two quadrant costant current generator - paid project

[ddavidebor]

Hi,

We're looking for someone that can design a bipolar current generator, voltage controlled.
We need the tested schematic (and routing suggestions if needed) but we can design the PCB.

It must meet the following specifications:
- powered by +-12v (if you want to go single-supply we can discuss it)
- +-50mA max output current
- 0.5% or better precision
- controlled by a 0-2.5v or a 0-4.096v source. A few digital pins are available if you need them.
- stable over three twisted pairs 1.5m long cables in parallel

The following control methods are both acceptable:
- one single voltage source, with output Iout=(Vin-offset)*K. Example: 0-2.5v input. 0V in, -50mA out. 2.5V in, +50mA out. 1.25V in, 0mA out.
- one single voltage source plus a polarity digital input. Iout=Vin*K*(-PinBinaryStatus). Example: 0V in, digital in low/high, 0mA out. 2.5V in, digital in low, -50mA out. 2.5V in, digital in high, +50mA out.

Secondary design goals:
- to be easily configured to a +-3A version for low resistance loads swapping a few components.
- Provide an Altium schematic

Two design suggestion:

It’s a bipolar current generator, voltage controlled.

The circuit was drafted by a physics professor with 30 years of experience in analog design, but due to personal reasons he was been unable to complete the job.

Of course it’s not complete. It’s necessary to pick the transistors (probably darlington or maybe FETs), the opamp and to do the stability analysis and compensate the feedback network.

The wire connecting the load connected to ground (tagged with “S”) is about 1 meter long, so it needs to be well stabilized.

It needs to operate in DC mostly, so better stable than fast.

The circuit is thought to operate in a broad range, from mA to A depending on Rref.

The circuit is supposed to power

Translations of the sentences in Italian:
* misura di i (measure of the current)
* campione hall (hall sample. you can consider it like a resistor with 2m of wire)
* generatore di corrente… (bipolar current generator with the load connected to ground on one side)

Hand drawn schematic attached.

Use costant current dummy load to limit the current of a costant-voltage source, with a similar design as Dave's dummy load. https://www.eevblog.com/2010/08/01/eevblog-102-diy-constant-current-dummy-load-for-power-supply-and-battery-testing/
Connect the load to a H-bridge made with FET (jfet or mosfets) isolated with optocouplers, which can be used to invert the current polarity.

The current used by the FETs is negligible. A circuit like this could be easily adapted to a broad range of source and output voltages and currents.

This offer is from Fermium LABS srl, my company. https://fermiumlabs.com https://frm.li/datifatt

You need to be able to produce a regular invoice, especially if you are in the EU.

To apply write an email to d@fermiumlabs.com with subject "costant current generator" or something like that.

Davide Bortolami,
Fermium LABS srl

[timb]

Email sent!

[Berni]

There are some things missing in the spec:
   - What is the required bandwidth for the voltage input signal.
   - What does the load look like. Is it resistive, what resistance, does it have capacitive or inductive components to it, is it non linear?
   - What voltage swing is required on the output?

[mikerj]

There are some things missing in the spec:
   - What is the required bandwidth for the voltage input signal.
   - What does the load look like. Is it resistive, what resistance, does it have capacitive or inductive components to it, is it non linear?
   - What voltage swing is required on the output?

Also output voltage compliance.

[Berni]

Also output voltage compliance.

That's what i meant by output voltage swing, tho i will admit voltage compliance is the more correct term for a current source.

[ddavidebor]

There are some things missing in the spec:
   - What is the required bandwidth for the voltage input signal.
   - What does the load look like. Is it resistive, what resistance, does it have capacitive or inductive components to it, is it non linear?
   - What voltage swing is required on the output?

Hi,

In the current application, we need to flip the current at about 10hz. We do not plan to go above 100hz.

For the load, that's not an easy question. The load is a Germanium sample with imperfect ohmic contacts. It behaves mostly as a resistor of a few hundred ohms. There are some parasitic diodes on the metal-germanium contacts, if that may be a problem we can discuss it with our physicist.

The power supply we have available is +-12v. The most one can do with the available supply lines, the better. Using both supply lines at the same time would be great, but a 10v voltage swing is also acceptable. I'd like to discuss this on a case-by-case basis depending on what circuit configuration one intends to design.

Email sent!

Thank you.

Also output voltage compliance.

That's what I meant by output voltage swing, tho I will admit voltage compliance is the more correct term for a current source.

If by voltage compliance you mean the isolation voltage, this current source is the only in the system, with a purely passive load.
If you mean the maximun voltage swing of the current source, see previous answer.

[Berni]

The thing can be made using just a single opamp and some passives to get about 12Vpp of output swing at 3A or if two opamps are used and the load can be floating that brings it up to 24Vpp. Adding an extra opamp dedicated to current sensing could bring it up to about 36Vpp.

[ddavidebor]

The thing can be made using just a single opamp and some passives to get about 12Vpp of output swing at 3A or if two opamps are used and the load can be floating that brings it up to 24Vpp. Adding an extra opamp dedicated to current sensing could bring it up to about 36Vpp.

Hi,

Do you have an example schematic for such configuration?

[Berni]

The typical opamp current source cirucit from all over the internet can be used by carefully selecting the right opamp that can provide such a strong output. To double the voltage swing a inverting unity gain circuit can be used to drive the other side of the load in the oposite direction(Full bridge) and using a separate current sense amplifier lets you meet the accuracy spec while using a smaller current sense resistor in order to reduce voltage lost on it.

A fully opamp approach is easier to keep stable for driving strange loads as there is less circuitry to add delay to its actions on the output. Not to say that discrete power output stages are unstable, just takes more careful design to grantee stability. The extra effort is mostly just worth it if you are mass producing the product and want to optimize the cost by using cheaper transistors or you need a very powerful output such as 10A and up.

A complete schematic i would count as actual design effort since i tend to simulate things before building to make sure it actually works.

[ddavidebor]

Yes you're right, that is probably the best approach and the one we ended up going with.

I wanted to updated you all: In the end we gave the job to a local engineer. he's doing a good job.

I can't pay those who offered but I hope some constructive criticism may be useful:

- Asking for 2k/€ for an opamp circuit BEFORE trying to understand the exact requirements or give prove of your own competence it's not a great idea
- Giving more answer than questions (working on suppositions on the requirements) did not appear to be a great idea either
- A couple of people were more interested in conditions than the job to do itself. that was kinda of a surprise...

Thank you to anyone who made their offer