Hi All,
I've a project on the go at the moment that needs a 5v power supply, its got a 16f1827 at its heart (enhanced mid range is good !). Its diving a coil and getting as much current through the coil as possible is important. As you can't fool ohms law I would like a 5v supply. It the past I would stuff a 7805 on the board and give it a nice battery. I think a maximum of 100mA should be enough (25mA sink/source)+ a bit.

I would like to use a small li po battery I got for another project but never used, however its 3.7v 560mAh. I can't seem to find any 5v stuff the same size (physically). This all fits in a nice wrist mounted enclousure.

The question is do I use a DC/DC converter chip to drive it up to 5v? Anyone have much experiance of using these or should i just search harder fro a suitable battery?

Cheers Geoff

 

The battery voltage depends on its chemistry, so there will not be much variation.

Your chip can work at 3.7V as well.

For battery-powered project you an get PIC16LF1827 (note "L" in the middle) which runs at 3.3V and consumes much less power.

 

Thanks for reply, I do need my pic to run at 5v , I tried to run it at 3.7 on the bench supply and it works but not as well as at 5v.
I was looking for any pitfalls if boosting voltage etc. Not done it before so any advice is welcome.
Cheers Geoff

 

Thanks for reply, I do need my pic to run at 5v , I tried to run it at 3.7 on the bench supply and it works but not as well as at 5v.
I was looking for any pitfalls if boosting voltage etc. Not done it before so any advice is welcome.
Cheers Geoff

Aside for relatively low efficiency (70% I think?) Due to how boost converter works, and higher PCB size, I'd say no. unless you knowingly avoid all the safety measures to ensure stability in the circuit.

I will up the previous post, use a 3.3V chip such as the L version of your chip and simply implement an Undervoltage Lockout circuit to ensure you don't destroy the battery. I have used several 3.3V MCUs with zero issues at all. Wearables usually even go for 1.8V instead of 3.3V, just saying.

 

I wish I could, the pic is driving a coil, 5v is a must!
I need to either find a 5v battery with the right dimensions (like hens teeth!) or go with the boost idea?
Is efficiency as low as 70%? Mind you I haven't done the sums for a 7805 burning of the excess
Cheers Geoff

 

I wish I could, the pic is driving a coil, 5v is a must!
I need to either find a 5v battery with the right dimensions (like hens teeth!) or go with the boost idea?
Is efficiency as low as 70%? Mind you I haven't done the sums for a 7805 burning of the excess
Cheers Geoff

It's pretty low but not LDO levels of low. I'd rather use a 7.2 V battery and a decent Zener configuration to achieve the smallest PCB size combined with a decent decoupling capacitor.

 

WHY is 5 volts a must?

Why did you select that relay? Why not select a relay that works with the voltage you have?

 

I wish I could, the pic is driving a coil, 5v is a must!

You can get 3V relays too.

 

Sorry, should have said its driving a RFID coil not a relay, to get the read range I need 5v is about the minimum, I got a nice enclosure and battery from farnell here in UK.
I can't seem to find a 5v battery the right dimensions, hence the question about boosting 3.7 up to five in order to maximize current through the coil.
I know there will be losses but current design is a 9 volt battery and a 7805 in a not so nice enclosure so I was hoping to improve!
I've not used a DC/DC before and was just wondering what pits I may fall in?
Cheers Geoff

 

I've not used a DC/DC before and was just wondering what pits I may fall in?

3.7V to 5V switching regulators are common. Just use an appropriate inductor and catch diode.

 

I agree with others that the design is backwards. You should have selected the coil so that the whole thing runs off the li po.

A boost converter is an unfortunate complication for that mistake.

 

I'm intending to use a mc33063a from Texas, has anyone used it before? Its nice and cheap here , thanks dannyf a bit of criticism and no real advice.
I really don't think I have a choice on the 5volts, I've made a lot of coils in the last few years, there are a lot of factors that affect read range, but its mostly down to amp/turns, can't dodge ohms law.
Cheers Geoff

 

can't dodge ohms law.

I would be surprised if "ohm's law" actually applies here if it is an RFID coil.

 

Yes, if this is a 125KHz setup then the normal way to drive the antenna is to make it part of an LC circuit which is driven by an external signal produced by a microcontroller. The input is at 5V (though if you could use more, it might have longer range) and I measured one once, and found the highest voltage between the coil and the capacitor was around 50V.

By the way, the 125KHz isn't a magic number. If your LC setup has a resonant frequency that's different by some moderate amount, it should work OK if you drive it at that frequency.

 

Thanks john,
That's the main reason I want 5volts, as you said, I could use more but the most the pic can handle is 5.5v. I have spent a fair bit of time tuning the lc and getting it resonant. I have seen 160v peak to peak from 9 volts on a previous setup.
Interestingly that didn't have the best read range, I think the bandwidth was too narrow.

I was hoping to have a discussion on DC/DC converters as opposed to RFID , but if I can learn something that's good!

Dannyf, when can ohms law not be applied to current flowing in a conductor? I know the ac resistance of my coil is different to the DC due to the skin effect, but its still resistance so surely v=ir is still valid. Therefore if read range is governed by amp turns the only way to increase it for a given coil is up the voltage? (So far as you can filter out the carrier anyway)
The tag needs a certain amount of energy transferred to it to start loading the reader coil so dropping the voltage drops the current therefore power.
If I've got it wrong please put me right, I'm here to learn!
Anyway the DC/DC chip should come tomorrow, will let you know how it goes.....
Cheers Geoff

 

Hi All,

The bits finally came yesterday and its working, sort of.
I used the circuit from the data sheet, here...

Substituting R2 for a 6.8k. With my bench supply set at 3.7v I have a 5v output.

But, I haven't used the optional filter so the 100kHz ish switching messes with my RFID, not a huge problem, just add the filter (I hope!)
The power consumption is quite another thing! My circuit running at 5v uses 46mA =0.23 w. Running from the step up dc/dc @ 3.7v uses 370mA =1.369w!!!!!!!!!!! 16% efficiant? and I thought 70% was bad.

Any suggestions for a different chip or circuit for this chip to improve things would be gladly received!
Cheers Geoff

 

Sounds like something is wrong with your step-up regulator!

1) build, test and debug your regulator as a separate block, measure the efficiency under DC conditions.
it's impossible (very hard) to tell what is going on when your load is something non-linear or pulsing like your RFID circuit.
Once you know its working correctly, then add the RFID stuff as the load.


2) choose a regulator configuration that runs at a frequency much higher than RFID.
The switching noise is going to screw things up if it's near the same frequency, and hard to filter out.

 

Thanks for the quick reply,
I think this will lead me away from that chip, the max oscillator frequency is 100khz, which is a bit too close to my RFID frequency.
I will have a play to see where all that power is going, load is still drawing around 46mA at 5v.
Any one got a favorite dc/DC chip?

Cheers Geoff

 

Just got round to having a look at this, no real news other than I added a resistor which draws 46mA ish from the boost circuit, it still draws around 370 mA from my bench supply set to 3.7 v, I will investigate this further but have been distracted by a new pic..... That's for another thread though!
Cheers Geoff

 

Another option is to run the PIC on 3.3 volts and drive the coil with an H-Bridge. This doubles the voltage across the coil. Note that the H-bridge must have MOS-FET's for the outputs to minimize losses. I think TI makes some that will work.