Bode plot measurement using only an oscilloscope

[Hammy]

For anybody looking for an inexpensive way to check BW and PM of your
PSU's I came across a method using just a scope and signal generator.
The test procedure is outlined in the application note below from
National.

www.national.com/an/AN/AN-1889.pdf

I don't understand why FRA's in the Hz to maybe 2MHz range aren't
available for a reasonable price. You could buy a nice used truck for
what Venable wants for his?

I know the accuracy depends on your scope but it should give you a
good idea and you can always check with a pulse load test for BW.

 

[Joerg]

For anybody looking for an inexpensive way to check BW and PM of your
PSU's I came across a method using just a scope and signal generator.
The test procedure is outlined in the application note below from
National.

www.national.com/an/AN/AN-1889.pdf

I don't understand why FRA's in the Hz to maybe 2MHz range aren't
available for a reasonable price. You could buy a nice used truck for
what Venable wants for his?

I usually do this with a network analyzer. Ok, not everybody has one but
many PC sound cards reach past 20kHz which is sometimes good enough.

I know the accuracy depends on your scope but it should give you a
good idea and you can always check with a pulse load test for BW.

You absolutely have to do a pulse load test. I still remember when I did
that on a "vetted" design last year because it didn't look quite kosher
to me. The results provoked words like "Oh s..t!".

 

[Hammy]

Time domain is a lot easier. Squirt a little bit of low-frequency
square wave into the sense node and scope the resulting output voltage
transient. As you note, you can get about the same information by
applying a square-wave load current, with is less invasive. All you
really need to know is response time and damping.

We have a Kikusui power supply load box, which can combine a static
load with a pulsed load. That's great for exploring the "load space"
dynamics. You can make the equivalent with a function generator, a
power fet, and a source resistor.

John

There are people who think you need FRA's to build reliable PSU's. My
thoughts were a pulsed load test should reveal any stability issues.
This is how I've always tested my PSU's.

For anyone else interested here is how.

"Overshoot as a Function of Phase Margin"

http://www.ele.uri.edu/~daly/535/margin.html

That and I leave them running all day at rated power in an elevated
ambient temp. I have no desire to burn my house down or anybody else's
for that matter.

The article was interesting and adds another inexpensive tool the box.

I still dont understand why they are so pricey? The HF VNA's for RF
frequencies yea but the ones that only go to a couple of MHz?

The cheapest one I could find is the Circuit Sleuth. Base model $5k.
0.1Hz to 40 MHz Frequency Range.

http://coretechgroup.com/products.htm

Even that is far more then what you need for testing PSU's.

 

[Joerg]

There are people who think you need FRA's to build reliable PSU's. ...

Those folks probably can't change a tire on their car and most likely
can't decently barbecue

... My
thoughts were a pulsed load test should reveal any stability issues.
This is how I've always tested my PSU's.

Yup. I cobble together the needed power resistors to create a max load
situation, a big FET and its gate goes to a Philips pulse generator.

For anyone else interested here is how.

"Overshoot as a Function of Phase Margin"

http://www.ele.uri.edu/~daly/535/margin.html

That and I leave them running all day at rated power in an elevated
ambient temp. I have no desire to burn my house down or anybody else's
for that matter.

That almost happened at a client. Someone else's design, and they left
it running overnight. Suddenly the Shepherd of the night guard began to
bark rather ferociously. Turned out the carpet underneath the power
supply had begun so smolder.

The only mishap I ever had was a tiny pinhole leak that sprung up it the
oil cooler pot for my load resistors. Well, it used to be a metal bucket
for honey and I guess it was too old. It made a mess, took a while to
clean that up, SAE10 everywhere.

[...]

 

[Joerg]

On Sun, 13 Dec 2009 09:47:02 -0800, John Larkin
[auto-snip]
There are people who think you need FRA's to build reliable PSU's. ...

Those folks probably can't change a tire on their car and most likely
can't decently barbecue

... My
thoughts were a pulsed load test should reveal any stability issues.
This is how I've always tested my PSU's.

Yup. I cobble together the needed power resistors to create a max load
situation, a big FET and its gate goes to a Philips pulse generator.

For anyone else interested here is how.

"Overshoot as a Function of Phase Margin"

http://www.ele.uri.edu/~daly/535/margin.html

That and I leave them running all day at rated power in an elevated
ambient temp. I have no desire to burn my house down or anybody else's
for that matter.

That almost happened at a client. Someone else's design, and they left
it running overnight. Suddenly the Shepherd of the night guard began to
bark rather ferociously. Turned out the carpet underneath the power
supply had begun so smolder.

The only mishap I ever had was a tiny pinhole leak that sprung up it the
oil cooler pot for my load resistors. Well, it used to be a metal bucket
for honey and I guess it was too old. It made a mess, took a while to
clean that up, SAE10 everywhere.

