Isolator/Amplifier

This is the place for any magazine-related discussions that don't fit in any of the column discussion boards below.
User avatar
MrAl
Posts: 3862
Joined: Fri Jan 11, 2002 1:01 am
Location: NewJersey
Contact:

Post by MrAl » Mon Jul 30, 2007 7:17 pm

Here is the new circuit so far...

Image
LEDs vs Bulbs, LEDs are winning.

simf14
Posts: 24
Joined: Sun Jul 29, 2007 2:00 pm
Contact:

Post by simf14 » Tue Jul 31, 2007 9:14 pm

Al your the greatest!

Will give it a try thank you!!!!!!

simf14
Posts: 24
Joined: Sun Jul 29, 2007 2:00 pm
Contact:

Post by simf14 » Wed Aug 01, 2007 2:23 am

Sir Al, Is there enough safety margin (isolation) between the DA converter and the rest of the circuit. If the 358 is merely measuring differential in the two input legs can we put higher resistance on both input legs, or even the isolation amplifier we talked about earlier??? That DA converter is $130 and I want to protect it from stupidity on my part...

User avatar
Chris Smith
Posts: 4325
Joined: Tue Dec 04, 2001 1:01 am
Location: Bieber Ca.

Post by Chris Smith » Wed Aug 01, 2007 8:13 am

Most inputs to the op amp are quite safe and low power driven.

To increase the input sensitivity level of the op amp to a level of "almost no draw" , the FET op amp is substituted in place of your ordinary op amp.

Perhaps something closer to the TL -081, LF -351, EGC -857 but you would need to cross reference to see if they are correct for your app.

User avatar
haklesup
Posts: 3048
Joined: Thu Aug 01, 2002 1:01 am
Location: San Jose CA
Contact:

Post by haklesup » Wed Aug 01, 2007 12:48 pm

While it is true Op amp inputs are very low to no current, it is when the device fails and the output or supply becomes shorted to the input where the concern lies. Also miswires during construction or hot plugging of a cable are all reasons to include a few additional components.

See the modified schematic. I have added A 1M resistor to limit overstress current in the event of a shorted Op amp and I also added two clamping diodes. The 1N4001 will conduct if the signla goes below ground and the zener will shunt if it goes above 5V. These will protect the DAC in the event of a short on the other end.

Image

Oops, the zener is backward, just mentally rotate it 180

User avatar
MrAl
Posts: 3862
Joined: Fri Jan 11, 2002 1:01 am
Location: NewJersey
Contact:

Post by MrAl » Wed Aug 01, 2007 3:15 pm

Hello again,

hackle, if you modify the schematic please put a note as to such.
My s name is on that drawing and i didnt draw it.


Here's the new schematic with a protection network and other
protection features...
(NOTE: Dont change the 1N4148 diodes to 1N4001 type!)

Also, a typical full electrical isolation circuit is shown too.

Image
LEDs vs Bulbs, LEDs are winning.

User avatar
haklesup
Posts: 3048
Joined: Thu Aug 01, 2002 1:01 am
Location: San Jose CA
Contact:

Post by haklesup » Wed Aug 01, 2007 4:30 pm

Oops my omition on the title block.

That is a bit better. The Cap C1 provides a shunt for fast changing transients like ESD while the diodes provide clamping for closer to DC signals. Not sure why you added D2 but it only increases the forward clamp voltage to about 6.3V or so which is still good protection.

Why the 1N4148? Lower capacitance and faster switching?

Lower series resistance is needed so that the signal is not slowed too much after the addition of C1. Is that what you had in mind?

Finally, Why D4? to provide undershoot protection at the output?

simf14
Posts: 24
Joined: Sun Jul 29, 2007 2:00 pm
Contact:

Post by simf14 » Wed Aug 01, 2007 6:01 pm

Thank you Hackle and AL, I buy the beer if ever in CA!!!

User avatar
MrAl
Posts: 3862
Joined: Fri Jan 11, 2002 1:01 am
Location: NewJersey
Contact:

Post by MrAl » Thu Aug 02, 2007 3:17 pm

simf14:
Ok get a case, i'll be there tomorrow afternoon :smile:

