Delta vs Y arrangement for 3-terminal transient voltage suppresors?

[Tim Shoppa]

When I open the catalogs I see both "Y" and "Delta" arrangements
for three-terminal TVS's. But I'm not entirely clear on where each
arrangement is most appropriate. My random thoughts:

The Y has two suppresors in series between any two terminals. Putting
them in series, I suppose, reduces capacitance, which may be important
at higher frequencies. But it also seems to double the impedance to
shunt a HV pulse.

The Delta seems to be better
at clipping lower voltages because you don't have to put two elements
in series.

But there must be something I'm missing here as to which is best used where.

My current application is rather low voltage (3.3V differential pair) at
moderate rates (a few Mbits) and it seems that the Delta configuration is
the best match.

Tim.

 

[Rich Grise]

When I open the catalogs I see both "Y" and "Delta" arrangements
for three-terminal TVS's. But I'm not entirely clear on where each
arrangement is most appropriate. My random thoughts:

The Y has two suppresors in series between any two terminals. Putting
them in series, I suppose, reduces capacitance, which may be important
at higher frequencies. But it also seems to double the impedance to
shunt a HV pulse.

The Delta seems to be better
at clipping lower voltages because you don't have to put two elements
in series.

But there must be something I'm missing here as to which is best used where.

My current application is rather low voltage (3.3V differential pair) at
moderate rates (a few Mbits) and it seems that the Delta configuration is
the best match.

Tim.

The first thing that you're missing is that the TVS's you're looking at
are for three-phase mains. The Y and Delta are two common connection
configurations - it's probably pretty easy to look up an explanation
for them, but I'm almost confident that the parts you're looking at are
not at all suitable for the app you're talking about. That's the
kind of deal where I'd put reverse-biased signal diodes between the
line(s) and Vss and line(s) and Vdd (or Vee/Vcc, or ground and +V,
or however you're calling it out. i.e.:


+3.3Vcc
|
k -
^ Diode
|
--------+-------- signal
|
k -
^ Diode
|
Gnd

Hope This Helps!
Rich

 

[legg]

When I open the catalogs I see both "Y" and "Delta" arrangements
for three-terminal TVS's. But I'm not entirely clear on where each
arrangement is most appropriate. My random thoughts:

The Y has two suppresors in series between any two terminals. Putting
them in series, I suppose, reduces capacitance, which may be important
at higher frequencies. But it also seems to double the impedance to
shunt a HV pulse.

The Delta seems to be better
at clipping lower voltages because you don't have to put two elements
in series.

But there must be something I'm missing here as to which is best used where.

My current application is rather low voltage (3.3V differential pair) at
moderate rates (a few Mbits) and it seems that the Delta configuration is
the best match.

Your current application is unlikely to benefit directly from the app
notes you are refering to.

In line power applications, Higher voltage suppressors are more
difficult to source and to apply.

Normally, if the input power is Y, you would use a Y configured
suppressor.

The delta configuration requires a higher voltage part, so the
temptation is still to use the Y configuration. As the energy has to
be limited by an external impedance, the impedance of the clamping
network is not particularly signifigant.

Y configured suppressors share the energy, doubling the absorption
capacity and halving the limiting impedance requirement on each line
for the same attacking wavefront, if the source is delta-configured.

Y configurations can also be used to clamp a system to a safety earth
potential, from the virtual reference node of the suppression network.
A significant family of impulses are presented as common mode (ie
lightning).

RL

 

[Jim Adney]

The first thing that you're missing is that the TVS's you're looking at
are for three-phase mains.

Maybe, but why would there be a THREE terminal Y configuration?

A 3-phase Y system would need a 4th terminal.

-

 

[Peter A Forbes]

When I open the catalogs I see both "Y" and "Delta" arrangements
for three-terminal TVS's. But I'm not entirely clear on where each
arrangement is most appropriate. My random thoughts:

As has already been explained by others, the Delta and Star (Y or Wye)
connection arrangements originate from 3-phase power supplies and consumers like
motors etc.

Delta is normally used for machinery supplies where a neutral is not needed,
similarly for HV (11kV to 256kV) transmission purposes.

Star or Y or Wye is used at local distribution level where three-phase and
neutral can be used to extract (for example) 240V single phase from a 440V
phase-phase 3-phase and neutral supply. In the USA it is 110V from 208
phase-phase supply IIRC.

Transfomers usually have Delta input and Star output for phasing and harmonics
reasons. We have just taken delivery (yesterday) of a 100kVA 415/440V Delta
input to twin secondary 58V and 583V for a special battery charger. The neutral
point on the secondary goes unused as we feed the secondaries into fully
controlled thyristor bridges. Tranny is about 700kg or nearly 3/4 of a ton. With
the two smoothing chokes the inductors are about 850kg total.

