Low-leakage triac or solid-state relay

[CodyDowling]

Hello,

I am looking for a SSR (or triac, since I don't need isolation) that has very low leakage (~nA's or, preferably, pA's) voltage requirement is low ( >50V), and current carrying capability is low (>1A-ish). The key care about is leakage and capacitance presented to the isolated circuit (<5pF either from input to output or from input to ground). I was looking at RF SSR (https://www.digikey.com/product-detail/en/panasonic-electric-works/AQY212FG2SX/255-3866-1-ND/3885576):which is pretty good but, of course, pretty pricey.

Has anyone come across something like what I'm describing? Digikey doesn't give me the ability to filter on leakage or capacitance, so I'm having to crack open and search the data sheet for each one and, so far, no dice.

Thanks in advance!
Cody

[Zero999]

That's what I would have suggested.

You could use a photovoltaic opto-isolator, such as the VOM1271T and two external MOSFETs, but that won't save you any money.

EDIT:
Here are some cheaper alternatives:
https://www2.mouser.com/ProductDetail/Panasonic-Industrial-Devices/APV2121SZ?qs=sGAEpiMZZMsAwfxLaIqwNwEEjpYKNg2Ndv3wF62dBks%3d
https://www2.mouser.com/ProductDetail/Panasonic-Industrial-Devices/APV2111V?qs=sGAEpiMZZMsAwfxLaIqwN8mKK5OKu01PlBRJ6K3A0lk%3d

[CodyDowling]

Thanks for the help Hero999!

On those below, it seems that the short circuit current they list in the data sheets found in links you sent below has more to do with the current "eaten" by the control circuit when the device is fully on, and not the current carrying capability- is that correct? They show a forward current of 1A, for instance, but short circuit current in the microamps...

In any case, the triac seems like a natural fit, but they have even less information regarding parasitic capacitance presented to the isolated circuit than do the SSR. For the SSR, I took a guess that something labeled for "RF" would be a bit more careful with stray capacitance, but I don't see a "RF" sort parameter on the triacs.

It seems that, if one didn't need to have the optical isolation, one could maybe "roll their own", right? I was thinking a couple of series mosfets (with low Ciss Coss specs) similar to a bi-directional switch for battery charging circuits, figure out how to drive the gates, and "Bob's your Auntie"... right?

[NiHaoMike]

Depending on how low is "low voltage" and "low current", an analog switch might be the solution.

[Kasper]

Its too bad digikey doesn't offer that as a filter option. I have typed "ctrl-f quiesc" in way too many datasheets.

[Zero999]

Thanks for the help Hero999!

On those below, it seems that the short circuit current they list in the data sheets found in links you sent below has more to do with the current "eaten" by the control circuit when the device is fully on, and not the current carrying capability- is that correct? They show a forward current of 1A, for instance, but short circuit current in the microamps...

They're just opto-couplers with an IR LED on the input and solar cell on the output. The peak forward current rating of the emitter is 1A, but the solar cell current is tiny.

The opto-coupler can't control the load circuit directly. The idea is to use it to drive a MOSFET. For an AC circuit, two MOSFETs are required.

In any case, the triac seems like a natural fit, but they have even less information regarding parasitic capacitance presented to the isolated circuit than do the SSR. For the SSR, I took a guess that something labeled for "RF" would be a bit more careful with stray capacitance, but I don't see a "RF" sort parameter on the triacs.

It seems that, if one didn't need to have the optical isolation, one could maybe "roll their own", right? I was thinking a couple of series mosfets (with low Ciss Coss specs) similar to a bi-directional switch for battery charging circuits, figure out how to drive the gates, and "Bob's your Auntie"... right?

You could use a TRIAC based opto-coupler. The capacitance isn't specified but it will be tiny, far below your requirements.
http://www.farnell.com/datasheets/97984.pdf

[perieanuo]

Hi,
Do you mentioned frequency involved?
Regards,pierre


Envoyé de mon iPad en utilisant Tapatalk

[Zero999]

Here's how to connect an opto-coupler with a photovoltaic output to a couple of MOSFETs, forming a solid state relay.

