Hi again,
rshayes:
Actually after i drew it i noticed that too, and was going to change it
but was getting a bit tired at that point. Later, i took a look at the
possible current distributions and they didnt look too much different
so i decided to leave it alone. It is true that they will be better with
the opposite ends connections, but it doesnt make an extremely
huge difference, that is, enough to cause one strip to overheat.
It doesnt affect calibration at all either...when connecting the sense
leads to the center strip...the heavy leads can be connected either
way.
While i was looking over the mechanical aspect of this design i
also realized that it is a bit much anyway. The eight separators
have to be cut from 1/8 inch stock and have a 1/4 inch clearance
hole drilled in each one. I realized there is a much simpler way,
which is the subject of this next drawing.
This shunt only requires two strips of brass and a little bending
and the holes should be easy to drill too. The current distribution
is nearly perfect (equal in both strips). Of course the 1/4-28 bolts
should be well tightened.
To calibrate, unsolder an amp lead and move it one way or the
other to get the reading to be more accurate at some known current.
The dimensions for this new shunt come about from the resistivity
of the brass that i had obtained from a hobby shop quite a while back.
If the brass isnt exactly the same, the dimensions might have to
change a little unless the calibration makes up for it ok.
up to ~300A current sensor
Hi again Kurt,
Oh nice! That should certainly handle 60 amps right? he he.
BTW, if you use an ADC at 5v and you want to boost the 50mv
signal up to 5.00 volts, that's a gain of 100 (im sure you know)
and using the LM358 with it's offset spec of 3mv might give an
output offset error of 0.3 volts. Now while that's only a 6 percent
error at full scale, it will be present even when measuring 60 amps,
and at 60 amps that would be a 50 percent error.
To help this situation a bit usually a low offset error op amp is
chosen, or else some means to adjust the analog offset is incorporated
into the circuit. The low offset amps are the best though, if you
feel like getting one.
It may help if you have a uC as you can subtract a constant (0.3v)
from the output thus obtaining the true value. The downside is
that there may be a temperature drift to work into the algorithm too.
Of course it depends what accuracy you are after as well.
With my circuit one time i used even more gain (200) and that
gave quite an offset with only 2mv actual offset. The way i compensated
was to use a pot between +V and ground and used the pot arm as
my 'meter' ground. I was using an analog meter on the output however
so it was easy, with a uC this wont be as easy. I was also lucky that
the output had a positive offset so it was adjustable with the pot.
They also make pretty good amps that are made to have a pot
connected to them to adjust the offset too, and these are cheaper
than the super low offset type. It's up to you though, what you want
to spend. The super low offset type usually have a much more
limited bandwidth too in case you are thinking about measuring
currents in AC circuits as well as DC. Probably ok for 60Hz though.
Just a few things i thought you would like to know just in case you
were not aware of them. If you were already, then ignore this
post entirely
Oh nice! That should certainly handle 60 amps right? he he.
BTW, if you use an ADC at 5v and you want to boost the 50mv
signal up to 5.00 volts, that's a gain of 100 (im sure you know)
and using the LM358 with it's offset spec of 3mv might give an
output offset error of 0.3 volts. Now while that's only a 6 percent
error at full scale, it will be present even when measuring 60 amps,
and at 60 amps that would be a 50 percent error.
To help this situation a bit usually a low offset error op amp is
chosen, or else some means to adjust the analog offset is incorporated
into the circuit. The low offset amps are the best though, if you
feel like getting one.
It may help if you have a uC as you can subtract a constant (0.3v)
from the output thus obtaining the true value. The downside is
that there may be a temperature drift to work into the algorithm too.
Of course it depends what accuracy you are after as well.
With my circuit one time i used even more gain (200) and that
gave quite an offset with only 2mv actual offset. The way i compensated
was to use a pot between +V and ground and used the pot arm as
my 'meter' ground. I was using an analog meter on the output however
so it was easy, with a uC this wont be as easy. I was also lucky that
the output had a positive offset so it was adjustable with the pot.
They also make pretty good amps that are made to have a pot
connected to them to adjust the offset too, and these are cheaper
than the super low offset type. It's up to you though, what you want
to spend. The super low offset type usually have a much more
limited bandwidth too in case you are thinking about measuring
currents in AC circuits as well as DC. Probably ok for 60Hz though.
Just a few things i thought you would like to know just in case you
were not aware of them. If you were already, then ignore this
post entirely
LEDs vs Bulbs, LEDs are winning.
Perhaps I don't need to re-invent the wheel here...
Anybody know of a product like the AD8210 that has a gain of 100 instead of 20?
http://www.analog.com/UploadedFiles/Dat ... AD8210.pdf
Anybody know of a product like the AD8210 that has a gain of 100 instead of 20?
http://www.analog.com/UploadedFiles/Dat ... AD8210.pdf
Kurt - SF Bay
Hi Kurt,
Well, that's interesting but i think your output is a max of 50mv
and the device has a full scale of 250mv so im not sure if that would
be a good choice anyway. If you are still interested, you might
check the Zetex site however as they have other current to
voltage devices like that.
You might also consider just a low offset amp like the AD620,
with it's gain being set with one resistor. It is also spec'd for
very low linearity error so it may be a good choice. Downside
is that it needs a negative supply so you'd have to get one of
those negitive supply generators to get a negative voltage
from your positive voltage. This would provide plus and minus
12v to the AD620. Not many parts required however.
