It will be installed as two switches in parallel.
Every switch must switch ON/OFF with an accuracy of 2 microseconds, the current will flow in both directions.
It must be switched ON and OFF at maximum voltage and minimum current, but it will pass through every switch 20 kiloamps.
Dear Javier,
this is a typical task for high-power thyristors.
As there are no 50kV / 20kA single switches, using paralleled devices because of the current and series connection to achieve the blocking voltage would be mandatory. This will lead to a further topic: synchronization and delay times.
I'd suggest to go for so-called Light-Triggered-Thyristors (LTT) which are available in up to 8000V / 2300A (click the link for the according datasheet).
Paralleling 10 would get you 23kA and series-connection of 7 would achieve 56kV.
Two of those arrangement lead to 140 devices, anti-parallel, to achieve bidirectional current flow - not even a big array compared to what is built in HVDC-Applications.
Due to using LTT; there is no trouble with synchronizing the trigger pulses but I assume a setup to achieve homogeneous current sharing in transient condition will be a challenging task.
http://www.infineon.com/dgdl/Infineon-T2563NH-DS-v11_00-en_de.pdf?folderId=5546d4694909da4801490a07012f053b&fileId=db3a304412b407950112b430fab552bb
I thought that it is not possible to switch off the thyristors when they have voltage in the anode, it is not a problem? (accordingly specs turn off time=550us)
I read at the specifications that the thyristor withstands 93kA surge
Dear Javier,
a thyristor will turn off naturally if the current goes down to zero and the trigger is off..
As your description said "turn-off at minimum current" I assumed you would turn off in zero current condition according to the diagram you added.
The datasheet value correlating to 10ms is given, because this refers to using the device in a 50Hz application. I'm sorry I didn't recognize the time scale in your figure was µs and not milliseconds.
Critical voltage slope for the device I linked to is 2kV/µs and with 7 in series, turns to about 14kV/µs. Thus, 50kV can be turned off in less than about 4µs.
A solution using High-Voltage IGBTs will be even more challenging and demand even more devices, series-connecting IGBTs is rarely done in the industry.
You would need 6.5kV switches, maximum current today is 750A.
Wow - at least 8 in series to achieve your 50kV but at least 27 in parallel will get you 216 modules per direction, 432 for the whole switch and controlling these will get even worse. The number of gate drivers alone scares me :-)
So yes - possible but challenging.
I am impressed by LTTs they will do the duty, what driver do you recommend for 14 of them (7 forward and 7 reverse)?
I think that the LTT are easily controlled by using fiber optics and laser only.
I found also this one T2251, that is an electrical controlled (also from your company), what driver do you recommend?
I am sorry for being unfamiliarized with thyristor driving
http://www.infineon.com/dgdl/Infineon-T2251N-DS-v06_00-en_de.pdf?folderId=db3a304412b407950112b408e8c90004&fileId=db3a304412b407950112b42fb26d4cfb
Dear Javier,
now we start reaching a point at the edge of my knowledge :-)
Please see your private messages, I forwarded an expert's contact.
He'll be the one to help you with the details. To my regret he is not participating in this forum.
Dear Martin,
Do you have a spice (or any other) model of these high voltage thyristors?
It is very important to users to have them
Well I found how to obtain by using analog behavioral model.
After simulation I realized that the Toff is too large (550us) is too large to switch off the SCR.
I have seen after simulation that period must be >8ms:
- Badges
- Science topic
- Similar topics
- Mathematics