[CodyDowling]

Depending on how low is "low voltage" and "low current", an analog switch might be the solution.

Nĭ hăo NiHaoMike!

Sorry, yes- those are a bit ambiguous. The voltage about +/-20V, current is +/- 1A. The switch- even a relay, would be an option but I would like to have the ability to control the speed at which the triac goes from inactive to active, and vice versa.

[wraper]

SSR has triac optocoupler between input/output if it's AC type. Capacitance depends on that optocoupler and PCB/package. I just grabbed some random optocoupler and it has 3pF between input/output. I guess SSR in SIP package can do better.

[CodyDowling]

Absolutely- chasing down specs in data sheets is no biggie when you have 3 or 4 of them- but the staggering amount of components that fit the application every other way *but* a few super important ones makes rifling through data sheets a right "pain in the cunning linguist" as AvE says.
 

[wraper]

For example this one has 3pF maximum. https://eu.mouser.com/ProductDetail/IXYS-Integrated-Circuits/CPC1966YX6?qs=sGAEpiMZZMsUriz2CNI3E27VrHTFQPrBI1CmPCUVKJJNRJywL0SmHw%3d%3d

[CodyDowling]

They're just opto-couplers with an IR LED on the input and solar cell on the output. The peak forward current rating of the emitter is 1A, but the solar cell current is tiny.

The opto-coupler can't control the load circuit directly. The idea is to use it to drive a MOSFET. For an AC circuit, two MOSFETs are required.

  Yes, certainly- I missed that difference. I guess that a MOSFET with the set of parameters I'm interested in should be easier to identify since they tend to have Ciss and Coss values more readily available.

You could use a TRIAC based opto-coupler. The capacitance isn't specified but it will be tiny, far below your requirements.
http://www.farnell.com/datasheets/97984.pdf

Great- I'll take a look. I probably should have attached this earlier in the thread, but here's a simplified block diagram what I'm aiming to do in the attachment.

As you can see, I have a ground referenced load driven by a  sort of push-pull / class-A output stage. The emitter resistors are large enough to handle the shunts to ground that will result from activating the triac (or mosfets). This will be only a momentary operating condition that occurs as a result of a fault condition. Soon after the shunt is applied the supplies will be turned off. Likewise when the supplies are turned back on, the shunt will release once the supplies and signal path are stable.

[CodyDowling]

SSR has triac optocoupler between input/output if it's AC type. Capacitance depends on that optocoupler and PCB/package. I just grabbed some random optocoupler and it has 3pF between input/output. I guess SSR in SIP package can do better.

Hi Wraper, thanks for the info. The 3pF sounds very reasonable- you mentioned that was from the triac's input to it's output, correct? I think in order to get the other leakage paths (input to ground, for example) I'll either have to get lucky with a data sheet or get some in and measure. As shown in the diagram I attached to Hero's response above, I would like to present as high of an impedance as possible- from say, 10Hz to 100kHz, to the node between the two 14 resistors, so understanding the leakage paths better will really help I think.

[CodyDowling]

Hi,
Do you mentioned frequency involved?
Regards,pierre
...

I didn't but, ideally, I'd like to be able to get very high impedance up to 100kHz. The RF components seem like they shouldn't start really leaking until much higher, so I thought that might be a good jumping off point.

[perieanuo]

Hi,
Do you mentioned frequency involved?
Regards,pierre
...

I didn't but, ideally, I'd like to be able to get very high impedance up to 100kHz. The RF components seem like they shouldn't start really leaking until much higher, so I thought that might be a good jumping off point.