Once you get a low offset op amp like this the rest is easy.
Well, that's interesting but i think your output is a max of 50mv
and the device has a full scale of 250mv so im not sure if that would
be a good choice anyway. If you are still interested, you might
check the Zetex site however as they have other current to
voltage devices like that.
You might also consider just a low offset amp like the AD620,
with it's gain being set with one resistor. It is also spec'd for
very low linearity error so it may be a good choice. Downside
is that it needs a negative supply so you'd have to get one of
those negitive supply generators to get a negative voltage
from your positive voltage. This would provide plus and minus
12v to the AD620. Not many parts required however.
Once you get a low offset op amp like this the rest is easy.
LEDs vs Bulbs, LEDs are winning.
The AD8210 looks like it is a good device to use.
Switchboard shunts are usually calibrated for either 50 mV or 100 mV at full scale current. A 100 mV shunt would produce 200 mV peak to peak (depending on current direction). This, times a gain of 20 would give a 4 volt peak to peak output. This is a reasonable level for a device operating on a single 5 volt supply, since the signal can be offset and there will still be some margin between the maximum output swing and the supply rails.
With a 50 mV shunt the signal will be half as much, and could be offset to cover a range of 1.5 volts to 3.5 volts. The A/D converter can usually be offset and scaled to cover this range, and will probably operate better than it would with a rail to rail signal. A level of 1.5 volts would represent negative 500 amps, 2.5 volts would be zero amps, and 3.5 volts would represent 500 amps.
Switchboard shunts are usually calibrated for either 50 mV or 100 mV at full scale current. A 100 mV shunt would produce 200 mV peak to peak (depending on current direction). This, times a gain of 20 would give a 4 volt peak to peak output. This is a reasonable level for a device operating on a single 5 volt supply, since the signal can be offset and there will still be some margin between the maximum output swing and the supply rails.
With a 50 mV shunt the signal will be half as much, and could be offset to cover a range of 1.5 volts to 3.5 volts. The A/D converter can usually be offset and scaled to cover this range, and will probably operate better than it would with a rail to rail signal. A level of 1.5 volts would represent negative 500 amps, 2.5 volts would be zero amps, and 3.5 volts would represent 500 amps.
Hi again,
I agree that the 8210 looks like a decent device. The price isnt
too bad either.
One thing to be aware of however, is that a 500 amp shunt that
puts out 50mv could be rated for 500 amps peak, not continuous.
The continuous rating could be more like 300 amps. It's best
to check out the part number on the manufacturers site first.
At 500 amps and 50mv the shunt has to dissipate 25 watts,
and at 100mv it would be 50 watts, so it's probably a 50mv shunt.
Maybe a little forced air cooling at 400 amps would do it.
I agree that the 8210 looks like a decent device. The price isnt
too bad either.
One thing to be aware of however, is that a 500 amp shunt that
puts out 50mv could be rated for 500 amps peak, not continuous.
The continuous rating could be more like 300 amps. It's best
to check out the part number on the manufacturers site first.
At 500 amps and 50mv the shunt has to dissipate 25 watts,
and at 100mv it would be 50 watts, so it's probably a 50mv shunt.
Maybe a little forced air cooling at 400 amps would do it.
LEDs vs Bulbs, LEDs are winning.
I asked the folks at Analog about the AD8210 and apparently it won't work. They recommended the AD623:
I'm on Digikey looking up the 623 now...This is a great device but it doesn't look like it's going to be a great fit for you. There is a mandatory gain of 3 that accompanies this device and that would mean you are going to need an extremely small and precise sense resistor (on the order of 3 milli-ohms) Now if you can do this and don't mind its a fine device to use, but if your looking to use this at a gain of 1than your going to need to use an instrumentation amp and set it up as a sense amp. A device like the AD623, or similar. There are no DIP packages for these, the next largest package would be an SOIC. Hope this helps.
Kurt - SF Bay
Al,
In looking back at the question I asked the AD people, it wasn't entirely clear. So, we can't beat 'em up too much for not answering well...but we can certainly snicker about them not knowing their product came in a DIP package.
I noticed the AD623 DIP and ordered one from Digikey earlier this afternoon. Also got a TPS6735IP for the -5v I need.
Actually ordered two of each for when I fry the first one.
--K
In looking back at the question I asked the AD people, it wasn't entirely clear. So, we can't beat 'em up too much for not answering well...but we can certainly snicker about them not knowing their product came in a DIP package.
I noticed the AD623 DIP and ordered one from Digikey earlier this afternoon. Also got a TPS6735IP for the -5v I need.
Actually ordered two of each for when I fry the first one.
--K
Kurt - SF Bay
Kurt
From a quick look at the circuit you may need to look at the common mode voltage that is present when useing high side monitoring .
your 50mV is sitting on top of the 14V from the battery
you need to either have the monitoring circuit floating with respect to the main battery or you could consider monitoring on the return or low side.
hope this helps.
Colin
From a quick look at the circuit you may need to look at the common mode voltage that is present when useing high side monitoring .
your 50mV is sitting on top of the 14V from the battery
you need to either have the monitoring circuit floating with respect to the main battery or you could consider monitoring on the return or low side.
hope this helps.
Colin
Who is online
Users browsing this forum: Bing [Bot] and 143 guests