Output diode for constant current source

[anvoice]

Hi, I've got two 50A constant current sources like this: https://www.ebay.com/itm/50A-DC-step-down-power-supply-module-constant-voltage-constant-current-LCD/273810462533?ssPageName=STRK%3AMEBIDX%3AIT&_trksid=p2057872.m2749.l2649

The info page says to add a diode in certain applications to protect the CC source. I'm planning to power inductive loads (a few turns of 16 AWG wire, air core, at 50A or less if temperature gets too high), so I'm guessing I need the diode in series with the load. However, I'm not sure which to choose. For safety's sake, my guess is I should use something like a 100A diode, but most of the diodes I'm finding in that range are stud diodes. I don't know how I would connect the threaded portion to my circuit. Is there a common type of diode for this application that I can connect to (e.g. by soldering or with screw terminals)? Also, should I watch for any special parameters, like Vf or maximum rated voltage? Thanks in advance!

Edit: forgot to include link. Fixed.

[Circlotron]

I expect the diode would be across the psu output terminals in the reverse biased direction, similar to how you put a diode across a relay coil. In series with the inductor is not going to do anything much.

[anvoice]

Ok, that seems reasonable. That also changes what kind of diode I need, but I still don't know what parameters to look out for. Is it Vf or something else? What is the current rating I will need?

[Circlotron]

The diode current will be no higher than what gets inadvertently diverted of of the inductor.  Seeing the current is only momentary and probably occasional, I'd say no heatsink is necessary. Seeing your power supply is capable of 50 amps, its voltage capability is probably not high. I'd go for a T0-247 case Schottky with both halves paralleled. Voltage rating is whatever the max voltage of your power supply is plus a bit of margin.

[anvoice]

Thank you! The power supply is 800W, so at 50A it would be 16V max. What is the purpose of having two halves though (dual, common-cathode?), as opposed to a single diode?

[Kleinstein]

The schottky diodes often come as doubles. So to get a higher current rating one could use both halves in parallel. With good thermal coupling inside the same case this is usually OK, especially for transient use.
A single large diode would be OK too - dual diode may be easier to get.

[anvoice]

Ah, I see. I'll have to do some digging to see which I can source more easily.

[mikerj]

Note that it's very easy to get stability issues when driving inductive loads with a current source, and at these high current levels it could get destructive pretty quickly.

[anvoice]

Note that it's very easy to get stability issues when driving inductive loads with a current source, and at these high current levels it could get destructive pretty quickly.

Would you mind elaborating? I don't want to burn anything down.

The load is a coil of copper wire, air core. The application is to create a magnetic field in the middle of the coil, so high currents are necessary. I was hoping the current source would actually give me better stability than, say, a voltage source where the current would change when resistance changes due to temperature. What are the dangers of doing that?

[mikerj]

An inductive load will introduce a phase shift that will reduce stability of a current source, this is analogous to a voltage source driving a capacitive load.  If the total phase (including feedback within the current source) approaches 180 degrees at a frequency where the loop gain is greater than unity it will oscillate.  To prevent this either the current source should include suitable compensation or the load itself must be compensated.  If you are using a commercial current source it would be worth speaking to the manufacturer about your application.

Many years ago I worked for a company making railway track signalling equipment and they used a very high current source (100's amps) to drive signals into a length of railway track to simulation traction currents, and they blew that thing up a few times before they got the compensation right.

[anvoice]

An inductive load will introduce a phase shift that will reduce stability of a current source, this is analogous to a voltage source driving a capacitive load.  If the total phase (including feedback within the current source) approaches 180 degrees at a frequency where the loop gain is greater than unity it will oscillate.  To prevent this either the current source should include suitable compensation or the load itself must be compensated.  If you are using a commercial current source it would be worth speaking to the manufacturer about your application.

Many years ago I worked for a company making railway track signalling equipment and they used a very high current source (100's amps) to drive signals into a length of railway track to simulation traction currents, and they blew that thing up a few times before they got the compensation right.

I probably should have mentioned that the current sources are DC. So phase shifts should be a non-issue for that particular part of the device.

I will also have an AC field perpendicular to the DC field, but that one will be nothing close to 50A. Thank you for the valuable info!

[Circlotron]

I think phase shift would be a result of not so much an alternating current or voltage, but simply a varying quantity.

Could a large capacitor be placed in parallel with the inductor to make it less inductive? Or would that make things just so much worse? Probably opening a can of worms here...

[anvoice]

I think phase shift would be a result of not so much an alternating current or voltage, but simply a varying quantity.

Could a large capacitor be placed in parallel with the inductor to make it less inductive? Or would that make things just so much worse? Probably opening a can of worms here...

Varying quantity of current or voltage? Doesn't DC mean it will stay the same, provided the power supply is doing its job?

About the second one I know next to nothing. If someone can weigh in on whether a capacitor could help here I'd appreciate it.

[Circlotron]

If your power supply is a constant current source it will vary the dc voltage to keep the current in the inductor constant with variations in temperature etc, as you said. Those variations in an otherwise constant voltage be viewed as an ac voltage with a dc offset.

[exe]

Doesn't DC mean it will stay the same, provided the power supply is doing its job?

There is noise from environment plus transient load (such as when connecting the load). DC power supply constantly counter-acts changes in the output voltage/current. Because of this, it needs proper compensation. Or, in other words, DC supply does a lot of job to maintain output stable .

Often circuits that are stable in Spice are unstable in practice. That's because Spice provides ideal environment without noise (unless explicitly added) and with ideal current and voltage sources.

[mikerj]

I think phase shift would be a result of not so much an alternating current or voltage, but simply a varying quantity.

Could a large capacitor be placed in parallel with the inductor to make it less inductive? Or would that make things just so much worse? Probably opening a can of worms here...

Varying quantity of current or voltage? Doesn't DC mean it will stay the same, provided the power supply is doing its job?.

An electronic current (or voltage) source is typically a closed control loop i.e. the output current will be measured, compared against some target value and then the output voltage adjusted accordingly to minimise the error.  This does not happen instantaneously, i.e. the adjustment is never exactly in phase with the measurement.  If you add an inductive load this delay increases even more since the inductor resists current change.  A fixed time delay equates to a phase shift that increases with increasing frequency.  At the same time loop gain falls with increasing frequency, but if you get enough phase shift before the loop gain has fallen to unity your current source turns into an oscillator. 

The fact that you are setting a DC current makes no difference, if you introduce enough phase shift it will very likely oscillate.  This is why I suggest you speak to the manufacturers regarding the application, they will (should) have an in depth knowledge of the performance of their device and should be able to tell you if your load is likely to be problematic.

[anvoice]

Thank you for the great info/suggestions everyone. I will try to get in touch with the manufacturer to see if they have any comments on this.

[Kleinstein]

The usual way to avoid oscillation with a current source driving an inductive load is to have an RC series element in parallel to the load or source. So not just a capacitor but with added loss in form of the resistor.

At the rather high power level of the source and the likely relatively slow switched mode current source this may need quite a large capacitor. How much is needed depends on the details of the current source and the inductance of the coil. One may be lucky and the needed damping is already inside, but I would not count on this.

[eblc1388]

First thing first, I would verify whether the power supply can really provide a stable 50A of current, simply by placing a direct short across its output terminals.