Ok.
So ssr is the only option, reed is too slow.me I always prefer at home building my own ssr takes some minutes but sure it's better than chinese ebay ones.at work, sil types for reasonable power or bigger ones for big heaters.
For ssr building the posts before explains very well, I won't repeat, just saying I use always moc's with zero switching capabilities.
Tell us what your work did in the end.
Regards,pierre


Envoyé de mon iPad en utilisant Tapatalk

[CodyDowling]

Ok.
So ssr is the only option, reed is too slow.me I always prefer at home building my own ssr takes some minutes but sure it's better than chinese ebay ones.at work, sil types for reasonable power or bigger ones for big heaters.

Forgive me, but- may I assume that by "sil" you mean silicon? or maybe single in-line?

For ssr building the posts before explains very well, I won't repeat, just saying I use always moc's with zero switching capabilities.
...

I definitely will update with what I come up with. When you mentioned "... the posts before." Did you mean the responses in this thread or others somewhere here on the Forums?

[perieanuo]

Forgive me, but- may I assume that by "sil" you mean silicon? or maybe single in-line?

I definitely will update with what I come up with. When you mentioned "... the posts before." Did you mean the responses in this thread or others somewhere here on the Forums?

sil like single in line like S202S12 ssr
sorry, the schematics for ssr are from other threads (the ones with optotriac moc3020+triac+stuff), but watch to choose a zero-switching optotriac...

[Zero999]

Hi,
Do you mentioned frequency involved?
Regards,pierre
...

I didn't but, ideally, I'd like to be able to get very high impedance up to 100kHz. The RF components seem like they shouldn't start really leaking until much higher, so I thought that might be a good jumping off point.

I'm confused.

Do you want to switch 100kHz current with this relay? Or do you just need it to not leak a 100kHz applied between the control and load circuits?

A TRIAC is no good for switching a 100kHz current and MOSFETs will leak a considerable amount, at that frequency. You need a reed relay to switch that frequency current.

[perieanuo]

Hi,
Do you mentioned frequency involved?
Regards,pierre
...

I didn't but, ideally, I'd like to be able to get very high impedance up to 100kHz. The RF components seem like they shouldn't start really leaking until much higher, so I thought that might be a good jumping off point.

I'm confused.

Do you want to switch 100kHz current with this relay? Or do you just need it to not leak a 100kHz applied between the control and load circuits?

A TRIAC is no good for switching a 100kHz current and MOSFETs will leak a considerable amount, at that frequency. You need a reed relay to switch that frequency current.

As I already excluded reed in my earlier post (I understood the signal needed to be switched at high freq,reed can't do quick, not that the signal itself has 100k).
I give up...
In fact hero999 is right if signal is HF and commutation speed is slow, very slow.
I need a coffee and clearer specs, which f for signal,which bandwidth for signal to be commuted...


Envoyé de mon iPad en utilisant Tapatalk

[Zero999]

Hi,
Do you mentioned frequency involved?
Regards,pierre
...

I didn't but, ideally, I'd like to be able to get very high impedance up to 100kHz. The RF components seem like they shouldn't start really leaking until much higher, so I thought that might be a good jumping off point.

I'm confused.

Do you want to switch 100kHz current with this relay? Or do you just need it to not leak a 100kHz applied between the control and load circuits?

A TRIAC is no good for switching a 100kHz current and MOSFETs will leak a considerable amount, at that frequency. You need a reed relay to switch that frequency current.

As I already excluded reed in my earlier post (I understood the signal needed to be switched at high freq,reed can't do quick, not that the signal itself has 100k).
I give up...
In fact hero999 is right if signal is HF and commutation speed is slow, very slow.
I need a coffee and clearer specs, which f for signal,which bandwidth for signal to be commuted...


Envoyé de mon iPad en utilisant Tapatalk

Yes, I did notice that, but it confused me.

Where did he say he wanted to switch the signal at 100kHz? It seems more like he wanted to switch a 100kHz signal, at a low frequency.

To chop a DC or AC signal up at 100kHz, with isolation, use a couple of MOSFETs driven by a pulse transformer.

You're right. We need a better specification.