Peter

 

[Tim Shoppa]

The first thing that you're missing is that the TVS's you're looking at
are for three-phase mains.

Nope, they aren't. When a manufacturer sells a 3-device sidactor network
that trips at 14V, it's not for mains power.

The Y and Delta are two common connection
configurations [for three-phase power]

Like everyone else, you're assuming that when I say "Delta" and "Wye"
that I'm talking about mains power, but I'm not. Those terms are
perfectly relevant for any connection of three devices to three terminals.

I suppose I would've had better success asking in a telecom-related
group about TVS's used in telco demarcs. When I ask about differential
Mbps-rate signal surge protection and everyone starts correcting me and
telling me I'm really asking about three-phase mains power, I must've asked
the question in the wrong place.

Tim.

 

[legg]

The first thing that you're missing is that the TVS's you're looking at
are for three-phase mains.

Nope, they aren't. When a manufacturer sells a 3-device sidactor network
that trips at 14V, it's not for mains power.

The Y and Delta are two common connection
configurations [for three-phase power]

Like everyone else, you're assuming that when I say "Delta" and "Wye"
that I'm talking about mains power, but I'm not. Those terms are
perfectly relevant for any connection of three devices to three terminals.

I suppose I would've had better success asking in a telecom-related
group about TVS's used in telco demarcs. When I ask about differential
Mbps-rate signal surge protection and everyone starts correcting me and
telling me I'm really asking about three-phase mains power, I must've asked
the question in the wrong place.

A recommended connection for three-terminal (delta) devices on data
lines is illustrated in ST AN821, and AN585. The illustration of a Y
connection is restricted to the example of discrete component
application.

http://www.st.com/stonline/books/pdf/docs/4351.pdf
http://www.st.com/stonline/books/pdf/docs/3597.pdf

other app notes

http://www.st.com/stonline/books/toc.old/an/59.htm
http://www.microsemi.com/support/micnotes.asp?MN=316

Multiple integrated TVS/clippers, tied to common points or split
between busses are also common low-voltage configurations.

Reverse-biased signal diodes are sometimes used to isolate the signal
line from the capacitance of clamping semiconductors and circuitry.

Perhaps if you can give references to the low-voltage Y configurations
that were the source of your innitial confusion, we can be of more
assistance.

RL

 

[Rich Grise]

"Rich Grise" <[email protected]> wrote in message

Like everyone else, you're assuming that when I say "Delta" and "Wye"
that I'm talking about mains power, but I'm not. Those terms are
perfectly relevant for any connection of three devices to three terminals.

I suppose I would've had better success asking in a telecom-related
group about TVS's used in telco demarcs. When I ask about differential
Mbps-rate signal surge protection and everyone starts correcting me and
telling me I'm really asking about three-phase mains power, I must've asked
the question in the wrong place.

Sorry. I stand humbly corrected.

Thanks,
Rich

 

[Tim Shoppa]

this is merely one application of star and delta connected networks. A "pi"
filter is a delta three-terminal network; conversely a "T" filter is a star.
The same transformations from star-to-delta and back again also apply to
filters. And any other suitable network and (IIRC) its dual....ultimately
its just network theory. And yes, it works at 50Hz, from mW to MW.....

Not to diss network theory, linear network theory is great stuff
.... but TVS's (Sidactors, MOV's, Gas discharge tubes,
etc.) are not linear devices when activated. They are quite nonlinear
and often exhibit hysteresis. I suspect that part of the reason for
preferring Y over Delta (or the other way around!) has something to do
with the nonlinear characteristics, but that's just a guess.

Tim.

 

[Tim Shoppa]

A recommended connection for three-terminal (delta) devices on data
lines is illustrated in ST AN821, and AN585. The illustration of a Y
connection is restricted to the example of discrete component
application.

http://www.st.com/stonline/books/pdf/docs/4351.pdf
http://www.st.com/stonline/books/pdf/docs/3597.pdf

Those are good examples... but I see them as Delta arrangement, not Y.

My definitions, tell me if I'm wrong. 1,2,3 = terminal; A,B,C = device

Delta: 1---A---2
\ /
B C
\ /
3

Y:
1 2
\ /
A B
\ /
+
|
C
|
3

Perhaps if you can give references to the low-voltage Y configurations
that were the source of your innitial confusion, we can be of more
assistance.

http://rocky.digikey.com/WebLib/Teccor/Web Data/pxxx3a_.pdf

shows some telecom-oriented Y configs. (Although the voltages
are not nearly as low as the individual elements are available.)

Now that I look, Teccor is claiming the patent on the "Y"
(US Patent 4905119), I'll have to download the patent and see what it
says. Usually I get so lost in the patent gobbledygook that I need
someone else to explain it to me .

Tim.

