[topher]

I thought someone might be interested in seeing
pictures of one technique that can be helpful in
using parts prone to oscillation.

I put up some pictures of old projects I did on bare copper board. These are harder than using
dips on copper plated perf board, but you can see
how long to make leads and how to mount caps.

I put the pictures at:
http://www.theriault.org

Higher res pictures are at:
http://www.theriault.org/opamp.jpg
and
http://www.theriault.org/switch.jpg

 

[Tomo]

Hello,

Wait this is high freq circuit designed operated like in 100MHz~10GHz. This is not what you need to do for audio amps. You can just use normal circuitry except you need to damp out the high freq/RF. No need to do crazy things.

Tomo

 

[Joobu]

Nice pictures. I thought I was crazy.

What type of coax is that? I'm looking for flexible small diameter coax.


I was playing with a copper clad OPA642 in the lab today (350 MHz GBP) and I could not for the life of me get it to work properly. There were some bizarre things going on. With the input grounded, the output would display a nice step (<10ns rise time) then decay and then go beserk. My guess is stray capacitance.

I'm going to start with a new OPA642 and if that doesn't work I'm giving up high speed for a couple of months.

 

[topher]

Tomo,
This is exactly what guarantees stability when
using opamps with BW's over 50MHz. Chip caps to
a gnd plane and really short leads are what helps.

Joobu,
That's called semi-rigid coax. Teflon or
polyethylene dielectric(sometimes foamed or
"Gore-Tex") and a solid copper jacket for an outer
conductor. It bends easily with your fingers.

That problem kind of sounds like a marginally
stable layout. When it gets off 0 volts, one side
get more voltage and hence a bit more bandwidth.

You might try a DC input and a SMALL capacitively
coupled square wave.

Start the DC at 0 and slowly ramp by hand. The
small AC (mV's) lets you see what's up and
stimulates problems if they are there to be seen,
but keeps the bias point constant.

 

[Tomo]

Hello,

Yea, I got it.

So who is up to make ultra high bandwidth headphone amp? You probably need to really damp high freq or your oscillation can really cause problems.

My 350MHz 470V/usec amp sound pretty good. I wonder how it is descrete version

Tomo

 

[topher]

You liked the sound from that Eval board right?

Well, if someone wants to follow your lead, say
with a dip opamp, they can build it like this:

Get a perf board with copper on one side.

Use a drill bit to remove the copper around the
holes that the pins go through.

Do the same with most resistors and capacitors.

For the feedback resistor, solder it from output
to input just underneath the dip, on the
non-copper side of the board.

For the power supply pins, solder 1206
polypropylene's or 0805 NPO/COG ceramics from the
pin at the package to the copper of the board.

If it comes in a dip, this will be stable with no
load.

If it is not stable with a load, put a 10 to 50
ohm resistor between the output and the load, but
put it right at the opamp output..

If you don't want the output impedance that high,
do what the amplifier companies do. Put an
inductor in parallel with the output resistor.

 

[Joobu]

Ya, there's some really strange stuff happening with that op-amp, I'm going to try and stabilize it later. I has pretty extensive bypassing, four 0805 ceramics and two tantalums.

Does anyone of a good way to create a copper void in the middle of a copper clad board? I don't really have any proper tools but I've tried cutting and sanding to little avail.

I'm looking for some extremely flexible coax, ie bends with gravity. I wonder if that type of stuff is available or not.

 

[Tomo]

Tophor,

I forgot to mention that there are three types of THS4022 available. One with PowerPad, one without, and one manufacture dependent DIP. The max possible power dissipation is different for different types. In fact, the max power diss for PowerPad Version is 2.14Watt while the other is below 1Watt.

I will add this info in the THS4022 amp thread. Sorry.

Looking at the board makes me frinch in fear of messing up or burning myself. I guess I was not meant for doing wideband amps in this way. I will just make a board for wideband amp. Probably worth got cost for me, hehehehe.

Non-Edible Butter Fingers,

Tomo

P.S. No! I do not have peanut butter in my anatomy.

 

[topher]

Joobu,
I think I have seen people cut the copper with an
exacto knife and then peel the copper within the
cuts...

Usually, we just put a piece of tape down.

If you want to remove the copper near the inputs
to cut capacitance to ground, can you just totally
cut away the board there?

 

[Apheared]

Do you want a hole there too? Or just a spot to land? They make "pad cutters" from the olden days, they're drill bits that sand off a donut around the hole they just made... quite nifty, but it's for thru-hole. Hmm, what about something abrasive on something? Say a carbide tip on a dremel flexi-shaft? What do you have access to?

topher, nice amp stage. How do you hold stuff? At those junctions where liquifying the solder looks like everything would adjust, that is?

 

[Apheared]

forget all this silliness...

Where can I get one of these?
http://lpkf.herrlich-ramuschkat.de/a...uppeID=7&ID=24

a .1mm resolution board with zero chemicals, in minutes? as many as I want? as simple as File, Print?!

The things' gotta cost hundreds of thousands of dollars and break/wear out alot of tools... else why doesn't one of you have one and offering to help me make crap?

 

[Joobu]

Ya, I really need a Dremel. What attachments would I need to sand/mill the copper off at specific places?

 

[topher]

Joobu,
if you play with a .25 inch drill bit I think
you will find that you can remove copper pretty
well by hand. It's extremely easy if you remove
the copper around a hole in perfboard, but not
hard even if you don't have a hole.

Apheared,
Do you HAVE to show us stuff like that protmat?
The drool gets over EVERYTHING!

You hold stuff with tweezers. Every wire has to
be tinned and you just touch long enough to
reflow. That said, people smarter than I bend the
leads of the 1/8w resistors at the package with
needle nose pliers for a mechanical joint. I am
typically too lazy and every once in a while I get
a group melt. Which means I had the soldering pin
on waaaay to long.

liquid rosin flux helps.