0
\$\begingroup\$

I am seeking some feedback in improving the reed switch snubber protection circuit shown below. The switch will drive a coil of an AC relay. Please note that the AC relay coil is not accessible. Any protection will have to be across the reed contact. Here is the relay part# Spec sheet: Relay spec Sheet

enter image description here

  • Relay coil operating current is 24mA, 115AC.
  • The reed switch SW1 is Mk23-90C-2

Existing RC values. R=100Ohms. C=5nF Any value close to 10nF will make the capacitor conduct and causes the relay to activate. reed Switch snubber Here is a screenshot showing 3 waveforms. On the common reed terminal I placed the probe ( blue waveform) 115 AC 400Hz . On the NO terminal I placed the yellow probe.

The purple waveform represents the voltage difference across the reed switch contacts ( the potential difference of the blue and yellow)

Here is a scope shot showing what happens when I activate the reed switch “SW1" with the snubber circuit above R=100, C=5nF. The voltage spike is 180V. The max contact rating on the reed switch is 175V. Scope Shot-With snubber

\$\endgroup\$
  • 2
    \$\begingroup\$ Have you considered a ZCS Opto Triac in series? All OMRON AC relays specify the optimum RC snubber, but ZCS will significantly eliminate the spikes. with near zero current off switching. \$\endgroup\$ – Tony Stewart Sunnyskyguy EE75 Nov 30 '16 at 17:41
  • \$\begingroup\$ @TonyStewart.EEsince'75 I didn't know about it. I will have to look into it, thank you. Which one would you suggest? \$\endgroup\$ – Rocky79 Nov 30 '16 at 17:48
  • \$\begingroup\$ What's the limiting voltage on the reed switch? \$\endgroup\$ – Andy aka Nov 30 '16 at 18:03
  • \$\begingroup\$ @Andyaka Max contact voltage is 175V Max current is 0.5A. Max power is 10W \$\endgroup\$ – Rocky79 Nov 30 '16 at 19:45
  • \$\begingroup\$ @Rocky79 Do you know that Reed switch life expectancy will be reduced many orders of magnitude with this snubber. What is Relay L, DCR & Cable length. Snubber must be across coil NOT the Reed switch and series R is needed. \$\endgroup\$ – Tony Stewart Sunnyskyguy EE75 Nov 30 '16 at 19:47
2
\$\begingroup\$

This is to identify the problems to avoid, in the solution.

Problems:

  • magnet powered Reed SPDT switch induces contact bounce 23kHz of 0.5ms
  • load is a reactive power relay coil something like this. from my interpretation
  • if coil is rated for 60Hz then at 400Hz coil current is reduced by increase in frequency ratio. (Warning ) Suspect L is close to 1H, DCR1~2kOhm

  • What the Reed spec doesn't tell you is horizontal time scale spans many decades. See below

enter image description hereenter image description here

To make a reliable Reed switch , never use more than 10% of rated Ipk

Adjust series R to limit current and protect input V with series R and TVS or MOV for 200Vp which is the Breakdown voltage spec of chosen Reed . I suggest 500V reed if you must use a reed. enter image description here enter image description here

Who would have ever thought an aircraft switch was so complicated? I prefer a better design than above , but too much info is missing.

enter image description here

Note that I have plotted their data on a LOG-LOG chart and added your operating point with conservatively is 1A but actually 115/100 Ohm =1.15A * 115 = 135VA for a switch rated for 10VA?????

\$\endgroup\$
  • \$\begingroup\$ Tony, thank you for the detailed reply, I am still trying to wrap my head around it. I have updated my original post to show the relay datasheet, I hope that helps. I will have to add a resistor and measure the current spike on reed contact. So based on the spike current value I will determine a value for R1 so that the reed switch peak current stays within specs? Ideally I would like to go to the higher spec reed but the application requires SPDT reed switch and that's the only one offered. \$\endgroup\$ – Rocky79 Nov 30 '16 at 22:28
  • \$\begingroup\$ Reed relays are rated for 1e6 mechanical operations. but if you use the 1A max contact current abs. max limit, the contacts may last 1e3 operations, if memory serves me. See other Reed supplier specs for MTBF vs contact current. Even this 10A Relays are derated down to 2A for lamps due to 10x surge current possible but 3x on average with random phase. ask supplier too standexelectronics.com/inquiry-form \$\endgroup\$ – Tony Stewart Sunnyskyguy EE75 Dec 1 '16 at 1:37
  • \$\begingroup\$ Tony, operating current of the relay is 24mA @ 115AC RMS not 1.15A. I am worried about the short peak voltage/ current during switching in micro to nano second spikes. I have added a 10 ohms resistor in series with the reed switch and removed the RC snubber circuit. I measured the differential voltage using a probe and found the voltage go out to 300v which is almost 30amps of current. I couldn't measure the pulse width because it was so small. \$\endgroup\$ – Rocky79 Dec 1 '16 at 3:16
  • \$\begingroup\$ Cable capacitance causes an instant short circuit on closure limited by your 10 Ohms. Before I said 115V/100 Ohms =1.15A for just snubber current. THe reed is rated for 10VA and it needs to be 1VA for long life on contact , e.g. < 10V at 100mA pulse not 10A. This can only be done with a ZCS with some small series e.g. 100R. Read my links for ZCS parts. Start to understand all impedances including cable capacitance, ESR ESL, etc Effectively it is the Joule energy that burns contacts so compute 1/2CV^2, 1/2 LI^2 for close & open. Consult with Reed factory if unsure \$\endgroup\$ – Tony Stewart Sunnyskyguy EE75 Dec 1 '16 at 8:33
  • \$\begingroup\$ or try my 330R >100R+47nF filter \$\endgroup\$ – Tony Stewart Sunnyskyguy EE75 Dec 1 '16 at 8:38
1
\$\begingroup\$

I suggest this Application Note will give you the best solution. OMRON are the best in the industry for Relays for reliability.

https://www.omron.com/ecb/products/pdf/precautions_ssr.pdf

I would choose a ZCS triac IC with optional load R and input & output RC filter for improved immunity from dV/dt noise false triggers, which is far more reliable.

If you do use a snubber with high dI/dt, you may want to add ferrite beads and use twisted pair wires to reduce egress but ZCS will do this too.

http://www.digikey.com/product-search/en/isolators/optoisolators-triac-scr-output/3736394?k=triac&k=&pkeyword=triac&pv41=719&pv1787=2&FV=fff40039%2Cfff8034a&mnonly=0&newproducts=0&ColumnSort=0&page=1&stock=1&quantity=0&ptm=0&fid=0&pageSize=500

\$\endgroup\$
  • \$\begingroup\$ I can’t use a triac in this case specially the one you have shown with the opto isolator which require a DC voltage to turn on/off the load. I only have a 115AC and return available. Thank you \$\endgroup\$ – Rocky79 Dec 1 '16 at 17:03
  • \$\begingroup\$ How hard is it to get 10mAdc from 115Vac? but I understand. This is not a flexible design environment. \$\endgroup\$ – Tony Stewart Sunnyskyguy EE75 Dec 1 '16 at 17:34

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.