Posted pics in A.B.S.E.

Well it's not exactly what I was looking for as
it seems to have some kind of hysteresis, although the
bottom halves of the curves appear to be close.

I'm feeding a 400V transformer from a variac
the half wave rectifying it and then onto the test triode
via a 47K resistor. From the anode to ground I have a
100:1 divider chain so 409.6 V maxes out my 12 bit A/D
unit. From cathode to ground I have a 100 ohm current
sensing resistor hooked up at the same time to get a max
of 40.96 mA.

The grid is connected to an LM317 circuit
arranged to give from -1.25V to -20V so I'll have to modify
it later so I can get lower voltages.

Mark Harriss

"Mark Harriss" schreef in bericht
...

Posted pics in A.B.S.E.

Well it's not exactly what I was looking for as
it seems to have some kind of hysteresis,

Do both channels from the aquisition module
take a sample at exactly the same moment?
If not, 1 dot represents a value of Va and Ia
of the curve that are dispersed in time.

Try connecting both A/D channels to the same
point, f.e. the output of the 100 : 1 divider.
Then run a test. You should get a straight 45° line.
(if both A/D's have the same sensitivity).

As a real time alternative, connect an o'scope
in X-Y mode to both outputs.
If any hysteresis is visible, suspect your test jig to
have parasitic capacitance.
Hard to imagine though they cause any trouble
at the applied low frequency.

although the
bottom halves of the curves appear to be close.

If channels sampling is NOT simultaneous, the
curve's bottom halve is as useless as the upper.


I'm feeding a 400V transformer from a variac
the half wave rectifying it and then onto the test triode
via a 47K resistor. From the anode to ground I have a
100:1 divider chain so 409.6 V maxes out my 12 bit A/D
unit. From cathode to ground I have a 100 ohm current
sensing resistor hooked up at the same time to get a max
of 40.96 mA.

The grid is connected to an LM317 circuit
arranged to give from -1.25V to -20V so I'll have to modify
it later so I can get lower voltages.

That's easy, see page 3 of :
http://www.national.com/an/AN/AN-181.pdf


Mark Harriss

Rgds,
Jan.

"Mark Harriss" wrote in message
...
Well it's not exactly what I was looking for as
it seems to have some kind of hysteresis, although the
bottom halves of the curves appear to be close.

Something to do with the DACs maybe?

I'm feeding a 400V transformer from a variac
the half wave rectifying it and then onto the test triode
via a 47K resistor. From the anode to ground I have a
100:1 divider chain so 409.6 V maxes out my 12 bit A/D
unit. From cathode to ground I have a 100 ohm current
sensing resistor hooked up at the same time to get a max
of 40.96 mA.

For AX and AT I'd use 10 to 50 ohms. 100 is certainly a bit much for the AT
since its Gm is so much higher. You're only 5 to 10 times under a typical
bias resistor value...

The grid is connected to an LM317 circuit
arranged to give from -1.25V to -20V so I'll have to modify
it later so I can get lower voltages.

Connect the cathode in series with two biased diodes to bump it down
to -1.2V.

You might also try the current sense resistor in the return lead of the HV
transformer, this way you escape the NFB effects on the tube and can also
measure pentodes and other positive grid currents. And speaking of positive
grid voltage and 12AX7, I would love to see a complete set of curves for
that thing up to like maybe 20Vg or so!

Tim

--
"That's for the courts to decide." - Homer Simpson
Website @ http://webpages.charter.net/dawill/tmoranwms

Hi Gang,

I've mentioned it before, transistor curve tracers, such as the
Tektronix 576 and 577 work very nicely for tracing tubes. I employ
my 577 with my IP-17 bench powersupply from time to time. It sweeps
up to 1KV, and 100W... which handles most of the tubes we use quite
well. The IP-17 provides filament, and screen voltages.

Also, it is normal for curve tracers to have a trimmer cap that nulls
out the hysterisis like effect that makes the return trace form a
loop with the forward trace. Even at these slow speeds, parasitic
capacitance comes into play.

