3
\$\begingroup\$

I just finished designing my rectifier module that has a built-in voltage monitor IC and an eFuse IC.

Its purpose is to convert the 3-phase AC output from my wind turbine and use it to power my USB type-C charger.

My biggest hurdle with the type-C charger is that it only turns on with a hard start, and won't function from a soft start. Unfortunately, since the power is coming from my wind turbine, it's going to get nothing but soft starts.

So in order to get around this, I made this circuit to force a hard start, allowing the charger to work correctly, yet also allows a range of hysteresis (about ~1.4V) between turning on and turning off.

I chose to use a TPS3700 voltage detector to drive an SR latch, which controls an NPN transistor that switches the Enable pin (EN/UVLO) on and off for the TPS2595. The TPS3700 sets the SR latch when the rectified turbine voltage is greater than 5.2V, and then resets the SR latch when the voltage falls below 3.8V

If my TPS3700 is wired correctly(I think it is), the OUTA pin will go high when the input voltage is more than 5.2V (INA+ pin senses more than 400mV). And, the OUTB pin will go high when the input is less than 3.8V (INB- pin senses less than 400mV). I used a voltage divider to make sure the 400mV threshold voltage at INA+ and INB- pins are met at the correct time.

According to the datasheet for the TPS2595, the Enable pin (EN/UVLO) has a turn-on threshold of 1.2V and a maximum threshold of 7V. Because of this, I used a voltage divider to keep the voltage under 7V, yet high enough that it will remain over 1.2V throughout operation. I remember someone mentioned a Zener diode could be used to prevent the threshold from exceeding 7V, but I am unsure which method is better than the other.

Also, I designed my SR latch in reference to various schematics found on google, so I just chose the resistor values according to what I saw. Is there anything I'm overlooking as far as the SR latch layout goes?

Same thing goes for capacitor layout, I just want to know if there is anything I'm overlooking. I used three 220uF capacitors to smooth the rectifier voltage, and a 100nF capacitor in front of each IC to absorb noise (the datasheets said it's good to do this). Are there any other locations I should insert a capacitor?

Any help is greatly appreciated, as well as all of the help I've received in the past! Thank you!

Here is the circuit, I circled the SR latch and the transistor that it controls.: I circled the SR latch and the transistor that it controls.

TPS2595 Datasheet: https://www.ti.com/lit/ds/symlink/tps2595.pdf?ts=1678518905267

TPS3700 Datasheet: https://www.ti.com/lit/ds/symlink/tps3700.pdf?ts=1678532115159&ref_url=https%253A%252F%252Fwww.ti.com%252Fsitesearch%252Fen-us%252Fdocs%252Funiversalsearch.tsp%253FlangPref%253Den-US%2526searchTerm%253DTPS3700DDCR%2526nr%253D13

\$\endgroup\$

2 Answers 2

3
\$\begingroup\$

One issue I see in your schematic: The OUTA, OUTB of TPS3700 need pullups because they are open drain outputs.

The other point is w.r.t FLT resistor. It seems the max voltage recommended is only 6V. Please ensure this. Since you have a voltage divider on EN pin to ensure that the voltage is lower than 6V, I am guessing that you have to do something similar here also. Also, in my view, the FLT resistor should be connected to an always ON supply not the output of TPS2595.

\$\endgroup\$
1
  • \$\begingroup\$ Thanks for catching the lack of pullups. And I'll go over the FLT resistor again and straighten that all out. I'm a beginner who's just getting started, so I really appreciate the second set of eyes; thank you. \$\endgroup\$
    – soconnoriv
    Mar 27 at 17:55
2
\$\begingroup\$

Why design your own latch, though? If you feel like you absolutely must make your own, discrete gates could be used - cross connected NANDs from CD4011, or even the 8-pin CD40107. '107 has open-drain outputs, so needs pull-ups on the output.

One concern I have with such circuits is parts availability. They are great if you want something made pretty much once in 10k quantity, since you can preallocate parts in advance, run production, and are done - often the problem of sourcing falls to whoever will be dealing with it next.

But for - presumably - smaller quantities, and especially for one- or two-offs, and especially if the design needs to be long-supported - I'd go for parts you can buy anywhere. That means a double comparator-based UVLO+eFuse. And through-hole construction you can put together without having to get a lot of soldering practice first.

About the only requirement on the comparators is that they work from 5V and have rail-to-rail I/O.

Due to current sensing, M2 needs to have a "reasonable" drain-source resistance. I'm not sure what kind of load the USB charger presents in your application, but presumably an amp or a couple. As long as V(DS) under overload is more than 20mV, it'll work OK. It also should be a logic-level type, or at least one that will be reasonably well conducting at 5V V(GS).

schematic

simulate this circuit – Schematic created using CircuitLab

CMP1 establishes the undervoltage lockout. R4 provides hysteresis needed for stability and clean response. R5 is the current threshold adjustment.

C2 establishes the time constant of voltage detection C2*(R1||R2).

C3 establishes the time constant for over-current "cool down" period C3*(R3+R4+R5+R6).

D1 limits the sense input to the comparator to about 0.5V. D1 forms an AND gate of overcurrent shutdown with UVLO.

To use the venerable LM393 with its open-collector outputs, R8, Q1 and Q2 have to be added:

schematic

simulate this circuit

\$\endgroup\$
3
  • \$\begingroup\$ To be honest, I figured my own SR Latch would be lower cost, and wasn't aware of better ICs like the ones you mentioned. That's pretty much the only reason I went with my own, But now that you mention it, I really like how your first schematic has an adjustable hysteresis and current limit all in one package. Overall I'll go ahead and make a circuit with either the MCP6542 or the LM393 to try it out. Thank you for the help! \$\endgroup\$
    – soconnoriv
    Mar 28 at 1:45
  • 1
    \$\begingroup\$ What I mentioned aren't necessarily "better" ICs, just general purpose workhorse parts that can be used for "anything you can imagine", almost. When making something in small quantities - as I presume you do - your time is most expensive. A CD4011BE is $0.65 quantity one from DigiKey in the US. That's what I'd consider a minimal FF building block, since that's a chip everyone should have in a parts bin - it's super flexible. Next sensible step is using an actual S-R latch, but in this case that wouldn't be necessary. The eFuse you chose is an excellent part. Good luck soldering it though. \$\endgroup\$ Mar 28 at 2:24
  • 1
    \$\begingroup\$ If you also need overvoltage protection like TPS2595 provides, you'd just need one more comparator - so a quad package would do. Both devices are available in quad versions: MCP6544-E/P and LM2901N, in through-hole packages. The mosfet could also be through-hole for completeness, I just didn't look for one :) Another quad push-pull comparator would be TLC3704CN. \$\endgroup\$ Mar 28 at 2:28

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge that you have read and understand our privacy policy and code of conduct.

Not the answer you're looking for? Browse other questions tagged or ask your own question.