One piece of equipment that the lab has been lacking is an isolated power supply. Such a supply is useful for testing circuits/devices such that they’re “isolated”, as the name of course suggests. A good overview of why one might do / needs to do this is provided on Dave Jones’ EEVBlog: EEVblog #279 – How NOT To Blow Up Your Oscilloscope!.

Overview

It’s not a very complicated device as shown by the circuit diagram in Figure 1. J1 is the input, J2 and J3 are the isolated outputs (any number could be provided) and T1 is a suitable transformer with a 1:1 winding ratio and which is specified for the supply voltage.

isolated power supply

Figure 1 – Circuit diagram of the isolated power supply

A thing to note at this stage is: the wiring of the earth pin on the outputs.

Typically you don’t want the output earth pin grounded to the same point as the input earth pin, because in certain configurations the isolated power supply becomes un-isolated. Saying this, some use cases do call for it. Often in isolated power supplies the output earth pin is simply left unconnected/floating.

The circuit in Figure 1 meanwhile has the output earth pin connected to to the centre tap on output winding of the transformer. The output supply is therefore a balanced isoalted power supply, which potentially has advantages in regard to noise rejection (if this was to supply an audio circuit for example) 1. The fact that the connection is to the centre tap of the transformer makes little difference, it’s just a reference, but it’s a different reference to the input earth pin.

Transformer selection

The criteria for the transformer to use basically depends on the following three points:

  • Power rating
  • Transformer type, i.e, toroidal, laminated core, etc.
  • Suitable isolation, per standard IEC 61558-2-4

For the design I chose a toroidal transformer because they exhibit lower external magnetic field and are typically smaller than a laminated core transformer for the given power rating. Specifically I chose the 1182M117 transformer from Hammond Manufacturing with a rating of 300 VA.

One point to note about toroidal transformers is their magnetic characteristics – they typically have a high magnetising current. This is even highlighted in the data sheet for the selected transformer and more on this point below(!):

Due to the superior magnetic properties of Toroidal transformers they will be susceptible to high magnetizing current when initially energized, only limited by the low DC resistance of the primary winding. Depending on where you are in the AC cycle when the transformer is energized dictates the chances of overloading the supply circuit. This is why the transformer may sometimes energize without a problem and other times it will blow the fuse or trip the circuit breaker. The duration of this overload is rarely longer than a half of a cycle. Therefore, you should consider using a slow-blow fuse, time delayed circuit breaker or other form of soft start circuitry for the supply line when using these high efficient Toroidal transformers.

Filtering and protection

I thought it appropriate to have fuse protection for the isolation transformer, and upon review of available receptacles there are good selections including inbuilt fuses and filtering. So why not? A little filtering never hurt anything apart from potentially the total parts cost, but the range of combo receptacles is pretty cost efficient. (If the isolation transformer is to be used for supplying medical devices then the specific filter used needs attention, because the IEC 60601 standard defines very exacting leakage current limits.)

This decision is now coming back to haunt the design… And it isn’t the filtering element, rather the fuse/circuit protection. The main problem is I ddin’t read the transformer datasheet well enough. Specifically the warning included (above) highlighting that high magnetising current was likely. I’m mean I read the data sheet. I read it multiple times. But I didn’t propoerly think through the implications…

The transformer has a primary DC resistance of 1.96 ohms per winding, with an overall DC resistance of 3.92 ohms. Meaning that the inrush current at switch on will be at least 58 A, momentarily, given that the supply voltage is 230 V. Also, the situation is at its “worst” when the system is unloaded. Crap… Even a slow blow fuse (of rating 2 A) is not going to survive this. And upon testing, it surely didn’t.

To be continued…

So what to do now? A current inrush limiter! And it would also be nice to confirm exactly what the input current actually is – therefore a mains voltage current measurement circuit! These both will be detailed in future postings.

1 For an overview of the advantages of balanced power supplies, see: http://www.equitech.com/articles/bpng.html