-Chuck Harris


Tim Williams wrote:
"Mark Harriss" wrote in message
...

Well it's not exactly what I was looking for as
it seems to have some kind of hysteresis, although the
bottom halves of the curves appear to be close.


Something to do with the DACs maybe?


I'm feeding a 400V transformer from a variac
the half wave rectifying it and then onto the test triode
via a 47K resistor. From the anode to ground I have a
100:1 divider chain so 409.6 V maxes out my 12 bit A/D
unit. From cathode to ground I have a 100 ohm current
sensing resistor hooked up at the same time to get a max
of 40.96 mA.


For AX and AT I'd use 10 to 50 ohms. 100 is certainly a bit much for the AT
since its Gm is so much higher. You're only 5 to 10 times under a typical
bias resistor value...


The grid is connected to an LM317 circuit
arranged to give from -1.25V to -20V so I'll have to modify
it later so I can get lower voltages.


Connect the cathode in series with two biased diodes to bump it down
to -1.2V.

You might also try the current sense resistor in the return lead of the HV
transformer, this way you escape the NFB effects on the tube and can also
measure pentodes and other positive grid currents. And speaking of positive
grid voltage and 12AX7, I would love to see a complete set of curves for
that thing up to like maybe 20Vg or so!

Tim

--
"That's for the courts to decide." - Homer Simpson
Website @ http://webpages.charter.net/dawill/tmoranwms

Well, how are you loading the ADC? You need something like a low leakage
opamp (i.e. not a 324) in the front end.

It also isn't a particularly good idea to use one of these things "bareback"
with high voltages as you can fry the serial input of your computer. While
this isn't necessarily catastrophic for a desk-top box, it can be a disaster
if you are using a laptop. DATAQ has a vague reference to this in their
product literature.

Jack

"Mark Harriss" wrote in message
...

Posted pics in A.B.S.E.

Well it's not exactly what I was looking for as
it seems to have some kind of hysteresis, although the
bottom halves of the curves appear to be close.

I'm feeding a 400V transformer from a variac
the half wave rectifying it and then onto the test triode
via a 47K resistor. From the anode to ground I have a
100:1 divider chain so 409.6 V maxes out my 12 bit A/D
unit. From cathode to ground I have a 100 ohm current
sensing resistor hooked up at the same time to get a max
of 40.96 mA.

The grid is connected to an LM317 circuit
arranged to give from -1.25V to -20V so I'll have to modify
it later so I can get lower voltages.

Mark Harriss

Hi,
Mark wrote:

Well it's not exactly what I was looking for as
it seems to have some kind of hysteresis, although the
bottom halves of the curves appear to be close.

While I haven't looked at the traces, this sort of problem is
mentioned in the Tektronix manuals as being due to stray capacitances,
and they went to some trouble to provide trimmers for balancing it
out.

73, Alan

John Walton wrote:

Well, how are you loading the ADC? You need something like a low leakage
opamp (i.e. not a 324) in the front end.

It also isn't a particularly good idea to use one of these things
"bareback"
with high voltages as you can fry the serial input of your computer.
While this isn't necessarily catastrophic for a desk-top box, it can be a
disaster
if you are using a laptop. DATAQ has a vague reference to this in their
product literature.

Jack


Hi Jack, I'm ashamed to admit it but I'm just using my DAC
bareback, albeit with a 100:1 divider chain which according to the DAC
manufacturer should allow it to survive up to 2KV. I then have the input
0-4V zener clamped to 5 volts. So something like a unity gain TL071 might
be a worthwhile buffer stage?.

Regards
Mark

Alan Douglas adouglasatgis.net wrote:

Hi,
Mark wrote:

Well it's not exactly what I was looking for as
it seems to have some kind of hysteresis, although the
bottom halves of the curves appear to be close.

While I haven't looked at the traces, this sort of problem is
mentioned in the Tektronix manuals as being due to stray capacitances,
and they went to some trouble to provide trimmers for balancing it
out.

