------Question for SGT8690500 Please Put your question below------
We found the datasheet online for Item Part Number 3667-SGT8690500-ND, but the datasheet does not have any information about the duty cycle or maximum cycles per hour that we saw. Can you get any information from the manufacturer about the duty cycle rating or more specifically the maximum rated cycles per hour this SSR is rated for?
Hello David_Lelen01 and welcome to the DigiKey TechForum.
I was unable to find this information listed anywhere on the manufacturer website, but I have reached out to our Product Manager to see if they are able to provide this information. We will update you when we have more information.
Okay, awesome, thank you! Looking forward to your response.
Devices of this sort are generally not rated in terms of duty cycle or maximum cycles per hour, in some part because user-chosen accessories have a large influence on what is or is not feasible in an end application, and in some other part because the limitations that apply arise from entirely different reasons than those which are relevant to electromechanical relays.
One can refer to the below chart in the datasheet to get an estimation of maximum permissible load current based on which factory-qualified heatsink is chosen and ambient temperature. Note that heatsink performance is affected by a number of factors, such as airflow/mounting position, surface contamination/debris buildup, humidity and atmospheric pressure, etc. In other words, take the chart with a grain or two of salt and leave oneself some generous margin.
This post outlines the basic process for making thermal calculations in electronics applications, and can be transferred to situations of this sort with minimal inference.
To a first approximation, solid-state relays are insensitive to switching frequency and limited in their carrying capacity only by the ability of the installation to remove heat that is generated as current passes through the device. The contact arcing phenomenon which is the primary wear mechanism of conventional relays is absent in the solid-state variety, but in trade SSRs have vastly higher conduction losses that require deliberate thermal management.
Thank you very much for that valuable input! So in our particular application, we are bumping a motor in forward and reverse roughly 1000 times per hour (there are 10sec dwell periods roughly every 40s) and the power on pulse lasts roughly 80ms, so the mechanical contactors we have now are basically only seeing the motor inrush current of 92Arms(130Apk) and are failing very rapidly. The steady state motor current is 8Arms. So if i am understanding the datasheet correctly, this SSR can handle repetitive surge currents of ~500A for 0.1s which is well above our expected 130A peak. The inductive RMS continuous rating of this SSR is 32A which is also well above our expected 8A. And doing a rough calculation using the on-state resistance of 2miliOhm and worst case130A current for 0.1s gives us 3.38J (33.8W during the 0.1s pulse) that must be dissipated before the next pulse in ~1.5s. So that would be an average power until the next pulse of 2.25W. So with even the worst heat sink on the chart of 6C/W, we would have an average deltaT of 13.5C. With an ambient room temp of 30C, we would be well within the limits of this SSR. Does all of that sound even remotely correct or am I just totally off in left field watching butterflies go by?
EDIT: Missed the formula for the power dissipation in the datasheet. Actual dissipated power looks to be ~100W from the pulse giving a deltaT of 82.9C. But it’s significantly less than 100W because this is an AC wave and its only 130A for about 0.04s over 3 cycles. So with the best heat sink and also now including the junction-case thermal resistance, it should still be ok, but by a small margin?
Not quite; the output stage is a bipolar rather than MOS device, so the first-order conduction loss model is a fixed voltage rather than a resistance; that 2 mΩ figure is a dynamic resistance, representing the variation in voltage drop that occurs with changes in current. For your worst-case scenario, the conduction loss model would be 130A*(1V+0.002Ω*130A)=164W, or 16.4J for a 0.1s duration.
I’m not quite clear on your repetition frequency, but the ratio between the above figure and your own would lead me to suggest more heatsink than less. There are switching losses that will be additive to the conduction losses and the dust bunny fudge factor to account for here also…
Speaking of switching, the part mentions snubber and clamp protection for the outputs but doesn’t quantify them much. One would be well advised to evaluate dV/dt, di/dt, and off-state voltage during switching events to ensure that limiting figures aren’t being crowded too closely.
Some additional points: one may wish to note that if one’s only running for 100ms at a time, the half-cycle delay/variability that can occur due to the zero-cross nature of the device might start looking significant. Also, SSRs leak and are unsuitable as a disconnect mechanism.
Overall, it looks like a reasonable device to consider for the application as described; it’s not the sort of use case that a mechanical contactor is suited for unless the contacts are the size of hamburger patties…
Okay, I see what is going on here now. That makes sense. That’s definitely good info to know, thank you!
Yeah, its a bit hard to explain because of all of the dwell periods it goes through. It pulses 80ms every 2 seconds for 4 times (~130Apk), then dwells like 5 seconds, does another set of 4 pulses, dwells for 10 seconds, turns on constantly for about 30s (so only pulling 8Arms at this time), dwells 20s again, turns on for 30s again (again 8Arms), dwells 10s, and restarts the process.
So this circuit is 400VAC 60Hz 3Ph… dV/dt seems to work out to 0.137v/us which is much less than 500V/us. di/dt works out to 0.031A/us which is also much less than 50A/us. Off state voltage is 400Vrms, so ~565Vpk which is also less than 1600Vpk rating. I think we should be good there. Only question is, can the SSR handle reverse voltage applied to the outputs? So we have a pair of contactors per motor, one for forward and one with two leads swapped to be in reverse so when the first relay is not active but the second one in the pair is active, the first one is seeing a reverse voltage. That would still be less than 1600Vpk though. Will this be okay?
I am thinking this will not be terribly important in this scenario since the current mechanical contactors take a certain amount of time to physically move and close the contacts, though I don’t know what that time is I would assume a few milliseconds. Also, the nature of this quick pulse is just to bump the motors in a direction, so the timing is not critical. We can adjust the pulse width if necessary to account for this though if needed.
