending on the output voltage, but I'd like to implement some hard limit on the buck converter, so I need to limit the max power output.
Ideas?

#### NiHaoMike

• Super Contributor
• Posts: 4619
• Country:
• "Don't turn it on - Take it apart!"
##### Re: buck converter power limit
« Reply #1 on: March 13, 2018, 01:52:06 pm »
Implement an input current limit.
Cryptocurrency has taught me to love math and at the same time be baffled by it.

Cryptocurrency lesson 0: Altcoins and Bitcoin are not the same thing.

The following users thanked this post: pitagoras

#### pitagoras

• Regular Contributor
• Posts: 115
• Country:
##### Re: buck converter power limit
« Reply #2 on: March 13, 2018, 02:05:26 pm »
Implement an input current limit.
Right. But the current limit has to be set such that output_voltage * max_current = 50W.
So when output voltage is 10V the current limit has to be set to 5A, etc...
I can surely do it in software, the question is how to do it in hardware.

#### NiHaoMike

• Super Contributor
• Posts: 4619
• Country:
• "Don't turn it on - Take it apart!"
##### Re: buck converter power limit
« Reply #3 on: March 13, 2018, 05:45:24 pm »
Have a separate current sense for the input side.
Cryptocurrency has taught me to love math and at the same time be baffled by it.

Cryptocurrency lesson 0: Altcoins and Bitcoin are not the same thing.

#### T3sl4co1l

• Super Contributor
• Posts: 11850
• Country:
• Expert, Analog Electronics, PCB Layout, EMC
##### Re: buck converter power limit
« Reply #4 on: March 13, 2018, 07:51:52 pm »
Why not use the TL494 for an average current mode controller, get 80-90% efficiency, and with the space and money you've saved on transistors and heatsinks, add a couple nice chokes and caps for even lower output noise?

Tim
Seven Transistor Labs, LLC
Electronic Design, from Concept to Layout.
Need engineering assistance? Drop me a message!

The following users thanked this post: pitagoras

#### pitagoras

• Regular Contributor
• Posts: 115
• Country:
##### Re: buck converter power limit
« Reply #5 on: March 13, 2018, 09:54:29 pm »
Why not use the TL494 for an average current mode controller, get 80-90% efficiency, and with the space and money you've saved on transistors and heatsinks, add a couple nice chokes and caps for even lower output noise?
Thanks.
I'm noob at switching PSUs. Where do I start? here? https://www.ti.com/lit/an/slua079/slua079.pdf

#### capt bullshot

• Super Contributor
• Posts: 1248
• Country:
##### Re: buck converter power limit
« Reply #6 on: March 14, 2018, 12:18:58 am »
There are other ways to achieve your goal, e.g. limiting the input current of the buck converter (as mentioned),
but using this chip you can monitor and limit the output power of the buck converter: The chip provides you with an output voltage proportional to power, you can feed this back into your regulator circuit to limit the power:
http://cds.linear.com/docs/en/datasheet/2940f.pdf
They have some application examples showing regulated or limited power circuits - use this as an inspiration. IMO it's a bit of overkill, I'd go for input current limiting.
Safety devices hinder evolution

The following users thanked this post: pitagoras

#### pitagoras

• Regular Contributor
• Posts: 115
• Country:
##### Re: buck converter power limit
« Reply #7 on: March 14, 2018, 12:45:23 am »
Ohhhhh I see now. Thank you everyone. I'll explore all options. Thanks.

#### T3sl4co1l

• Super Contributor
• Posts: 11850
• Country:
• Expert, Analog Electronics, PCB Layout, EMC
##### Re: buck converter power limit
« Reply #8 on: March 14, 2018, 06:07:49 am »
Why not use the TL494 for an average current mode controller, get 80-90% efficiency, and with the space and money you've saved on transistors and heatsinks, add a couple nice chokes and caps for even lower output noise?
Thanks.
I'm noob at switching PSUs. Where do I start? here? https://www.ti.com/lit/an/slua079/slua079.pdf

Yes. The formulas and associated terminology (poles and zeroes) are probably too advanced for you, but understand that:
- The TL494 contains all the circuitry from current sense amp to driver,
- The error amps need to be compensated somehow, and this will be done by starting with an overly generous RC time constant, then trying smaller and smaller values until the output step response starts to become unstable,
- You have to add the voltage error amp yourself, probably a TL431 or TLV2372.  The TL494 has two error amps, but they're wired in parallel, not cascade, so one has to be disabled (tie +in to GND, -in to VREF).

Also, you may want TL598 for totem pole outputs, instead of TL494.  Or, just add a bootstrap gate driver IC.  (Hm, I suppose TL598 could be wired to a bootstrap supply, as long as your +V supply is 10-20V.  Maybe not too useful above that.)

Tim
Seven Transistor Labs, LLC
Electronic Design, from Concept to Layout.
Need engineering assistance? Drop me a message!

The following users thanked this post: pitagoras

#### David Hess

• Super Contributor
• Posts: 7919
• Country:
• DavidH
##### Re: buck converter power limit
« Reply #9 on: March 14, 2018, 07:44:11 am »
My current plan is to use a TL494 based buck converter, controlled by the psu controller (e.g. if user sets 20V, it will ask say 24V from the buck converter).

Why control the buck regulator with the controller?  Why not have it track the output?

The following users thanked this post: pitagoras

#### pitagoras

• Regular Contributor
• Posts: 115
• Country:
##### Re: buck converter power limit
« Reply #10 on: March 14, 2018, 07:49:19 am »
Why control the buck regulator with the controller?  Why not have it track the output?
Yes, agreed, rather that.