[...]

I used to smoke Douwe Egbert's Amphora...

http://www.flickr.com/photos/wiless/2466065784/

About a pound a week, until about 1993 ;-)

A pound a week? Yikes! As a student I had the occasional pipe with
"Scandinavic Mixture" in there but a small can lasted many months.

So I had an ample supply of tins for "load oil", and other bread-board
needs.

Way too small. Anything under half a gallon would have fried up in most
of my cases. I found that the honey bucket could take a kilowatt for
about 10-12 minutes before the metal handle would become uncomfortably
toasty and the whole thing began to smell. Sadly, recently I had to
throw it out because it was too far gone. To me that was like when a
little kid gives up his woobie ... <sniffle>

 

[Tim Williams]

Way too small. Anything under half a gallon would have fried up in most of
my cases. I found that the honey bucket could take a kilowatt for about
10-12 minutes before the metal handle would become uncomfortably toasty
and the whole thing began to smell. Sadly, recently I had to throw it out
because it was too far gone. To me that was like when a little kid gives
up his woobie ... <sniffle>

I use a five gallon pail of water for my induction heater tests. At 1kW
power output (which might be 300W dissipated), it hardly gets warm, with
only a gallon of water inside. As my 10kW tests progress, I suppose I'll
want to fill it up to two gallons or so. And as the duration increases, I
can patch in an oil cooler sized radiator. No biggie.

Water isn't handy if you've got live voltage on hand, but it's not
completely useless. If you've got aluminum body resistors, you can water
cool them quite nicely. I suppose you could seal a water jacket onto a
vitreous enamel resistor, that'd be pretty sweet (100W resistor will
probably handle over 1kW continuous?).

Tim

 

[legg]

That almost happened at a client. Someone else's design, and they left
it running overnight. Suddenly the Shepherd of the night guard began to
bark rather ferociously. Turned out the carpet underneath the power
supply had begun so smolder.

In coordinated development of source and platform, there are more
tales of the platform self-destructing, with the power supply
surviving, in my book. But then, power supplies are simple.....

RL

 

[Joerg]

I use a five gallon pail of water for my induction heater tests. At 1kW
power output (which might be 300W dissipated), it hardly gets warm, with
only a gallon of water inside. As my 10kW tests progress, I suppose I'll
want to fill it up to two gallons or so. And as the duration increases, I
can patch in an oil cooler sized radiator. No biggie.

I did think about that but didn't want to risk it with a pump because
all that could leak some day.

Water isn't handy if you've got live voltage on hand, but it's not
completely useless. If you've got aluminum body resistors, you can water
cool them quite nicely. I suppose you could seal a water jacket onto a
vitreous enamel resistor, that'd be pretty sweet (100W resistor will
probably handle over 1kW continuous?).

Problem with any sort of seal is that it inhibits a temperature exchange
to the outside. Meaning like my honey pail it heats up over xx minutes
and then needs forever to cool off.

 

[Joerg]

In coordinated development of source and platform, there are more
tales of the platform self-destructing, with the power supply
surviving, in my book. But then, power supplies are simple.....

Talking about platform self-destruction do you remember the *KABLAM* ad
from an EE magazine in the late 80's? Rocket test, turned sideways, hit
a building and exploded in a fireball. Li'l circle in the picture around
a speck of all the flying debris with the words "That's our power
supply". Further down they wrote that it fired right up on the bench.

 

[ehsjr]

Water heater elements. Series/parallel them to make a useful resistance.
If the energy dissipated gets pretty high, just run a garden hose in and
siphon the hot water out.

Maybe do a load of laundry with it.

"Dear, can I use the washing machine for a dummy load?"

"No! And who you calling 'dummy'?"

You should put your dummy load inside another bucket.
Ask me how I know. :-(

Ed

 

[Hammy]

legg wrote:

Talking about platform self-destruction do you remember the *KABLAM* ad
from an EE magazine in the late 80's? Rocket test, turned sideways, hit
a building and exploded in a fireball. Li'l circle in the picture around
a speck of all the flying debris with the words "That's our power
supply". Further down they wrote that it fired right up on the bench.

By comparision to today. Drop one two inches onto a rug and it rattles
like Santas bells, plug it in and phoof smoke!

Gotta love mass produced Chinglish shit.

You can always tell on forums who is from China.

Heres an example subject.

" Give me schematic art work bom for 5kw UPS NOW PLEASE URGENT"

 

[JosephKK]

Talking about platform self-destruction do you remember the *KABLAM* ad
from an EE magazine in the late 80's? Rocket test, turned sideways, hit
a building and exploded in a fireball. Li'l circle in the picture around
a speck of all the flying debris with the words "That's our power
supply". Further down they wrote that it fired right up on the bench.

Sure do. That was Abbot IIRC.