0- 10v PWM OpAmp Problem

[ianscott]

Hey guys

Im making a arduino 0-10V controller at the moment with this schematic (attached).

But I'm only getting values between 4-9V  on the output where I really need a least a minimum of 1.5-1.8V if I can't get to 0V.

I'm using a LM348 instead of LM324 at the moment which I think is the cause of the problem (I'm using it because its all I have got  at the moment and Jaycar is closed on Sundays). Also the VIN supply is 12V.
 

What can I do to fix this problem?

Thanks
Ian

[codeboy2k]

To reach exactly 0V, you'll need a negative supply, and likely a larger positive supply, and a better opamp then either of those two opamps.

The LM348 says max output is +/-12V when VCC is +/- 15V.  So that means it can get to within 3 volts of the power rails.  It gets worse as output load goes down.

The LM324 output high voltage can reach Vcc - 1.5V and the output low voltage can reach to within 0.5V of GND.

So you see neither of these can reach the top rail or ground rail on their outputs, so you won't be able to use it to go 0 - 10V.  What you need to use is what's called a RRIO OpAmp (Rail-to-Rail on Input-and-Output) .  Then the input can go 0 - 5V and the output can reach your required approx. 0 - 10V.   RRIO opamps still can't reach all the way down to 0 volts when using a single supply, but some of them can get closer to 0 volts than others.

There are plenty of RRIO opamps.  I just went to Digikey, used the parametric search, chose RRIO, Vsupply range min 12V max 60V and, max Vos of 5mV and sorted by price, and I found the Micrel LMC7101, it can get to within about 300mV of either rail.  The Texas Instrument's TL971 is also a RRIO opamp,  has max Vos= 4mV, and can get to within 100mV of each rail.  Both chips are about $1.00 for QTY 1.   Both of these can operate with a 12V supply.

In this thread:
  https://www.eevblog.com/forum/projects/recommended-single-supply-rail-to-rail-jellybean-op-amp/msg448870/#msg448870
@poorchava recommends the LM358 and LM2904, both of which can swing their output to within 20mV of GND, but can only go to within 1.5V of the upper rail.  So if you need 10V out, you need to run the Vsupply at around 12V or more to give enough headroom. If you can get the supply to 12V or more, then you can use these much cheaper opamps, they are about $0.55 and $0.41 respectively.

That thread also recommends OPA192 but it will be expensive at $4 each.

[lukaq]

That thread also recommends OPA192 but it will be expensive at $4 each.

Or OPA4188

[Kleinstein]

The LM348 is not working with inputs and outputs very close to the rails.

The LM324 is OK with this, though the output can not deliver much current to GND below about 100 mV. A resistor at the output to GND might help a llittle. The LM2904 is essential a slightly lower quality (lower maximum voltage) version of the LM324, but still OK. Both are rather cheap - about the cheapest OPs you can get.  There is also the LM358 as the dual instead of 4 OP version.

If more output current and real Rail-Rail operation is wanted, something like the TS914 is a good alternative, still not that expensive, but a maximum of 16 V.

[Richard Head]

If you use an LM324 you can configure the spare op-amp as an RC oscillator and with some small steering diodes and a couple of capacitors create a -Vcc generator. I've never done it this way before but I'm sure it'll work.
Dick

[Zero999]

TS272 is good for this.

[Kleinstein]

The TLC272, TS272  (essentially the same from ST instead of Ti) and others from that family should be taken with a grain of salt: they may not behave well when input and output voltage is close to the neg. supply range at the same time. So for this circuit this may be a problem.

[Zero999]

The TLC272, TS272  (essentially the same from ST instead of Ti) and others from that family should be taken with a grain of salt: they may not behave well when input and output voltage is close to the neg. supply range at the same time. So for this circuit this may be a problem.

I didn't know about that. What sort of issues have you experienced with the TLC272?

[Kleinstein]

I personly did not havereal problems with the TLC272, except low diver strength at the negative side - but that was calling for to much without reading the DS: The datasheet (TI) for the TLC272 has special cures for the miniumum output voltage at low common mode voltage. So in application like a follower something like 300-500 mV might be the lower useful limit, especially if current in the mA has to flow.

I have not found this note for the TS272, but otherwise this is just the 1:1 replacement - so at least I would be careful and expect trouble.

The TS912/TS914 is much better in that respect, due to the stronger output stage.

[LukeW]

LM358/LM324 can get fairly close to ground on a unipolar supply, but they'll start clipping their output about 1.5V below the positive supply rail.

However, you should be able to deal with this by making that supply a little bit higher than the maximum output required - say 12V.

[tree]

What's your PWM frequency?
What output current are you looking for?

[Dave]

What's your PWM frequency?

Considering the corner frequency of that lowpass filter in the front (just 0.7Hz), the frequency of PWM really doesn't matter here.