hakle:
Yes, that's pretty much it. I had hoped the diode D2 would
help to increase the slope of voltage vs current through
the zener, but after a simulation i see it doesnt do enough
to make it worthwhile for that purpose. It does however
raise the threshold voltage for turn on and so that might
help with a real world zener to help reduce conduction
when the input goes up to the full 5 volts (which could
go to 5.1 perhaps).
The series resistance R5 limits current getting to all three
diodes in the event of a short on pins 3 and 8, while the
other resistor R4 limits current back to the DA converter +
terminal if either the zener is conducting (6 or so volts)
or D3 is conducting (minus 0.7 volts). No matter what
happens the current getting to the DA is going to be very tiny,
so 50k should be good enough for R4, and 50k for R5 means
pin 8 of the LM358 would have to go above something like
2500v to cause a problem. This should be good enough i think.
Yeah, the output diode D4 was added to protect against
reverse current if the output goes negative during circuit
turn off if the output rings, and since he said he was driving
a solenoid i thought it would be a good idea to have this catch
diode.
R6 is there to make sure there is some impedance for the 50k
to work into in the event that for voltages presented to
the + DA output that are more positive, just in case it has
difficulty handling reverse currents (as if it goes open).
With the DA going open and a short pins 3 to 8, the highest
voltage that can reach the DA is about 1/50 ^th of the zenering
voltage or about 6/50 of a volt. For negative voltages,
the diode clamps at -0.7 volts but the limiting action of
the resistor R4 keeps the current very low even for DA out
equal to 5 volts.
R6 might also might help keep the output of the DA more linear in
that it will be putting out some decent current rather then
microamps.
Yes, C1 is there to make the response to a problem on pin 3 very
fast, and also to reduce stray noise pickup.

The reason for making the diode D3 a 1N4148 diode rather
than a 1N4001 (or similar) is that the 1N4001 diode has
more reverse leakage, which affects the gain of the entire
circuit. The reverse leakage causes the gain to go down
which affects accuracy so for 5 volt input rather than
8 volt output it might be as low as 7.5 volts. This is
especially true because i was also going to say that it
might be possible to increase R4 if he was too paranoid
about using 50k. With a 1Meg resistor for R4 for example,
for a 5 volt input the output is only around 7.1 volts
using a 1N4001 diode. Using a 1N4148 diode avoids this
kind of problem. BTW, the theoretical accuracy of this
circuit im pretty sure is perfect in that a 5.000 volt input
produces a 8.000 volt output, but only if the leakage
of the diodes is zero.

What i am also wondering is if he has any feedback from the
output back to the computer, to monitor what is actually
happening with the physical process.
LEDs vs Bulbs, LEDs are winning.

simf14
Posts: 24
Joined: Sun Jul 29, 2007 2:00 pm
Contact:

Post by simf14 » Fri Aug 03, 2007 6:24 am

Here is a less desirable circuit that does have a feedback loop...


Image

simf14
Posts: 24
Joined: Sun Jul 29, 2007 2:00 pm
Contact:

awesome

Post by simf14 » Fri Aug 03, 2007 9:47 am

Perfect

Thank you so much Al. Really sweet.

The R3 gets a little warm, would it be as easy as just going to a 1.0 or 2.0 watt??

In case I want change the max output from 0-8V to 0-6V which resistor(s) would that be?

You have a friend for life in Orange County Ca. / [email protected]
Free Root Canal!!! :smile: :grin:
Curtis Out

User avatar
haklesup
Posts: 3048
Joined: Thu Aug 01, 2002 1:01 am
Location: San Jose CA
Contact:

Post by haklesup » Fri Aug 03, 2007 3:07 pm

Thanks Mr. Thats just the kind of discussion about component selection I find useful.

2500V protection is a bit optimistic though. At that voltage the Diodes will fuse open pretty fast but in reality you only need good protection to 120V or so if a power supply shorts to a DC output (worse case in most projects).

The most essential protection component is R4 to limit overstress current into the DAC but if the application were different and the DAC had to output more current, you would be limited to using only the diode clamps.

As an alternative to a zener, one could also use the 1n4148 again but instead of grounding it, connected to VCC. Now any voltages greater than VCC+0.7V or so will be shunted to VCC (instead of ground)

So he can adjust R6 to change the apparent load the DAC sees without changing the gain of the circuit. One might make a case for 50 ohm depending on the DAC output specs.

The Gain: R1+R2/R1= 8/5 (is that it, I kinda guessed :) )

User avatar
MrAl
Posts: 3862
Joined: Fri Jan 11, 2002 1:01 am
Location: NewJersey
Contact:

Post by MrAl » Fri Aug 03, 2007 11:02 pm

Hi again,

simf14:
If the resistor gets a little too warm then yes simply raise the power
rating.
To get 6 volts out at 5 volts in, change R1 to 10k rather than 30k.
You can also use R1 as a calibration adjustment by making it a pot.
The circuit should be pretty accurate as is after adjustment, with
the only problem coming up at very low inputs of 100mv or less
when the offset of the cheap op amp starts to have a bit of an
affect on the gain accuracy, but i dont anticipate you needing high
accuracy at very low voltages like that. In case you do, check
out the other thread "op amp needed" where haklesup recommended
using an LT1077 op amp that has very low offset. That op amp
would probably work for your app too if you really need it.
Free root canal? Sounds good :smile:

haklesup:
Yes you are right, 2500v is not correct. 250v is a better limit.
The 1/4 watt resistors probably cant take more than that, i think
that's what you meant. The diode will be forward biased so it
will only drop 0.7 volts or so even with -250v at pin 3. Sound ok?
Because the LM338 is in the feedback loop the gain of the entire
circuit can be calculated as a regular non inverting buffer amp,
which is:
G=R1/R2+1
Note that's the same as the typical non inverting op amp circuits seen
in many manufacturers design notes.

