Although doubling the transformer secondary voltage makes for a more useful Zener tester, capable of testing >15V Zeners, depending on the transformer's regulation factor, you are likely to be pushing the abs. max voltage limits of a LM317 regulator. A mains transformer's output voltage is normally specified at its full rated load current (or VA), and rises when unloaded or minimally loaded. For very small transformers the rise may be extreme, maybe even to double the nominal voltage. A LM317 is normally never operated with more than 35V across it from In to Out, and has an abs. max. limit of 40V. Allowing for the drop across the current set resistor, diode drops in the bridge rectifier etc. if the unloaded secondary voltage exceeds 29V, there is a significant risk of LM317 failure. There are a couple of alternatives here:
- Use a LM317HV, which is good up to 60V (abs.Max.)
- Use a JFET current source, good up to VDS(max) of the JFET (subject to power & SOA limitations)
- Use a BJT current source, similar to a JFET, but needs a stabilised base voltage
- Don't try to maintain a constant current, just use a simple dropper resistor!
The LM317HV offers best current regulation in a simple circuit. The JFET and BJT options are less precise, and tend to drift with temperature, + require the current set resistor to be initially adjusted to get the desired current due to device to device parameter variations. The simple resistor doesn't make any pretence at current regulation thus the Zener current will decrease with increasing Zener voltage. This isn't necessarily a bad thing unless you need to test if a particular Zener meets specification (for which you would need an adjustable current source anyway so it could be set to the specified Zener current) as higher voltage Zeners are typically specified at a lower current than the 5mA typical of <1W low voltage ones. If you go down this route, I'd choose a resistor that gave about 6mA with the probes shorted, E24 6K2 1/2W should do nicely. That will drop the current to 5mA at around 6.2V Zener voltage, and to 2mA at around 24V to suit lower current high voltage Zeners.Personally, I'd probably go 'whole hog' and use that transformer to build an 'octopus' curve tracer, or even an Arduino based digital curve tracer rather than a simple Zener tester, but if you've got thousands of unmarked Zeners to test and sort, YMMV.A 100uF capacitor is suitable if you reduce the power on LED current considerably by increasing its series resistor - the ripple at the expected Zener current will only be about half a volt pk-pk. Modern high efficiency red LEDs are plenty bright at only a couple of mA If. However if you keep the high LED current, you'll need at least 220uF to keep the ripple under 1V