#### pitagoras

• Regular Contributor
• Posts: 115
• Country:
##### Re: buck converter power limit
« Reply #11 on: March 19, 2018, 04:19:35 am »
Ok I've been simulating an average current mode buck converter around TL494, as suggested by T3sl4co1l.
This is what I've come up so far.
It still lacks input current limit, and the output CC.

From the TL494 I'm using only one opamp as current error amplifier.
(btw this link helped me too: https://www.maximintegrated.com/en/app-notes/index.mvp/id/3939)
U2 is the current sense amplifier and U3 is the voltage error amplifier.
I use a simple booster (https://www.re-innovation.co.uk/docs/open-charge-regulator/charge-controller-project-power-switching/) around C5, D2, Q1 to triggering the N-MOSFET. In the simulation seems to work fine, we'll see...

One problem I found is calculating the compensation for max GCA (as in the app note), because it depends on the output voltage (which is variable in this case). I guess if will be sub optimal but still work.

I'll breadboard this version to understand a couple of things and measure real performance.

#### Rog520

• Regular Contributor
• Posts: 89
• Country:
##### Re: buck converter power limit
« Reply #12 on: March 19, 2018, 06:07:25 am »
I know you have some time invested now with the TL494....but it was designed for voltage mode control. If you want to do current mode, why not build around a more modern IC that was designed for current mode and eliminate a lot of the external parts that you're using? If you want to stick with TI, they have a bunch of them:

http://www.ti.com/power-management/non-isolated-dc-dc-switching-regulator/step-down-buck/buck-controller-external-switch/products.html#p1077=Current Mode

For more innovative stuff, have a look at Linear Tech's line:

http://www.linear.com/parametric/external_power_switch_buck_controllers

The following users thanked this post: pitagoras

#### pitagoras

• Regular Contributor
• Posts: 115
• Country:
##### Re: buck converter power limit
« Reply #13 on: March 19, 2018, 09:33:18 am »
I know you have some time invested now with the TL494....but it was designed for voltage mode control. If you want to do current mode, why not build around a more modern IC that was designed for current mode and eliminate a lot of the external parts that you're using?

Well, I was going to use voltage mode. In that context, I opted for TL494 because it's widely available from many sources, it's cheap, and has a high max frequency. For a PSU I'd like it to be easily serviceable for long time. Adding an input current limit solves my OP and this would end the thread.

But since T3sl4co1l suggested building a PSU directly in switching mode, and suggested average current mode (which I was not aware of), I was interested in learning more about it.
In my current understanding I'm not quite sure of what I'm saying, but average current mode is different from "plain" current mode, for the later there is no ramp generated externally for PWM,  as it's generated by the current sensing and "cut" by the voltage control loop. From what I read, plain current mode has several problems (variable slope compensation, noise).
For the avg current mode, there is an external ramp generator as with voltage mode. Thus using a TL494 is not that awkward for an average current converter as it includes the ramp generator (if my understanding is vaguely correct).
« Last Edit: March 19, 2018, 09:44:38 am by pitagoras »

#### T3sl4co1l

• Super Contributor
• Posts: 11850
• Country:
• Expert, Analog Electronics, PCB Layout, EMC
##### Re: buck converter power limit
« Reply #14 on: March 19, 2018, 01:19:08 pm »
I wouldn't say "plain", I'd say "peak".  That's where you're sensing inductor current, ramping up or down over time, and using that to switch PWM on and off.  (There's also hysteretic control, where the positive and negative peaks trigger turn-off and turn-on respectively.)  But yes, that's basically correct.

Tim
Seven Transistor Labs, LLC
Electronic Design, from Concept to Layout.
Need engineering assistance? Drop me a message!

#### Rog520

• Regular Contributor
• Posts: 89
• Country:
##### Re: buck converter power limit
« Reply #15 on: March 20, 2018, 04:36:46 am »
Well, I was going to use voltage mode. In that context, I opted for TL494 because it's widely available from many sources, it's cheap, and has a high max frequency. For a PSU I'd like it to be easily serviceable for long time. Adding an input current limit solves my OP and this would end the thread.

But since T3sl4co1l suggested building a PSU directly in switching mode, and suggested average current mode (which I was not aware of), I was interested in learning more about it.
In my current understanding I'm not quite sure of what I'm saying, but average current mode is different from "plain" current mode, for the later there is no ramp generated externally for PWM,  as it's generated by the current sensing and "cut" by the voltage control loop. From what I read, plain current mode has several problems (variable slope compensation, noise).
For the avg current mode, there is an external ramp generator as with voltage mode. Thus using a TL494 is not that awkward for an average current converter as it includes the ramp generator (if my understanding is vaguely correct).

You can certainly use the TL494 to achieve current limiting. You actually *can* use both error amps at the same time. They're arranged so that the amp demanding the lowest duty cycle takes priority. Normally you'd use one amp for voltage feedback and the other for current limiting. You could *probably* achieve current limiting on the input using the same error amp, using a sense resistor and differential amp with suitable filtering to condition the input...but the input current on the buck is quite choppy so you could run into some stability problems. Maybe set up a prototype and play around with it. That's the only way to know for sure.

Current mode really doesn't offer you any advantage in your application because you're going to have a linear post-reg set the current limit of the supply's output anyway, and that will give you the fastest transient response as well as some benefit with noise reduction.

Smf