Diode sampler (again)

But they sure can make life easier. Of course, if you need this for mass production they'd have to be made in China or it'll get expensive.

I have made transformers from little ferrite cores that I had bought by the pound, a bag full for a few bucks.

[...]

--
Regards, Joerg

http://www.analogconsultants.com/

"gmail" domain blocked because of excessive spam.
Use another domain or send PM.

to come

can't

range,

how

bandwidth.

fast one

this

what I

breaking the

so on.

Transmission line transformers *are* pretty nearly magic, and they're usually easy to design too. For broadband impedance matching, the basic idea is to wire the coax in series at one end, and in parallel at the other--and use the ferrite core to choke out the inevitable DC shorts this causes. There's no net magnetization in the core, either, so it can be just about as lossy as it likes without hurting performance at all. You can get all sorts of ratios by doing series-parallel combinations.

I've never done a DC isolated one, though--for samplers, I gather you drive the two ends of the shield like a normal solenoid and take the output across the two ends of the centre conductor? That sounds pretty vulnerable to the inductance of the pigtails, so there must be a clever approach that I'm missing. Care to enlighten us?

Cheers

Phil Hobbs

l)

e:

rst) to come

hat I can't

on

ltage range,

his is how

bandwidth.

ne some

ly fast one

und this

get, what I

reaking the

and so on.

arge

cad for

ng in

aybe.

I

I can

ay be

r on.

ery,

ed

t

When it comes to electronics, we seem to agree much more often than not ...

Thanks for the European source of surface-mount core formers.

-- Bill Sloman, Nijmegen

Really a good book. I bought a copy a couple of years ago, and it was quite enlightening, especially the necessity of using a sampling loop, i.e. pre-biasing the sampling cap with the previous sample's value, which helps linearity considerably.

Cheers

Phil Hobbs

The trick is not to have pigtails. Where they come off the transformer they need to be tucked close together and possibly twisted (but sometimes you can't). The difference between a transformer with sloppy pigtails and the same one with clean connections can be huge.

--
Regards, Joerg

http://www.analogconsultants.com/

"gmail" domain blocked because of excessive spam.
Use another domain or send PM.

[...]

Your are welcome. My wife would be eternally grateful if you could send us some global warming in return, could need it right now :-)

No, no, this is a joke, no AGW thread. Cuz I ran out of popcorn yesterday ;-)

--
Regards, Joerg

http://www.analogconsultants.com/

"gmail" domain blocked because of excessive spam.
Use another domain or send PM.

to come

can't

range,

is how

bandwidth.

some

fast one

this

what I

breaking the

so on.

for

Most modern sampling scopes (ie, after about 1963) use a 2-diode sampler that has no transformers. A few of the slower ones used a

4-diode bridge sampler, which sometimes involved a balun in the bridge drive. There were also some 6-diode transmission-line samplers, no transformers.

You might find a transformer somewhere in the SRD drive circuit, especially the ones that used an avalanche transistor to drive the SRD. Unnecessary.

John

What grade of ferrite do you use for this stuff (or doesn't it matter)?

I guessed so, but it seems you need an equal positive and negative pulse with pretty sharp edges to get the sampler to work.

I have other projects to attend to (and money to make). I already bought an etching machine and so on to make some quick prototypes. This project just requires more time than I have right now :-)

--
Failure does not prove something is impossible, failure simply
indicates you are not using the right tools...
nico@nctdevpuntnl (punt=.)
--------------------------------------------------------------

Mostly 43 material, sometimes 61 or another type from the 60's range. A clean multi-filar winding technique often matters more than the core material though.

If you can get away with 43 this has the advantage that it is ubiquitous and cheap.

--
Regards, Joerg

http://www.analogconsultants.com/

"gmail" domain blocked because of excessive spam.
Use another domain or send PM.

Which is often more expensive to do than placing a little transformer. A transformer also has the advantage that it cannot introduce an extra DC error.

[...]

--
Regards, Joerg

http://www.analogconsultants.com/

"gmail" domain blocked because of excessive spam.
Use another domain or send PM.

I guess you're not concerned with the Q? Type 43 is too lossy for resonance above 1MHz, but we aren't talking resonance here, just a single pulse.

Is it that the losses are hysteretic, *after* your single pulse has got through unimpeded?

What does the core actually do in this case then?

I don't get ferrites yet.

For transformers that are driven hard Q doesn't matter much. I would not use 43 material for anything resonant.

I found distortion through such ferrite transformers to be negligible. It really would have mattered in ultrasound frontends but everything comes through clean. Else they wouldn't use them in communications receivers.

Mainly two things:

a. Stretch the bandwidth of the transformer at the lower end of the spectrum.

b. Keep most of the magnetic field inside. Old saying: "Toroids don't talk" :-)

Try some stuff, it's fun and it's cheap. But learn how to do multi-filar winding or you'll be frustrated with performance.

--
Regards, Joerg

http://www.analogconsultants.com/

"gmail" domain blocked because of excessive spam.
Use another domain or send PM.

) :

el)

te:

irst) to come

that I can't

on

oltage range,

this is how

e bandwidth.

ine some

lly fast one

ound this

get, what I

breaking the

, and so on.

harge

pcad for

ing in

y

maybe.

. I

I can

may be

er on.

very,

sed

d t

Thanks, I needed that.