Again not a problem, small leakage current is acceptable and we have separate safety disconnects on this machine.
If I did math right, The average rms amperage over the duration of a sample pulse we measured is 59Arms for a duration of 0.06s. Plugging that into your equation gives 59A*(1V+0.002Ω*59A)=66W, or 3.96J for 0.06s duration. So average power until the next pulse would be roughly 2W, and using our best case heat sink with a value of 0.55C/W we get 1.1C temp rise which is great. Even considering absolute worst case here of 66W continuous, that gives a temp rise of 36.3C + 30Cambient = 66.3C which is still within allowable limits. So we should be perfectly fine using this SSR, sound correct-ish?
So one thing I read in the link to the thermal calculations you attached earlier was the breakdown and degradation of the thermal junctions, increasing thermal resistance. Is that something we will need to worry about with this type of device also? Of is that specific to the type of device referenced in your attachment?
It’s more the values observed during turn-off and -on that are of concern. Thyristors turn off when current flow through them crosses zero, which is near the peak of the AC line waveform when supplying a highly inductive load. As a result the voltage across your switch might go from ~0 to ~600v in some microseconds, which could well approach the stated dV/dt limits.
As for di/dt, that’s an issue during turn-on. Not all regions of the semiconductor device enter conduction at the same instant, with the result that fast-rising current waveforms can overwhelm the portions that get out of bed first and damage the device. An RC snubber placed across the switch to mitigate dV/dt is a likely culprit…
I’d rather see a person achieve the desired end using a 3PDT relay for directional control and a single contactor for on/off, due to the substantial risk of shorting the source by triggering both contactors simultaneously. So long as the directional relay isn’t switched under load, the wear to its contacts should be minimal.
I’m still fuzzy on the operating profile, but I get the sense that you’re in reasonable territory from a thermal management standpoint.
Possibly. The sort of “bumping” operation described is the sort of thing that one would expect to result in the silicon die to experience temperature changes relative to its package, potentially weakening that interface over time, though I’d expect better longevity relative to mechanical contacts. As for the interface between package and heatsink, it looks like you’ll probably have ample temperature margins to allow for some dry-out over time.
First of all, I would like to sincerely thank you again for your very detailed and extremely helpful replies! This is all very valuable information to learn.
Yeah, I was thinking about this all completely wrong and independently. I was thinking of a purely resistive load too. Oops. Okay, so I really need to get some simultaneous measurements of voltage and current to see what the power factor is and be able to accurately calculate the dV/dt and di/dt. My assumption of the bumps is the motor is acting mostly resistive, but during the steady state operation it would be highly inductive.
I don’t quite follow right now how the voltage could jump 600V in a few microseconds if the turn on/off time according to the datasheet is 0.5 cycles which would be ~0.008 milliseconds? What am I missing here?
Again I don’t think I am following, this one. If this relay is zero crossing, I understand that means it turns on when voltage is near 0 so current should also be near zero and rise as fast as the voltage rises as a function of the circuit resistance? So unless we have a capacitor in the circuit with a very low impedance, the di/dt should not be a concern, should it? I suppose “near” zero is the key though, if voltage is 10V when it turns on and the series resistance is 0.01Ohm. That could give a 1000A rise, but over what time period? The way I understand the datasheet, probably mistakenly, would be over 0.5 cycles or 0.008s?
That’s an excellent idea and I’m honestly a tad embarrassed we hadn’t thought of that already.
Power factor is more concerned with looking at the pretty sinusoids; the dV/dt and di/dt issues involve the spiky nasties occurring in the microsecond-scale window where the semiconductor devices are changing their conduction state.
The Basic Idea for many motors is that a spinning rotor generates a voltage referred to as a back EMF which opposes the supply; it is the difference between the supply and back EMF that forces current flow through the windings. Apply a mechanical load, motor slows down, back EMF decreases, and current draw increases. 'Aint quite accurate for induction motors, but enough so to be useful.
The reason one gets such ugly startup currents is because the rotor isn’t spinning when it’s not spinning, therefore making no back-EMF and leaving the winding resistance in the motor windings and supply cabling as the main thing limiting current flow. The actual inductance of the windings is there regardless, but its role often isn’t as significant as one might think.
That’s talking about the delay between de-assertion of the control signal and when the load turns off. Due to the nature of thyristors, that can be anything up to an AC half-cycle. The actual process of the output devices transitioning from their conducting to non-conducting state is a different concept that occurs more on the microsecond time scale, and where that dV/dt thing can be dicey.
As for di/dt, it’s indeed a lesser concern with zero-cross compared to random-fire devices. Still worth mentioning though, because it’s one of those limiting values on the datasheet that a person should understand. If you were to buy a random-fire device some day, one might see the capacitor on the RC snubber charged to, say, 500v when a person hit the button to trigger the relay, and that is definitely the sort of thing that’ll make for a fast-rising current waveform.
You’re welcome–hopefully others will find the discussion useful. Back in the golden age of electronics communication, there was a little company called Teccor (now part of Littelfuse) that produced an excellent series of application notes on the class of power switches in question titled AN100(1-?) that I’d highly recommend if a person wants to understand the things. Search queries such as “Teccor AN1001” still summon those, but I might have to make an archive somewhere for posterity…
Well, I suppose I truly meant the lag angle instead of power factor, but I believe power factor is analogous to lead/lag angle. I figure with that information, I could see the voltage when the current crosses zero and then be able to calculate dV/dt.
The only problem calculating dV/dt then is the dt part… where do I find that time window without having a part to measure? I don’t see that info in the datasheet. I feel like i remember seeing something like that on datasheets for MOSFETs and transistors and such, but not on a SSR datasheet.
I found it with a quick google search. That’s quite a detailed document! I’ll read through it as I have time. I love learning stuff. Thanks!