 

[Terry Given]

On 5 May 2004 09:42:27 -0700, [email protected] (Tim Shoppa) wrote:

As has already been explained by others, the Delta and Star (Y or Wye)
connection arrangements originate from 3-phase power supplies and consumers like
motors etc.

this is merely one application of star and delta connected networks. A "pi"
filter is a delta three-terminal network; conversely a "T" filter is a star.
The same transformations from star-to-delta and back again also apply to
filters. And any other suitable network and (IIRC) its dual....ultimately
its just network theory. And yes, it works at 50Hz, from mW to MW.....

Terry

 

[R.Legg]

Those are good examples... but I see them as Delta arrangement, not Y.

The Y application, as previously indicated, is illustrated only using
discrete components, specifically in fig#3 on page 2 of the latter
3597 reference.

My definitions, tell me if I'm wrong. 1,2,3 = terminal; A,B,C = device

Delta: 1---A---2
\ /
B C
\ /
3

Y:
1 2
\ /
A B
\ /
+
|
C
|
3


http://rocky.digikey.com/WebLib/Teccor/Web Data/pxxx3a_.pdf

shows some telecom-oriented Y configs. (Although the voltages
are not nearly as low as the individual elements are available.)

The lowest voltage device in your reference has 130V breakdown, the
highest is 420V. These parts are not aimed at logic-level data
communication line protection applications, except possibly for basic
telephone (ie ISDN), which is ~60Vac on a 48VDC bus.

Now that I look, Teccor is claiming the patent on the "Y"
(US Patent 4905119), I'll have to download the patent and see what it
says. Usually I get so lost in the patent gobbledygook that I need
someone else to explain it to me .

The patent is quite specifically for a three-terminal integrated
device, composed of SIDAC's, when the device is applied to telephone
lines.

A SIDAC is, in itself, an integrated combination of a triac and a
small-signal voltage graded diac (or four-layer diode). A SIDAC is not
a voltage limiter, it is a voltage transient-triggered crowbar. In
discrete form, it has been marketed by Motorola for over twenty years.
(prefix MKP)The full application note is included in the general
library under the title 'Thyristor Device Data' from ON Semi DL137
revD.

http://www.onsemi.com/pub/Collateral/DL137-D.PDF

A common challenge is to define the original words used to form the
acronym that became the devices name - SIDAC. I believe Motorola even
offered a prize once, to anyone who could 'guess' - nobody won it, as
I recall.

RL

 

[Tim Shoppa]

The Y application, as previously indicated, is illustrated only using
discrete components, specifically in fig#3 on page 2 of the latter
3597 reference.

Thank you, I missed that the first time around.

The lowest voltage device in your reference has 130V breakdown, the
highest is 420V. These parts are not aimed at logic-level data
communication line protection applications, except possibly for basic
telephone (ie ISDN), which is ~60Vac on a 48VDC bus.

I suppose it's possible that the Y arrangement lets them make a
monolithic device that wouldn't be possible in a delta arrangement
that meets the telco needs.

Most of the more industrial/military data line protectors are a
high-voltage limiter/crowbar on the front end (usually gas discharge
tubes), followed by maybe some L filtering, then a much lower
voltage (few volts to few tens of volts depending on needs)
solid-state protection.

Bourns has some Y-network three-thyristor networks with much lower voltages...

http://www.bourns.com/2/pdfs/TISP70xxL1.pdf

The patent is quite specifically for a three-terminal integrated
device, composed of SIDAC's, when the device is applied to telephone
lines.

Yeah, I get that, and it's specifically for the Y arrangement as far
as I can read. But they don't say why they use the Y arrangement
vs the Delta arrangement.

A SIDAC is, in itself, an integrated combination of a triac and a
small-signal voltage graded diac (or four-layer diode). A SIDAC is not
a voltage limiter, it is a voltage transient-triggered crowbar.

The Gas Discharge Tubes have a similar crowbar (holding voltage less
than triggering voltage) although there is much less range between
holding and triggering.

Tim.

 

[legg]

Yeah, I get that, and it's specifically for the Y arrangement as far
as I can read. But they don't say why they use the Y arrangement
vs the Delta arrangement.

They don't have to.

Please note that they are not offering to sell you three independent
sidacs in the same body size. The vendor is obviously picking circuits
that are practically realisable on the same semiconductor substrate.
App notes will reflect what they've got to sell you, not what is
necessarily best in an application.

Even dual integrated devices have one common terminal.

This will have an inadvertent advantage in that the entire energy
absorption capability of the package is available to an individual
element for asymmetrical strikes.

The mfr is quite specific as to the waveforms intended for
suppression. Most applications are equally specific regarding the
waveforms that will be requiring suppression. Suggest you compare the
two lists carefully.

When used in conjunction with other devices and impedance, they should
assist nicely in promoting survivability of attached equipment

RL