On a sunny day (Wed, 23 Nov 2011 08:39:31 -0800) it happened John Larkin wrote in :
This is sort of what I have been doing today, see my other posting, this PIC has several 10 bit ADC channels, and 2 hardware comparators. So I have now input and output voltages, I can also control the output voltage as I use the PIC DAC output as reference. It has a stable programmable internal reference too (1.024 V, x2, x4).
I like the RS232 interface, you do not normally see that on switcher chips :-)
Thank you. Yes, this will, once the final circuit is determined, go on those little eurocards with holes.
I have build a lot of stuff like this, spider web, we call it 'haystack', but I try to stay in 2 dimensions, even did a short wave receiver that took a whole table top, with CA3020 diff amps and CA3028 audio amp, or was it the other way around, XF9B crystal 9 MHz filters. Was a VERON (radio amateur club) design, I improved it a bit with better input stage. That thing was incredibly sensitive, took it apart again when it worked, just a fun experiment.
It looks really nice, dremel work? That is nice for RF.
My design is not optimized, and the 1 ohm resistor was added only for=20 convenience in simulation. I don't have the specs for what you are=20 designing, so I just wanted to see how a cheap, simple design compared. = I=20 think it may be useful for some purposes, especially when trying to use = a=20 single cell battery or very low voltage to provide enough voltage to = drive a=20 MOSFET and power 3.3V or 5V logic and PICs.
In that case, if these are really needed, your design is well justified.
I agree, and I chose my PIC design for similar reasons. But maybe even = more=20 important was my fascination with PICs and the fact that they can be=20 reprogrammed to achieve various different behavior without major = hardware=20 changes.
I need to check your other thread to see what the original design was, = along=20 with the requirements. I "jumped the gun" a bit.
I have not used SPI (yet), and I have not used that PIC, but that is = good to=20 know. Thanks!
either.
The design I was working on usually seemed OK, but sometimes "burped". = It=20 was a high power flashlight which used 12 VDC from an SLA battery, = driving a=20 bank of either seven or thirteen high power white LEDs at about 3W each. = So=20 battery current could be as high as 5 amps, and output voltage could be = as=20 much as 60 volts, and this was on a tiny 1" x 2.5" PCB. I monitored = battery=20 voltage, output voltage, and (primarily) output current. I also had a = low=20 and high brightness setting which was selected just by switching input = power=20 on and off. There were to be no other controls, as this was a sealed = unit=20 for diving purposes.
I worked on this design on and off for a couple of years, and my = customer=20 always seemed to find new LEDs that were brighter and/or more efficient, = so=20 his specs kept changing. I started with a PIC16F684 design, then, on = advice=20 from a friend with more experience with switchers, I made a design using = an=20 LT1247. Finally I made a third design using a PIC16F616, but still had=20 problems. Then when the economy tanked the demand for such diving = equipment=20 dried up and the project was put on hold.
Now my customer is using a design, made by a Chinese company, that is = much=20 more complex and (I think) not as good, but they did it for cheap and he =
seemed happy enough when they finally got it working. But he came to me = many=20 times complaining that they didn't understand what he wanted. I think = part=20 of the problem is that it is difficult to make a 40 watt boost converter = in=20 a volume of 2 cubic inches, especially with efficiency greater than 85%.
It becomes much more difficult when it must be boosted from a low = battery=20 voltage of about 11 VDC to an output of 55 VDC, using a single inductor = (and=20 not a step-up transformer). The peak currents through the inductor,=20 capacitors, and MOSFET became very high, and I think there may have been =
some RF interference into the PIC that caused erroneous readings from = the=20 ADCs that caused instability. It was also a learning experience for me, = and=20 not exactly my area of expertise.
I can post a schematic and PIC code if anyone is interested. And I have = a=20 bunch of PC boards and parts if anyone can use them.
I would be interested to know how you arrived at your efficiency numbers. I have run your circuit and I cannot get the same results.
I used the LTSpice ability to get the power in R2, averaged it, then got the (negative) power in V1, averaged it, then divided output by the input. I got about 69.8% with V1 at 3V.
"John S" wrote in message news:jajot1$qg2$ snipped-for-privacy@dont-email.me...
numbers.=20
I have changed the circuit a bit now, but I run the simulation long = enough=20 for the output to stabilize (5-10 mSec), and then I select the last 5 = mSec=20 or so for the calculation. I use I(V1)*V(IN) for input power, and=20 I(R3)*V(Vout) for output power.
I just ran the new circuit at 2 VDC with 5.25 Vout (292 mW), and = efficiency=20 is 81.8% between 10 mSec and 15 mSec.
At 3 VDC I got 88.3% efficiency with 5.49 Vout (341 mW). At this voltage = the=20 circuit operates in burst mode.
At 1.5 VDC I got 73% efficiency with 4.75 Vout (226 mW)
It even works down to 0.7 VDC with 1.78 Vout (31.4 mW), at 62% = efficiency.
I'm sure it can be optimized to do a bit better, and an actual circuit = may=20 not do as well, but I have built a similar circuit and it worked OK. I = even=20 built a more powerful version using a MOSFET and a gate driver, and I = think=20 it worked OK down to about 6 volts and as high as 16 volts, which were=20 beyond what I needed for the 12 VDC battery application. Adding a good=20 voltage reference and an op-amp with some hysteresis in the feedback,=20 controlling the MOSFET driver, should give tight regulation and high=20 efficiency, at the cost of some output ripple.
But mostly I think this circuit is handy to boost nominal 1.5 VDC to = about 5=20 VDC for logic and maybe 8 VDC for MOSFET gate drive.
Here is a variation of the circuit, which puts out a (roughly) constant=20 current of about 2 amps at up to 35 watts. It uses only 15 parts, = including=20 the load, which I represent as a 15 volt zener instead of a string of = high=20 power LEDs. I read efficiencies of about 87% from 10 to 15 volts, but = the=20 output actually drops at higher voltage. This circuit needs optimizing, = but=20 shows the possibilities for a very simple design.
He has supplied something of substance especially compared to your posts of almost indecipherable verbiage. What have you posted of any value other than your ill-informed criticism? Please point to something.
I thought there were some problems with the design, but I was not = motivated=20 enough to analyze it. I think it's cool to have some of the features = that=20 the PIC offers, but the basic circuit seemed overly complex for whatever = the=20 specs may be. I looked at a previous thread on "Favorite boost = converter"=20 but there were no details. But if you really just need a power supply = with=20 normal specs, there are many available for cheap. I just bought an = LM2596=20 step-down converter on eBay for $3.25 including shipping:
formatting link
ame=3DADME:L:OC:US:1123
The same company has many other small boards available cheap, such as = this=20 LM3577 boost converter for $0.99 + $4.90 shipping:
formatting link
h=3Ditem3a6d0b77c2
This one is rated for an input of 0.9 to 6V, and an output of 3.3-9V, = for=20 $3.25 and free shipping.
formatting link
rs&hash=3Ditem3cbed8d3fc
I love to design my own circuits and add special features, but when you = just=20 need a particular solution for a need, there's usually something = available=20 better and cheaper than you can do it yourself. Faster? Not so much.=20 Shipping from China is about 3 weeks.
ElectronDepot website is not affiliated with any of the manufacturers or service providers discussed here.
All logos and trade names are the property of their respective owners.