73, Alan


Hi Alan, I just hooked up a resistor instead of a triode
and still get a similar squared off oval shape instead of a straight
line. How are the trimmer caps connected?. I guess I'll have to
hook it up to a CRO and see if I get a straight line.

Thanks for the info
Mark

Chuck Harris wrote:

Hi Gang,

I've mentioned it before, transistor curve tracers, such as the
Tektronix 576 and 577 work very nicely for tracing tubes. I employ
my 577 with my IP-17 bench powersupply from time to time. It sweeps
up to 1KV, and 100W... which handles most of the tubes we use quite
well. The IP-17 provides filament, and screen voltages.

Also, it is normal for curve tracers to have a trimmer cap that nulls
out the hysterisis like effect that makes the return trace form a
loop with the forward trace. Even at these slow speeds, parasitic
capacitance comes into play.

-Chuck Harris



Thanks for that avenue to follow up Chuck, I'm going to
eliminate DAC sampling problems first if thats the cause. Your
suggestion is next cab off the rank, how is the trimmer connected?

Regards
Mark Harriss

Mark --

the leakage current can back-bias the resistive divider -- I use LT1013 and
LT1014's from Linear Tech (TI also makes the same unit number), and some
other TI units. clamping diodes are a good idea as you utilize. i have
also used Agilent analog opto-isolators -- these are one of the neatest
products to play with -- i am trying to figure out how to implement them in
another power supply.

tubes are by their nature noisy, you might want to put a routine which
repeatedly samples a data point and take the average

i couldn't tell from the data sheet whether you can "start-stop" the
data-logger. if you go to the DATAQ website, www.dataq.com there is a link
to a third-party software provider -- done some communications control work
for them i guess. you may be able to employ their routines with the unit
you have.

i will continue to work with StampPlot --you can use it with any device
which has a serial out.

Jack

"Mark Harriss" wrote in message
...
John Walton wrote:

Well, how are you loading the ADC? You need something like a low
leakage
opamp (i.e. not a 324) in the front end.

It also isn't a particularly good idea to use one of these things
"bareback"
with high voltages as you can fry the serial input of your computer.
While this isn't necessarily catastrophic for a desk-top box, it can be
a
disaster
if you are using a laptop. DATAQ has a vague reference to this in their
product literature.

Jack


Hi Jack, I'm ashamed to admit it but I'm just using my DAC
bareback, albeit with a 100:1 divider chain which according to the DAC
manufacturer should allow it to survive up to 2KV. I then have the input
0-4V zener clamped to 5 volts. So something like a unity gain TL071 might
be a worthwhile buffer stage?.

Regards
Mark

oh, and while i am at it... learn how to work with surface mount devices --
life get's a hell of a lot cheaper when you can buy 1K of high quality
opamps for $30, or 100 Linear Tech ADC's for $9.99.

the other thing to remember is that standard carbon resistors are non-linear
above a certain voltage -- for your divider you can series several
resistors, or use precision wire-wounds.

with respect to stray capacitances -- this was one of the "bugaboos" of ham
radio linear amplifiers -- you had to null the stray capacitance elsewise
you'd have a 2kW oscillator on your desk.

Jack

"Mark Harriss" wrote in message
...
John Walton wrote:

Well, how are you loading the ADC? You need something like a low
leakage
opamp (i.e. not a 324) in the front end.

It also isn't a particularly good idea to use one of these things
"bareback"
with high voltages as you can fry the serial input of your computer.
While this isn't necessarily catastrophic for a desk-top box, it can be
a
disaster
if you are using a laptop. DATAQ has a vague reference to this in their
product literature.

Jack


Hi Jack, I'm ashamed to admit it but I'm just using my DAC
bareback, albeit with a 100:1 divider chain which according to the DAC
manufacturer should allow it to survive up to 2KV. I then have the input
0-4V zener clamped to 5 volts. So something like a unity gain TL071 might
be a worthwhile buffer stage?.

Regards
Mark