Hello,
Can anyone check the schematic below and tell me if it’s fine (especially the part inside the ride circle).
Thanks.
Half the fun is figuring out what the circuit is, lol. I can’t read all the fuzzy part numbers though. It did make my brain hurt :roll:
It’s a 5V 5A switcher followed by 20F ultracap bank charger followed by some 5V 2A UPS switcher?
Q2 is P-channel mosfet? it’s drawn backwards, swap D and S. R4 will turn Q2 on during power up, which I don’t think you want. Run two switches instead of one following the other, as Q1 turns everything on/off anyhow.
You don’t want Q3 to act as a switch, instead a constant current source? R8 is big. I would add a 1.5 ohm emitter resistor on Q3 and lower R8 to 1k or so. Why a 5A switcher but only 0.4A charging current? Why not use another mosfet like Q2?
For the ultra capacitors, you can connect them in a bank but Maxwell recommends cell balancing somehow, as each cap isn’t identical (+/-20%) and you might get 2V and 3V split instead of the textbook 2.5V each. Usually you add small resistors across each cap and maybe tie the cap midpoints together too. BCAP0010 is 0.03mA min. leakage current. http://www.maxwell.com/pdf/uc/applicati … ancing.pdf
U3 Maxim uses 2.2uH for L2. Pin 28 ONB need to be low to enable it, not to +5V.
You need a fuse before D1, and D1 is only a 1 Amp part, a bit small. At 35V R1 will smoke. U2 needs ground.
Hi man thanks for your answer.
In fact, this is a 5V 20W UPS. It’s a 5V 5A regulator followed by 10F ultracap bank charger followed by some 5V 4A UPS switcher.
Q2 is a P-channel mosfet. In fact, I used this application note: http://www.maxim-ic.com/app-notes/index.mvp/id/553 to do the circuit so I dunno if I have to swap the D and S of the mosfet. Do I use another P-channel mosfet instead of Q3? The current handled by the ultracap is 3.5A. What do u mean by R4 will turn Q2 on during power up?
Do I add 4 parallel resistors to the 4 ultracaps (1 resistor for each cap)? How can I figure out their value? And why do I have to tie the caps midpoint? (see attached figure)
Reading the data sheet of U3:
-
Pin 28: Shutdown input. When ONB/ = high and ONA = low the device turns off.
-
L2 = 1uF.
If you tie the ultracap’s midpoints together, there’s no difference but then you only need 2 resistors instead of 4 and only one voltage point to check. The resistors need to be low enough to override leakage current differences and high enough to not waste any power. Maxwell recommends 10x-100x the ultracap’s leakage current (BCAP0010=0.03mA) so that’s 2.5V/0.03mA10 or 100 2caps = 416 ohms to 4.16k ohms. Try 1k ohm each.
I had the wrong datasheet for the MAX1709, so the other stuff is good 
If Q2 was flipped into the “usual” configuration, R4 would turn it on. But not to worry, from the Maxim app note 533:
“Q2 is a low-rDS(ON), p-channel MOSFET that drops only 60mV at 1A. Its connections-drain to the main supply and source to the uninterruptible 5V output-are backwards with respect to the usual configuration for p-channel high-side switches. The connections shown prevent Q1’s body diode from draining the battery when the main supply fails. Also, this diode conducts when the main supply initially turns on, which assures a gate drive sufficient to turn the MOSFET fully on”
Let us know how the circuit does, I’m not sure how long the ultracaps last compared to batteries but they look pretty good.
They even work from -40C to +65C and up to 150F! Wow, 150 farads at 193A short circuit current… I’m gonna ask SparkFun to stock some.
[edit] Oops I see they already sell supercapacitors: http://www.sparkfun.com/commerce/produc … cts_id=746
This is what I am trying to do with this hardware configuration:
-
If V = 5V => /RST = 1 => Q2 on => (Q1 and Q3) on and U3 off.
-
If V < 4.65V => /RST = 0 => Q1, Q2 et Q3 off => U3 on.
Is the configuration good?
Yes, it looks fine. You will have to tweak R8 for 0.4A charging current. Good luck
I didn’t got the point, 11K is wrong?
I’m saying your circuit’s ultra-cap charging current could vary depending on the transistor’s gain and you should measure it.
Math says Q3 should get about 0.4mA of base current (5-0.7V/11k). FZT788B has a current gain 400-1500, depending on temperature and desired saturation voltage. That results in collector (charging) current 0.16A to 0.6A
Still can’t see where is the problem do you mean I should vary the value of R8 so Ic could get a maximum value of 0.4A? Why for? Can you give the solution so I could see things better.
I think redwire is concerned with the inrush current into the supercap upon power up. He suspects that the transistor Q3 , with about 0.5 ma of current into its base when on, will be running in a linear mode momentarily and controlling that inrush. That current limiting would depend on the Q3 transistor beta (gain).
Maybe something else in the circuit, such as U2, limits the current upstream, so that exceeding what looks like a max rated 400 ma will not occur. I don’t know.
Reading all the posts I don’t think he meant that. He previously suggested to reduce the value of R8. Or maybe I am missing the point I don’t know. Anyway a solution could clarify everything.
This is a diagram of what the circuit is meant to do.
When you first power up the UPS, the ultracaps are empty and draw max. charging current. Things would “hiccup” or oscillate without any current limiting. Once U2 turns on Q3 the voltage would drop due to the empty ultra caps, so then U2 turns off Q3, voltage comes up… and this repeats… :roll:
In Maxim app note 533, they run Q3 in a linear mode (not fully switched on) by giving it just a little base current so the ultracap charging current doesn’t overload the switcher U1 and Q2. It’s okay to do this but transistor beta varies so much (300%) that I mentioned tweaking R8.
ElieGW, if you had a bag of 100 Q3’s, their current gain spec can be all over the place, from 400 to 1500. So R8 needs to be tailored for what part you have. Start with R8=11k. If it’s too slow to recharge the ultracaps, lower it. If it charges too fast, Q3 will run hot during charging.
I wasn’t sure how fast the UPS recovery time needed to be. I think 10F at 0.4A charges in about two minutes. If it was lower say 0.2A would 4 minute charge time be okay? At 1A charge, Q3 dissipates 5W which drops off as the caps charge, but is too much heat for Q3.
To charge the ultracaps we will need about 2min 40sec. So you are saying that I need to change the value of R8 based on the transistor’s beta value to be able to control the value of Ic who will charge the ultracaps.
For example if I have a transistor with beta = 600.
5 - 0.7 / 11k = 0.4mA
0.4mA * 600 = 0.24A
And I test if this value is good for charging the ultracaps. If not I change the value of R8 and repeat the same test.
Yes, you got it
Hello again,
Well I was wondering if I can have a 12Vdc rail beside the 5Vdc rail using this same schematic off course I will have to add/modify some components. Any idea of how to do it?
Thanks!