If I understand correctly, then having a transformer there may be a significant help from a safety standpoint. This is especially true if the transformer steps down the voltage.
I have not spent the time to try to reverse engineer the schematic and I do not know what is the function. For that reason I don’t know why a toroidal transformer would be preferable to one built with conventional EI laminations.
A few years ago I contracted with Microsoft as a test engineer. We bought a 1KVA toroidal transformer for mains isolation for some tests. But almost all of the time, when this transformer was connected to the mains, it would INSTANTLY (by human perception) trip the power service 15A circuit breaker. I fixed this by adding an inrush limiter. The problem is that the toroidal core has negligible magnetic gap and very low reluctance. The hysteresis loop is shaped such that the core can retain significant residual magnetic flux after the transformer is unplugged. If the phase of the AC mains is not quite right the next time power is applied, the magnetic core can saturate and (being a 1 KVA transformer) a huge current spike can result. By comparison, a transformer which is built with EI laminations tends to have more core gap and will have less residual flux so this problem may be less severe. Also a toroidal transformer will almost always be more expensive.
This transformer couples a carrier signal (~12 kHz) into the mains, which is likely why it is a toroid and not a conventional EI-core transformer. The core is about as big around as a US ‘silver dollar’.
Well the frequency certainly reduces the size of the transformer. But probably the main reason to go with a toroid is “self shielding”. But this is important: The self shielding characteristic is mainly “far field” and relies on the windings being evenly distributed around the toroid core. If you were to bunch the turns in 90 degrees of the circle (for example) you would get very little self shielding benefit.
I have made quite a few transformers for switching power supplies but those operate all above 20 kHz and mostly above 50 KHz. At 12 KHz you could use ferrite core but there may be some type of iron lamination which would work. For winding a transformer by hand, I have found that the easiest thing is to use a round center leg core and bobbin such as a PQ core. Ferrite would work at 12 KHz although some other material might permit higher flux density and a smaller finished transformer. I think that Digikey sells Ferroxcube ferrite cores.
Don’t forget about creepage across the body of the component from where the legs enter the body. Probably why @Naib mentioned flooding conformal coating (or my edit: potting compound) under the part…
Incidentally, I’m glad I have been making a conscientious effort to try and make my updated version of this piece of equipment safer than the original… there were a number of kludges and shortcuts taken in the original, one of which is the PC board set is ‘hot chassis’!
Once my copy is done, the original is heading to a museum where it belongs!
Anyway… I went to make an EPROM swap to validate the ones I’d burned were good. The unit was powered down but still plugged in… nope! Got a nice little ‘nip’!
(I unplugged the unit for the swap and finished what I set out to do afterwards)
So how old was this instrument? Old enough to have a hot chassis but new enough to have an EPROM?Was it commercially made?
When I was a child I had a tube type “portable” phonograph which was capable of administering shocks. But this thing was probably built in 1960 or earlier. Thinking about it, it probably needed a power transformer so that it could light the vacuum tube filaments but the chassis was probably coupled to one side of the non-polarized plug. (In the USA these days many appliances with 2-prong plugs have a polarized plug in which the neutral blade is wider than the hot blade. I am not sure when this practice started; pretty sure some time after 1960; maybe in the 1970s.)
The equipment dates to the mid-'80s. Given the brand name on the outside (not naming it), this major OEM really should have known better! I think the small target audience and the low production requirement of this equipment played a part in the shortcuts taken.
The following standards can be applied for calculating creepage distances for AC voltages:
IPC-2221 Generic Standard on Printed Board Design
IEC-60950-1 (2nd edition)
UL61010-1: Electrical Equipment for Laboratory Use
The formulae given in these standards are not available for free but you may opt for some useful calculators available online. You can directly use them for creepage distance calculations depending on the voltage of your design.
You need to check isolation for the components coming in direct contact with the power supply (in your case, pin 1 of the relay K1 and pin 4 of K1). It will appear like pin 4 is open but it will be connected to the main supply when the relay is not connected to pin 3. At the input of the transistors Q5 and Q4, there is a connection coming to pin 1 from the relay (12V). So, you need to check the isolation for this line too.
Check for reinforcement categories (isolation between high voltage and low voltage line) also. These categories are based on material type, pollution degrees, and type of environment.
This project is intended as a one-off hobbyist build, and not intended to be sold, else I’d pay extra attention to these standards.
It DID prompt me to take a look at the ‘stackup’ of the boards connector-wise, and it did show I’d chosen an unfortunate location for one of the AC connectors. Some minor adjustment is in order…!
That changes a lot. My home project of the last several years is a potentially dangerous product which (as it is) would likely be cause for a serious lawsuit if it were to go into production. If you are comfortable with the risks of what you are doing then go for it!
