PTH08T221W - 16A 88W w/ UVLO...

[mamu]

Features:
16A, 88W
Input: 4.5 - 14V
Output: 0.69 - 5.5V
Adjustable UVLO

2 versions - PTH08T220W and PTH08T221W. I chose to work with PTH08T221W (ceramic cap version).

Datasheet here: PTH08T221W

Tiny profile for a 16A module (23mm x 20mm x 9mm)...

Size comparison with the OKR-T10, OKL2-T20, and Raptor 20A...

So I go to solder wire to the pins to get it breadboarded and start by tinning the pins on the board. The instant I touched the pin with the soldering iron - the friggin pin loosened and tilted sideways. All of them, except pin 3, immediately loosened and tilted when I touched the soldering iron to it. gawd lol.

I spent the next half hour or so using flux and a desoldering braid to remove all solder in all the pin holes. It actually turned out to be an advantage that the pins came off because now I have a through hole module for wiring and with even a lower height.

Height between the through hole version compared with as it comes from the manufacturer...

All wired up but holy moly - see how close some of those onboard components are to the pins? This would not be recommended for someone new to soldering...

I'm not using Track, Sync, and TurboTrans. As per datasheet, connected Track to Vin, Sync to ground, and TurboTrans open. All other pins of this converter are in use and wired as shown in the above pic.

I connected Track on the front of the board...

and Sync to the back of the board... I lined the back of the board with several layers of Kapton tape first...

Added 3x 100uF (parallel) input and output ceramic caps to the back of the board... I'll either secure these caps with epoxy or change where I placed them before putting the converter in a mod...

Now breadboarded...

I'm using a 1K POT plus 4.7 ohm trim resistor for an output voltage range of 3.5 - 5.6V.

An NC fire switch is required to turn on/off inhibit - the inhibit of this converter requires a switch that must be closed in order to turn OFF output and open in order to turn ON output. That's the short layman's version - read the datasheet for the reason behind this.

I removed the NC switch and took the pic above with an NO tact switch - didn't have to press the button.

I'm not keen on NC switches plus there is a very limited supply and type. If I can't find a work around, what I'll probably do is use a FET to handle current and wire in a non-rated NO switch on Vin. The datasheet indicates that I should still have UVLO with this configuration with the appropriate resistor placed between inhibit and ground.

I've emailed TI tech support with some questions and am waiting to hear back.

Next step is to test the UVLO feature. I bought resistors to test at different voltage cutoffs - 6.0V, 6.2V, 6.4V and 6.5V.

 

[mamu]

Update: a work around for not having to use an NC switch and instead use an NO switch has been tested and is working aok. Thanks so much to MJarvis for his help in locating the schematic that shows this - and of all things with an Evercool board.

I should call this converter Evercool on steroids. Both datasheets have some sections almost exactly the same word for word.

Inverting the logic of the inhibit pin on TI PTR08100W

Now I have only 2 concerns with this, not with the N-FET, but with the 10K resistor. One is will the 10K interfere with the internal pull-up resistor (datasheet says an external pull-up resistor should never be used with the inhibit pin) and will this interfere with the UVLO monitoring at inhibit. I'll be testing and watching.

I didn't have the part listed in the schematic so instead used the N-FET I bought for this converter in case I needed it. Here it is breadboarded. Working like a champ - no firing when not pressing the NO tact switch and fires when pressing the switch.

 

[Tnumbers]

Ive been wanting to build one of these. If i ever get the money to would you be willing to help me out with the schematics you used?

 

[bapgood]

I got my stuff ordered yesterday

Using a n channel fet is exactly what I was thinking (even though I was reading the data sheet wrong for the TI board). Did you try it without the pull up resistor?

 

[mamu]

Ive been wanting to build one of these. If i ever get the money to would you be willing to help me out with the schematics you used?

It's still a work in progress.

I got my stuff ordered yesterday

Using a n channel fet is exactly what I was thinking (even though I was reading the data sheet wrong for the TI board). Did you try it without the pull up resistor?

FET with no pull-up resistor causes firing with and without pressing the switch.

FET + pull-up resistor with N.O. switch works, and no FET with N.C. switch works.

 

[ecat]

Now I have only 2 concerns with this, not with the N-FET, but with the 10K resistor. One is will the 10K interfere with the internal pull-up resistor (datasheet says an external pull-up resistor should never be used with the inhibit pin) and will this interfere with the UVLO monitoring at inhibit. I'll be testing and watching.

Looking at the datasheet for the PTH08T221W, I'm guessing you have wired something that looks like Figure 24 on page 24?

If your 10k resistor connects to the line that's labelled '1 = Inhibit' and Vin there should not be a problem, your pullup is on the gate of the fet and not on pin 11.

One thought that does come to mind: The output is inhibited when the fet is ON and enabled when the fet is OFF, what happens if you connect a resistor from pin 11 to ground ie in parallel with the fet? When the fet is ON pin 11 is effectively connected to ground, when the fet is OFF pin 11 would be connected via the resistor to ground as shown in figure 21. The on/off operation should still work and looking at table 7 on page 22 connecting a suitable value resistor should give you UVLO to protect the batteries from over discharge as well

 

[mamu]

Hi ecat! Thanks so much for helping.

I already have pin 11 wired with the UVLO resistor between it and ground as that's the only way to get UVLO regulation (see pic below).

Edited to make this the short version.

The converter is working aok with the N.O. switch how I have it wired with the NFET, but I am having a problem with UVLO.

The actual cutoff voltage is not matching the calculated voltage - at all. Example: I used a 28.7K UVLO resistor. Calculated cutoff voltage is 6.5v, but the actual cutoff voltage is 4.4v. I used a 15K UVLO resistor. Calculated cutoff voltage is 8.4v, actual cutoff voltage is 5.8v. There is a 0.5v hysteresis to account for some of this, but more than 2v like this is really off.

The results are the same when using a (FET + 10K pull-up resistor + N.O. switch) and (N.C. switch with no pull-up or FET).

 

[mamu]

After tinkering with various value resistors and seeing what actual cutoff voltage is vs calculated cutoff voltage, I think we're lucky that we have a narrow cutoff voltage range and that range seems to be in a linear portion because the fudge factor of calculated vs actual is the same for our cutoff voltage range. Outside of that range, the fudge factor changes.

So now I need to buy various UVLO 1% resistors in the range of what I want actual cutoff voltage to be. I was testing with the resistors I have on hand and some are 1% and some are 5% and some I had to series to get the value I was testing.

Off to mouser, then to bed. Maybe I can get some sleep now.

 

[mamu]

Three 221 versions are available - I ordered one of each from TI. Gotta love getting free samples.

 

[mamu]

Getting ready to put this converter in a mod. I chose a 6.4v low voltage cutoff (resistor sets unloaded cutoff). All the resistors, caps, and N-FET are tied to the board. Makes a nice small 1" x 1" package...

 

[jyjarvi]

Hi

I have also used this circuit and found that it is good and efficient for my mods. (and cheap TI sample)

- JJ

 

[jyjarvi]

Hi

here is the wiring diagram of PTH08T220 series circuit
the chart is written in Finnish, sorry about that

- JJ