Boost converter wisdom

The EMC spec is DC-2MHz. ;)

Cheers

Phil Hobbs

Ah, now I understand where you're going. I didn't get that you had a linear after the switcher. I do that fairly often for fixed supplies where I care about ripple. I haven't had need for a variable supply, though.

So use two cap multipliers and two low-side switches as your DPDT switch.

Yeah, I'm familiar with that management edict, "fix it but don't change anything!"

So use a 2.2MHz switcher? How much current do you need?

I don't recall did you give us a set of specs? Output voltage(s) and maximum current(s)?

You do have to make sure that the class D waveform doesn't get as far as th e TEC. The TEC response is non-linear and cycle-to-cycle variation in the c urrent going through it generates more ohmic heating than a more nearly sta ble current.

Sloman A.W., Buggs P., Molloy J., and Stewart D. ?A microcontroller

-based driver to stabilise the temperature of an optical stage to 1mK in th e range 4C to 38C, using a Peltier heat pump and a thermistor sensor? ? Measurement Science and Technology, 7 1653-64 (1996)

used a thumping great toroidal common mode inductor with capacitors to filt er the class D switching component out of the current going through the TEC . I also messed with the switching waveform to minimise the lower frequency harder-to-filter components.

A couple of amps at maybe +-3V, depending on the TEC.

Cheers

Phil Hobbs

That should be easily do-able with a 2MHz FSw, which would get around your EMI problems (shouldn't be a big deal if it's laid out properly). You should be able to get away with 2520 or maybe 4-5mm inductors, so it should be pretty small.

Are there any high-current DPDT solid-state switches, or SSRs?

oller

nit

ts.

of our

re, as

erter.

of

he

t I've

so

lower

mum

something like this?

Something like that.

I was thinking of a tiny buck switcher (I like TPS54302) then a noise filter then the h-bridge into the TEC.

ntroller

g unit

ducts.

ad of our

there, as

onverter.

itch

ts of

o the

but I've

ath

the

and so

and lower

inimum

. I

er

ion

re's

e in

It seems a bit silly to have a set of four switches in the H-bridge and add more to drive a separate buck converter. If you use the H-bridge switches to activate say, a Cuk converter, and filter the output (whose polarity wou ld depend on which switches you activated) before it goes into the TEC you should end up with the same result from fewer switching elements.

That circuit would have to be designed, rather than evolved by glomming tog ether conceptually simpler elements, so it isn't surprising that John Larki n couldn't come up with it.

Sorry, I was unclear. That's the spec for the buck driving the Class H amp. Pop option #1 will be to run that off the +5V, but pop option #2 (hopefully) will be to run it off the boost converter, which will put out +12 to +14. That way I can use either kind of TEC.

Cheers

Phil Hobbs

mandag den 27. august 2018 kl. 01.41.55 UTC+2 skrev John Larkin:

ntroller

g unit

ducts.

ad of our

there, as

onverter.

itch

ts of

o the

but I've

ath

the

and so

and lower

inimum

. I

er

ion

re's

e in

that DRV8839 is a h-bridge and the datasheet specify 0-11V motor supply and separate 1.8-7V logic supply

Right. The biggest part would be the inductor.

Color me stupid and ignorant, but what eXplicitly wrong with using your standard, known-characteristics 24V supply?

It isn't available in the instruments they want to retrofit.

Cheers

Phil Hobbs