Reverse polarity protection circuit & output backward flow control

Question

I am building my own (2 channel) lab bench. This is the conversion board I'm using to convert 12V (input) to 1-24V (one for each channel).

In the product description they say

Input reverse connect protection: No (If necessary, please input series schottky diode)
Output control flow backward: No (If used for battery charging or load is bring electricity load, please on the output side series schottky diode)

This does mean that there is no reverse polarity input protection (that is not needed anyways), but does this mean that there is reverse polarity output protection?

In the datasheet of the LTC3780 I found the following pin descriptions in the pinout:

SENSE+ (Pin 3/Pin 1):
The + Input to the Current Sense
and Reverse Current Detect Comparators. The ITH
pin voltage and built-in offsets between SENSE– and SENSE+ pins,
in conjunction with RSENSE, set the current trip threshold.
SENSE– (Pin 4/Pin 2):
The (–) Input to the Current Sense
and Reverse Current Detect Comparators.

Is this the reverse polarity protection feature and output backward flow control I'm looking for?

If yes, I don't have to design the circuitry to protect the output, if no, could you provide a solution which can accept variable voltages?

This is one solution I found very interesting due to the indicator LED and the 'no power loss' (claim).

Due to inefficiency of a schottky diode I am only going to use it if necessary.

Thanks in advance for helping me out!

jw0752 · Accepted Answer

The reverse polarity circuits are not designed to protect your bench supply. They are intended to protect what ever equipment you are hooking up to the bench supply. If accidentally you hook the positive and negative of the bench supply up to the incorrect terminals of a piece of equipment without polarity protection you may damage it. With polarity protection the piece of equipment would be OK. The polarity protection is part of the circuit or equipment you are experimenting with and not part of the bench supply. Here are the three simplest I could find. https://provideyourown.com/2012/reverse-polarity-protection-circuits/ John

michaelkellett · Answer

This board will make a truly dreadful bench supply - unless you only want it to work light bulbs or heaters then please think again. I recently helped a customer trouble shoot a new design where they had used the LTC3780 - the network connections would not work dues to power supply noise. I made two different versions of test boards, bought a board a bit like yours from Amazon and also tested the Linear Technology demo board - all of the boards were still very very noisy - quite unusable for a general purpose bench supply. The conversion efficiency that can be achieved with this chip is great - but the noise is a huge issue. We moved over to using buck only switchers on the original board (since we didn't have to have the LTC3780's boost capability) and this worked much better. For a general purpose bench supply you can't beat a decent linear design - just learn to live with the size, weight and heat MK

rachaelp · Answer

Michael Kellett wrote: For a general purpose bench supply you can't beat a decent linear design - just learn to live with the size, weight and heat MK You can also do a hybrid design to try and get the best of both worlds (with a little compromise). If you use a switcher to get the voltage to a few volts above your desired output and then use a linear regulator to step down to your desired output voltage, you get a reduction in switching noise because the power supply rejection ratio (PSRR) of the linear regulator is high without too much heat in the linear regulator as the switcher reduces the voltage drop across it. Best Regards, Rachael

michaelkellett · Answer

Hello Tim,

The noise is un-fixable by any reasonable means - it consists of bursts of very large amplitude 50-130 MHz decaying sine bursts as the power devices switch. We put at least 160 hours of engineering time, two pcb iterations and 2 test board designs into this. We added every possible external filter and tried countless different capacitor and inductor filter combinations. We had a 1GHz scope and a spectrum analyzer to help as well. We just could not get the power devices to switch cleanly.

The pictures are our test board, issue 1 and 2 (note the increase in filtering) and the cheap board from Amazon (which we tweaked to no avail). I don't have a picture of the official demo board. The best one is our own with the extra filters (450mV p-p at about 75 MHz) but the the high frequency noise couples into pcb ground planes and finds its way round filters. The LT demo board gave about 800mV p-p at 90MHz. In both cases the noise takes the form of short (a few cycles) bursts of rapidly decaying HF sine waves.

In the end we took the 'easy' option and completely re-designed the whole power supply

MK

michaelkellett · Answer

Probably not, but you can experiment for free.

Download the free LT Spice programme and do some simulation - there is a good model of the LTC3780 with suitable power MOSFETs.

You may well find that if you put active and/or non-linear components inside the feedback loop the regulator is unstable with some kinds of load.

Remember to simulate resistive, current sink and dynamic loads, as well as adjusting the output voltage over the working range.

MK

michaelkellett · Answer

Tim Koers wrote:

This is a similar board that I ordered. (see link in the first post). It's not just the LTC3780. Does this one have noise issues too?

Oh yes - every board I've seen with this chip has noise issues - the different layouts and switching devices alter the ringing frequencies a bit but the output noise is huge - the Chinese board from Amazon was the worst (and had other issues as well) but all 6 boards I've seen (including LT's own) were dreadful.

MK

jw0752 · Answer

Hi Tim, While the LTC3780 may support some protections the advertising indicates that there is no protection. Schottky diodes are relatively inexpensive and if you feel that there is any danger of reverse voltages being applied it won't hurt to put a diode in parallel for the input and in series for the output. You will only loose 0.4 volts on the output. There is no loss for the use of the polarity protection diode in parallel on the input. John

rachaelp · Answer

You basically cascade the regulators so you have Vin -> Switching Regulator -> Linear Regulator -> Vout and then you take your feedback from Vout and generate appropriate control voltages to the two regulators to ensure the switcher output remains a few volts above the final output voltage.

I think it's unlikely that you'll get a clean enough output for your supply by just trying to filter the output with adding capacitors. You could try to create an LC filter on your output to suppress the noise but really, if you are trying the create a nice clean supply the best way is to have a linear power supply design.

The board you are planning to use supports up to 10A or 7A for long term use but what current do you really need?

Best Regards,

Rachael

Reverse polarity protection circuit & output backward flow control

Top Replies