I was thinking about that:
As an alternative to a zener, one could also use the 1n4148 again but instead of grounding it, connected to VCC. Now any voltages greater than VCC+0.7V or so will be shunted to VCC (instead of ground)
The only thing i didnt like about that idea was that the zener is there for
protection if the circuit goes into a failure mode, and a possible failure
mode might include Vcc going to a very high value, in which case the
diode wouldnt conduct which of course means the DA would not be
protected. The zener is like it's own little reference, which doesnt
depend on the failure level of Vcc.
LEDs vs Bulbs, LEDs are winning.

User avatar
haklesup
Posts: 3048
Joined: Thu Aug 01, 2002 1:01 am
Location: San Jose CA
Contact:

Post by haklesup » Mon Aug 06, 2007 11:58 am

The diode will be forward biased so it
will only drop 0.7 volts or so even with -250v at pin 3
No, only if the 250V supply had no more than 1A to give (current limited), then it might collapse to 1V or less. In the case where the supply can deliver lots of current, the voltage drop across the diode will obey ohms law.

To simplify, a diode acts like a low resistance resistor once forward biased. 1 to 30 ohms is reasonable for most diodes. Say 2 ohms was Rf for your chosen diode. at 250V sustained, you would have almost 125A. Most diodes would fry instantly. Fortunately we are not making a rectifier. Add a 100mA fast blow (instrument) fuse between the protection diodes and the op amp and at any current level, you should be OK, the fuse should blow faster than the diode but don't be suprised if the diode blows too. Fuses on the DC and AC supplies won't hurt either (both not much more than what is needed to run the load and circuit).

Well, spec sheets don't give Rf because its not linear, they just spec Vf at the max rated current. For the 1n4001, that is 1.1V and 1.0A implying Rf is about 1 ohm at that point. For the1n414 the max current is 300mA and Vf is 1.0V at 100mA.

Even the current in a Zener must be controlled so that you don't destroy the device. Zener currents are generally lower than permitted forward currents. The 1n752 is rated for 65mA max and has a zener resistance of 11 ohms (more current, further from zener voltage).

In reality, protection to the VCC voltage of 12V is more often what is needed. In this case a combination of a fuse, a current limited supply and the protection components (series resistor, ground clamp and supply clamp diodes) will give you very good protection of the DAC output.

User avatar
MrAl
Posts: 3862
Joined: Fri Jan 11, 2002 1:01 am
Location: NewJersey
Contact:

Post by MrAl » Mon Aug 06, 2007 3:24 pm

haklesup wrote:
The diode will be forward biased so it
will only drop 0.7 volts or so even with -250v at pin 3
No, only if the 250V supply had no more than 1A to give (current limited), then it might collapse to 1V or less. In the case where the supply can deliver lots of current, the voltage drop across the diode will obey ohms law.

To simplify, a diode acts like a low resistance resistor once forward biased. 1 to 30 ohms is reasonable for most diodes. Say 2 ohms was Rf for your chosen diode. at 250V sustained, you would have almost 125A. Most diodes would fry instantly. Fortunately we are not making a rectifier. Add a 100mA fast blow (instrument) fuse between the protection diodes and the op amp and at any current level, you should be OK, the fuse should blow faster than the diode but don't be suprised if the diode blows too. Fuses on the DC and AC supplies won't hurt either (both not much more than what is needed to run the load and circuit).

Well, spec sheets don't give Rf because its not linear, they just spec Vf at the max rated current. For the 1n4001, that is 1.1V and 1.0A implying Rf is about 1 ohm at that point. For the1n414 the max current is 300mA and Vf is 1.0V at 100mA.

Even the current in a Zener must be controlled so that you don't destroy the device. Zener currents are generally lower than permitted forward currents. The 1n752 is rated for 65mA max and has a zener resistance of 11 ohms (more current, further from zener voltage).

In reality, protection to the VCC voltage of 12V is more often what is needed. In this case a combination of a fuse, a current limited supply and the protection components (series resistor, ground clamp and supply clamp diodes) will give you very good protection of the DAC output.

Hi haklesup,

Im not sure what you are talking about. Are you looking at the most
current schematic earlier in this thread? There is a 50k resistor in
series with the diode and zener. Any current getting to that thing
has to go through that resistor (R5).
LEDs vs Bulbs, LEDs are winning.

Post Reply

Who is online

Users browsing this forum: No registered users and 45 guests