Upload & Sell: Off
Brian you are quoting out of context and editorializing and moderating about the advice of others again based on your own narrow interpretation of the question. You say I'm correct but just don't understand the context of the information I'm trying to convey because you can't get past your point of view....
Although I didn't quote the OP I was responding to his last statement where he said... " I had read in the day when there was less technology .... " which is why I referenced the Vivitar 285HV to explain the differences, not with respect to the narrow topic of thyristors but in practical usage of auto thyristor type flashes vs. camera TTL control.
"The difference between thyristor style like a Vivitar 285HV and a modern digital camera is how the exposure calculation is made...." With older Thyristor flashes like that one the exposure calculation was made entirely in the flash as a real time feedback loop independent of any camera control vs. today's digital cameras where pre-flash read and flash output controlled by the results of the camera with today's digital cameras.
So to clarify I was pointing out meter was done on flash vs. camera as the main difference between today and "when there was less technology".
A thyristor doesn't "shunt the electricity .... BACK to the capacitor" it more accurately simply switches off the flow of current from the capacitor to the flash tube which preserves the charge remaining in the capacitor.
With a flash the gap between the electrodes of the xenon flash tube act as the "ON" switch. To fire the flash a charge is sent to a ring, wire or shield around the flash tube which changes the state of the xenon in the tube enough to make it conduct a charge between the electrodes creating a "short circuit" across the electrodes which is what actually unleashes the charge from the capacitors through the open gate of the Thyristor to energize the xenon into a plasma state. A "thyristor style" speedlight signals the thyristor gate to close when it sees sufficient light, stopping the flow to the capacitor. The gate of the thyristor must then be reset to the open position for the next exposure.
Very early flash designs prior to the introduction of semi-conductor based Thyristors switching flash power was controlled by banking capacitors because there was no "OFF" switch. A flash with four capacitors could be banked to fire at Full power (all 4), 1/2 power (2 of 4) or 1/4 power (1 of 4). The trigger circuit would energize the flash tube and make the xenon conductive, closing the circuit and firing the flash until all the current in the connected capacitors was depleted.
The semi-conductor Thyristor technology added the "off" switch which allowed self-contained automatic metering on the flash and multiple flashes per capacitor charge. The auto sensor on the Vivitar flashes consisted of a simple sensor with ND filters which are placed over it. The ND filters clicked in place over the sensor would transmit less light vs. when no filter was over the sensor, which would cause the flash to stay on longer before the metering circuit triggered the closing of the Thyristor gate.
As for practical usage and comparison of flash control methodology (flash metering control vs. camera metering control)?
I used pairs of Vivitar Thyristor flashes since the early 70's, first with a pair of model 263 and then when switching to digital in 2000 a pair of 285HVs. In 2005, after using the 285HVs for a year on my Canon 20D I switched to a pair of 580ex.
The simple averaging on-the-flash metering on the Vivitar worked quite well with negative film for single flash use because negative film has about a 2 stop latitude for overexposure. So you could just err on the side of over-exposure and a good print could be made. Outdoors I'd set the camera at f/11 @ 1/250th and the flash on the "f/8" color code for ISO 100, shoot into the shadow side and get flash fill that was slightly under the ambient.
The auto mode didn't work with dual flash because the light of one flash would affect the output of the other. Since I use a pair of flashes as fill and key most of the time, most of the time I used my Vivitars in their manual power mode.
When switching to digital manual power mode on the Vivitars offered the same precise control it had with film but I found that in "auto thyristor" mode simple averaging metering cell on the flash didn't work a well because digital requires more careful exposure control for the highlights. The fact the flash sensor was metering the FOV similar to a 50mm lens created exposure variables when I used zoom lenses. I found it was was cumbersome to use the Vivitars in auto mode on a digital camera. While in theory you would set the lens to an f/stop corresponding to the color code on the flash sensor control ring such as f/8, in practice the metering wouldn't correctly read the scene requiring adjustment of the actual f/stop to obtain correctly exposed highlights.
Switching to the 580ex allowed me to utilize the zone based evaluative metering of my 20D in ETTL mode, or when more advantageous switch to M mode. It also freed me from the ball and chain of the sync limit outdoors, which was actually my main motivation for making the switch from the Vivitars.
Learning to use Canon ETTL flash effectively was an acquired skill. First it was necessary to understand how the camera metering of ambient and flash works on in the various modes and what the limitations are. I tested all 18 combinations of ambient and flash metering on my 20D because I was curious which, if any, produced the best results. The test results are here: http://photo.nova.org/Canon/TTL/ I concluded that evaluative ambient and evaluative flash metering did the best job overall and have used that combination with good results.
Part of my learning curve was how the flash operated at the higher ISOs possible with digital cameras. I discovered, as the OP did, that the limitation is a physical one: the minimum flash duration. As the OP discovered it is possible to raise the ISO to the point where the flash can't trigger the flash then quench it quick enough to avoid underexposure. In terms of getting good results understanding strategies to work around that limitation are more important IMHO than whether or not the physical limitation of the flash is the result of it utilizing Thyristor or IGBT switching circuit.