Let's start from square one...

[Sean_Goddard]

Deleted my comment.

[golfnut]

Boost switcher noise

" Turned out to be more difficult than I thought, as a majority of the switchers I looked at created enormous amounts of supply noise, especially at the high boost (low voltage) end."

Had same trouble in a commercial radio Tx design - The only way to remove the Tx modulation spurs from the switcher was to boost a little higher than required - then turn it off for the Tx duration and freewheel for Tx slot. This gave silent operation when I needed it.

It may be possible to synch the Boost to cut off for the Rx period - bit of a long shot. I had the luxury of 7 of 8 timeslots to recover the rail with the switcher on. S

[golfnut]

There is one more gotcha with Lipos and it is this. If you let a cell voltage go below their lower limit - they can swell on recharge and you need to dispose of them. Plus if you are real unlucky they can cause fire. S There are cheap dedicated chargers for lipos - Aero modellers are Lipo experts they were early adopters of these cells. If you need to know, ask an electric flight modeller.

[der_fisherman]

LiPo technology - cheap light and powerful.
Yep the only catch is dont over discharge them.

You are right, but there are chips around that do most of the clever stuff for you.....so it should not be a big deal.....

[6666]

LiPo technology - cheap light and powerful.
Yep the only catch is dont over discharge them.

[ODM]

Layouts are critical with switchers, as well as external switch components if there are any. Integrated ones are often nicer, as low gate charge mosfets for the efficient buck boosters seem to be a fickle thing to find in constant production. LT simpleswitchers will hopefully stay as a "staple" product for them.

Just wanting to stress the layout for the DIY building perspective, but plated 2/4 layer boards are growing to be pretty commonplace for hobby projects already.

[golfnut]

On Batteries Im a convert to LiPo technology - cheap light and powerful.

The orange one is 11v (3cell) weight 65gram
White one is 7v (2cell) weight 32gram

My next unit will have 2x 7v units for the op amps pos and neg rails.

S

[Carl-NC]

Resurrecting this is pretty good timing as I have recently done some work evaluating switchers, though for an entirely different project. I want to run a metal detector either from a 9V battery or a 2-AA pack, so I needed a buck-boost converter that would run from ~1.6V up to ~12V. Turned out to be more difficult than I thought, as a majority of the switchers I looked at created enormous amounts of supply noise, especially at the high boost (low voltage) end. But I did find that PWM-style switchers running in modest circumstances (even non-synchronous) pretty much have zero noise impact, at least on the detector I was testing. So I would feel a lot better using a switcher to simplify the power supply design. What I can say is, don't think for a minute you can choose a switcher based on the data sheet... you absolutely need to get a bunch of eval boards and do comparative testing to find one that works.

[ODM]

I'd be cautious about switchers with any measurement setup. It doesn't take an opamp "able to see" 1MHz or 10MHz or what frequency it happens to be, it can affect through nonlinear interference, think rectification. I would be especially wary with sampling systems, both having bumped into it personally a few times, and seeing it happen often while consulting at work even if we live in the era of "fun and easy fast low noise switchers"

On the bench there's always (hopefully!) some clean linear PSUs, and first rounds in the garden can be done with a bucket of batteries. I don't see the sense in designing a power supply until the rest of the circuit has been specified down, along with its power needs, at least for the majority of it? Starting out with a linear power supply will also allow establishing some known performance reference.

[GT Blocker]

Soooo....

Where are we since I was last on the board? Not very far, it seems

My BIG question is: Has any of this PWS problem been solved? It has been a few years now since the last post to this thread and the tech and times have moved on.

Tinkerer, BB Sailor, Reg, et.al. ... are there any answers to this question yet?

I am just finishing what I hope is the answer for my HHd. But the requirements for DSP and uPs sets a new bar.

Anyone care to fill in the blanks on this thread? We left this all hanging with no consensus (or even close to one).

I tend to agree with Carl about the linear thing. There are now extremely low noise linears on the market for audio apps. I do still lean toward the dual battery solution, although a solution with solid isolation between coil, analogs, and digitals with a single supply is not out of reach by any means (not to mention the re-charging nightmares of two different cappacity Li batts).


Jump in, 'gurus',

GTB

[der_fisherman]

Hi Andy,

the PSU is the heart of the metal detector. An efficient, low noise design is a must for a good detector.
I see that your recommendations are all for step down converters. In fact we need step up and inverters too.
Low noise is of extreme importance. An additional LC output filter would seem to help a lot in reducing the output ripple.

Monolith

Thats a fair statement.

Also, most of the chips use quite high frequencies, well away from the the frequencies that most MDs use... so with some good smoothing, not a problem!

regards

Andy

[Monolith]

(I know I am replying to an old thread!)

I believe you are talking about "Switched Mode Power Supplies" when you or someone mentions "switchers", am I right?

May I mention that a correctly designed SMPS has a very high efficiency, usually 85% or even more, which implies a far longer battery life for metal detectors. Or a smaller battery load to carry.

I personally use "Step Down regulator Chips", which are very similar, but not technically a "SMPS" as they use a DC input, but they are easy to work with and very efficient. But whether they are useful in a MD I cannot say, but it may prove interesting to look at as they handle voltages from 5 - 60 volts and the result looks REALLY clean on a scope trace.....

I am adding a .zip file with several different chips described, some of which are maybe useful for someone designing a modern MD, as here. Basically they are DC to DC converters, some can also step up as well.....

Regards

Andy

PS The ones I have used personally are the 5 amp LT1074, which I haven't included as its probably overkill for an MD.....but its also a great chip, which can be used in a dual circuit to deliver up to 10 amps.....if anyone needs more details, just ask.

Hi Andy,

the PSU is the heart of the metal detector. An efficient, low noise design is a must for a good detector.
I see that your recommendations are all for step down converters. In fact we need step up and inverters too.
Low noise is of extreme importance. An additional LC output filter would seem to help a lot in reducing the output ripple.

Monolith

[der_fisherman]

Switched Mode Power Supplies

(I know I am replying to an old thread!)

I believe you are talking about "Switched Mode Power Supplies" when you or someone mentions "switchers", am I right?

May I mention that a correctly designed SMPS has a very high efficiency, usually 85% or even more, which implies a far longer battery life for metal detectors. Or a smaller battery load to carry.

I personally use "Step Down regulator Chips", which are very similar, but not technically a "SMPS" as they use a DC input, but they are easy to work with and very efficient. But whether they are useful in a MD I cannot say, but it may prove interesting to look at as they handle voltages from 5 - 60 volts and the result looks REALLY clean on a scope trace.....

I am adding a .zip file with several different chips described, some of which are maybe useful for someone designing a modern MD, as here. Basically they are DC to DC converters, some can also step up as well.....

Regards

Andy

PS The ones I have used personally are the 5 amp LT1074, which I haven't included as its probably overkill for an MD.....but its also a great chip, which can be used in a dual circuit to deliver up to 10 amps.....if anyone needs more details, just ask.

[WM6]

Have you ever looked at the many advantages of separate TX and RX coils?

Tinkerer

Acording my experience all this advantages are more theoretical than practical.

Now I am trying to far avoid any PI schematic which does not allow mono coil.

[Tinkerer]

Sorry Tinkerer, but this sounds me like: why simple if we can complicate.

Mono coils are (comparable) still deepest in PI design.

You can build 5 different mono testing coils in time of building only one of others design.

At the end I can not imagine bigger PI coil in other than mono design.

Unshielded mono work very well too and deeper.

If there is some specific drawback by mono coil it does not outweigh advantages. Vote for mono.

Have you ever looked at the many advantages of separate TX and RX coils?

Tinkerer